JPH03186061A - Picture binarizing processor - Google Patents

Abstract

PURPOSE:To prevent picture quality from being degraded at the time of reading a regular density picture or a fixed density picture by changing binarizing threshold value data to be compared with a current picture signal at random within a prescribed range. CONSTITUTION:Line scanning is executed to a half tone picture and picture signals to be inputted are converted to digital data, and compared with the binarizing threshold value data to be supplied form a ROM 13 by a comparator 12 and binarizing data are formed to the current picture signal and outputted. Further, these data are temporarily stored and fed back as the binarizing data in the past in a binarizing processing and a prescribed weighting processing is executed so as to prepare the binarizing threshold value data to the current input picture signal. Random digital data from a random digital data generator 14 selects the binarizing threshold value data with slightly various density levels prepared in the ROM 13 at random. Thus, when reading the regular density picture or the fixed density picture, the picture quality can be prevented form being degraded.

Description

[Detailed Description of the Invention] [Object of the Invention] (Industrial Field of Application) The present invention provides an image 21ifl conversion device useful when reproducing images with halftone levels using a binary recording device such as a dot printer. Regarding.

(Prior art) Recently, simple 2! Il! 2. Description of the Related Art Recording devices are used to display and output images having halftone levels, such as photographic images. Since this type of recording apparatus can only display binary images of black and white, it is necessary to effectively convert the halftone image to 21II. This 2
Regarding value processing, for example, in photographs published in newspapers, etc., a screen is used to create shading effects by utilizing the mesh size, but it is extremely difficult to apply this method directly to dot printers, etc. Have difficulty.

Therefore, in the past, attempts have been made to binarize halftone images by typically introducing binarization processing such as a textured dither method or a minimum average error method. In the tissue dither method, an appropriate threshold value is determined in advance according to the pixel position, and pixel signals are discriminated using this.
Although this is achieved by a simple process of converting into values, it has the disadvantage that the image quality is generally poorer than the above-mentioned minimum average error method.

On the other hand, the minimum average error method determines binarization so that the average difference between the binarized image and the original image within a small area made up of multiple pixels is small, and the image quality is sufficiently improved. However, it has drawbacks such as the binarization process being complicated and requiring a large-scale system configuration.

Considering these city conditions, we first published JP-A No. 57-104369.
In this issue, we introduce a 26fl halftone image conversion device called the average density approximation method, which can obtain a binarized image with the same quality as that obtained by the minimum average error method and can be realized at low cost with a simple hardware configuration. is proposed. This method performs binarization based on the area ratio occupied by black dots in a small area, that is, the average density. This average density is calculated for the dot that will be binarized from now on and all the dots around it that have been binarized, and the black and white of the dot that will be binarized from now on is determined by the average density when it is black and white. This is done by calculating the average density at the time of , checking which one is closer to the input signal, and selecting the closer one.

This method will be further explained with reference to FIGS. 3 and 4.

Figure 3 shows the relationship between intermediate and fourth images 1 and pixel 2, which is obtained by raster scanning the intermediate image &1 and separating the pixels by sampling. b) shows a part of it enlarged. Now, if the input double image is r (1, j) and the binarized data in this double one quadrant position (1, j) is defined as g (1, J), it can be expressed as follows. In other words, the input 1iTii image r(Lj) at both image positions (1, j) is the normalized 0~
The level of 1 is set to H, and the binarized data g is set to this level.
It shows that (1, J) takes 0 or 1. Now, in FIG. 3(b), it is assumed that the pixel 2a indicated by diagonal lines is the data position to be processed, and that the pixel 2b has already completed the binarization process and the binarized data has been determined. . Note that 2C indicates an unprocessed pixel. 1. The pixel 2a shown in FIG. 3(b).

2b and 2c are the extracted areas around the pixel 2a to be processed, and the set of pixels 2b for which processing has been completed is Ql, and the set consisting of this set Q and pixel 2a to be processed is Q
′, the algorithm for binarization processing using this method is shown as follows.

g(M,N)-Flr(M,N)1g(1,j)[(1
, j) εQ]l... (2) However, the coordinates (M, N) indicate the pixel position to be processed, and F is a function. In other words, the binarized data g(M,
N) is based on the input image signal data r(M, N) at the same pixel position and the data g(1, j) that has already been converted into 2tn of the neighboring pixel value if (1, j) included in the set Q. , determined by some weighting function F. This function F is, for example... (3) x-tΣα(M-1,N-j)l・f(M,N)(f,
j)εQ.

