JPH06231301A - Optical character reader - Google Patents

Abstract

PURPOSE:To provide an optical character reader which can highly precisely read a small character such as a printing character and which can read at high speed a slip including only a comparatively large character such as a handwritten character and a comparatively simple character such as the printing character of a number. CONSTITUTION:Image data from ADC 8 on the CCD imager 4-side of high resolution is selected when the character mentioned in the slip 1 is the small one such as the printing character, and image data from ADC 9 on the CCD imager 5-side of low resolution is selected when the character is the comparatively large one such as the handwritten character so as to recognize the character in a signal switching circuit (MPX) 10 by using the CCD imager 4 of high resolution with the large number of picture elements and the CCD imager 5 of low resolution with the small number of the picture elements.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical character reader (hereinafter abbreviated as OCR (optical character reader)).

[0002]

2. Description of the Related Art FIG. 7 is a block diagram showing a configuration example of a conventional OCR. In the figure, according to an instruction from the CPU 15, the form 1 is conveyed by the conveying unit 13, and an image of the form 1 is imaged by the optical system 2 on the image pickup surface of the CCD imager 4 which is a photoelectric conversion means to perform photoelectric conversion. Then, the output signal of the CCD imager 4 is transferred to the signal amplifier (A
After being amplified by the MP) 6 and further digitized by the analog-to-digital converter (ADC) 8, all the image data of the form 1 is stored in the image buffer 11.

Next, the character recognition unit 12 cuts out characters from the image buffer 11 to perform character recognition.
Memory 1 for form format data based on these characters
4, and character recognition is performed from information such as the position of the character to be cut out in this form format data and the character type, and the result is stored in the file device 17 by the file control unit 16 or an external interface (I / F) 18 to send to an external device. This OC
In R, in order to increase the reading speed, conventionally, not only is each component speeded up, but also the character recognition unit 12
A plurality of character recognition units 12 are provided and these character recognition units 12 are simultaneously operated.

[0004]

However, in the conventional OCR having the above structure, in order to read a small character such as a printed character with high accuracy, a CCD imager 4 having a large number of pixels is used to increase the resolution. As a result, the maximum conveyance speed of the form 1 is determined by the scanning speed of the CCD imager 4, so that relatively large characters such as handwritten characters and relatively simple characters such as printed characters of numbers are used. At the time of reading, there was a problem in that the reading speed could not be increased. Further, even when reading a large form, it is necessary to use the CCD imager 4 having a large number of pixels, and the maximum conveyance speed of the form is determined by the scanning speed of the CCD imager 4, so reading with a small form is possible. There was also a problem that the speed could not be increased.

Therefore, the present invention is capable of reading small characters such as printed characters with high accuracy, and is high speed for a form including only relatively large characters such as handwritten characters and relatively simple characters such as printed characters of numbers. It is intended to provide an excellent OCR that can be read by a user. Another object of the present invention is to provide an excellent OCR capable of reading a large form and reading a small form at high speed.

[0006]

The OCR according to the present invention comprises:
First and second photoelectric conversion means having different resolutions for reading the image information of the form and converting it into an electric signal, and the first and second photoelectric conversion means depending on the size of the characters written on the form to be read.
The second photoelectric conversion means includes a selection means for selecting either one of the output signals and a character recognition means for recognizing a character based on the output signal selected by the selection means. . Also, another OCR according to the present invention
Are first and second photoelectric conversion means having different sizes in the scanning direction for reading the image information of the form and converting it into electric signals, and the first and second photoelectric conversion means depending on the size of the form to be read.
The second photoelectric conversion means includes a selection means for selecting either one of the output signals and a character recognition means for recognizing a character based on the output signal selected by the selection means. .

