JPH0461476A - Picture processor - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image processing device applied to binarization processing of grayscale images such as photographs.

[Prior Art] For example, in a facsimile machine, image information on a document is read as an electrical analog signal by an image sensor, etc., and this analog signal is compared with a reference signal.
Image processing is performed to obtain a binary signal from the magnitude relationship.

Conventionally, methods such as systematic dither method, error diffusion method, minimum average error method, etc. are known as means for binarizing an image when the ratio of input pixels to output pixels is 1=1. ing.

The systematic dither method compares an analog signal of image information read from a document with a signal obtained by converting dither pattern digital information prepared in advance into an analog signal using a D/A converter to obtain a binary signal. It is widely used because it does not require complicated calculations and is relatively easy to configure.

However, because the outline of this image may move or disappear, in the case of a shading image containing text, etc.
Highly accurate binarization processing is difficult, and if binarized display is carried out as it is, there are disadvantages in which text parts may be missing or outlines may be blurred.

On the other hand, the error diffusion method and the minimum average error method are superior to the systematic dither method described above in terms of resolution and can perform highly accurate binarized display, but both of these methods require algorithms. It is complex, and if you try to process it with software, there will be a problem with the processing speed, and if you try to process it with hardware, you will need digital processing, so for example, you will need an A/D converter, a line buffer, a multiplier, an adder, a comparator, etc. However, the disadvantage is that the circuit structure is large and the device becomes large and expensive.

[Problems to be Solved by the Invention] As described above, the method used for binarizing gray scale images including characters, etc., which adopts the systematic dither method, has a relatively simple configuration, but has a low accuracy. On the other hand, methods that employ the error diffusion method and the minimum average error method can perform highly accurate binarization processing, but
The disadvantages were that the circuit configuration was large, large, and expensive.

The present invention has been made in view of the above circumstances, and provides an image processing device that is capable of good binarization of grayscale images including characters, has a simple circuit configuration, and is inexpensive. The purpose is to provide.

[Means for Solving the Problems] The image processing g M of the present invention reads image information on a document, outputs an electrical signal in pixel units from a document reading means, and outputs a signal output from the document reading means. is compared with a reference value, and a binarized signal obtained from the magnitude relationship is output from the binarizing means, and the binarized data output from the binarizing means is divided into at least one line of the image. 2.
A reference value is determined for a pixel to be subjected to digitization processing based on the binarized data of the pixel stored in the storage means constituting the neighborhood area of the target pixel and the binarized data of the pixel immediately before the target pixel. This is given to the value conversion means.

[Operation] As a result, according to the present invention, for a pixel to be binarized, a threshold value is determined using the binarized signal of a pixel that has already been processed in the vicinity of the target pixel, and this is Since the target pixel is binarized as a reference value, high-quality 2ffi conversion processing that is faithful to the density of the original image can be realized even in the case of a grayscale image containing characters or the like. In addition, a circuit that calculates the threshold value from a binary signal in the vicinity of the target pixel can be realized with a resistor and an analog switch. Since binarization can be realized by analog processing, the circuit configuration can also be simplified.

[Example] An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 shows the circuit configuration of the same embodiment. In the figure, 1 is a document, and on the first side of this document is an LED light source.
The light output from the array 2 is irradiated. Then, the reflected light from the first surface of the original is applied to the image sensor 4 via the SELFOC lens array 3. Here, a CCD is used as the image sensor 4, and the original 1
An analog signal corresponding to image information is output from the reflected light from the surface.

The analog output signal of the image sensor 4 is sent to a sample and hold section 6 via an amplifier 5.

Here, the amplifier 5 is for amplifying the analog output signal of the image sensor 4, and the sample hold section 6 removes reset noise etc. from the analog output signal amplified by the amplifier 5, and only necessary data is used. It is for taking out.

The output of the sample hold section 6 is given to one input terminal of an analog comparator 7. This comparator 7 is provided with the output of the nonlinear D/A converter 8 or a threshold reference voltage at the other input terminal, and outputs a binary signal based on the magnitude relationship of these inputs.

