JPH07226840A - Binarization device - Google Patents

Abstract

PURPOSE:To prevent the occurrence of a notch without an abuse such as the thinning and the collapse of a fine line and to prevent the increase of code quantity at the time of compressing/encoding information quantity after binarization. CONSTITUTION:An contour judgement means 15 for detecting a part where the notch occurs is constituted of a means for obtaining a binary threshold from the density levels of plural picture element areas with an attentional picture element as a center, a means for binarizing the plural picture element areas by the threshold and a means for detecting whether the binary pattern is a specified pattern or not. Furthermore, a means 16 for selecting a slice level binarizing the attentional picture element by referring to adjacent binary data in vertical direction or lateral direction which are previously binarized by the signal sent from the means 15 is provided.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for converting multi-valued image information into binary image information.

[0002]

2. Description of the Related Art In a conventional facsimile apparatus, read multi-valued image information is subjected to image processing or the like to be binarized by a certain binarization threshold value and converted into binary image information. If the density level of image data is unstable near the binarization threshold due to the influence of noise, etc., the binarization result becomes discontinuous (hereinafter referred to as notch), which causes the problem of image quality deterioration. there were.

Therefore, conventionally, as described in JP-A-63-233673, the threshold value is not fixed at the time of binarization, but the threshold value is processed by the previous pixel. The binarization is performed by providing a hysteresis characteristic having a predetermined fluctuation width that makes it easy to hold the value of the binarized data that has been converted, thereby preventing notches and improving the image quality. .

For example, if the binarized data of the previous pixel is white, the threshold value for the pixel of interest is set to a value such that the pixel of interest tends to be white, and conversely if the binarized data of the preceding pixel is black, the pixel of interest is selected. The threshold value for is set to a value such that the pixel of interest tends to become black, so that the occurrence of notches can be prevented to some extent even when the density level is unstable near the threshold value.

[0005]

However, JP-A-63-23
In the method described in Japanese Patent No. 3673, the notch generated in the vicinity of the threshold value where the density level of the image is greatly changed is
This can be prevented to some extent, but in areas such as thin lines where changes in density level are small, changes in density level are absorbed by the hysteresis characteristics, and blurring occurs or white between characters becomes black on both sides. There was a bad effect of doing.

An object of the present invention is to provide a binarization device which prevents notches and which does not cause the thin lines to be blurred or crushed.
This is not only to improve the image quality, but also to prevent an increase in the code amount when the information amount compression encoding is performed after the binarization.

[0007]

According to the present invention, as means for detecting a portion where a notch occurs (contour determination means), means for obtaining a binarization threshold value from the gray level of a plurality of pixel areas centering on a pixel of interest. And a means for binarizing a plurality of pixel regions with the threshold value and a means for detecting whether or not the binarized pattern is a specific pattern. A unit for selecting a slice level for binarizing the pixel of interest is provided by referring to the binarized data of the already binarized pixels that are adjacent in the vertical or vertical direction.

[0008]

In the contour determining means of the present invention, a binarization threshold value is obtained from the gray level of a plurality of pixel areas centering on the pixel of interest, and the plurality of pixel areas are binarized by the threshold value to obtain a plurality of pixels. The portion where the change in the density level is large in the area is detected. Further, in order to detect only the horizontal and vertical contours of the image in which the notches are conspicuous, it is detected whether or not the pattern obtained by binarizing a plurality of pixel regions centered on the pixel of interest is a specific pattern. Then, in the threshold value selecting means, in the case of the contour portion in the horizontal direction, the pixel of interest selects the binarized slice level that tends to be the same value as the binarized data of the already binarized pixel that is laterally adjacent, In the case of the contour portion in the vertical direction, by selecting the binarized slice level that tends to be the same value as the binarized data of the already binarized pixel in which the pixel of interest is vertically adjacent, the error in the vicinity of the threshold value is selected. Even at a stable density level, the continuity of binary data can be maintained and the occurrence of notches can be prevented. Further, since the notch removal processing is performed only on the portion detected by the contour determination means, it is possible to obtain a high-quality image with almost no adverse effects such as thin line blurring and crushing.

[0009]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows a schematic construction of a facsimile apparatus of the present invention, and FIG. 2 shows a construction of an image reading section of the present invention. Hereinafter, the present invention will be described with reference to FIGS. 1 and 2.