[0007]

Of the two photoelectric conversion means having different resolutions, when the character written on the form is a small character such as a printed character, the output signal of the high resolution photoelectric conversion means is changed to a relatively large value such as a handwritten character. In the case of a character, character recognition is performed using the output signal of the low-resolution photoelectric conversion means. Thus, when a CCD imager is used as the photoelectric conversion means, the resolution and the number of pixels correspond to each other, and the low resolution one has a smaller number of pixels than the high resolution one.
Since the time required for scanning is short, it is possible to quickly read a form that includes only relatively large and large characters such as handwritten characters and printed characters of numbers.

Also, the sizes in the scanning direction are different from each other.
Character recognition is performed using the output signal of the photoelectric conversion means having a large size in the scanning direction in the case of a large form and the output signal of the photoelectric conversion means having a small size in the scanning direction of the small form. I do. Thereby, when a CCD imager is used as the photoelectric conversion means,
The size in the scanning direction corresponds to the number of pixels, and the smaller size has a smaller number of pixels than the larger size, and therefore the time required for scanning is short, so that a small form can be read at high speed.

[0009]

Embodiments of the present invention will now be described in detail with reference to the drawings. FIG. 1 is a schematic block diagram showing an embodiment of the present invention. In the figure, the written characters of the form 1 to be read are imaged by the optical systems 2 and 3 on the image pickup surfaces of, for example, CCD imagers 4 and 5 which are photoelectric conversion means. The CCD imager 4 has, for example, 20 pixels.
A high resolution of 00 pixels is used, and the CCD imager 5 having a low resolution of 1000 pixels is used.

Since the CCD imager 5 has a smaller number of pixels than the CCD imager 4 and thus has a smaller size in the scanning direction, when the same optical system 4 as the optical system 3 is used, it is clear from FIG. As described above, the optical separation distance from the form 1 is set to be longer than that of the CCD imager 4. Each image signal obtained by photoelectric conversion by these CCD imagers 4 and 5 is a signal amplifier (AM).
P) 6, 7 is amplified, and is converted into a digital signal by the analog / digital converters (ADC) 8, 9 and supplied to the signal switching circuit (MPX) 10.

In response to a command from the host device, the signal switching circuit 10 outputs a signal from the analog / digital converter 8 on the high resolution CCD imager 4 side when the characters written on the form 1 are small characters such as printed characters. If the image data is a form that includes only relatively large and large characters such as handwritten characters and numbers, the low resolution C
The image data from the analog / digital converter 9 on the CD imager 5 side is selected and supplied to the image memory 11, and all the image data of the form 1 is stored.
The character recognition unit 12 cuts out characters from the form image data on the image memory 11 and performs character recognition. At this time, the position where the character is cut out is determined by the common format data read in advance in the memory 14.

The common format data is stored in the memory 14 together with the form format data by the following two storage methods. The first storage method is a method of receiving from an external device (not shown) by an external interface (I / F) 18 and storing it in the memory 14. Another storage method is a method of reading from the file device 17 and storing it in the memory 14 by the file control unit 16. Here, as the contents of this common format data, the form type code (hereinafter, I
The data indicating the position (abbreviated as D), the number of characters, the character type, and the like are included.

The character recognizing unit 12 searches the memory 14 for the form format data based on the cut-out characters, and performs character recognition from the information such as the position of the cut-out character in the form format data and the type of the character. The result is stored as the reading result of the form 1 in the file device 17 by the file control unit 16 or sent to an external device (not shown) by the external interface (I / F) 18. 2A and 2B are diagrams showing the relationship between the CCD imager and the form. FIG. 2A shows the case of the CCD imager 4 having a large number of pixels, and FIG. 2B shows the case of the CCD imager 5 having a small number of pixels. . In the figure, light emitted from a light source (not shown) illuminates the form 1, and the reflected light is guided to the CCD imagers 4 and 5 by the optical systems 2 and 3, so that the characters written on the form 1 are written to the CCD imager 4. 5
An image is formed on the image pickup surface of the CCD imager 4, 5
Is photoelectrically converted at.

