JPH05103196A - Image reader - Google Patents

Abstract

(57) [Summary] (Correction) [Purpose] To provide an image reading apparatus capable of accurately correcting dark output. A first holding means 16a, 16b for pixel output of only a transfer section without a photosensitive section, and a second holding means 17a, 17b for pixel output of which a light shield is provided on the photosensitive section, among dark output signals. Output signal normalization means 21a, 2
1b, first storage means 23a and 23b for dark output signals, and second storage means for storing pixel outputs whose light shields are provided on the photosensitive portions.
Storage means 24a, 24b, first dark time output signal generation means 25a, 25 for generating a dark time output signal generated in each pixel
b, second dark-time output signal generating means 26a, 26 for generating the dark-time output signal output from the photoelectric conversion elements 5a, 5b
b, 27a, 27b, subtraction means 28a, 28b for subtracting this output from the output at the time of image reading, and rearrangement for outputting two data strings as one serial data string arranged in order from the end data of the document image. Means 32 and are provided.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image reading device using a CCD line sensor, which is capable of high-speed and high-definition reading.

[0002]

2. Description of the Related Art Recently, a scanner capable of high-speed and high-definition reading has been desired. Also, due to advances in microfabrication technology, CCD line sensors capable of high-speed operation and high-density CC
D-line sensors are emerging. An image reading device using a CCD line sensor is, for example, Japanese Patent Publication 1-53.
As disclosed in Japanese Patent No. 538, a method has been developed in which a document is divided into a plurality of regions, and a plurality of line sensors corresponding to the respective regions are arranged in the main scanning direction to read a document image. Is trying to increase the reading speed and reading density.

[0003]

However, in the image reading apparatus using one CCD line sensor which is a conventional technique, it cannot be said that the image reading apparatus is high in speed and high in definition, and it is forcibly read in high speed and in high definition. In order to achieve the above, it is required that the CCD drive circuit and the output signal processing circuit can operate at high speed, which is not only disadvantageous in terms of cost, but also the accuracy of the pixel information to be read may be deteriorated. Further, in the conventional example, when reading an original smaller than the maximum readable original size, all pixel information including unnecessary pixel information among the pixel information of the CCD line sensor is read,
The reading efficiency is poor because it is being processed.

[0004]

According to a first aspect of the present invention, an illumination light source for illuminating a document, an image forming unit for forming an image of reflected light from the document, and photoelectrically converting the image-formed reflected light. In an image reading apparatus including two photoelectric conversion elements, a driving unit that drives these photoelectric conversion elements, and a signal processing unit that processes the output signal of the photoelectric conversion elements,
A light-shielding plate arranged in front of the photoelectric conversion element so as to shield light reflected from other than the original, and the original position corresponding to the pixel read out first of the photoelectric conversion element is the original area. At the center of the main scanning direction, and the original positions corresponding to the pixels read subsequently are opposite to each other, and only the pixels corresponding to the original area are read out and the photosensitive portion of the dark output signal over one line is read. First holding means for holding the output portion of the pixel only in the transfer portion, and second holding means for holding the output portion of the pixel in which the light shield is provided on the photosensitive portion, and the first and second holding means. 1 based on the output value of
Normalizing means for normalizing the dark output signal across the line,
First storage means for storing the normalized dark-time output signal over one line, second storage means for storing the output of a pixel having a light shield on the photosensitive portion, and output of the first storage means And the output of the second storage means to generate a dark output signal generated in each pixel, and an actual dark output signal generating means and an actual output based on the output of the first dark output signal generating means. Second dark output signal generating means for generating a dark output signal output from the photoelectric conversion element, and subtracting means for subtracting the output of the second dark output signal generating means from the output when the image is read. In the signal processing means, rearranging means for outputting two data strings output in parallel in opposite directions from the central portion of the document image as one serial data string arranged in order from the end data of the document image. Set up in.

According to a second aspect of the invention, an illumination light source for illuminating the original, an image forming means for forming an image of reflected light from the original, and two photoelectric conversion elements for photoelectrically converting the formed reflected light. In an image reading apparatus including a driving unit that drives these photoelectric conversion elements and a signal processing unit that processes the output signal of the photoelectric conversion elements, the photoelectric conversion is performed so as to shield reflected light from other than the original. A light-shielding plate disposed in front of the element, and a document position corresponding to a read pixel read at the beginning of the photoelectric conversion element is substantially centered in the main scanning direction of the document area and is read subsequently. The positions of the originals corresponding to the read pixels are opposite to each other, and only the pixels corresponding to the original area are read out and a storage unit for storing a dark output signal over one line and a light shield on the photosensitive section are provided. A first holding means for holding the output of the pixel, and a first output for multiplying the output of the holding means and the output of the storage means read out over one line to generate a dark output signal generated in each pixel. The dark output generating means, the second dark output generating means for actually generating the dark output signal output from the photoelectric conversion element based on the output of the first dark output generating means, and the second dark output generating means. Subtracting means for subtracting the output of the dark output generating means from the output when the image is read, and two data strings output in parallel in opposite directions from the central portion of the original image are arranged in order from the data at the end of the original image. Reordering means for outputting as a serial data string arranged side by side is provided in the signal processing means.

According to the third aspect of the invention, an illumination light source for illuminating the original, an image forming means for forming an image of reflected light from the original, and two photoelectric conversion elements for photoelectrically converting the formed reflected light. In an image reading apparatus provided with a driving means for driving these photoelectric conversion elements and a signal processing means for processing the output signal of the photoelectric conversion elements, the photoelectric conversion is performed so as to shield light reflected from other than the original. A light-shielding plate arranged in front of the element, and the original position corresponding to the read pixel read at the beginning of the photoelectric conversion element is substantially at the center in the main scanning direction of the original area and is read subsequently. Holding means for reading out only the pixels corresponding to the original area and holding the output of the pixels of only the transfer section without the photosensitive section; Peak hold means for holding the peak value of the output signal when the white reference plate is read over the in, and white reference plate read signal and image read signal for one line based on the output values of the hold means and the peak hold means. With a first normalizing means for normalizing, a storage means for storing the normalized white reference plate read signal which is the output of the first normalizing means over one line, and an output of the first normalizing means. Second normalizing means for normalizing a certain normalized image read signal by the output of reading the contents of the storage means over one line, and two normalizing means output in parallel in opposite directions from the central portion of the original image. Arrangement means for outputting a data string as one serial data string arranged in order from the end data of the original image is provided in the signal processing means.

According to a fourth aspect of the invention, an illumination light source for illuminating a document, an image forming means for forming an image of reflected light from the document, and two photoelectric conversion elements for photoelectrically converting the image-formed reflected light. In an image reading apparatus including a driving unit that drives these photoelectric conversion elements and a signal processing unit that processes the output signal of the photoelectric conversion elements, the photoelectric conversion is performed so as to shield reflected light from other than the original. An original position corresponding to a read pixel read at the beginning of the photoelectric conversion element, which includes a light-shielding plate disposed in front of the element, an original size detecting unit, an illumination light source control unit, and a sub-scanning speed control unit. Is approximately the center of the original area in the main scanning direction, and the directions of the original positions corresponding to the read pixels read subsequently are opposite to each other, and the original area corresponds to the original area depending on the size of the original. First holding means for reading out only the read pixels, controlling the intensity of the illumination light source and the sub-scanning speed, and holding the output portion of the pixels of only the transfer portion without the photosensitive portion in the dark output signal over one line; The second that holds the output part of the pixel where the light shield is provided on the photosensitive part
A holding means, a normalizing means for normalizing the dark output signal over one line with reference to the output values of the first and second holding means, and a dark time over one line normalized by the normalizing means. First storage means for storing the output signal, second storage means for storing the output of the pixel whose light shield is provided on the photosensitive portion, and output of the second storage means and one line of the first storage means. First dark output signal generating means for generating a dark output signal generated in each pixel by multiplying the read output, and the photoelectric conversion element actually based on the output of the first dark output signal generating means. Second dark output signal generation means for generating a dark output signal output from the second dark output signal, subtraction means for subtracting the output of the second dark output signal generation means from the read output, and a central portion of the original image. Output in parallel in opposite directions The two data rows from the data of the end of the document image and a sorting means for outputting as a single serial data ordered sequence provided in said signal processing means.

According to a fifth aspect of the invention, an illumination light source for illuminating a document, an image forming means for forming an image of reflected light from the document, and two photoelectric conversion elements for photoelectrically converting the image-formed reflected light. In an image reading apparatus including a driving unit that drives these photoelectric conversion elements and a signal processing unit that processes the output signal of the photoelectric conversion elements, the photoelectric conversion is performed so as to shield reflected light from other than the original. An original position corresponding to a read pixel read at the beginning of the photoelectric conversion element, which includes a light-shielding plate disposed in front of the element, an original size detecting unit, an illumination light source control unit, and a sub-scanning speed control unit. At the center of the original area in the main scanning direction, and the directions of the original positions corresponding to the read pixels read subsequently are opposite to each other. A storage unit that reads out only the read pixels, controls the intensity of the illumination light source and the sub-scanning speed, and stores the output signal in the dark at one line, and a hold that holds the output of the pixels whose light shield is provided on the photosensitive unit. Means, the output of the hold means and the contents of the storage means
First dark output signal generating means for generating a dark output signal generated in each pixel by multiplying the output read across the line, and the actual output based on the output of the first dark output signal generating means. Second dark output signal generating means for generating a dark output signal output from the photoelectric conversion element, subtracting means for subtracting the output of the second dark output signal generating means from the output when the image is read, and the original document Arrangement means is provided in the signal processing means for outputting two data strings output in parallel in opposite directions from the central portion of the image as one serial data string arranged in order from the end data of the document image. It was

According to a sixth aspect of the present invention, an illumination light source for illuminating an original, an image forming unit for forming an image of reflected light from the original, and two photoelectric conversion elements for photoelectrically converting the formed reflected light. In an image reading apparatus provided with a driving means for driving these photoelectric conversion elements and a signal processing means for processing the output signal of the photoelectric conversion elements, the photoelectric conversion is performed so as to shield light reflected from other than the original. An original position corresponding to a read pixel read at the beginning of the photoelectric conversion element, which includes a light-shielding plate disposed in front of the element, an original size detecting unit, an illumination light source control unit, and a sub-scanning speed control unit. Is substantially the center of the original area in the main scanning direction, and the directions of the original positions corresponding to the read pixels read subsequently are opposite to each other, and correspond to the original area depending on the size of the original. Hold means for reading out only the read pixels, controlling the intensity of the illumination light source and the sub-scanning speed, and holding the output of the pixels only in the transfer section without the photosensitive section, and the output when the white reference plate is read over one line Peak holding means for holding the peak value of the signal, and 1 based on the output values of the holding means and the peak holding means.
First normalizing means for normalizing the white reference plate read signal and the image read signal over the line, and a memory for storing the normalized white reference plate read signal which is the output of the first normalizing means over the one line. Second means for normalizing the normalized image reading signal output from the first normalizing means by the output of reading the contents of the storage means over one line.
The normalizing means and the rearranging means for outputting two data strings output in parallel in opposite directions from the central portion of the document image as one serial data string arranged in order from the end data of the document image. It is provided in the signal processing means.