This is shown as However, α(1, J) is a weighting function having weights as shown in FIG. 4, for example. As shown in this algorithm, by simply predetermining a weighting function, the binarized data for the above image signal data can be easily created from the past binarized data and the input image signal data at the current coordinate position. , this can be output. Moreover, since all the past data required for this binarization process is two-strand data, it is easy to accumulate and store it.

However, according to the image binarization processing device that adopts the above method, when reading artificially created standard gray scale images or constant density images,
Since the regular reading level continues, the output dot pattern also becomes the same, resulting in a problem of deterioration of image quality.

(Problems to be Solved by the Invention) As described above, the above device has the problem that image quality deteriorates when reading artificially created regular gradation images or constant density images. was there. Therefore, the present invention alleviates this problem and provides an image binarization processing device that does not cause image quality deterioration even when reading an artificially created standard IQ grayscale image or an image with a constant density. The purpose is to provide.

[Structure of the Invention] (Means for Solving the Problems) This invention provides means for inputting an image signal obtained by raster scanning an image, and a means for inputting an image signal obtained by raster scanning an image, and a means for inputting an image signal obtained by raster scanning an image, means for creating binarized threshold data for the current input image signal by performing a predetermined minimization process on the small amount of past 211m data converted to 2f11; Means for forming binary data for the current image signal by comparing with the signal, and when outputting this binary data, it is stored and fed back as binary data of errors in the binarization process. The double-image binarization processing device includes means for randomly changing binarization threshold data to be compared with the current image signal within a predetermined range.

(Function) In this invention, by randomly changing the binarized threshold data to be compared with the current image signal within a predetermined range, the binarized threshold data can be made minute in the range of the two-image level that humans cannot perceive. 11 to prevent the standard binarization pattern from continuing. As a result, it is possible to provide an image binarization processing device that does not cause image quality deterioration even when reading a standard gray-scale image created manually or an image with a constant density. .

(Embodiments) Hereinafter, embodiments of the present invention will be described with reference to the drawings.

First, the principle of this invention will be explained. In the method disclosed in JP-A-57-104369 described above, as is clear from the above-mentioned equation (3), the above-mentioned value X is obtained, and its mx and coefficient α(0,0) are Binarized data g(M, N) is obtained by comparing. In other words, (4) was determined 11 times. Therefore, if we pay attention to this equation (4), we can obtain the following equation (5), and if we set Lami 11, we can directly obtain the binarized data g(M, N). In this way, the 71st side of the above equation is expressed as a function of past data g(i, j).

By the way, according to this method, if there is regularity in the read halftone level or if a uniform constant density continues, the left and right sides of Equation 2 will take values that are very close to each other. Therefore, similar patterns appear repeatedly, making them noticeable in images and leading to deterioration in image quality.

Therefore, in the present invention, the bottom of the groove is randomly added or subtracted from the minute density level δ as shown in the following equation (7) to the right side of the above equation (6). With this configuration, the above relationship collapses, the same pattern does not appear repeatedly, and no deterioration in image quality occurs.

... (7) Here, ... (8) α(M-1, N-J) is the condition for the minimum value to be obtained.

FIG. 3 is a block diagram 4j of an embodiment of an image binarization processing apparatus according to the present invention that implements the above algorithm. An image signal inputted by line-scanning a halftone image is converted into n-bit digital data via, for example, an A/D converter 11, and then captured. This n-bit image signal data is stored in a ROM (described later) in a comparator 12.
It is compared with the n-bi-soto binarized threshold data supplied from the read-only memory 13, and the result is output as binarized digital data corresponding to the image signal. Needless to say (A/D converter 11 and comparator 12
If necessary, an image correction circuit such as a shading correction circuit or an automatic gain correction circuit may be provided between the two. The ROMI 3 receives 12-bit binary digital data, which will be described later, and m-bit random digital data from the random digital data generator 14 as address data, and receives n-bit data from the ROM 13 based on the algorithm described earlier. The binarized threshold data of is output.

The binary 1-bit data thus output from the comparator 12 is supplied to a predetermined output device W (not shown).
are delayed by two sampling times through the shift register 15, and then sequentially transferred to the shift registers 16, 17, 18, .
19. The shift registers 16 and 17 are configured in series to delay the above-mentioned 1-bit data by 1 line scanning period, and from each tap of the shift register 17 in the subsequent stage, 1 bit 2 at an adjacent position of, for example, 5 samples (5 pixels) is delayed. Valued data is extracted in parallel.

Similarly, the shift registers 18.1 are arranged in series to delay 1-bit data by 1 line scanning period, and 1-bit binary data at 5 sample positions are taken out in parallel from the subsequent stage. A total of 12 bits of parallel data taken out in parallel from these shift registers 15, 17, and 19 is used as address data for the ROM 13.