Here, for example, the CCD imager 4 has 2000 pixels and the CCD imager 5 has 1 pixel.
If the CCD imager 4 and the CCD imager 5 have the same sensitivity and the same scanning speed as 000 pixels, the CCD
The time required to scan the imager 5 once is about half the time required to scan the CCD imager 4 once, and the transport speed can be doubled. Therefore, when reading a form containing only relatively large characters such as handwritten characters and relatively simple characters such as printed characters of numbers, the photoelectric conversion is performed by the CCD imager 5 and the reading is performed.
It is possible to realize a reading speed about twice as high as that when the CCD imager 4 is used.

Next, a series of operations of the OCR according to the present invention will be described with reference to the schematic structural diagram shown in FIG. When the form 1 is set in the hopper 21, the hopper 21 is lifted and the form 1 comes into contact with the pull-in roller 22 according to a command from the CPU 15 (see FIG. 1). The pull-in roller 22 is driven to rotate in the direction of the arrow in the figure by a motor (not shown) to convey the document 1 to the separation rollers 23a and 23b. The separation rollers 23a and 23b are rotationally driven by a motor (not shown) together with a roller (not shown),
The form 1 is conveyed while preventing its overrun.

When the conveyed form 1 arrives under the light source 24, the light emitted from the light source 24 is reflected by the form 1,
Rays 25 and 26 become mirrors 27a, 27b, 27
c and the mirrors 28a, 28b, 28c, the lens 2
CCD imager 4 and CC through lens 9 and lens 30
The image is projected on each imaging surface of the D imager 5. The mirror 2
7a, 27b, 27c and lens 29, and 28a, 28
b, 28c and the lens 30, the optical system 2 in FIG.
And the optical system 3 is configured. CCD imager 4,
Each image projected on 5 is converted into an electric signal, amplified by signal amplifying units 6 and 7, digitized by analog / digital converting units 8 and 9 and supplied to signal switching circuit 10 as shown in FIG. To be done.

The signal switching circuit 10 selects image data from one of the CCD imagers 4 and 5 and supplies it to the image buffer 11 to store all the image data of the form 1. The selection in the signal switching circuit 10 is made in advance by an instruction from the host device. Specifically, when the written characters on the form 1 are small characters such as printed characters, the image data from the analog / digital converter 8 on the high resolution CCD imager 4 side is selected to print handwritten characters or numbers. In the case of a form that is relatively simple and contains only large characters, such as
The image data from the analog / digital converter 9 on the CD imager 5 side is selected.

Next, the character recognition unit 12 cuts out characters from the form image data on the image buffer 11 and recognizes them, and the result is sent to the file device 17 via the file control unit 16 or externally via the external interface 18. Output to the device. Next, the read form is discharged to the accept stacker 32 or the reject stacker 33 by switching with the blade 31 in FIG. 3, and a series of operations is completed. In addition, in FIG.
FIG. 4 is a schematic structural diagram of the OCR having the structure of FIG. 3 in which one set of CCD imagers 4 and 5 and optical systems 2 and 3 including mirrors and lenses are configured as a contact type image sensor. The operation is the same as in the case of FIG.

FIG. 5 is a schematic block diagram showing another embodiment of the present invention. In the figure, the same parts as those in FIG. 1 are designated by the same reference numerals, and the description thereof will be omitted to avoid duplication. In this embodiment, a CCD imager 4 having a large number of pixels in the scanning direction is used, and a CCD imager 5 having a small number of pixels in the scanning direction is used. Then, the signal switching circuit 10
In accordance with an instruction from the host device, either one of the image data from the analog / digital converters 8 and 9 is selected according to the form size and is supplied to the image buffer 11. FIG. 6 is a diagram showing the relationship between the readable range of the CCD imager and the size of the form,
X is the readable range of the CCD imager 4, Y is CC
The readable range of the D imager 5 is shown. In the figure, 41 is a small form, 42 is a large form,
Reference numeral 43 is an ID of a small form, and 44 is an ID of a large form.