The output of the comparator 7 is applied to a previous line buffer 9 and a reference area window buffer 10. Here, as shown in FIG. 2, the front line buffer 9 stores a binary signal corresponding to one line of pixels constituting an image, and the binary signal corresponding to each pixel is It is designed to be shifted and inputted by clock CKI. Further, the reference area window buffer 10 stores two pixels a, b, and cd of four pixels a, b, and cd in the neighboring reference area shown by hatching with respect to the target pixel f shown in FIG.
Valued signals are written to the pixels a, b, and pixel stored in the previous line buffer 9 as shown in FIG.
The binarized signal of pixel c is given, and the comparator 7 also gives the binarized signal of the previous pixel d.

In this case, the binary signals of these pixels a, b, c, and d are
It is designed to be latched by clock CK2.

Then, the output of the reference region window buffer 10 has nonlinearity D
/A converter 8. This nonlinear D/A converter 8
generates different voltages according to the signal values of a, b, c, and d, and supplies this to the analog comparator 7 as a threshold voltage for binarization. This nonlinear D/A converter 8 has resistors RO, R
1, a first resistor group R5-R2 each having an analog switch in parallel, a second resistor group R2-R5 similarly having an analog switch, and resistors R6 and R7 are connected in series. configured, the first resistance group R5-R2
and the second resistor group R2-R5 are connected to the analog comparator 7.
is connected to the input terminal of

By applying the voltage V to the end of the resistor RO and grounding the end of the resistor R7, the divided voltage of the voltage V divided by these resistors is input to the comparator 7 as a threshold value. Further, the resistors RO and R7 are variable and set reference values for the black level and white level. Data of a, b, c, d of the reference area window buffer 10 is given as an open/close signal to the analog switch 5ll-814 provided in the first resistance group, and the analog switch 5ll-814 of the second resistance group
1-524 has an inverter V for a, b, c, d, respectively.
21-V24 as an opening/closing signal. a,
By the combination of rOJ and rlJ of b, c, and d, a threshold voltage based on the average concentration approximation method is input to the comparator 7.

Here, the average concentration approximation method will be briefly explained using FIG.

In the average density approximation method, the pixel to be binarized is f, and the pixels in the vicinity of f (for the previous line, the upper left, right upper, and upper right pixels are a, b, c, and the pixel immediately before the same line) d
This method looks at a small area made up of (2), and binarizes f at a point close to the average density considering the weight of each pixel in the small area, in order to obtain a feel close to the original image. a, b,
For each pixel c and d, at the time of processing f, it is already "
0 (white) or "1 (black)" has been determined, and if f is assumed to be "0", the average concentration of the small area is mO,
Letting the average concentration in ml when f is “1”, (mO+m
1) The idea is that if f is binarized using 72 as a threshold value, a more faithful reproduction of the original image can be achieved.

In Figure 3, a-1, b-2, c=1°d-2,
Each pixel is weighted as f-4. a, b, c
, d are all "1" and f is rOJ, then mo
-6/10, and assuming f is "1", m 1-
It becomes 10/10. Therefore, if a, b, c, and d are all black, the threshold value is set to 5h-8/10. In order to provide this threshold value to the analog comparator 7, a, b
, c, and d-1, the resistance value of each resistor of the nonlinear D/A converter 8 should satisfy the following equation.

R1+R2+R3+R4+R5+R610...(1) On the other hand, if a, b, c, or d are all "0", set f to "
If it is assumed that
Assuming that, it becomes ml-4/10.

Therefore, if a, b, c, and d are all white, the threshold s
h = 2/10. In order to provide this threshold value to the analog comparator 7, a, b, c.

Under the condition of d-0, the resistance value of each resistor of the nonlinear D/A converter 8 should satisfy the following equation.

...(2) Similarly, in the case of a, b, cml, d-0, a, c,
In the case of d-1, b-o, b, d-1, a.

In the case of c-0, m O = 4 / 10, m
Since 1-8/10 and s h = 6/10, the resistance value should satisfy the following equation.

Solving the above equation gives R1:R2:R3:R4:R5:R6 -2: 1 :2: 1 :2+2, and by setting the resistance value with this ratio, the nonlinear D/A converter 8 will generate thresholds according to the states of a, b, c, and d.