In other words, now is timing t. When the output from the A/D converter is n-bit image signal data (C6, CI, C2
~C, -+), the same timing t. The shift register 15 outputs binary data D one sample before, and binary data D-2 two samples before. Furthermore, when the number of samples for one line scan is L, the parallel output of the shift register 17 is the binary data D-
(L-21+D-T1.-11+ Dl, + D-(L
”., D-tt+2., and the parallel output of the shift register 19 is the binary data D, 2L-H, D-(
2L-11, D-2L, D-(2t, +11 r
It becomes D-+2L+21. Therefore, the ROM 13 stores data D1, D,, D-, L-2, ~D-(L+2
1+ D-(21,-2)~D-,22. Naru 12
bit data and m bits of random digital data from the random digital data generator 14 (12
+tn) bit data.

Here, the data Co ""' Cm −+ 'A corresponds to r(M, N) in the above-mentioned algorithm, and is Dl, D2+D-(L-21+++Df1.+
21+D-+21. -21+'D-+21.12+ 12-bit data is g(1,j)[1,jE
It corresponds to Ql.

Further, the n-bit binarization threshold data registered in advance in the ROM 13 is calculated and created based on the above algorithm, and is generated by the random digital data generator 14.
The m-bit random digital data from
This is data for randomly selecting binarized threshold data with slightly different density levels created in the . This m
Bit random digital data is data that changes for each pixel, and the method of generation and the size of m are arbitrary.

Further, in the above embodiment, a shift register is used for line delay of binary data, but this may be stored in a RAM (random access memory) and controlled by an address controller. In addition, although the comparator 12 has compared n-bit digital data, an analog comparator is used to input the input image data directly without using an A/D converter, and n-bit data from the ROM 13 is compared. The digital data may be converted into an analog value using a D/A converter and then compared.

FIG. 2 shows another embodiment of the invention. This embodiment is constructed by replacing the ROM 13 and comparator 12 shown in FIG. 1 with a ROM 20.

In the case of this embodiment, a total of 12 bits of parallel data taken out in parallel from each of the shift registers 15, 17, and 17, m bits of random digital data from the random digital data generator 14, and random digital data of m bits from the A/D converter 11. The ROM 20 is addressed by a total of (12+m+n) bits of output n-bit data, and a desired binary image signal is output. The other configurations are similar to those shown in FIG.

Note that this invention is not limited only to the above embodiments. For example, the past 2 that is fed back to the 2m conversion process
The pixel area Q and the number of pixels of the unified data, the weighting function α, and the minute density level δ may be determined according to the specifications. Furthermore, although the minute concentration level δ is randomly added or subtracted, it may be added or subtracted randomly only when a constant concentration value continues, or it may be added or subtracted using a predetermined method, or it may be added or subtracted using a predetermined value instead of addition or subtraction. It may be configured to produce a state of affairs. Also, it is of course possible to configure the system so that the calculation data is not stored in the ROM in advance, but is processed each time. Furthermore, although the minute density level δ is added to or subtracted from the binary threshold value data, it is of course possible to add or subtract this from the input image signal.

[Effects of the Invention] As explained above, according to this invention, the ffti is relatively low.
To provide a dual-image 2ffi processing device that can be realized at low cost with a lightweight configuration and that does not cause image quality deterioration even when reading artificially created regular gradation images or constant density images. can do.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing one embodiment of this invention, FIG. 2 is a block diagram showing another embodiment of this invention, FIG. 3 is a diagram explaining the concept of pixel units and image processing, and FIG. 4 FIG. 2 is a diagram illustrating an example of a weighting function. 11... A/D converter, 12... Comparator, 13.2
0...ROM, 14...Random digital data generator, 15, 16.17, 18.19 shift register. (a) (b) Figure 3 Figure 4

Claims (1)

[Claims] Means for inputting an image signal obtained by raster scanning an image, and a means for inputting an image signal obtained by raster scanning an image, and a means for inputting an image signal obtained by raster scanning an image; means for performing predetermined weighting processing to create binarization threshold data for the current input image signal; and binarizing the current input image signal by comparing the binarization threshold data and the current input image signal. a means for forming data; and a means for outputting the binarized data and temporarily storing it.
In the image binarization processing apparatus, the image binarization processing apparatus includes means for feeding back as past binarization data in the digitization process, further comprising means for randomly changing binarization threshold data to be compared with the current input image signal within a predetermined range. An image binarization processing device comprising: