JPS60206372A - Picture signal binary-coding system - Google Patents

Abstract

PURPOSE:To respond faithfully even to the variation of a dynanic ground density level, and to execute optimum binary-coding operation by executing the optimization of a threshold level from the dimension of a picture element unit. CONSTITUTION:A comparing circuit 101 compares density levels of two picture elements which are inputted simultaneously and continued in the main scanning direction, and detects an ascending and descending variation state of the density level. A feature extracting circuit 105 discriminates a picture element to become clearly black or white from this variation state, and generates a forced signal for showing black or white. A picture element selecting circuit 107 selects a density level of a picture element for which the forced signal is generated, from a buffer memory 108, and supplies it to a threshold operating circuit 111. The circuit 111 sets an initial threshold values as a threshold level as it is until the first forced signal is generated from the head picture element of each line, and after the first forced signal is generated, an arithmetical mean value of a density level of a picture element corresponding to the forced signal, and a density level of another picture element corresponding to the initial threshold value or the forced signal is fed as a threshold level to a binary-coding circuit 110, and the binary-coding circuit 110 binary codes a density level inputted through a buffer memory 109.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field] The present invention relates to an image signal binarization method for binarizing an image signal obtained by scanning an original with a pixel-resolving scanner.

[Prior art]

Conventionally, when converting a single image signal into a binary signal, a threshold level is obtained by dividing a white level peak hold voltage. This method generally uses the charging and discharging of a capacitor to obtain a peak hold voltage, so the degree of response is determined by the time constant. I only get responses.

Therefore, when the density distribution of the document changes dynamically, when focusing on a certain local area, the threshold level becomes extremely inappropriate, resulting in a disadvantage that the probability of a binarization error is quite large.

〔the purpose〕

The present invention has been made in order to eliminate the drawbacks of the prior art described above, and its purpose is to optimize the threshold level from the pixel-by-pixel dimension, thereby maintaining fidelity even to dynamic changes in the background density level. An object of the present invention is to provide an image signal binarization method that enables optimal binarization operation in response to the above.

〔Example〕

FIG. 1 shows a block configuration of an apparatus according to an embodiment of the present invention. A digital image signal (hereinafter simply referred to as an image signal) corresponding to the density of each pixel is input to an input terminal 100 from a scanner (not shown). This image signal is transmitted to the comparator circuit 101
The signal is applied directly to one input of the comparison circuit 101, and after being delayed by one pixel by the latch 102, it is applied to the other input of the comparison circuit 101. In other words, the density levels of two consecutive pixels in the main scanning direction, that is, in the moving direction of the pixel of interest, are
They are simultaneously input to the comparison circuit 101. Comparison mouth 1!8]
01 compares the density levels of two pixels and latch 10
If the density level of the pixel directly input is higher than the density level of the pixel input via 2 (increase in density), the code LL I II is output, and in the opposite case (decrease in density), the code LL I II is output. A code IL O11 is output, and no code is output when the density levels of two pixels are equal. Further, the comparison standard circuit 101 outputs the density level of the pixel directly inputted from the input terminal 100 as it is.

Reference numeral 103 denotes a code suppression circuit, to which the density level and code are input from the comparison circuit 101, and the prohibited area setting value is input through the input terminal 104.

In this embodiment, the input image signal is a 16-value signal from level 0 to level 15, and the prohibited area setting value specifies level 2 or lower and level 13 or higher. When the density level input from the comparison circuit 101 is below level 2 or above level 13, the code suppression circuit 103 performs
The code from comparison circuit 101 is suppressed. If the density level is level 3 or higher or level 12 or lower, the code input from the comparison circuit 101 is output as is.

The reason for performing such code suppression will be described later.

Therefore, when an image signal having a density level as shown in the upper part of FIG. 2 is input, the code suppression circuit 103 outputs a code to each pixel as shown in the middle part of the figure. Note that the x symbol in the figure indicates no code.

Reference numeral 105 denotes a feature extraction circuit, which extracts the characteristics of changes in the density level of successive pixels from the code string inputted from 03, and discriminates pixels that should clearly be black or white. , generates a forced signal that clearly indicates black or white.

31 In this embodiment, the feature extraction circuit 105 considers four pieces of code data as one block while shifting the sequentially input code data one by one.

Then, there are three or more codes “1” in each block,
If code par 0''' is not mixed, a forced signal W is generated to treat the pixel corresponding to the first code data of the block as a white pixel, and the pixel corresponding to the last code data of the block is treated as a black pixel. Generates a forcing signal B that is regarded as .

Also, if the block contains three or more codes II O11 and no codes II I 11 are mixed.