As shown in FIGS. 1 and 2, the flow of the image signal of the facsimile apparatus in the present invention is as shown in FIG.
First, the light source 11 and the line sensor 12 that irradiates the original 10 with the light from the light source 11 and detects the reflected light reads the object to be read, and the A / D converter 13 converts it into a digital signal. The result processed by the binarization processing unit is sent to the coding unit, coded by the coding unit and sent to the communication line through the modem 5. On the contrary, the signal sent from the communication line is decoded by the decoding unit 4 via the modem 5 and printed by the recording unit 3. Note that the recording unit has almost the same structure as the reading unit, except that the flow of signals is opposite.

As shown in FIG. 2, after the read image signal is A / D converted, the gamma correction unit 14 performs gradation correction suitable for characters and fine lines, and the contour determination unit 15 outputs the image signal horizontally and vertically. It is detected whether or not it is located in the contour portion of the, and the threshold value selecting unit 16 selects a binarization threshold value according to the detection result. Then, by using the binarization threshold value, the image signal edge-enhanced by the one edge enhancement unit 15 is converted into the binarization unit 1.
It is binarized in 8 and output to the encoding unit.

Next, the details of the contour determination unit 15 will be described with reference to FIG.
Will be explained.

As shown in FIG. 3, for the signal input to the contour determining section 15, first, the average value calculating section 151 calculates the average value of the density levels of the peripheral pixels centering on the pixel of interest. Then, the binarizing unit 152 binarizes the peripheral pixels around the pixel of interest with the value of 3/4 of the obtained average value as the threshold value. That is, among the peripheral pixels, the pixel having a large change in the density level is obtained by comparing with the average value of the density levels of the peripheral pixels. Next, the pattern detection unit 153 detects whether the pixel of interest is located in the horizontal contour portion or in the vertical contour portion. Even if the notch occurs in the diagonal line, it is not very noticeable and does not lead to deterioration in image quality, but the notch that occurs in the vertical line and horizontal line is very noticeable, so only the vertical line and horizontal line are detected, and the vertical line is detected. ， The process of removing the notch is performed only for the horizontal line. As described above, by performing the processing for removing the notch only in the case of the vertical line and the horizontal line in which the change in the density level is large, it is possible to reduce the adverse effects of the notch removal.

An embodiment of the processing of the present invention will be described below with reference to FIGS.

FIG. 4 shows a detailed configuration example of the average value calculation unit 151 of FIG. As shown in the figure, in this example, a pixel area of 3 pixels in the main scanning direction × 3 pixels in the sub scanning direction is used, and the average value of the density information of A1 to A8 of the peripheral pixels excluding the pixel of interest A is obtained. Then, the average value of the density information of the peripheral pixels is output as 151a, and the density information of the peripheral pixels is output as 151b.

The output of the average value calculation unit 151 is the binarization unit 15.
Entered in 2. FIG. 5 shows the details of the binarization unit 152.
The average value 151a of the density information of the peripheral pixels is multiplied by 3/4, and the comparator 1521 binarizes the pixel area of 3 pixels in the main scanning direction × 3 pixels in the sub scanning direction by using the value as a threshold value. When it is larger than the threshold value, it is black (“1”), and otherwise it is white (“0”).

The pattern detecting unit 153 has a binarizing unit 152.
It is detected by pattern matching whether or not the output result of 1) applies to the contour pattern of FIG. 6, and "1" is output only in the case of the contour pattern, and "0" is output in other cases. Further, the signals are separately output depending on the case of the horizontal contour pattern and the case of the vertical contour pattern. Then, a signal whose output is "1" in the case of a horizontal contour pattern is a horizontal edge signal, and an output is "1" in the case of a vertical contour pattern.
Signal is a vertical edge signal. These two signals are
When both are "0", the notch removal processing is not performed. Like the contour pattern of FIG. 6, pixels (black) darker than 3/4 of the average value of the surrounding pixels are continuous in the horizontal and vertical directions,
In addition, when all the pixels on one side with a line continuous in the horizontal direction and the vertical direction as a boundary have pixels thinner than 3/4 of the average value of the peripheral pixels, it is determined as a horizontal or vertical contour.

