JPH0456508B2 - - Google Patents

Description

[Detailed description of the invention]

[Industrial Field of Application] The present invention provides an analog image signal obtained by reading an image with a reading device such as a CCD element, for example, black:
Regarding a binary processing device that converts into a binary image signal such as "1", white: "0", etc., it is possible to obtain a record with gradation, especially when reproduced and recorded with a recording device such as a printer or plotter. The present invention relates to a binarization processing device that also performs binarization processing. [Prior Art] For example, in an analog facsimile, since an analog image signal is transmitted, an image with gradation is reproduced on the receiving side. However, in a digital facsimile that transmits a binary signal after being encoded and compressed, images having gradation are usually not reproduced.
Therefore, even when reproducing a binarized image signal,
In order to create gradation, conventionally, storage devices such as ROM indicate threshold levels that are regularly or randomly distributed in a continuous or discontinuous manner in two dimensions in association with the image reading pixel divisions. Data (threshold level data) is stored, the read address of the storage device is determined based on the two-dimensional address of the incoming analog image signal, the threshold level data is read out, and this data is converted into an analog signal (threshold level data). Convert it to , compare the level of the analog image signal with this,
A binary image signal is obtained from the height relationship between the two. Japanese Patent Laid-Open No. 55-80958 discloses a first threshold with a rate of change of 1, in which m x n pieces of threshold level data with values from 0 to m x n-1 are assigned to an m x n pixel matrix. The level data group and the low-threshold level data are duplicated, and the high-threshold level data is deleted accordingly. Using the second threshold level data group where the rate of change is 0 in the high level area where the change occurs, the analog image signal is simultaneously binarized using the threshold level data of the two groups, and the data adjacent to the image area being converted is Compare the average values of the binary patterns of the areas to determine whether the image area being converted is an image edge area, and if it is an image edge area, select a binary signal based on the threshold level data of the second group. However, if this is not the case, an image edge enhancement processing technique is disclosed in which a binarized signal based on the first group of threshold level data is selected and output. [Problems to be Solved by the Invention] This type of binarization gradation processing expresses the density gradation of the entire m×n pixel area.
When the original image is a text image such as characters or illustrations, there is a problem in image processing that the reproduced recorded image becomes blurred. Furthermore, even if the original image is a halftone image, there are many variations in its gradation expression, such as an image with a lot of solid black, or an image with a narrow density distribution range in the middle density range but with subtle changes in density. , images may be pale in low-density areas, and with the above-mentioned binarization gradation processing, images with a lot of solid black will be further emphasized, and images in the intermediate density range will have subtle density differences disappear. There are problems such as when the image is put away or a pale image becomes an even fainter image. Therefore, it is possible to bias-shift the threshold level to high or low, or to divide the ROM threshold level data into several groups, and each group's threshold level data is the data of the other groups shifted to high or low. Conceivable. For example, when four pixels are classified into one section in both the main and sub-scanning directions, the first group of threshold level data as shown in FIG. 1a, and the threshold level data of the first group as shown in FIG. 2 groups of thread level data and 1c
As shown in the figure, the third group of threshold level data further shifted by 2 is stored in the ROM in advance, and one of these groups is selected. However, the rate of change in the thread level height array (1-16, 4-19, 6-21) of each group is the same, and as shown in Figure 1d, (1st group: Figure 1a), b ( 2nd group: Figure 1b) and c (3rd group: Figure 1c),
The density difference (contrast) remains unchanged. Also, since there is only a level shift between groups, depending on the combination of the image and the threshold level data selected for binarization, the gradation may change, such as light black areas disappearing or deep black areas becoming solid black. If the density becomes low, image reproducibility is impaired, and it is difficult to set the binarization density. According to the image edge enhancement processing technology disclosed in the above-mentioned Japanese Patent Application Laid-Open No. 55-80958, it is presumed that the blurring of the reproduced recorded image is reduced to some extent when the original image is a text image such as characters or illustrations. 2
This method causes more blurring, white spots, crushed blacks, etc. than in the case of digitalization, and is less effective for text images. In addition, when the original image is a halftone image with relatively clear image edges, the image edges are further emphasized by image edge enhancement processing, resulting in image processing with low fidelity due to the contrast between white and black gradations being too strong. I get used to it. The present invention provides a binarization processing device that can substantially eliminate blurring, etc. in image processing for text images, and can more faithfully reproduce the halftone expression of the original image for halftone images. The purpose is to provide. [Means for Solving the Problems] The binarization processing device of the present invention includes a processing density specifying means for specifying a binarization processing density, a plurality of threshold level data for simple binarization processing, and Each group consists of threshold level data in which the highest and lowest values are the same among the groups and the height arrangement is continuously distributed non-linearly,