A forcing signal B is generated for the pixel corresponding to the first code data of the block, and a forcing signal W is generated for the pixel corresponding to the last code data of the block.

If the first code data of the block is 1°t and all the remaining code data are ``0'', then a forcing signal B is generated for the first pixel of the block, and conversely, the first code of the block If the data is "0"' and the remaining code data is #] II, a forcing signal W is generated for the fourth to first pixel of the block.

For the input image signal in FIG. 2, a forced signal is generated as shown in the lower part of the figure.

As described above, since the code data is divided into blocks while being shifted one by one, forced signals B and W may be generated in duplicate for one pixel, as shown by the arrows in the lower part of Figure 2. be. In this case, the forced signal for that pixel needs to be suppressed.

A means for suppressing this forced signal is the forced signal suppression circuit 106. The forced signal is input to the pixel selection circuit 107 via the forced signal suppression circuit 106.

Here, the reason for providing the code suppression circuit 103 will be explained. Considering both 4-bit (16 gradation) signals,
Around level 2 is considered to be the background level of the original, but forced signal B may be generated for that pixel (level 2 is the lowest density level at which forced signal B can be generated).
. However, in practice, it is common that it cannot be determined that level 2 is clearly a black pixel, and therefore it is necessary to place a limit on the density level of the pixel to which the forced signal can be applied.

Therefore, in this embodiment, for pixels of level 2 or lower, the code II
We are trying to prevent 11 from occurring.

Similarly, a forced signal W may be generated for a pixel at level 13 (level 13 is the maximum density level at which the forced signal W can be generated).9 For such pixels at level 13 or higher, In order not to generate the forced signal W, the code II OH for pixels of level 13 or higher is suppressed in the code suppression circuit 103.

Note that the code suppression circuit +03 can be omitted by appropriately selecting the reference voltage of the analog/digital converter for the image signal. That is, of the two reference voltages of the analog/digital converter, the lower level reference voltage is set higher than normal, so that the image signal, which is normally quantized to about level 2, is quantized to level 0, Also, the high-level reference voltage is set lower than usual, so that an image signal that is normally quantized to level 13 is quantized to level 15. It is clear that if this is done, the same effect as provided by the code suppression circuit 103 can be expected. moreover,
Since the background level and high-frequency noise level are substantially removed and the effective density range is quantized more finely, it is also possible to obtain the effect that the density level waveform can be grasped more accurately.

Referring again to FIG. 2, the image signal input to the input terminal 100 is temporarily stored in the buffer memory 108, and after being delayed for a certain period of time by the buffer memory 109, it is input to one input of the binarization circuit 110. Given. The pixel selection circuit 107 selects the density levels of the pixels in which the forced signal B and the forced signal W are generated in the papier memory 1 in a predetermined order.
08 and supplies it to the threshold calculation circuit 111.

An initial threshold value is applied to this threshold value calculation circuit Ill via an input terminal 112.

This initial threshold value can be set to a fixed value, but in this embodiment, it is determined from the peak hold voltage of the white level of the image signal, as in the prior art. and,
The time constant is determined so that the peak hold sample region 7- is the region near the beginning of the line.

The threshold calculation circuit Ill applies the initial threshold value as it is to the other input of the binarization circuit 110 during the period from the first pixel of each line until the first forced signal B or W is generated. After the first forced signal is generated, the binarization circuit uses the arithmetic average value of the density level of the pixel corresponding to the forced signal and the initial darkness value or the density level of another pixel corresponding to the forced signal as the threshold level. 110 to the other input.

The binarization circuit 11O binarizes the density level input via the buffer memory 109 using the dark level provided by the threshold calculation circuit 111, and outputs a binarized signal to the output terminal 112.

A specific example of the binarization operation described above will be explained with reference to FIG. Assume that an image signal whose density changes from the beginning of the line as shown by the solid line in (a) of the figure is input. In the case of this line, the output side of the code suppression circuit 103 is as shown in (a) in the figure.
The code string shown directly below occurs. Note that the x symbol is 8- without a code. For this code string, a forced signal as shown in (b) of the figure is sent out from the forced signal suppression circuit 106.

At the time when the density levels of pixels before the first pixel a where a forced signal is generated are input to the binarization circuit 110,
The threshold calculation circuit 111 supplies the initial threshold TH, as it is, to the binarization circuit 110.

Therefore, for those pixels, the initial threshold TH,
It is binarized by

When the density level of pixel a is input to the binarization circuit 110, the density level of pixel a is read from the buffer memory 108 and input to the threshold calculation circuit 111. Then, the arithmetic mean value of the initial threshold value THo and the density level of the image number a is determined by the threshold value calculation circuit 111 and provided to the binarization circuit 110.