In the invention described in claim 7, claims 1, 3,
In the inventions described in Nos. 4 and 6, the accumulation time at the time of reading out the dark output signal over one line from the two photoelectric conversion elements is
It is set to be longer than the storage time per line when reading an image.

[0011]

According to the first aspect of the present invention, the two photoelectric conversion elements are independently and simultaneously driven by the driving means, and the light shielding plate shields light reflected from other than the original and is output from the photoelectric conversion element. The two signals are output from almost the center of the original image, and the output directions are opposite to each other, and only the pixels corresponding to the original image area are output, and these two output signals are stored in advance for dark output correction. A dark line output pattern for one line, which is normalized by the output signals of the pixel having no light receiving portion of the photoelectric conversion element and the pixel which is optically shielded, and the average or typical output value of the pixel of the light shield portion Multiply to generate the dark output generated in each pixel, generate the dark output actually output from the photoelectric conversion element from the dark output of each pixel, subtract that value from the output value of the photoelectric conversion element, and ,output Over data are sorted as single serial data string arranged from the data of the end of the original image in order. As a result, one line can be read at twice the speed when the driving frequency of the photoelectric conversion element is the same and the reading density is the same as in the case where only one photoelectric conversion element is used. Further, when the driving frequency of the photoelectric conversion element is the same and the reading speed of one line is the same, it is possible to read at a double reading density, and there is a dark output remaining on the preceding main scanning line. Even in this case, it is possible to accurately correct the dark output. Furthermore, since the output data is output as one serial data lined up in order from the data at the edge of the document image, it is possible to obtain the output data in the same output format as when using one photoelectric conversion element.

According to the second aspect of the present invention, the two photoelectric conversion elements are independently and simultaneously driven by the driving means, the light shielding plate shields the reflected light from other than the original, and is output from the photoelectric conversion element. The two signals are output from almost the center of the original image, and the output directions are opposite to each other, and only the pixels corresponding to the original image area are output, and the output signals are the output value output from the storage means and the light shield portion. The average or typical output value of each pixel is multiplied to generate a dark output generated in each pixel, and a dark output actually output from the photoelectric conversion element is generated from the dark output of each pixel. The value is subtracted from the output value of the photoelectric conversion element, and the output data is rearranged as one serial data string arranged in order from the data at the edge of the document image. As a result, in the case where the driving frequency of the photoelectric conversion element is the same and the reading density is the same as in the case where one photoelectric conversion element is used, the speed is doubled at one time.
The line can be read. When the driving frequency of the photoelectric conversion element is the same and the reading speed of one line is the same, it is possible to read at a double reading density, and when there is a dark output remaining on the preceding main scanning line. However, it is possible to accurately correct the dark output. Further, since the output data is output as one serial data lined up in order from the end data of the original image, the output data can be obtained in the same output format as when one photoelectric conversion element is used.

According to another aspect of the present invention, the two photoelectric conversion elements are independently and simultaneously driven by the driving means, and the light shielding plate shields light reflected from other than the original and is output from the photoelectric conversion element. The two signals are output from almost the center of the original image, and the output directions are opposite, and only the pixels corresponding to the original image area are output. The output signals are subjected to dark output correction, then the white reference plate read signal The image reading signal is normalized by the output from the storage unit that stores the data, and the output data is rearranged as one serial data string arranged in order from the data at the edge of the document image. Thus, when one photoelectric conversion element has the same driving frequency and the same reading density, one line can be read at twice the speed as compared with the case where one photoelectric conversion element is used. Further, when the driving frequency of the photoelectric conversion element is the same and the reading speed of one line is the same, the reading can be performed with double the reading density, and there is a dark output remaining on the preceding main scanning line. However, it is possible to accurately perform output correction in the dark. Further, even if there is variation in sensitivity between the pixels of the photoelectric conversion element or variation in the intensity of the illumination light source, the image can be read accurately, and the output data is the original image. Since it is output as one serial data lined up in order from the end data, it is possible to obtain output data in the same output format as when using one photoelectric conversion element.

According to the present invention, the two photoelectric conversion elements are independently and simultaneously driven by the driving means, and the light shielding plate is a light shielding plate for shielding the reflected light from other than the original depending on the size of the original. Moving and lighting light source control means,
The sub-scanning speed control means controls the illumination light source and the sub-scanning speed according to the size of the original, and the two signals output from the two photoelectric conversion elements are output from substantially the center of the original image and the output direction is opposite. Therefore, only the pixels corresponding to the original image area are output, and the output signals are output from the pixels which are stored in advance for dark output correction and have no light receiving portion of the photoelectric conversion element and the pixels which are optically shielded. The dark output of each pixel is generated by multiplying the one-line dark output pattern normalized by the signal by the average or typical output value of the pixel of the light shield portion, and the dark output of each pixel is generated. From the photoelectric conversion element, the dark output is actually generated, and the generated value is subtracted from the output value of the photoelectric conversion element. Further, the output data is one data arranged in order from the end data of the document image. It is sorted as the column data column.
As a result, one line can be read at double speed when the driving frequency of the photoelectric conversion element is the same and the reading density is the same as in the case of using one photoelectric conversion element. Drive frequency is the same and 1
If the line read densities are the same, it is possible to read at twice the read density. Further, when reading a small original, higher speed reading can be performed, and even when there is a dark output remaining on the previous main scanning line, the dark output can be accurately corrected. Furthermore, since the output data is output as one serial data string arranged in order from the end data of the document image, the output data can be obtained in the same output format as when one photoelectric conversion element is used.

According to the present invention, the two photoelectric conversion elements are independently and simultaneously driven by the driving means, and the light shielding plate is a light shielding plate for shielding the reflected light from other than the original depending on the size of the original. Moving and lighting light source control means,
The sub-scanning speed control means controls the illumination light source and the sub-scanning speed according to the size of the original, and the two signals output from the two photoelectric conversion elements are output from substantially the center of the original image and the output direction is opposite. Then, only the pixels corresponding to the original image area are output, and the output signal thereof is multiplied by the dark output value output from the storage means and the average or typical output value of the pixels of the light shield portion, and each pixel is output. The dark output generated in step 1 is generated, the dark output actually output from the photoelectric conversion element is generated from the dark output of each pixel, and the output value is subtracted from the output value of the photoelectric conversion element. Further, the output data is rearranged as one serial data string arranged in order from the end data of the document image. As a result, as compared with the case where one photoelectric conversion element is used, when the driving frequency of the photoelectric conversion element is the same and the reading density is the same, it becomes possible to read one line at twice the speed, and the driving frequency of the photoelectric conversion element is In the same case, when the reading speed of one line is the same, it is possible to read at a double reading density, and it is possible to read at a higher speed when reading a small original. Further, even when there is a dark output remaining on the preceding main scanning line, the dark output can be accurately corrected. Further, since the output data is output as one serial data string arranged in order from the end data of the document image, the output data can be obtained in the same output format as when one photoelectric conversion element is used.

In the sixth aspect of the present invention, the two photoelectric conversion elements are independently and simultaneously driven by the driving means, and the light blocking plate is a light blocking plate for blocking the reflected light from other than the document depending on the size of the document. Moving and lighting light source control means,
The sub-scanning speed control means controls the illumination light source and the sub-scanning speed according to the size of the original, and the two signals output from the two photoelectric conversion elements are output from substantially the center of the original image and the output direction is opposite. Then, only the pixels corresponding to the original image area are output, and the output signal is corrected for the output at dark, and then the image reading signal is normalized by the output from the storage unit that stores the white reference plate reading signal. Further, the output data is rearranged as one serial data string arranged in order from the data at the edge of the document image. As a result, as compared with the case where one photoelectric conversion element is used, it becomes possible to read one line at twice the speed when the driving frequency of the photoelectric conversion element is the same and the reading density is the same, and the driving frequency of the photoelectric conversion element is increased. In the same case, if the reading speed of one line is the same, it is possible to read at a double reading density, and further, when reading a small original, it is possible to perform higher speed reading. Further, even when there is a dark output remaining on the preceding main scanning line, the dark output can be accurately corrected, and when there is a sensitivity variation between each pixel of the photoelectric conversion element and the illumination. Even if there is a variation in the intensity of the light source, the image can be read accurately. Further, since the output data is output as one serial data string arranged from the data at the end of the original image, the output data can be obtained in the same output format as when one photoelectric conversion element is used.