a storage means for storing a plurality of threshold level data groups in which the conversion rate of the density represented by the binarized signal with respect to the level change of the analog image signal differs between the groups; and simple binarization processing and halftone processing. processing mode specifying means for selectively specifying a processing density; and when simple binarization processing is specified by the processing mode specifying means, a simple binary value corresponding to the processing density specified by the processing density specification means; reading out threshold level data for gradation processing from the storage means, and when halftone processing is designated by the processing mode designation means;
reading means for reading out from the storage means threshold level data of a group associated with the processing density designated by the processing density designating means; and the threshold level data read by the reading means and the incoming image signal. binarization processing means for comparing the levels of the image signals and binarizing the image signals; [Operation] When the original image is a text image, when the processing mode specifying means specifies simple binarization processing, and the processing density specification means specifies a desired binarization processing density that matches the original image, the readout means reads the threshold level data for simple binarization processing corresponding to the designated processing density from the storage means, and the binarization processing means receives the threshold level data read by the reading means. The levels of the image signals are compared and the image signals are binarized. As a result, an extremely clear text image without blurring etc. can be reproduced, and the density of the reproduced image can be selected by the processing density specifying means to increase the reproduced image density of a text image with a low density, or the density of a text image with a background stain can be increased. It is also possible to select options such as erasing dirt. When the original image is a halftone image, the processing mode specifying means specifies halftone processing, and the processing density specifying means,
When a desired processing density, that is, a halftone processing characteristic that matches the gradation expression of the original image is designated, the reading means reads out the threshold level data of the group associated with the designated processing density from the storage means; The binarization processing means binarizes the image signal by comparing the threshold level data read out by the reading means and the level of the incoming image signal. Therefore, depending on the density of the original, select a threshold level data group using the processing density specifying means to obtain the desired gradation and smooth gradation transition from the lowest to the highest gradation. , a binary image signal with good halftone reproducibility can be obtained. Since the thread level data of each group is distributed non-linearly and continuously, rich gradation expression can be obtained. for example,
From the midtone expression of high-density image enhancement, which has a low rate of change in the low-density area and high rate of change in the high-density area, to the intermediate tone expression of low-density image enhancement, which has a high rate of change in the low-density area and a low rate of change in the high-density area. Since it is possible to select a wide range of intermediate tone expressions, including tone expression, in which the gradation transition from the lowest to the highest tone is substantially continuous and smooth, the gradation expression is enriched. According to the present invention, it is possible to obtain binarized image signals with good reproduction characteristics that bring out the respective image characteristics for both text images and halftone images. In addition, it is possible to select, for example, to obtain an image that is clearer than the original image. Other objects and features of the present invention will become apparent from the following description of embodiments with reference to the drawings. [Embodiment] FIG. 4a shows an embodiment of the present invention. In FIG. 4a, the pixel addressing circuit ACC is composed of a sub-scanning counter CO1, a main-scanning counter CO2 and an inverter IN1. In this example, the image is read by a shuttle method, as shown in FIG. 4b, in which an image having a width of 32 bits (32 pixels) in the sub-scanning direction is read in one main scan. That is, an image of 32 pixels in the sub-scanning direction is read during movement of one pixel pitch in the main scanning. There are 4 key switches OP1 to OP on the scanning board.
4, and OP1 to OP3 are mechanically coupled so that one of them is closed and the other two are open, but key switch OP4 is independent of these. When the key switch OP4 is closed, the gradation control signal, that is, the mode instruction signal becomes L, and this L specifies gradation processing. When the key switch OP4 is open, the gradation control signal is H, which specifies simple binarization processing. Each of the key switches OP1 to OP3 specifies the concentration. When OP1 is closed, the concentration 1 instruction signal is L (density 1 specified), and when OP2 is closed, the concentration 2 instruction signal is L (density 2 specified). designation)