The contour selection unit 15 is included in the threshold selection processing 16.
The output of is input. The algorithm of the threshold value selection processing 16 is shown in FIG. When the horizontal edge signal from the contour determination unit 15 is “1”, the binarized slice level is set so as to easily become the same value as the binarized data of the already adjacent binarized pixels in the horizontal direction. . That is, if the horizontally adjacent pixels among the slice levels shown in FIG. 8 are white,
The slice level is set to th1, and when it is black, it is set to th2, which easily causes black. When the vertical edge signal from the contour determination unit 15 is “1”, the same idea as the horizontal edge is used.
The binarized slice level is set so that the binarized data is likely to be the same as the binarized data of the already binarized pixels that are adjacent in the vertical direction. Of the slice levels shown in FIG. 8, when the vertically adjacent pixel is white, the slice level is th1, and when it is black, it is th2. However, the horizontal edge signal,
If either of the vertical edge signals is "1" and the binarized data of the already binarized pixel is white in both the horizontal and vertical directions, it is highly possible that it is a thin black line, The standard level is th3. Also, when both the horizontal edge signal and the vertical edge signal are "0", the slice level is standard and is th3.

The gamma correction unit 14 can be realized by a RAM table, and a gamma curve suitable for binarization processing or contour determination can be written in the RAM table.

FIG. 9 shows an embodiment of the edge emphasis processing section 17. The filter coefficient of the filter unit 171 is obtained by using the input signal at the present time and the input with the 1-line delay 172 in the pixel area formed by the signal from the gamma correction 14 in the main scanning direction 3 pixels × the sub scanning direction 2 pixels. To perform convolution calculation. The 1-line delay 172 uses a line memory. In this case, it is a filter coefficient that is emphasized in order to display characters and thin lines.

The image signal that has passed through the edge enhancement processing section 17 is input to the binarization section 18. An embodiment of the binarization unit 18 is shown in FIG.
It shows in 0. The image signal is binarized by the binarized slice level 1561 selected by the threshold selection unit 16.

Further, other embodiments and modified examples will be described. In the above embodiment, the reading section of the facsimile is described, but it can be applied to a scanner device or a copying machine.

[0023]

According to the present invention, the occurrence of notches can be prevented with almost no adverse effects such as thin lines blurring or crushing, and the amount of codes at the time of binarized information amount compression coding can be prevented. It is possible to prevent the increase of.

[Brief description of drawings]

FIG. 1 is a block diagram of a facsimile device.

FIG. 2 is a block diagram of a reading unit and notch removal binarization of a facsimile device.

FIG. 3 is a block diagram of a contour determination unit according to the present invention.

FIG. 4 is an explanatory diagram showing an example of an average value calculation unit.

FIG. 5 is an explanatory diagram showing an embodiment of a binarization unit.

FIG. 6 is an explanatory diagram showing an example of a contour pattern.

FIG. 7 is an explanatory diagram showing an example of an algorithm of a threshold selection unit.

FIG. 8 is an explanatory diagram showing an example of threshold values.

FIG. 9 is an explanatory diagram showing an example of edge enhancement.

FIG. 10 is an explanatory diagram showing an embodiment of a binarization unit.

[Explanation of symbols]

1 ... reading unit, 2 ... encoding unit, 3 ... recording unit, 4 ... decoding unit, 5 ... modem.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical indication H04N 1/409 G06F 15/68 350 H04N 1/40 101 C

Claims (3)

[Claims] 1. An image processing apparatus for binarizing image data, wherein a binarization threshold value is obtained from a gray level of a plurality of pixel areas centering on a target pixel, and the plurality of pixel areas centering on the target pixel are set. Means to binarize with the threshold,
It is characterized by comprising means for detecting whether or not the binarization result is a specific pattern, and means for changing the binarization threshold value according to the pattern when the binary pattern is matched with the specific pattern. Binarizer. 2. The means for selecting a threshold value for binarizing the pixel of interest according to claim 1, being adjacent to the determination result from the detecting means in the already binarized vertical and horizontal directions. A binarization device that selects a threshold value according to binarized data of a pixel. 3. The binarization threshold selecting means according to claim 1, wherein the binarization has already been completed when the pixel of interest is determined to be located in a lateral contour portion from the detecting means. When a threshold value that tends to be the same value as the binarized data of the pixels adjacent in the horizontal direction is selected and it is determined by the detection means that the pixel is located in the vertical contour portion, the adjacent pixels are already binarized in the vertical direction. Binarization device that selects a threshold value that tends to be the same value as the binarized data of the pixel to be processed.