On the other hand, the 8 units of comparator 7 are serial/parallel 1m51
Sent to 1. This serial/parallel converter 11 is
It converts a binary serial signal into parallel. The output parallel-converted by the serial/parallel converter 11 is sent to the memory 12. This memory 12 stores the parallel-converted binary signal from the serial/parallel converter 11 as image data. The image data stored in the memory 12 is sent to a thermal head 14 via a thermal head drive section 13 and printed on a recording paper 15.

In addition, 16 in the figure is a control unit that controls the operation of each circuit described above, and here, it includes a sample hold unit 6, a previous line buffer 9, a reference area window buffer 10, a serial/parallel converter 11, a memory 12, and a thermal head. Drive N1
Control commands are given to each of the three.

Next, the operation of the embodiment configured as described above will be explained.

Now, when the first side of the original is irradiated with light from the LED array 2, the reflected light from the first side of the original is given to the image sensor 4 via the SELFOC lens array 3, and an analog signal corresponding to the image information is output. be done.

The analog output signal of the image sensor 4 is sent to the sample and hold unit 6 via the amplifier 5, where it is amplified, reset noise, etc. are removed, and only necessary data is extracted.

The output of the sample hold section 6 is given to one input terminal of an analog comparator 7. The other input terminal of this comparator 7 is provided with the previously explained output of the nonlinear D/A converter 8 as a reference voltage.

Now, if we consider a case where certain pixel information is binarized by the comparator 7, the binarized signals corresponding to pixels a, b, c, and d written to the reference area window buffer 10 at this point are non-linear. If the pixels are sent to the digital D/A converter 8, then the pixels a, b, c, d
If the binary signals of all of the digitized signals are black "1", the analog switches 811, 512, 513 . S14 is on, analog switch S21. Since S22, 52B, and S24 are turned off and the first resistance group R2 to R5 is shorted,
The base r4 voltage 8/ is applied as a threshold to the other input terminal of the comparator 7.
IOV is now given, and the input corresponding to the target pixel f given to one input terminal of the comparator 7 is 21ii! from the magnitude relationship with respect to the reference voltage. It will be converted into a signal and output.

On the other hand, when the binary signals of pixels a, b, c, and d are all white "0", the analog switches S11 and S12
．． S13. S14 is off, analog switch 52
1. S22. S23. S24 is turned on and now the second
Since the resistance group R2 to R5 becomes short-circuited, 2/IOV of the reference voltage is applied as a threshold to the other input terminal of the comparator 6, and the target pixel applied to one input terminal of the comparator 7 The input corresponding to f is converted into a binary signal and output based on the magnitude relationship with respect to the reference voltage.

If a, b, c are black and d is white, or a, c, d are black, b is white, or if a, d are black, a, C are white, then a, b, c in each case , d signals determine the short circuit location, both of which generate 6/IOV as a threshold, and f is binarized based on this.

After that, in FIG. 3, the processing progresses and the target pixel becomes f'
When moving to , pixel a / as the nearby reference area,
Binarized signals of b/, ct, and dt are used, and among these, pixels al, b L, c
The binarized signal of pixel t is written from the previous line buffer 9 to the reference area window buffer 10, and the binarized signal of pixel d' is written as
The binarized signal from the comparator 7 regarding the target pixel f described above is used and written into the reference area window buffer 1o, and in this state, the same operation as described above is performed.

The output from the comparator 7 is sent to the serial/parallel converter 11 according to instructions from the control unit 16, where the binary signal is converted to a parallel signal, and the memory 12
is memorized. Thereafter, when the image data in the memory 12 is read out according to instructions from the control section 16, it is sent to the thermal head 14 via the thermal head drive section 13, and is sent to the recording paper 14.
5 will be printed.

Therefore, in this way, for the pixel f to be processed, a non-linear analog switch Sll of the digital D/A converter 8;
S12,813°514 and S21, S22. S23
．． Since the on/off state of S24 is set and the threshold value is obtained with the corresponding resistance value, the binarized signal of the target pixel f determined based on this threshold value is the optimal one that is close to the density of the original image. Even in the case of grayscale images containing text, etc., it is possible to achieve high-quality binarization processing that is faithful to the density of the original image.

In addition, the non-linear D/A converter 8 that determines the threshold value of the target pixel f is composed of a resistor and an analog switch, and furthermore, the threshold value obtained from this converter 8 is an analog signal, and is based on this threshold value. Since the binarization of the target pixel can be realized by analog processing, the entire circuit configuration can be simplified and the price can be reduced.