When OP3 is closed, the concentration 3 instruction signal is L (density 3 designation). The comparison level data for gradation processing is shown in a two-dimensional array as shown in FIG.
Bit (pixel), 8 bits (pixel) in sub-scanning direction y
64 pixels associated with each of the 8 x 8 = 64 pixels of
Setting density 1 for each group (OP4,
1 group (Figure 2a) used in OP1: Closed), density 2 settings (OP4, OP2: Closed)
There are 3 groups in total: 1 group (Figure 2b) used for setting density 3 (OP4, OP3: closed) and 1 group (Figure 2c) used for setting density 3 (OP4, OP3: closed), for a total of 64 x 3 pieces. is stored in. The first group uses the data shown in Figure 2a as the fourth
The data is applied to each of the four 4×4=16 matrix areas in figure c, and similarly, the second group has the data shown in figure 2b, and the third group has the data shown in figure 2c. It is an adapted form. The threshold level data of these groups (FIGS. 2a to 2c) have different rates of change in the level arrangement for each group, and the highest and lowest values are the same. That is, the first group of threshold level data has a minimum level of 20 and a maximum level of 200, as shown in FIG. 2a, and the height arrangement is the rate of change of a in FIG. 2d. The second group of thread level data has a minimum level of 20 and a maximum level of 200, as shown in FIG. 2b, and the height arrangement has a rate of change as shown in FIG. 2d, b. The third group of threshold level data has a minimum level of 20 and a maximum level of 200, as shown in FIG. 2c, and has a rate of change as shown in c in FIG. 2d. In addition, 1 of the 1st to 3rd groups
The height arrangement of the data may be such that the rate of change is shown in d of FIG. 2d. The analog image signal is now at the 100 level (Figure 3a)
, the reproduced image when binarized using the data group in Figure 2a selected when "Density 1" is specified is "Density 2" as shown in Figure 3B.
The reproduced image when binarized using the data group in Fig. 2b selected when specified is as shown in Fig. 3c, and the reproduced image in Fig. 2c selected when “density 3” is specified. The reproduced image when binarized using data groups is shown in Figure 3d.
Black and white will be distributed. In other words, the number of black dots is in the first group (density 1: Figure 2a)
9 based on the data of the second group (Concentration 2: Figure 2b), 7 based on the data of the third group (Concentration 3: Figure 2c), and 5 based on the data of the third group (Concentration 3: Figure 2c). Thus, images with three different densities are obtained. However, even if the density is different, the readable density range is the same, so the gradation level does not change, but the difference in density (contrast) changes. Therefore, it is possible to select a data group according to the density of the original and obtain a binarized image with good gradation and good halftone reproducibility. A group is specified according to the designated signal of density 1 to 3, and each data of that group is divided into 8×8=64 images.
In order to read out each pixel repeatedly and use it as a reference level, the sub-scanning counter CO1 counts up the sub-scanning pixel synchronization pulse CLOCK, and also in synchronization with the main-scanning pixel synchronization pulse CLOCK ENABLE (that is, the progress of one pixel in the main scanning In the end, the output of counter CO1 becomes 0, 1, 2, ... 7, 0, 1, in synchronization with the sub-scanning pixel synchronization pulse.
It cycles as 2,...7,0,.... counter CO
2 is reset once for every 8 main scanning pixel synchronization pulses, and its output is 0, 1, 2, ... 7, 0, 1, 2, ... 7, in synchronization with the main scanning pixel synchronization pulses.
It cycles as 0,... The ROM DMD2 stores 3 pieces of comparison level data for simple binarization processing in addition to the comparison level data (64 pieces x 3) for gradation processing mentioned above.
The relationship between input address data and read data is as shown in Table 1. As shown in Table 1, when the gradation control signal is L (gradation processing designation), density 1=L, density 2, 3
A ₆ of DMD2 input address data at =H
= 1, _{A 7} = 0, and _the gradation processing reference level data (2nd a
) is specified, and when density 2=L and density 1, 3=H, A ₆ =0 of the input address data of DMD2,
The reference level data for gradation processing (Figure 2b) of the group (64 pieces) associated with density 2, which is specified by A ₇ = 1, is specified, density 3 = L, density 1,
When 2=H, DMD2 input address data A ₆
Reference level data for gradation processing (FIG. 2c) of a group (64 pieces) associated with density 3, which is specified by A ₇ =1 and A 7 =1, is specified. When the gradation control signal is H, indicating no gradation control, the input address data of DMD2
Both A ₆ and A ₇ are "0", resulting in the input address and read data shown in the lower three columns of Table 1. Such address data is created by the aforementioned pixel addressing circuit ACC and address switching circuit AEC. The address switching circuit AEC is composed of a multiplexer MPL1, NOR gates NOR1 and NOR2, an inverter IN2, and AND gates AN1 to AN4, and the first set (1
A to 4A) 4-bit input terminal has a sub-scanning counter.
The 3-bit output of CO1 and the 1-bit output of the first digit of main scanning counter CO2 are applied. The 4-bit input terminals of the second set (1B to 4B) are connected to the outputs of NOR gates NOR1 and NOR2 and the ground potential (L=
"0") is applied, a gradation control signal (instructing gradation processing with L) is applied to the output selection terminal of multiplexer MPL1, and when it is L, multiplexer MPL1 selects the first group (A to 4A). The input of the second set (1B to 4B) is connected to the output terminal (1Y to 4Y) when the input is H. Output terminal of multiplexer MPL1 (1Y to 4Y)
is the 4 digits at the address input end of ROM DMD2