According to the seventh aspect of the invention, the accumulation time for reading the dark output signal over one line from the photoelectric conversion element is made longer than the accumulation time for one line for image reading, so that each pixel has a longer accumulation time. Since the dark output is integrated, white noise when capturing the dark output can be reduced, and since the dark output can be captured at a high level, deterioration during signal processing can be reduced. Therefore, it is possible to correct the dark output with high accuracy.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the invention described in claims 1, 3, 4, and 6 will be described with reference to FIGS. In this example,
High-speed and high-definition reading is possible by using two photoelectric conversion elements, and the output format is the same as that of one photoelectric conversion element, and the number of pixels read is smaller than the number of pixels of the photoelectric conversion elements. The present invention provides an image reading device capable of accurately correcting the output during dark even when performing the above.

First, the overall structure of the image reading apparatus will be described with reference to FIG. A document 2 is placed on the contact glass 1, and the document 2 is linearly illuminated by an illumination light source 3. The reflected light from the center of the original 2 to the left side from the center of the original 2 is imaged on the CCD line sensor 5a as the photoelectric conversion element via the imaging lens 4a as the left imaging means, and the reflected light from the center of the original 2 to the right is An image is formed on a CCD line sensor 5b as a photoelectric conversion element via an image forming lens 4b as the right image forming means. In this case, the output portions of the two CCD line sensors 5a and 5b are arranged outward so that the image signal of the central portion of the linearly illuminated image is first output. Further, the light shielding plates 6a and 6b shield the reflected light reflected from other than the reading area. In addition, this FIG.
In addition to the optical system shown in FIG. 2, a method using a mirror, a half mirror, or a roof mirror may be used as long as it is an optical system capable of dividing an original image into two in the main scanning direction. .

Next, the states of the photoelectric conversion system, drive system, and signal processing system inside the circuit of FIG. 2 described above will be described with reference to FIG. The CCD line sensors 5a and 5b include photodiode arrays 7a and 7b and CCD shift registers 8a and 8b.
b, and output units 9a and 9b. A CCD drive unit 10 as a drive unit is connected to the CCD line sensors 5a and 5b. The CCD driving unit 10 includes a document size detecting unit 11 as a document size detecting unit.
Are connected. A sub-scanning speed control unit 12 as a sub-scanning speed control unit, an illumination light source control unit 13 as an illumination light source control unit, and a signal processing unit 14 as a signal processing unit are connected to the document size detection unit 11. There is. Further, the light blocking plates 6 a and 6 b are controlled by the light blocking plate control unit 15. In the present embodiment, the internal configuration of the signal processing unit 14 is characterized, but a detailed description of this internal configuration will be given later.

In such a configuration, the CCD line sensors 5a and 5b photoelectrically convert the incident light by the photodiode arrays 7a and 7b, and receive the shift pulse from the CCD driving unit 10 for the accumulated signal charges to receive the shift pulses in parallel. Send to 8a, 8b. Next, the CCD shift registers 8a and 8b receive the transfer pulses of the CCD driving unit 10 for the respective pixel signals and sequentially output the signals to the output units 9a and 9b, and then the signals for one pixel, and then the reset pulse RS of the CCD driving unit 10. The charge is cleared by the number of pixels for one sensor line, and the pixel signals for one sensor line can be read. The CCD line sensors 5a and 5b are independently and simultaneously driven by the CCD driving unit 10, and the output signal is the signal processing unit 14.
With cyle hold, A / D conversion, dark output correction,
Two data strings output in parallel in opposite directions from the central portion of the original image are output as one serial data string arranged in order from the edge data of the original image. Rearrangement, shading correction, and other processing are performed. . Note that a detailed description of the operation processing of the signal processing unit 14 will be given later. Then, the illumination light source 3 and the CCD line sensors 5a, 5
b, the imaging lenses 4a and 4b are integrally installed on a carriage (not shown), and when the reading of one line is completed, it moves to the next reading position. FIG. 4 shows the shift pulse SH and the clock pulse φ output from the CCD driver 10.
1, φ2, and the timing of the reset pulse RS.

Here, when reading the original 2 smaller than the maximum readable original size, the light-shielding plates 6a and 6b shield the portions where the reflected light having the image information from the original 2 does not enter,
The pixels from the first to mth read pixels having the original image information are read out, and the pixels of the original 2 having no image information are not read out. For this reason, the document size detection unit 11 automatically detects the document size (main scanning direction X) or detects the document size (main scanning direction) designated by a person, and from that size, the CCD line sensor 5 is detected.
The required number of pixels m of a and 5b is set. Information on the set number of pixels is sent to the CCD driving unit 10, the signal processing unit 14, the illumination light source control unit 13, the sub-scanning speed control unit 12, and the light shielding plate control unit 15. The light-shielding plate controller 15 outputs m and outputs φ1 and φ2 so that only m pixels of the CCD line sensors 5a and 5b are read out according to the sent pixel number information, and then outputs the shift pulse SH. The signal processing unit 14 performs signal processing assuming that the number of pixels in one line is 2 m. The sub-scanning speed controller 12 determines the sub-scanning speed according to the sent information on the number of pixels, and performs sub-scanning. In this case, since the reading time for one line is m / n times, the sub-scanning speed is increased by n / m as compared with the case where all pixels n are read. The illumination light source control unit 13 determines the intensity of the illumination light source 3 according to the sent information on the number of pixels, and drives the illumination light source 3. In this case, since the reading time for one line is m / n times, the exposure amount of the CCD line sensor is also reduced to m / n times as it is. Therefore, the illumination light source 3 is controlled so that the illuminance is n / m times as compared with the case of reading all the pixels n.

As described above, the CCD line sensor 5
By using two a and 5b independently and simultaneously driving, compared with the case where one CCD line sensor 5a and 5b is used conventionally, CCD line sensors 5a and 5b are used.
When the driving frequency is the same and the reading density is the same, one line can be read at twice the speed. When the CCD run sensors 5a and 5b have the same driving frequency and the reading speed of one line is the same, the reading density can be doubled, and the data output from the central portion of the original image can be used as the data output from the central portion of the original image. By rearranging the data in the order from, it is possible to obtain output data in the same output format as when using one CCD line sensor as in the related art. Further, since the pixels which do not contribute to the reading are not read by using the light shielding plates 6a and 6b, the reading speed can be increased when reading a small document. The power consumption for driving the elements can be reduced by multiplying the clock speed of the CCD line sensors 5a and 5b by m / n without changing the brightness of the illumination or the sub-scanning speed. Also, the brightness of the illumination, the sub-scanning speed, the CCD line sensor 5
The total data amount can be reduced by not switching the clock speeds of a and 5b at all and reading only the pixels related to the image reading, and not reading the unrelated pixel portions without sending the clock.

Next, in the present embodiment, the signal processing unit 1
Although the internal processing operation of No. 4 is characterized, the reason why the signal processing unit 14 has the characteristics will be described before entering the description. In the image reading apparatus as described above, C
The outputs of the CD line sensors 5a and 5b appear in the form of the sum of the output corresponding to the exposure amount and the dark output which occurs regardless of the exposure amount. Further, by reading out only the pixels involved in image reading among the pixels of the CCD line sensors 5a and 5b, the photoelectric conversion output corresponding to the exposure amount includes the dark output generated at that pixel and the previous line. Appears in the form of the sum with the dark output generated in the non-read pixels during reading. FIG. 5 shows how dark outputs overlap when 25% of the pixels of the CCD line sensors 5a and 5b are read. CCD line sensors 5a and 5b having CCD shift registers 8a and 8b for n pixels in the case where there is no change in the output in the dark during one reading period of the original image.
When only m pixels are read from, the j-th pixel output dout (i, j) on the i-th line is expressed by the following equation.

Dout (i, j) = dmo (i, j) + Σ {dbk (j + m (h-1))} (1) where dmo (i, j): i-th line, j-th image Signal dbk (k): dark output generated at the k-th pixel of the sensor, when k> n, dbk (k) = 0 Therefore, in order to accurately convert the information of the original 2 into an electric signal. , Some correction for reducing the output in the dark for the output signals of the CCD line sensors 5a, 5b, and C
Sensitivity variations between the pixels of the CD line sensors 5a and 5b,
Shading correction for correcting non-uniformity of the intensity distribution of the illumination light source 3 is required. Further, in this embodiment, it is necessary to rearrange the output data together with the dark output correction and the shading correction. For this reason, the signal processing unit 14 has a configuration described below.

Therefore, the internal structure of the signal processing unit 14 will now be described with reference to FIG. The image reading apparatus as shown in FIGS. 2 and 3 described above has a circuit configuration as described below. That is, this device has
S / H (sample and hold) circuits 15a and 15b, and a first hold for reading out only the pixels corresponding to the original area and holding the output portion of the pixels of the transfer portion only in the dark output signal over one line S / H circuits 16a and 16b as means, S / H circuits 17a and 17b as second holding means for holding an output portion of a pixel whose light shield is provided on the photosensitive portion, and peak hold circuits 18a and 18b. Amplifier circuits 19a and 19b, switch circuits 20a and 20b, and S / H circuits 16a and
A as a normalizing means for normalizing the dark output signal over one line with the output values of 16b, 17a and 17b as a reference.
/ D conversion circuits 21a and 21b and switch circuits 22a and
22b and a dark output memory 23 as a first storage means for storing the normalized dark output signal over one line.
a and 23b, the addition and averaging circuits 24a and 24b as the second storage means for storing the outputs of the pixels whose light shields are provided on the photosensitive portions, and the dark output memories 23a and 23b.
Of the above-mentioned addition and averaging circuits 24a and 24b are multiplied to generate a dark-time output signal generated in each pixel. Multiplication circuits 25a and 25b as first dark-time output signal generating means.
And adder circuits 26a and 26b as second dark time output signal generating means for actually generating dark time output signals output from the CCD line sensors 5a and 5b based on the outputs of the multiplying circuits 25a and 25b, and memories. 27a, 27b, subtraction circuits 28a, 28b as subtraction means for subtracting the outputs of the addition circuits 26a, 26b and the memories 27a, 27b from the output when the image is read, and the switch circuits 29a, 2
9b, the white reference memories 30a and 30b and the divider 31a connected to the subsequent stages of the switch circuits 29a and 29b.
31b and a rearrangement circuit as rearrangement means for outputting two data strings output in parallel in opposite directions from the central portion of the original image as one serial data string arranged in order from the end data of the original image. And 32 are provided.