Here, for example, the CCD imager 4 has 5000 pixels and the CCD imager 5 has 2 pixels.
If the CCD imager 4 and the CCD imager 5 have the same sensitivity and the same scanning speed as 000 pixels, the CCD
The time for scanning the imager 5 once is about 2/5 of the time for scanning the CCD imager 4 once, and the transport speed can be increased about 2.5 times. Therefore, for a form having a size within the readable range Y of the CCD imager 5, photoelectric conversion is performed by the CCD imager 5 to perform reading, which is about 2.5 times as large as when the CCD imager 4 is used. A reading speed can be realized.

Next, a specific example of selection in the signal switching circuit 10 will be described. The upper device knows the size of the form to be read in advance, and the signal switching circuit 10 determines that the form size is CC by the instruction from the upper device.
When it is within the readable range of the D imager 5, the image data from the analog / digital converter 9 is supplied to the image buffer 11. If the size of the form is unknown, the image data from the analog / digital converter 8 is supplied to the image buffer 11 for the first sheet, and if the size is within the readable range Y of the CCD image data 5. The image data from the analog / digital converter 9 is supplied to the image buffer 11 from the form to be read next.

In each of the above-described embodiments, the case where the CCD imager is used as the photoelectric conversion means has been described, but the invention is not limited to this, and the image of the form such as another solid-state image pickup device or a pickup tube may be displayed. It can be used for all the things that can be read and converted into an electric signal.

[0023]

As described above, according to the present invention,
By using two photoelectric conversion units having different resolutions and selecting one of the output signals of the two photoelectric conversion units according to the size of the written characters on the form for character recognition, low resolution is achieved. Since the photoelectric conversion means of has a smaller number of pixels than the photoelectric conversion means of high resolution, the time required for scanning is short, so that a form containing only relatively large characters such as handwritten characters and printed characters of numbers can be used. In this case, the form can be read at high speed by selecting the output signal of the low-resolution photoelectric conversion means.

Further, the sizes in the scanning direction are different from each other.
By using one photoelectric conversion means and selecting one of the two photoelectric conversion means according to the size of the form for character recognition, the small photoelectric conversion means becomes the large photoelectric conversion means. Since the number of pixels is small compared with the above, the time required for scanning is short. Therefore, in the case of a small form, the small form can be read at high speed by selecting the output signal of the photoelectric conversion means having a small size.

[Brief description of drawings]

FIG. 1 is a schematic block diagram showing an embodiment of the present invention.

FIG. 2 is a diagram showing a relationship between a CCD imager and a form, and FIG. 2A shows a CCD imager having a large number of pixels,
Each of (b) shows a case of a CCD imager having a small number of pixels.

FIG. 3 is a schematic structural diagram showing an example of an OCR according to the present invention.

FIG. 4 is a schematic structural diagram showing another example of the OCR according to the present invention.

FIG. 5 is a schematic block diagram showing another embodiment of the present invention.

FIG. 6 is a diagram showing a relationship between a readable range of a CCD imager and a size of a form.

FIG. 7 is a schematic block diagram showing a configuration example of a conventional OCR.

[Explanation of symbols]

 1 Form 4, 5 CCD imager 8, 9 Analog-to-digital converter (ADC) 10 Signal switching circuit 11 Image buffer 12 Character recognition unit

Claims (2)

[Claims] 1. A first photoelectric conversion unit and a second photoelectric conversion unit having different resolutions for reading image information of a form and converting it into an electric signal, and the first photoelectric conversion unit according to a size of a character written on a form to be read. , A selection means for selecting one of the output signals of the second photoelectric conversion means, and a character recognition means for performing character recognition based on the output signal selected by the selection means. Optical character reading device. 2. A first and second photoelectric conversion means having different sizes in the scanning direction for reading image information of a form and converting it into an electric signal, and the first and second photoelectric conversion means according to the size of the form to be read. 2. An optical system comprising: a selection unit that selects one of the output signals of the photoelectric conversion unit 2; and a character recognition unit that performs character recognition based on the output signal selected by the selection unit. Type character reader.