Since the forced signal B opposite to that of pixel a is generated for the pixel two pixels after pixel a, when the density level of the next pixel after pixel a is input to the binarization circuit 110, it is different from that of pixel a. The density level of the pixel is read from the buffer memory 108. The threshold calculation circuit 111 calculates the arithmetic average value of the density levels of the pixels a and b, which is set as the threshold value 2.
It is applied to the value converting circuit 110.

For the next pixel C, the opposite forced signal W is applied to the next pixel C.
Therefore, when the density level after the pixel is input to the binarization circuit 110, the pixel

The density level of C is read from the buffer memory 108. The threshold calculation circuit 111 calculates the pixel value.

The arithmetic average value of the density levels of C is taken as a threshold value, and is given to the binarization circuit 110.

Since the opposite forced signal B is generated for the pixel e two pixels after the pixel C, the density level after the pixel d is the same as that of the binarization circuit 1.
10, the density level of pixel C1e is read out from buffer memory 108. The arithmetic average value of the density levels of those pixels is used as a threshold value in the dark value calculation circuit 111.
and is applied to the binarization circuit 110.

Similarly, when the density level of pixel e is input to the binarization circuit 110, the arithmetic average value of the density levels of pixels d and e is given to the binarization circuit 110 as a threshold value. Since a forced signal W opposite to that of pixel f is generated at pixel g, which is two pixels after pixel f, when the density levels from 1 pixel f onwards are input to the binarization circuit 110, the density levels of 1 pixel f and g are The pixels are read out from the buffer memory 108, and the arithmetic average value of the density levels of those pixels is given to the binarization circuit 110 as a threshold value.

In this way, the threshold level is dynamically set as shown by the dotted line in FIG. 1(a). Therefore, for the image signal of this example, a binarized signal shown in (c) of the figure is obtained. In this embodiment, the binarization circuit 110 binarizes the density level above the threshold into black.

(d) of the figure shows a binary signal obtained when the image signal is binarized using a conventional fixed threshold value T.

As mentioned above, the two pixels (
Since the threshold level is set between the density levels of clearly black pixels and clearly white pixels, even if the density level amplitude band of the effective pattern changes, it will always be possible to slice within that band at an appropriate level. Since the image signal is binarized, an optimal binarized signal can be obtained even when the density distribution of the document changes dynamically.

In the above embodiment, the feature extraction circuit 105 divides four consecutive pixels into one block, but the number of pixels included in a block is not limited to four, but may be 3, 5, 6, etc.

For example, when dividing blocks into blocks every three pixels, if two identical codes are included, they are subject to feature extraction; in that case, the code '11+1. Even if #Q11 exists, it will be considered. Specifically, when the code "'1"("O") exists, the forced signal W (B) is given to the leftmost pixel. Then, both forced signals B and W are given to one pixel. In this case, the forced signal for that pixel is suppressed, as in the case of dividing every four pixels into blocks.

Further, in the above embodiment, the arithmetic average of the density levels of two pixels is calculated and used as the threshold value, but the threshold value can also be determined using other calculation methods.

12-Furthermore, the prohibited area in which code suppression is performed may be changed or made variable as appropriate.

The comparison direction for giving the code can be the sub-scanning direction, and the algorithm can also be changed so that isolated points are removed in combination with the main-scanning direction.

Additionally, the present invention can also be implemented using a microcomputer.

〔effect〕

As explained above, according to the present invention, the threshold value is optimized from the pixel-by-pixel dimension so as to appropriately slice each effective pattern, and the dark value can faithfully respond to dynamic background level fluctuations. Therefore, it is possible to obtain a more effective and appropriate binarized signal than in the past.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a device according to an embodiment of the present invention, FIG. 2 is a conceptual diagram for explaining the operation of the device in FIG. 1, and FIG. 3 is a detailed diagram of the device in FIG. It is a waveform diagram for explanation. 100... Image signal input terminal, lot... Comparison circuit,
102... Latch, 103... Code inhibit circuit. 105...Feature extraction circuit, 106...Forced signal suppression circuit, 107...Pixel selection circuit, 108°109...
・Buffer memory, ■IO...Binarization circuit, 111・
...Threshold calculation circuit, 112...Binarized signal output terminal.

Claims (1)

[Claims] (1) In the image signal binarization method that binarizes the image signal,
It detects the rise and fall of the density level for each pixel in the direction of movement of the pixel of interest, determines which pixels should clearly become black or white based on the change, and sets the threshold level for binarizing the single pixel signal mutually. An image signal binarization method that dynamically sets the density level to a level between the density level of a pixel that should be black and the density level of a pixel that should be white that are close together.