The operation processing of the signal processing unit 14 of FIG. 1 having such a configuration will be described. One line of the outputs of the CCD line sensors 5a and 5b is shown in FIG. 6 (a), for example.
The waveform is as shown in (b). DS of FIG. 6 (a)
(E) is a pixel output only for the transfer section without a photosensitive section, DS
(S) shows the output of the pixel having the light-shielded photosensitive portion, followed by the read output of the original 2. The waveforms of these one line are guided to the S / Hs 15a and 15b, the waveforms of the transfer clocks of the CCD line sensors 5a and 5b are removed, and the waveform of FIG.
The waveform is as shown in (b). In this embodiment, the write address and read address of each memory are reset before all signal processing. Next, a dark output for one line is generated and stored. Therefore, the switch circuits 20a, 20
b is switched to the side opposite to the side shown in FIG. 1, and the S / H circuit 15
The outputs of a and 15b are S / H circuits 16a and 16b, S / H.
Circuits 17a and 17b and A / D conversion circuits 21a and 21b
Be led to. CCD line sensor 5 over 1 line
When the exposure amounts of a and 5b are set to zero, the output for one line becomes a waveform as shown in FIG. 7A through the S / H circuits 15a and 15b. That is, DS (S) and the reading area all have the output level in the dark, and the values vary from pixel to pixel. This is led to S / H circuits 16a and 16b, and DS
The value of (E) is sampled and held, and the A / D conversion circuit 21
It is set to the upper limit reference VrefH of a and 21b. Also, S
/ H circuits 17a and 17b, sample and hold the average value or representative value of DS (S), and amplify circuits 19a and 19b.
It is amplified by b and set as the lower limit reference value VrefL of the A / D conversion circuits 21a and 21b. That is, the A / D conversion 21
The relationship between the references VrefH and VrefL of a and 21b and the signal input (S / H circuits 17a and 17b) is as shown in FIG. As a result, A / D conversion is performed, and the outputs are output to the dark output memories 23a and 23b with the switching circuits 22a and 22b at the lower switching positions (having a memory area for one line or more of the CCD line sensors 5a and 5b). One line (only the number n of pixels possessed by the CCD line sensor) is stored. As a result, the acquisition of the dark-time output pattern for one line based on the average value or the representative value of the pixels DS (S) of the light-shielded photosensitive section is completed. This dark output pattern is: dbk (j) / dbk (2) where dbk (j): dark output signal of the j-th pixel during dark output reading dbk: DS (S during dark output reading) ) Is a representative value or an average value, and is a value indicating how many times the dark output level of each pixel on each reading line is higher than the output of a pixel having a light-shielded photosensitive section. After that, the output memories 23a and 23b in the dark
The write address and read address of are reset.

Next, the white reference of one line is taken in. At the time of capturing the white reference, all pixel signals of the CCD line sensors 5a and 5b are output at the time of capturing the dark output pattern, so that there is no residual dark output. When the switch circuits 20a and 20b for capturing the white reference are set to the switching positions shown in FIG. 1, the outputs of the S / H circuits 15a and 15b are input to the peak hold circuits 18a and 18b, and when the white reference is first read, one line The peak value of is held and used as the lower limit value VrefL of the A / D conversion circuits 21a and 21b. The next time the white reference is read, the S / H circuit 15
The outputs of a and 15b are led to S / H circuits 16a and 16b, and D
A / D conversion circuit 2 samples and holds with the value of S (E)
The upper limit reference value VrefH of 1a and 21b is set. Thereby, the reference V of the A / D conversion circuits 21a and 21b
The relationship between refH, VrefL and signal input (= output of S / H 15a, 15b) is as shown in FIG. 7 (b). A / D conversion is performed with such settings, and the output is dwo (j) / (VrefH−VrefL) + dbk (j) / (VrefH−VrefL) = {dwo (j) + dbk (j)} / (VrefH− VrefL) (3) However, dwo: photoelectric conversion output when the white reference is read. This is the position where the switching circuits 22a and 22b are shown in FIG. 1, and the addition and averaging circuits 24a and 24b and the subtraction circuit 28a,
You are led to 28b. The averaging circuits 24a and 24b have a function of sequentially adding input pixel outputs to obtain an average value when the input control signal is ON, and hold the value as it is after obtaining the average value. To do. In this case, DS (S)
A control signal that is turned on when the value is is input. Since the photoelectric conversion output dwo (j) during this period is zero, the output of the averaging circuits 24a and 24b becomes the average value of DS (S) or the representative value dbk, and this value is obtained by the following white reference reading. It is input to the multiplication circuits 25a and 25b. The other input of each of the multiplication circuits 25a and 25b is an output (equation (2)) from the dark output memories 23a and 23b. As a result, the output of the multiplication circuits 25a and 25b is {dbk / (VrefH-VrefL )} · {Dbk (j) / dbk} = dbk (j) / (VrefH−VrefL) (4) This output is input to the adder circuits 26a and 26b (having a memory area for one line or more of the CCD line sensors 5a and 5b), and the sum of the output and the outputs of the memories 27a and 27b is output from the adder circuits 26a and 26b. However, in this case, the memories 27a and 27b are reset first and the recorded data are all zero. Therefore, the adder circuits 26a and 26b are connected to the multiplier circuit 2
The outputs of 5a and 25b (equation (4)) are output as they are. This output is led to the negative side inputs (-) of the subtraction circuits 28a and 28b. On the other hand, the equation (3) is input to the positive side inputs (+) of the subtraction circuits 28a and 28b. As a result, the outputs of the subtraction circuits 28a and 28b are dwo (j) / (VrefH-VrefL) (5) according to the equations (3)-(4). Each calculation of the white reference reading is performed by the number of read pixels (m pixels by each sensor). This output (Equation (5))
The switching circuits 29a and 29b are recorded in the white reference memories 30a and 30b as switching positions opposite to those shown in the figure. This completes the capture of the white reference for one line (m pixels in each sensor). After that, the write address and the read address of the dark output memories 23a and 23b, the memories 27a and 27b, and the white reference memories 30a and 30b are reset.

Next, after reading the white reference for one line and before inputting an image, the CCD line sensors 5a, 5
All pixels of b are output. As a result, immediately before image input, all the pixels of the CCD line sensors 5a and 5b have no residual dark output. Image input is A
The steps up to the output of the / D conversion circuits 21a and 21b are performed in the same manner as the white reference fetch. That is, the relationship between the two references VrefH and VrefL of the A / D conversion circuits 21a and 21b and the input is shown in FIG.

The difference between the white reference input and the image input will be described below. The control signal input to the averaging circuits 24a and 24b is turned on only when the value of DS (S) at the input of the first line is reached. Therefore, the averaging circuits 24a, 24
The value held and output by b during image reading is the average value of DS (S) dbk / (VrefH-VrefL) during reading of the first line. Next, the dark output memories 23a and 23b sequentially output signals from the dark output pattern (Equation (2)) from the 1st to the nth, but the number of pixels possessed by one CCD line sensor 5a, 5b (n pixels) ) Only outputs the signal and then outputs zero. Therefore, the outputs of the multiplication circuits 25a and 25b are n
Equation (2) is used until a signal for pixels is output, and n + 1
It is zero after the pixel. On the other hand, the memories 27a, 2
As the 7b read operation, the read address is reset after the first to mth signals, which are the number of read pixels of one CCD line sensor, are output. As a write operation, the outputs of the adder circuits 26a and 26b are sequentially written from the first to the m-th, and after the m-th is written, the write address is reset. Therefore, Σ {dbk (j + m (h-1))} (6) is introduced to the negative side inputs of the subtraction circuits 28a and 28b. This is the dark output of the j-th pixel when reading the i-th line. On the other hand, on the positive side inputs of the subtraction circuits 28a and 28b, dmo (i, j) / (VrefH−VrefL) + Σ {dbk (j + m (h−1))} / (VrefH−VrefL) = [dmo (i, j) + Σ {dbk (j + m (h-1))}] / (VrefH−VrefL) (7) is derived. Therefore, the outputs of the subtraction circuits 28a and 28b are dmo (i, j) / (VrefH-VrefL) (8), and only the image signal from which the dark output has been removed can be taken out. Next, the output is switched to the switch circuit 29a,
29b is the switching position shown in FIG. 1, and the division circuits 31a,
31b, and white reference memories 30a and 30 for each line
Division is performed by the output that reads b. As a result, the outputs of the division circuits 31a and 31b are calculated by the equation ((8)) / ((5)
From the equation, dmo (j) / dwo (j) is obtained, and the data subjected to the dark output correction and the shading correction are output. The two data output in parallel from the center of the original image in the opposite direction are input to the rearrangement circuit 32, and output as one serial data string arranged in order from the data at the end of the original 2. That is, the data sequence before the rearrangement is shown in FIGS.
Although the two parallel data strings are arranged as shown in FIG. 8, they are converted and output in the rearrangement circuit 32 as shown in FIG. 8C.

As described above, in this embodiment, 1
Quantization without being affected by the dark output even when the dark output before the previous line reading remains when the dark output and the light source intensity do not fluctuate during one reading period.
Shading correction can be performed, and by rearranging, output data can be obtained in the same output format as when one CCD line sensor is used.

Next, a second embodiment of the invention described in claims 1, 3, 4, and 6 will be described with reference to FIG. In this embodiment, high-speed and high-definition reading is possible by using two photoelectric conversion elements, and the output format is the same as that of one photoelectric conversion element. Provided is an image reading device capable of accurately correcting dark output even when reading is performed with a small number of pixels.

That is, in this embodiment, the signal processing unit 14
However, only the signal processing unit 14 will be described here. The description of the same parts as those in the first embodiment is omitted, and the same parts are designated by the same reference numerals.

First, the main components of the signal processing unit 14 in this embodiment will be described with reference to FIG. In the signal processing unit 14, only the pixels corresponding to the original area are read out, and the dark output memories 23a and 23b serving as a storage unit for storing the dark output signal over one line, and the pixels having the light shield on the photosensitive unit. The averaging circuits 24a and 24b as holding means for holding the output are multiplied by the outputs of the averaging circuits 24a and 24b and the outputs of the dark output memories 23a and 23b read over one line. Multiplier circuits 25a and 25b as first dark output generating means for generating a dark output signal generated in each pixel.
And an adder circuit 26 as a second dark time output generating means for actually generating a dark time output signal output from the photoelectric conversion elements 5a and 5b based on the outputs of the multiplication circuits 25a and 25b.
a, 26b and memories 27a, 27b, subtraction circuits 28a, 28b as subtraction means for subtracting the outputs of the addition circuits 26a, 26b and memories 27a, 27b from the output when the image is read, and the central portion of the original image. A rearrangement circuit 32 as a rearrangement circuit that outputs two data strings output in parallel in opposite directions in parallel as serial data strings arranged in order from the end data of the original image, as shown in FIG. It is provided by. Note that description of other circuit configurations is omitted.

In the signal processing unit 14 of FIG. 9 described above, the circuit configuration up to the subtractors 28a and 28b is the same as that of the signal processing unit 14 of FIG. 1 described above, but the configuration of the subsequent circuit parts is different. That is, the outputs of the subtraction circuits 28a and 28b are input to the rearrangement circuit 32, and the two parallel data strings are rearranged as one serial data string arranged from the end data of the original image. Therefore, the switching circuit 29,
One white reference memory 30 and one division circuit 31 are provided. Further, in the signal processing unit 14 of FIG. 9, shading correction is performed after the rearrangement circuit 32, and the other parts have the same configuration as in FIG. Therefore, also in this embodiment, it is possible to obtain the same effect as that of the first embodiment described above, and it is possible to simplify the circuit configuration for performing the shading correction.

Next, a third embodiment of the invention described in claims 1, 3, 4, and 6 will be described with reference to FIG. In this embodiment, by using two photoelectric conversion elements, high-speed and high-definition reading is possible, and the output format is the same as that of an image reading apparatus using one photoelectric conversion element. The present invention provides an image reading apparatus capable of accurately correcting variations in sensitivity between pixels and variations in intensity distribution of an illumination light source.

That is, in this embodiment, the signal processing unit 14
However, only the signal processing unit 14 will be described here. The description of the same parts as those in the first embodiment is omitted, and the same parts are designated by the same reference numerals.

First, the main components of the signal processing unit 14 in this embodiment will be described with reference to FIG. In the signal processing unit 14, S / H circuits 16a and 16b as holding means for reading out only the pixels corresponding to the original area and holding the output of the pixels only in the transfer unit without the photosensitive unit,
Peak hold circuits 18a and 18b as peak hold means for holding the peak value of the output signal when the white reference plate is read over one line, and one line based on the output values of the hold means and the peak hold means. An A / D conversion circuit 21a as first normalizing means for normalizing the white reference plate read signal and the image read signal,
21b, a white reference memory 30 as storage means for storing the normalized white reference plate read signal output from the A / D conversion circuits 21a and 21b over one line,
A division circuit 31 as a second normalization means for normalizing the normalized image read signal output from the D / D conversion circuits 21a and 21b by the output obtained by reading the contents of the white reference memory 30 over one line; A rearrangement circuit 32 as rearranging means for outputting two data strings output in parallel in opposite directions from the central portion of the document image as one serial data string arranged in order from the end data of the document image. Figure 10
It is provided by the circuit configuration shown in. In addition,
Description of other circuit configurations will be omitted.

In the signal processing section 14 shown in FIG.
The circuit configuration up to the / D conversion circuits 21a and 21b is the same as that of the signal processing unit 14 in FIG. 1, but the circuit configuration after that is different. That is, the A / D conversion circuits 21a and 21b
Output from the rearrangement circuit 3 via the switching circuits 22a and 22b.
2. The two parallel data strings input to the memories 33a and 33b are rearranged as one serial data string arranged from the end data string of the original image. in this case,
If the data is directly input to the rearrangement circuit 32,
As shown in (c), the DS (S) information is destroyed. Therefore, in order not to destroy the information of DS (S), the CCD line sensor 5 of each memory 33a, 33b is stored.
Information on the DS (S) output from a and 5b is stored, and after rearranging the data, two DS (S) are inserted in the select circuit 34 and converted into data as shown in FIG. 13 (d). And output. Further, in this circuit, the dark output correction and the shading correction are performed in a process after the rearrangement circuit 32. Other processes are performed in the same manner as the signal processing unit 14 of FIG. Therefore, in the case of the present embodiment, it is possible to obtain the same effect as that of the first embodiment of FIG. 1, and by adopting such a configuration, a circuit configuration for performing dark output correction and shading correction It can be simplified.

Next, a first embodiment of the invention according to claims 2, 3, 5 and 6 will be described with reference to FIG. In this embodiment, by using two photoelectric conversion elements, high-speed and high-definition reading can be performed, and the number of read pixels of the photoelectric conversion elements is changed according to the reading area. By changing the
Further, the output format is the same as that of one photoelectric conversion element, and even when the number of pixels smaller than the number of pixels of the photoelectric conversion element is read, it is possible to accurately correct the dark output. Is provided.

That is, in this embodiment, the signal processing unit 14
However, only the signal processing unit 14 will be described here. In addition, claim 1, 3,
The description of the same parts as those in the first embodiment of the invention described in 4 and 6 is omitted, and the same reference numerals are used for the same parts.

First, the main components of the signal processing unit 14 in this embodiment will be described with reference to FIG. In the signal processing unit 14, S / H circuits 16a, 16 as first holding means for holding an output portion of a pixel having no photosensitive portion and only a transfer portion in a dark output signal over one line.
b, peak hold circuits 18a and 18b as second hold means for holding the output portion of the pixel where the light shield is provided on the photosensitive portion, and the S / H circuit 16a,
A / D conversion circuits 21a and 21b as normalizing means for normalizing the dark output signal over one line with reference to the output values of 16b and the peak hold circuits 18a and 18b.
Dark output memories 23a and 23b as first storage means for storing a dark output signal over one line normalized by the A / D conversion circuits 21a and 21b, and a pixel having a light shield on the photosensitive portion. And the outputs of the addition averaging circuits 24a and 24b as the second storage means for storing the output of the above, and the outputs of the dark output memories 23a and 23b read over one line. Then, based on the outputs of the multiplying circuits 25a and 25b as first dark output signal generating means for generating the dark output signal generated in each pixel, and the output of the multiplying circuits 25a and 25b, the CCD is actually used.
Adder circuit 2 as second dark output signal generating means for generating dark output signals output from the line sensors 5a and 5b.
6a, 26b and memories 27a, 27b, and a subtraction circuit 2 as subtraction means for subtracting the outputs of the addition circuits 26a, 26b and memories 27a, 27b from the output when read.
8a, 28b, and two data strings output in parallel in opposite directions from the central portion of the original image as a rearranging means for outputting as one serial data string arranged in order from the end data of the original image. The replacement circuit 32 is provided by the circuit configuration as shown in FIG. It should be noted that description of other circuit parts is omitted.

In the signal processing section 14 of FIG. 11 described above, instead of the dark output memories 23a and 23b of FIG.
Read-only (ROM) dark output memory 35a,
35b is used. Therefore, there is no operation for capturing the dark output, and the elements necessary for this are omitted. The storage in the dark output memories 35a and 35b can be processed in the same way as in the case of the configuration of FIG. Further, the signal processing unit 14 can be applied to the case where the dark output pattern is stored in each of the CCD run sensors 5a and 5b with little change with time and stored to some extent. Therefore, in the case of this embodiment, the same effect as that of the signal processing unit 14 of FIG. 1 can be obtained, and further, the dark output storing operation, the switching circuit, etc. are not necessary, so that the circuit configuration can be simplified.

Next, a second embodiment of the invention described in claims 2, 3, 5 and 6 will be described with reference to FIG. In this embodiment, by using two photoelectric conversion elements, high-speed and high-definition reading can be performed, and the number of read pixels of the photoelectric conversion elements is changed according to the reading area. By changing the
Further, the output format is the same as that of one photoelectric conversion element, and even when the number of pixels smaller than the number of pixels of the photoelectric conversion element is read, it is possible to accurately correct the dark output. Is provided.

That is, in this embodiment, the signal processing unit 14
However, only the signal processing unit 14 will be described here. In addition, claim 1, 3,
The description of the same parts as those in the first embodiment of the invention described in 4 and 6 is omitted, and the same reference numerals are used for the same parts.

First, the main components of the signal processing unit 14 in this embodiment will be described with reference to FIG. In the signal processing unit 14, dark output memories 35a and 35b as storage means for storing a dark output signal over one line.
And an averaging circuit 24 as a holding means for holding the output of the pixel whose light shield is provided on the photosensitive portion.
a, 24b, the outputs of the averaging circuits 24a, 24b and the output of the dark output memories 35a, 35b read over one line are multiplied to generate a dark output signal generated in each pixel. Multiplying circuits 25a and 25b as one dark output signal generating means, and these multiplying circuits 25a and 25b
Adder circuits 26a, 26b, 27a, 2 as second dark time output signal generating means for actually generating the dark time output signals output from the CCD line sensors 5a, 5b based on the output of
7b and the adder circuits 26a, 26b, 27a, 27b
Of subtraction circuits 28a and 28b as subtraction means for subtracting the output of the original image from the output when the image is read, and two data strings output in parallel in opposite directions from the central portion of the original image to the data at the end of the original image. The rearrangement circuit 32 as rearrangement means for outputting as one serial data string arranged in order from the above is provided by the circuit configuration as shown in FIG. It should be noted that description of other circuit parts is omitted.