Note that the present invention is not limited to the above-mentioned embodiments, but can be implemented with appropriate modifications within the scope without changing the gist.

For example, in the embodiment described above, a reference area is set with five pixels including the pixel to be processed, and a threshold value is determined based on the average density approximation method. However, in order to improve gradation expression, For example, prepare a line buffer for two lines, and
As shown in the figure, the reference area may be configured by a pixel n1 to be processed and 12 pixels a1 to p1 in the vicinity of the target pixel n1. In this case, as shown in FIG. 5, a previous line buffer 21 and a line buffer 22 are provided as line buffers, and the contents of these buffers 21 and 22 are sequentially applied to the reference area window buffer 23, and the target pixel n1 and 12 Pixels a1 to g1 constitute a reference area. Then, the contents of the reference region window buffer 23 are converted to nonlinearity D
/A converter 24, where a threshold value is determined, and this is applied to a comparator 25 as a reference voltage. Here, the nonlinear D/A converter 24 has resistors R2' to R13'
First series circuit of R1', RO' and resistors R2' to R1
3', R14', R15' second series circuit, analog switches S11' to S22'531' to S4
2' and inverters V31' to V42'. Moreover, in order to further improve the smoothness and gradation expression of the results of the above-described embodiments, a systematic dither circuit may be added to vary the threshold value.

[Effects of the Invention] The image processing device of the present invention reads image information on a document, outputs an electrical signal for each image from the document reading means, and uses the signal output from the document reading means as a reference value. A binarized signal obtained from the comparison and the magnitude relationship is outputted from a binarization means, and the binarized signal outputted from the binarization means is made to correspond to at least one line of pixels constituting the image. Then, regarding the pixel to be binarized, the binarized signal of the pixel stored in the storage means as the binarized signal of the pixel in the vicinity of the reference pixel and the binarized signal of the target pixel are stored in the storage means. Since the binarized signal of the immediately preceding pixel is taken in, and the threshold value is determined by the average density approximation method based on these binarized signals and given to the binarization means as a reference value, the binarization process is performed. Regarding the target pixel, it is possible to use the binarized signal of the already processed pixel in the vicinity of the target pixel to find the threshold value, and use this as the reference value to perform the binarization of the target pixel. Even in the case of grayscale images containing text, etc., it is possible to achieve high-quality binarization processing that is faithful to the density of the original image. In addition, a circuit that determines the threshold value based on the binarized signal in the vicinity of the target pixel can be realized with a resistor and an analog switch, and the threshold value obtained by this is an analog signal, and the target pixel based on this Since the binarization can be realized by analog processing, the circuit configuration can be simplified and the price can be reduced.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing a circuit configuration of an embodiment of the present invention, FIG. 2 is a block diagram showing a front line buffer and a reference window buffer used in the embodiment, and FIG. 3 is a block diagram showing the circuit configuration of an embodiment of the present invention. FIGS. 4 and 5 are diagrams for explaining the principle of the average density approximation method adopted in , and are diagrams for explaining other embodiments of the present invention. 1... Original, 2... LED array, 3... SELFOC lens array, 4... Image sensor, 5...
- Amplifier, 6... Sample hold section, 7... Comparator, 8... Non-linear D/A converter, 9... Previous line buffer, 10... Reference area window buffer, 11.・・・
Serial/parallel converter, 12... memory, 13.
- Thermal head drive unit, 14... thermal head,
] 5... Recording paper, 16... Control section.

Claims (1)

[Claims] Document reading means for reading image information on a document and outputting an electrical signal in pixel units; and comparing the electric signal output from the document reading means with a reference value to determine the magnitude relationship between the electrical signals. Binarization means for binarizing the electrical signal; Storage means for storing the binarized data output from the binarization means in correspondence with at least one line of pixels constituting an image; 2. Regarding the pixel to be the target pixel of the digitization process, based on the binarized data of the pixel stored in the storage means constituting the neighborhood area of the target pixel and the binarized data binarized immediately before the target pixel. An image processing apparatus comprising: a reference value generating means for determining a reference value and supplying the determined reference value to the binarizing means.