JPS6288479A - Picture processing method - Google Patents

Abstract

PURPOSE:To obtain a picture which is excellent in gradation, less in blur, and inconspicuous in moire even when a dithering method is used, by finding the dot period of a dot picture by detecting the extremal value of picture signal levels. CONSTITUTION:The uppermost line of a picture is used as the remarkable line 31 as the initialization in the auxiliary scanning direction. Then a smoothing processing 32 is performed for removing noises from input picture signals obtained when the remarkable line is read and the position P(i) of the picture element representing the maximal value is detected by taking the difference between adjacent picture elements of the smoothing-processed picture signals. Thereafter, the interval (l) between the position of the remarkable maximal- value picture element and the next maximal-value picture element position when advancing in the main scanning direction (36). The interval (l) corresponds to the dot period and the frequency characteristic of a filter is changed in accordance with the interval (l) and picture signals corresponding to the processing section remove the frequency components of the period corresponding to the interval (l) 37 and 38.

Description

Detailed Description of the Invention (Industrial Field of Application) The present invention relates to an image processing method used in facsimiles, etc., in which an image is divided into pixels and processed, and in particular, a method for reading pseudo-halftone images such as halftone photographs and the like. The present invention relates to an image processing method for converting an image back into a binary image with pseudo halftone display using a systematic dithering method.

(Prior Art) A systematic dither method has been known as a method of pseudo-binary expressing a halftone image.

When this image processing method using the dither method processes images that include periodic images such as lattice patterns or halftone photographs, moiré occurs due to interference between the period of the periodic image and the period of the tither matrix. This has the disadvantage that image quality deteriorates due to this.

As a method for reducing the occurrence of moire, there is a method disclosed in Japanese Patent Laid-Open No. 111471/1983. This method detects the average period of the input image from both signal sequences of the input image signal, selects a teaser matrix according to the average period, and reduces moiré even for input images with strong periodicity. This is the method of trying.

Further, as another method, there is a method disclosed in US Pat. No. 41.theta.4221. This method detects whether the input image is a periodic image or not, and if the input image is periodic, a low-pass filter is used to
This is a method of re-dotting after removing periodicity by removing the - signal.

(Problems to be Solved by the Invention) However, the method disclosed in Japanese Unexamined Patent Publication No. 1-59-111471 requires calculating the average period of the input image, and Since pattern ROMs and the like for a plurality of teaser matrices must be prepared in advance, there is a problem in that the hardware configuration becomes complicated. Furthermore, since the teaser matrix size, which should normally be determined based on the required number of gradations at the design stage of an image processing device, is switched according to the cycle of the input image, the There was a problem that the key changed.

Furthermore, in the method disclosed in U.S. Pat. No. 4,184,221, the fundamental waves and harmonics of a plurality of types of halftone dots are removed using different types of low-pass filters, resulting in a characteristic with a large cutoff range. .

FIG. 14 shows an example where halftone dot frequencies of 3/mm or more are cut off. If the cutoff area is large like this, even if the resolution of the input image is 85 lines/inch (however, 1 inch is about 2.54 cm, hereinafter the same applies), even if the image is obtained with a coarse mesh, the resolution of the input image will be 150 lines/inch. Since even images obtained with a fine mesh such as inches are processed using the same low-pass filter, there is a problem in that the image after this processing becomes a very blurred image.

The purpose of this invention is! − Solving the problems mentioned above,
An object of the present invention is to provide an image processing method capable of obtaining an image with excellent gradation, little blur, and no moiré even when using a dither method.

(F stage for solving the problem) In order to achieve this object, according to the present invention, a halftone image is read pixel by pixel by a reading means, and an image signal of this image signal is In an image processing method that performs pseudo-halftoning after attenuating the signal level, the pixel position that exhibits an extreme value is detected from the image signal level of each pixel, and the pixel position that exhibits an ai ratio and other pixel positions are The area between other pixel positions that exhibit extreme values is defined as a processing section, and based on the halftone period determined from the length of this processing section, a period comparable to this halftone period is extracted from the image corresponding to this processing section. It is characterized by attenuating the image signal level of the image signal.