〔Effect of the invention〕

According to the present invention, it is possible to obtain binarized image signals with good reproduction characteristics that bring out the respective image characteristics for both text images and halftone images. In addition, it is possible to select, for example, to obtain an image that is clearer than the original image.

[Brief explanation of the drawing]

FIGS. 1a, 1b, and 1c are plan views showing various threshold level data for binarization gradation processing, and FIG. 1d is a graph showing the high-base arrangement of these data. Figures 2a, 2b, and 2c represent the ROM DMD 2 shown in Figure 4a.
FIG. 2d is a plan view showing thread level data groups stored in a two-dimensional array corresponding to image distribution, and FIG. 2d is a graph showing the height arrangement of each data. FIG. 3a is a plan view showing the analog image signal level in pixel units, and FIG. 3b,
Figures 3c and 3d are the three
When the image signal of the level shown in figure a is binarized,
FIG. 3 is a plan view showing black and white pixel distribution. FIG. 4a is a block diagram showing one embodiment of the present invention;
Figure b is a plan view showing the image capture scanning direction of this embodiment, and Figure 4c is the dither matrix of Figure 4a.
FIG. 2 is a plan view showing one group of threshold level data stored in the ROM DMD2. CPA...Comparator, OP1 to OP4...Key switch, ACC...Pixel address designation circuit, AEC...
...Address switching circuit, DMD2...Dither matrix ROM, AMP1 to AMP4...Amplifier,
DA...D/A converter, MPL...multiplexer.

Claims (1)

[Scope of Claims] 1. Processing density designation means for specifying binarization processing density, a plurality of threshold level data for simple binarization processing, and each group has a maximum value and a minimum value between the groups. It is the same thread level data with a non-linear continuous distribution of heights,
a storage means for storing a plurality of threshold level data groups in which the density change rate represented by the binarized signal with respect to the level change of the analog image signal differs between the groups; and simple binarization processing and halftone processing. processing mode specifying means for selectively specifying a processing density; and when simple binarization processing is specified by the processing mode specifying means, a simple binary value corresponding to the processing density specified by the processing density specification means; reading out threshold level data for gradation processing from the storage means, and when halftone processing is designated by the processing mode designation means;
reading means for reading out the threshold level data of the group associated with the processing density specified by the processing density specifying means from the storage means; and the threshold level data read by the reading means and the incoming image signal. A binarization processing device comprising: binarization processing means for binarizing the image signal by comparing the levels of the image signals.