In the signal processing unit 14 of FIG. 12 described above, the dark output memories 23a, 2 in the signal processing unit 14 of FIG. 9 are used.
Instead of 3b, a dark output memory 35a in ROM,
35b is used. As a result, in the case of the present embodiment as well, as in the case of FIG. 11, it is possible to eliminate the storage operation of the dark output. Also, output memory 3 in the dark
The storage and adaptation conditions in 5a and 35b are the same as in the embodiment of FIG. Therefore, the signal processing unit 14 of FIG.
The same effect as that of the signal processing unit 14 of FIG. 9 can be obtained,
Moreover, since the dark output storage operation and the switching circuit are unnecessary, the circuit configuration can be simplified.

Next, a third embodiment of the invention described in claims 2, 3, 5 and 6 will be described with reference to FIG. In this embodiment, by using two photoelectric conversion elements, high-speed and high-definition reading can be performed, and the number of read pixels of the photoelectric conversion elements is changed according to the reading area. By changing the
Another object of the present invention is to provide an image reading device capable of accurately correcting the variation in sensitivity between photoelectric conversion element pixels and the intensity distribution of the illumination light source.

That is, in this embodiment, the signal processing unit 14
However, only the signal processing unit 14 will be described here. In addition, claim 1, 3,
The description of the same parts as those in the first embodiment of the invention described in 4 and 6 is omitted, and the same reference numerals are used for the same parts.

First, the main components of the signal processing unit 14 in this embodiment will be described with reference to FIG. In the signal processing section 14, the S / H circuit 1 as a holding means for holding the output of the pixel only in the transfer section without the photosensitive section.
6a and 16b, peak hold circuits 18a and 18b as peak hold means for holding the peak value of the output signal when the white reference plate is read over one line, the S / H circuits 16a and 16b, and the peak hold. First for normalizing the white reference plate read signal and the image read signal over one line with reference to the output values of the circuits 18a and 18b
A / D conversion circuits 21a and 21b as normalizing means,
The white reference memory 30 as a storage means for storing the normalized white reference plate read signal output from the A / D conversion circuits 21a and 21b over one line, and the outputs of the A / D conversion circuits 21a and 21b. And the division circuit 31 as the second normalization means for normalizing the normalized image reading signal by the output of reading out the contents of the white reference memory 30 over one line, and the directions opposite to each other from the central portion of the original image. The rearrangement circuit 31 as rearrangement means for outputting the two data strings output in parallel to each other as one serial data string arranged in order from the data at the end of the original image by the circuit configuration as shown in FIG. It is provided. It should be noted that description of other circuit parts is omitted.

In the signal processing unit 14 of FIG. 13 described above, the dark output memory 35 in ROM is used instead of the dark output memory 23 of the signal processing unit 14 of FIG. As a result, even in the case of the present embodiment, FIG.
As in the case of 1, the storage operation of the dark output can be eliminated. Also, storage in the output memories 35a and 35b in the dark,
The adaptation conditions are the same as in the case of the embodiment shown in FIG. Therefore, the signal processing unit 14 of FIG. 13 is the same as the signal processing unit 1 of FIG.
The effect similar to that of No. 4 can be obtained, and since the dark output storage operation and the switching circuit are unnecessary, the circuit configuration can be simplified.

Next, an embodiment of the invention described in claim 7 will be described. In this embodiment, the dark time output is more accurately set by setting the accumulation time when reading the dark output signal over one line from the photoelectric conversion element to be longer than the accumulation time per line when reading an image. An image reading device capable of performing correction is provided.

That is, in each of the signal processing units 14 shown in FIGS. 1, 9 and 10, among the embodiments described above,
Capture in advance by setting the accumulation time for reading the dark output signal from the CCD line sensors 5a, 5b to be longer than the accumulation time for each line when reading an image.
It is possible to increase the output level of the line dark output signal, which can reduce the deterioration during signal processing. Moreover, by storing the level as a dark output pattern, it is possible to obtain a more accurate dark output. The output can be corrected.

[0054]

According to the first aspect of the present invention, an illumination light source for illuminating an original, an image forming means for forming an image of reflected light from the original, and two photoelectric conversions for photoelectrically converting the formed reflected light. In an image reading apparatus including an element, a driving unit that drives these photoelectric conversion elements, and a signal processing unit that performs signal processing of an output signal of the photoelectric conversion element, the image reading apparatus is configured to block reflected light from other than the original. And a light-shielding plate disposed in front of the photoelectric conversion element, wherein a document position corresponding to a pixel read out first of the photoelectric conversion element is substantially at the center of the document region in the main scanning direction, and subsequently. The original positions corresponding to the pixels to be read out are opposite to each other, and only the pixels corresponding to the original area are read out. First holding means for holding, a second holding means for holding an output portion of a pixel having a light shield on the photosensitive portion, and a dark line for one line with reference to output values of the first and second holding means. Normalizing means for normalizing the time output signal, and the normalized 1
First storage means for storing a dark output signal across the line;
Second storage means for storing the output of the pixel having a light shield on the photosensitive portion, the output of the first storage means and the second storage means.
First dark output signal generation means for multiplying the output of the storage means to generate a dark output signal generated in each pixel, and the photoelectric conversion actually based on the output of the first dark output signal generation means. Second dark output signal generating means for generating a dark output signal outputted from the element, subtracting means for subtracting the output of the second dark output signal generating means from the output when the image is read, and the original image The signal processing means is provided with rearranging means for outputting two data strings output in parallel in opposite directions from the central portion as one serial data string arranged in order from the end data of the document image. High-speed and high-definition image reading is possible by using two photoelectric conversion elements and driving them independently and simultaneously. In addition, a light-shielding plate is used to shield reflected light from other than the original and to contribute to image reading. Only read And makes it possible to reduce the power consumption of the photoelectric conversion element driving, the total data amount by the original size. In addition, a dark line output pattern of one line that has been captured in advance is multiplied by the DS (S) value during image reading, and the result is used to generate a dark line output signal that is actually output from the photoelectric conversion element. Since the subtraction is performed from the image signal output from the conversion element, accurate dark-time output correction can be performed even when the number of pixels smaller than that of the photoelectric conversion element is read. Further, since two data strings output in parallel in opposite directions from the central portion of the document image are output as one serial data string arranged in order from the end of the document, a photoelectric conversion element having one output format is used. It can be the same as when used.

According to a second aspect of the present invention, an illumination light source for illuminating a document, an image forming means for forming an image of reflected light from the original, and two photoelectric conversion elements for photoelectrically converting the image-formed reflected light. In an image reading apparatus provided with a driving means for driving these photoelectric conversion elements and a signal processing means for processing the output signal of the photoelectric conversion elements, the photoelectric conversion is performed so as to shield light reflected from other than the original. A light-shielding plate arranged in front of the element, and the original position corresponding to the read pixel read at the beginning of the photoelectric conversion element is substantially at the center in the main scanning direction of the original area and is read subsequently. The direction of the original position corresponding to the read pixel is opposite to each other, and only the pixel corresponding to the original area is read out and a storage unit for storing the output signal in the dark over one line and a light shield on the photosensitive unit are provided. First holding means for holding the output of the stored pixel, and the output of this holding means and the output of the storage means read out over one line to generate a dark output signal generated in each pixel. A dark output generating means, a second dark output generating means for actually generating a dark output signal output from the photoelectric conversion element based on an output of the first dark output generating means, and a second dark output generating means. Subtracting means for subtracting the output of the dark output generating means from the output when the image is read, and two data strings output in parallel in opposite directions from the central portion of the original image are sequentially arranged from the end data of the original image. Since the rearrangement means for outputting as a serial data line arranged side by side is provided in the signal processing means, two photoelectric conversion elements are used and driven independently and at the same time,
High-definition image reading becomes possible, and the light-shielding plate is used to block the reflected light from other than the original and only the pixels that are involved in image reading are read out. It is possible to reduce the amount. In addition, a memory for storing the dark output pattern is provided, and the DS output during image reading is used for this read output.
Since the dark output signal actually output from the photoelectric conversion element is generated by multiplying the value of (S) and the result is subtracted from the image signal output from the photoelectric conversion element,
Even when the number of pixels that is smaller than the number of pixels that the photoelectric conversion element has is read, accurate dark output correction can be performed. Further, since two data strings output in parallel in opposite directions from the central portion of the document image are output as one serial data string arranged in order from the end of the document, the output format is one photoelectric conversion element. The same as in the case of.

According to a third aspect of the present invention, an illumination light source for illuminating an original, an image forming means for forming an image of reflected light from the original, and two photoelectric conversion elements for photoelectrically converting the formed reflected light. In an image reading apparatus provided with a driving means for driving these photoelectric conversion elements and a signal processing means for processing the output signal of the photoelectric conversion elements, the photoelectric conversion is performed so as to shield light reflected from other than the original. A light-shielding plate arranged in front of the element, and the original position corresponding to the read pixel read at the beginning of the photoelectric conversion element is substantially at the center in the main scanning direction of the original area and is read subsequently. The direction of the original position corresponding to the read pixels is opposite to each other, and only the pixels corresponding to the original area are read out and the holding means for holding the output of the pixels only in the transfer section and the one-row unit. Peak hold means for holding the peak value of the output signal when the white reference plate is read over the entire range, and the white reference plate read signal and the image read signal for one line based on the output values of the hold means and the peak hold means. With a first normalizing means for normalizing, a storage means for storing the normalized white reference plate read signal which is the output of the first normalizing means over one line, and an output of the first normalizing means. Second normalizing means for normalizing a certain normalized image read signal by the output of reading the contents of the storage means over one line, and two normalizing means output in parallel in opposite directions from the central portion of the original image. The image reading apparatus according to claim 1, further comprising rearrangement means for outputting a data string as one serial data string arranged in order from the end data of the original image in the signal processing means. By using the output of dark output correction obtained by the device to create the standard of quantization and the standard of shading correction, by performing the quantization and shading correction on the image reading signal for which dark output correction has been performed, the accuracy can be improved. It is possible to perform high quantization and shading correction.

According to a fourth aspect of the present invention, an illumination light source for illuminating a document, an image forming unit for forming an image of reflected light from the document, and two photoelectric conversion elements for photoelectrically converting the image-formed reflected light. In an image reading apparatus including a driving unit that drives these photoelectric conversion elements and a signal processing unit that processes the output signal of the photoelectric conversion elements, the photoelectric conversion is performed so as to shield reflected light from other than the original. An original position corresponding to a read pixel read at the beginning of the photoelectric conversion element, which includes a light-shielding plate disposed in front of the element, an original size detecting unit, an illumination light source control unit, and a sub-scanning speed control unit. At the center of the original area in the main scanning direction, and the directions of the original positions corresponding to the read pixels read subsequently are opposite to each other, and the original area corresponds to the original area depending on the size of the original. First holding means for reading out only the pixels to be taken out, controlling the intensity of the illumination light source, the sub-scanning speed, and holding the output portion of the pixels of only the transfer portion without the photosensitive portion in the dark output signal over one line; The second which holds the output part of the pixel where the light shield is provided on the photosensitive part
A holding means, a normalizing means for normalizing the dark output signal over one line with reference to the output values of the first and second holding means, and a dark time over one line normalized by the normalizing means. First storage means for storing the output signal, second storage means for storing the output of the pixel whose light shield is provided on the photosensitive portion, and output of the second storage means and one line of the first storage means. First dark output signal generating means for generating a dark output signal generated in each pixel by multiplying the read output, and the photoelectric conversion element actually based on the output of the first dark output signal generating means. Second dark output signal generation means for generating a dark output signal output from the second dark output signal, subtraction means for subtracting the output of the second dark output signal generation means from the read output, and a central portion of the original image. Output in parallel in opposite directions Since the two data columns from the data of the end of the document image and a sorting means for outputting as a single serial data ordered sequence provided in said signal processing means,
Since two photoelectric conversion elements are used and driven independently and simultaneously, the pixels of the photoelectric conversion elements that do not contribute to reading depending on the size of the original image are shielded, the illumination light source is controlled, and the sub-scanning speed is controlled. , High-definition reading is possible,
Especially when reading a document of a small size, high speed can be exhibited. In addition, a dark line output pattern of one line that has been captured in advance is multiplied by the DS (S) value during image reading, and the result is used to generate a dark line output signal that is actually output from the photoelectric conversion element. Since the subtraction is performed from the image signal output from the conversion element, it is possible to accurately correct the dark-time output even when the number of pixels smaller than the number of pixels of the photoelectric conversion element is read. Further, since two data strings output in parallel in opposite directions from the central portion of the document image are output as one serial data string arranged in order from the end of the document, the output format is one photoelectric conversion element. The same as in the case of.

According to a fifth aspect of the present invention, an illumination light source for illuminating a document, an image forming means for forming an image of reflected light from the original, and two photoelectric conversion elements for photoelectrically converting the image-formed reflected light. In an image reading apparatus including a driving unit that drives these photoelectric conversion elements and a signal processing unit that processes the output signal of the photoelectric conversion elements, the photoelectric conversion is performed so as to shield reflected light from other than the original. An original position corresponding to a read pixel read at the beginning of the photoelectric conversion element, which includes a light-shielding plate disposed in front of the element, an original size detecting unit, an illumination light source control unit, and a sub-scanning speed control unit. At the center of the original area in the main scanning direction, and the directions of the original positions corresponding to the read pixels read subsequently are opposite to each other, and the original area corresponds to the original area depending on the size of the original. Storage means for reading out only the pixels to be taken out, controlling the intensity of the illumination light source and the sub-scanning speed, and storing the output signal in the dark at one line, and the hold for holding the output of the pixels having the light shield on the photosensitive portion. Means, the output of the hold means and the contents of the storage means
First dark output signal generation means for multiplying the output read across the line to generate a dark output signal generated in each pixel, and the actual output based on the output of the first dark output signal generation means Second dark output signal generating means for generating a dark output signal output from the photoelectric conversion element, subtracting means for subtracting the output of the second dark output signal generating means from the output when the image is read, and the original document Arrangement means is provided in the signal processing means for outputting two data strings output in parallel in opposite directions from the central portion of the image as one serial data string arranged in order from the end data of the document image. Therefore, by using two photoelectric conversion elements and independently and simultaneously driving, the pixels of the photoelectric conversion elements that do not contribute to reading due to the size of the original image are shielded, the illumination light source is controlled, and the sub-scanning speed is controlled. ,high speed High-definition reading is possible, it is possible to exert a high speed when reading especially of small-size document. In addition, a memory for storing the dark output pattern is provided, and the DS output during image reading is used as the read output.
The value of (S) is multiplied and the result is used to generate a dark output signal actually output from the photoelectric conversion element, and subtraction is performed from the image signal output from the photoelectric conversion element. Even when the number of pixels having a small number of pixels is read, accurate dark output correction can be performed.
Furthermore, two data strings output in parallel in opposite directions from the central portion of the original image are arranged in order from the end of the original.
Since it is output as one serial data string, the output format can be the same as that of one photoelectric conversion element.

According to a sixth aspect of the present invention, an illumination light source for illuminating a document, an image forming means for forming an image of reflected light from the document, and two photoelectric conversion elements for photoelectrically converting the image-formed reflected light. In an image reading apparatus including a driving unit that drives these photoelectric conversion elements and a signal processing unit that processes the output signal of the photoelectric conversion elements, the photoelectric conversion is performed so as to shield reflected light from other than the original. An original position corresponding to a read pixel read at the beginning of the photoelectric conversion element, which includes a light-shielding plate disposed in front of the element, an original size detecting unit, an illumination light source control unit, and a sub-scanning speed control unit. At the center of the original area in the main scanning direction, and the directions of the original positions corresponding to the read pixels read subsequently are opposite to each other, and the original area corresponds to the original area depending on the size of the original. Hold means for reading out only the pixels to be taken, controlling the intensity of the illumination light source and the sub-scanning speed, and holding the output of the pixels only in the transfer section without the photosensitive section, and the output when the white reference plate is read over one line Peak holding means for holding the peak value of the signal, and 1 based on the output values of the holding means and the peak holding means.
First normalizing means for normalizing the white reference plate read signal and the image read signal over the line, and a memory for storing the normalized white reference plate read signal which is the output of the first normalizing means over the one line. Second means for normalizing the normalized image reading signal output from the first normalizing means by the output of reading the contents of the storage means over one line.
The normalizing means and the rearranging means for outputting two data strings output in parallel in opposite directions from the central portion of the document image as one serial data string arranged in order from the end data of the document image. Since it is provided in the signal processing means, the quantization output and the shading correction reference are created by using the dark output correction output obtained by the image reading device according to the fourth and fifth aspects, and the dark output correction is performed. By performing quantization and shading correction on the image read signal, highly accurate quantization and shading correction can be performed.

The invention according to claim 7 is the invention according to claim 1, 3, or 4.
Alternatively, in the invention described in 6, the accumulation time when the dark output signal over one line is read from the two photoelectric conversion elements is
Since it is set to be longer than the accumulation time per line during image reading, it is possible to increase the output level of the dark output of one line that is captured in advance, thereby reducing deterioration during signal processing. Further, by storing the level as a dark output pattern, it is possible to perform more accurate dark output correction.

[Brief description of drawings]

FIG. 1 is a circuit diagram showing a state of a signal processing unit which is a first embodiment of the invention according to claims 1, 3, 4, and 6.

FIG. 2 is a side view showing the overall configuration of the image reading apparatus.

FIG. 3 is a block diagram showing a configuration of a control unit of the image reading apparatus.

FIG. 4 is a timing chart showing the states of various pulse waveforms output from the CCD drive unit.

FIG. 5 is a state explanatory view showing how the dark outputs overlap when 60% of pixels of the CCD line sensor are read.

FIG. 6A is a waveform diagram showing an output state of one line of the CCD line sensor, and FIG. 6B is a waveform diagram showing an output state of one line after the transfer clock waveform is removed.

FIG. 7A is a waveform diagram showing an output waveform for one line when the exposure amount of the CCD line sensor is zero, and FIG.
Is a waveform diagram showing the relationship between the reference of the A / D conversion circuit and the signal input, and (c) is a waveform diagram showing the relationship between the two references of the A / D conversion circuit and the signal input.

8A and 8B are state diagrams showing a state of a data sequence before rearrangement, and FIG. 8C is a state diagram showing a state of a data sequence after rearrangement in a rearrangement circuit. , (D)
FIG. 6 is a state diagram showing a state of a data string converted and output in a select circuit after data rearrangement.

FIG. 9 is a circuit diagram showing a state of a signal processing unit which is a second embodiment of the invention described in claims 1, 3, 4, and 6.

FIG. 10 is a circuit diagram showing a state of a signal processing unit which is a third embodiment of the invention described in claims 1, 3, 4, and 6.

FIG. 11 is a circuit diagram showing a state of a signal processing unit which is a first embodiment of the invention according to claims 2, 3, 5, and 6.

FIG. 12 is a circuit diagram showing a state of a signal processing unit which is a second embodiment of the invention according to claims 2, 3, 5, and 6.

FIG. 13 is a circuit diagram showing a state of a signal processing unit which is a third embodiment of the invention according to claims 2, 3, 5, and 6.

[Explanation of symbols]

2 original 3 illumination light sources 4a, 4b image forming means 5a, 5b photoelectric conversion elements 6a, 6b light-shielding plate 10 driving means 11 original size detecting means 12 sub-scanning speed control means 13 illumination light source control means 14 signal processing means 16a, 16b 1st Hold means 17a, 17b Second hold means 21a, 21b Normalization means 23a, 23b First storage means 24a, 24b Second storage means 25a, 25b First dark output signal generation means 26a, 26b, 27a, 27b Second dark Time output signal generating means 28a, 28b subtracting means 32 rearranging means

Claims (7)

[Claims] 1. An illumination light source for illuminating a document, an image forming unit for forming an image of reflected light from the document, two photoelectric conversion elements for photoelectrically converting the image-formed reflected light, and these photoelectric conversion elements. In an image reading apparatus provided with a driving unit for driving and a signal processing unit for processing an output signal of the photoelectric conversion element, the image reading apparatus is provided in front of the photoelectric conversion element so as to shield reflected light from other than the original. And a light-shielding plate arranged so that a document position corresponding to a pixel read out first of the photoelectric conversion element is substantially at the center in the main scanning direction of the document region and a document position corresponding to a pixel read subsequently. Are opposite to each other, and only the pixels corresponding to the original area are read out, and the first hold hand for holding the output portion of the pixels of only the transfer portion without the photosensitive portion in the dark output signal over one line Stage, second holding means for holding the output portion of the pixel having a light shield on the photosensitive portion, and a dark output signal over one line based on the output values of the first and second holding means. Normalizing means for normalizing, a first storing means for storing the normalized dark-time output signal over one line, and a second storing means for storing the output of a pixel having a light shield on the photosensitive portion, A first dark time output signal generating means for multiplying the output of the first memory means and the output of the second memory means to generate a dark time output signal generated in each pixel, and the first dark time output signal generating means. Second dark output signal generating means for actually generating a dark output signal output from the photoelectric conversion element based on the output of the second photoelectric conversion element, and an output of the second dark output signal generating means when an image is read. Means for subtracting from the output of Rearranging means for outputting two data strings output in parallel in opposite directions from the central portion as one serial data string arranged in order from the end data of the original image are provided in the signal processing means. An image reading device characterized by. 2. An illumination light source for illuminating a document, an image forming means for forming an image of reflected light from the document, two photoelectric conversion elements for photoelectrically converting the image-formed reflected light, and these photoelectric conversion elements. In an image reading apparatus provided with a driving unit for driving and a signal processing unit for processing an output signal of the photoelectric conversion element, the image reading apparatus is provided in front of the photoelectric conversion element so as to shield reflected light from other than the original. And a light-shielding plate arranged so that a document position corresponding to a read pixel read first of the photoelectric conversion element is substantially at the center of the document region in the main scanning direction and corresponds to a read pixel read subsequently. Storage means for reading out only the pixels corresponding to the original area and storing the output signal in the dark for one line, and the output of the pixels with the light shield on the photosensitive portion And a first dark output generating means for generating a dark output signal generated in each pixel by multiplying an output of the holding means by an output read out over one line of the contents of the storage means. , This first
A second dark output generating means for actually generating a dark output signal output from the photoelectric conversion element based on the output of the dark output generating means and an image of the output of the second dark output generating means are displayed. Subtracting means for subtracting the output from the reading and rearranging for outputting two data strings output in parallel in opposite directions from the central portion of the original image as serial data strings arranged in order from the data at the end of the original image. And an image reading unit provided in the signal processing unit. 3. An illumination light source for illuminating an original, an image forming unit for forming an image of reflected light from the original, two photoelectric conversion elements for photoelectrically converting the formed reflected light, and these photoelectric conversion elements. In an image reading apparatus provided with a driving unit for driving and a signal processing unit for processing an output signal of the photoelectric conversion element, the image reading apparatus is provided in front of the photoelectric conversion element so as to shield reflected light from other than the original. And a light-shielding plate arranged so that a document position corresponding to a read pixel read first of the photoelectric conversion element is substantially at the center of the document region in the main scanning direction and corresponds to a read pixel read subsequently. The manuscript positions are opposite to each other, only the pixels corresponding to the manuscript area are read out, and the holding means for holding the output of the pixels only in the transfer part without the photosensitive part and the white reference plate for one line are read. Peak hold means for holding the peak value of the output signal when taken, and a first normalization for normalizing the white reference plate read signal and the image read signal for one line with the output values of the hold means and the peak hold means as references. Normalization means, storage means for storing the normalized white reference plate read signal output from the first normalization means over one line, and normalized image reading output from the first normalization means. Second normalization means for normalizing the signal by the output of reading out the contents of the storage means over one line, and two data strings output in parallel in opposite directions from the central portion of the original image are used to output the original image. An image reading apparatus, characterized in that rearrangement means for outputting as one serial data string arranged in order from the end data are provided in the signal processing means. 4. An illumination light source for illuminating an original, an image forming unit for forming an image of reflected light from the original, two photoelectric conversion elements for photoelectrically converting the formed reflected light, and these photoelectric conversion elements. In an image reading apparatus provided with a driving unit for driving and a signal processing unit for processing an output signal of the photoelectric conversion element, the image reading apparatus is provided in front of the photoelectric conversion element so as to shield reflected light from other than the original. An original position corresponding to a read pixel read at the beginning of the photoelectric conversion element is the main area of the original area. At the center of the scanning direction, the positions of the originals corresponding to the read pixels read subsequently are opposite to each other, and only the read pixels corresponding to the original area are read depending on the size of the original. Controlling the intensity of the illumination light source and the sub-scanning speed, and holding the output portion of the pixel of the transfer portion only without the photosensitive portion of the dark output signal over one line,
The holding means, the second holding means for holding the output portion of the pixel having the light shield on the photosensitive portion, and the dark-time output signal over one line based on the output values of the first and second holding means are normal. Normalizing means for normalizing, a first storing means for storing a dark-time output signal over one line normalized by the normalizing means, and a second storing means for storing an output of a pixel having a light shield on the photosensitive portion. Storage means,
First dark output signal generation means for multiplying the output of the second storage means and the output read out over one line of the first storage means to generate a dark output signal generated in each pixel, and the first dark output signal generation means. Second dark time output signal generating means for actually generating a dark time output signal output from the photoelectric conversion element based on the output of the dark time output signal generating means, and output of the second dark time output signal generating means And subtraction means for subtracting from the output when the document is read, and 2 output in parallel in opposite directions from the central portion of the original image.
An image reading apparatus, wherein the signal processing means is provided with rearranging means for outputting one data string as one serial data string arranged in order from the end data of the original image. 5. An illumination light source for illuminating a document, an image forming unit for forming an image of reflected light from the document, two photoelectric conversion elements for photoelectrically converting the image-formed reflected light, and these photoelectric conversion elements. In an image reading apparatus provided with a driving unit for driving and a signal processing unit for processing an output signal of the photoelectric conversion element, the image reading apparatus is provided in front of the photoelectric conversion element so as to shield reflected light from other than the original. An original position corresponding to a read pixel read at the beginning of the photoelectric conversion element is the main area of the original area. At the center of the scanning direction, the positions of the originals corresponding to the read pixels read subsequently are set to be opposite to each other, and only the read pixels corresponding to the original area are read depending on the size of the original. Storage means for controlling the intensity of the illumination light source and the sub-scanning speed to store the dark output signal over one line, a holding means for holding the output of the pixel whose light shield is provided on the photosensitive portion, and this hold First dark output signal generating means for generating a dark output signal generated in each pixel by multiplying the output of the means and the output read out over one line of the contents of the storage means, and the first dark output signal Second dark output signal generating means for actually generating a dark output signal output from the photoelectric conversion element based on the output of the generating means, and an image of the output of the second dark output signal generating means are read. The subtraction means for subtracting the output from the original image and the parallel output of the two data strings output in parallel in opposite directions from the central portion of the document image as one serial data string arranged in order from the end data of the document image. Alternative Image reading apparatus being characterized in that provided in the signal processing means. 6. An illumination light source for illuminating a document, an image forming unit for forming an image of reflected light from the document, two photoelectric conversion elements for photoelectrically converting the image-formed reflected light, and these photoelectric conversion elements. In an image reading apparatus provided with a driving unit for driving and a signal processing unit for processing an output signal of the photoelectric conversion element, the image reading apparatus is provided in front of the photoelectric conversion element so as to shield reflected light from other than the original. An original position corresponding to a read pixel read at the beginning of the photoelectric conversion element is the main area of the original area. At the center of the scanning direction, the positions of the originals corresponding to the read pixels read subsequently are set to be opposite to each other, and only the read pixels corresponding to the original area are read depending on the size of the original. The intensity of the illumination light source and the sub-scanning speed are controlled, and the holding means for holding the output of the pixel only in the transfer section without the photosensitive section and the peak value of the output signal when the white reference plate is read over one line are displayed. Peak holding means for holding, first normalizing means for normalizing the white reference plate read signal and the image read signal over one line with reference to the output values of the hold means and the peak hold means, and the first normalization means. The storage means for storing the normalized white reference plate read signal output from the means over one line, and the normalized image read signal output from the first normalizing means in the storage means 1 Second normalizing means for normalizing the output read out over the line and two data strings output in parallel in opposite directions from the central portion of the original image are used as data at the end of the original image. Image reading apparatus characterized by a rearranging section provided in said signal processing means for outputting as a single serial data sequence arranged in order from. 7. A storage time for reading out a dark output signal over one line from two photoelectric conversion elements is set longer than a storage time for one line during image reading. The image reading apparatus according to claim 1, 3, 4, or 6.