In carrying out the present invention, it is preferable that the position of a pixel exhibiting an extreme value be detected by changing the sign of the difference between the image signal levels of adjacent pixels.

In carrying out the present invention, it is preferable to attenuate the image signal level using a one-dimensional non-recursive (FIR) filter in the horizontal scanning direction.

Further, in carrying out the present invention, it is preferable to remove noise contained in the image signal and then detect a pixel exhibiting an extreme value.

(Operation) According to such an image processing method, frequency components exhibiting a period comparable to the halftone dot period of the processing section are removed from the image of the processing section based on the period obtained from the length of the processing section. Ru.

Therefore, the occurrence of moiré due to interference between the halftone dot period and a signal having a period comparable to the halftone dot period included in the halftone dot image is prevented.

Furthermore, since it is possible to attenuate only the frequency components that are effective in preventing the occurrence of moire, blurring in the image can be reduced even when pseudo halftoning is performed.

(Embodiments) Hereinafter, embodiments of the present invention will be described with reference to the drawings. It should be noted that these figures are only schematic illustrations to the extent that the present invention can be understood, and the algorithm described in the embodiments and the means for realizing this algorithm are not limited to the small examples in the figures. Further, in these figures, the constituent components of the circuit are indicated by the symbol of the circuit.

In addition, in the embodiment described below, the halftone image is 16 trees/m.
It is assumed that reading is performed at a reading pixel density of m. Further, the main scanning direction refers to scanning from the top of the image to the right, and the sub-scanning direction refers to scanning from I- to the bottom of the image. Furthermore, there are two virtual white pixels to the left of the leftmost pixel (leftmost pixel) and to the right of the rightmost pixel (rightmost pixel) in each row of the image. do.

FIG. 1 is an operational flowchart showing the processing order when the image processing power V of the present invention is processed by a microprocessor sensor (hereinafter referred to as a microcomputer). The image processing method of the present invention will be explained with reference to FIG.

First, as the initial setting for the sub-scanning direction, set the last row of the image to Note 1.
] line (step 31).

An example of an input image signal (image signal) obtained by reading the last line of a halftone photograph pixel by pixel with the reading means having a pixel density of 16 mm/mm is shown in Fig. 2 (A). show. Second
In figure (A), the horizontal axis represents individual pixels in the main scanning direction,
The image signal level for each pixel is plotted on the vertical axis. As is clear from FIG. 82(A), the local maximum values 21 appear at intervals corresponding to the halftone dot period T. The image processing method of the present invention detects the halftone dot period using extreme values including the maximum value 21. However, the input image signal also includes a maximum value 23 caused by noise independent of the halftone period as shown in FIG. 3(A). Therefore,
In order to reproduce halftone dots well from this input image signal, first,
Smoothing processing is performed to remove this noise from the input image signal (step 32 in FIG. 1).

This processing power V can be determined by a known method. For example, C
If (i) is the image signal level (pixel density) of the pixel number 1, then the image signal level after 41 smoothing processing (ga degree information)
C'(+) can be determined by equation (1) of Toh's equation.

C'(+)=C(i)/2 + (c (i+1) 10 (+-11...(1)
The spatial filter that performs this W-digesting process has characteristics that block a wide area as shown in FIG.

As shown in FIG. 2(B), the human image after the smoothing process maintains the maximum value 4 (ti 21) caused by the dot period, and removes the maximum value 23 caused by noise.

Image - After completing the smoothing process for the material part, next, calculate the extreme value from the smoothed values! Find the pixel position at t.

In this case, the case of finding the maximum value will be explained.

Both signal levels C'(+) of the pixel of interest and the pixel signal level C'(i+1) of the pixel one step ahead in the main scanning direction from the pixel of interest.
) and calculate the difference D(i) using the following equation (2) (
Figure 1 step 33).

D (i) = C(i+1) −C(i)...(2)
Here, the difference D(+) between the image signal levels of adjacent pixels is a positive value immediately before the pixel where the image signal reaches the maximum value, but the difference immediately after the maximum value is reached is a negative value. Become. Therefore, the pixel position P(+) exhibiting the maximum value can be detected (step 34).

Furthermore, the pixels to the right of the leftmost pixel and the rightmost pixel in each tth row are set to maximum value pixels.

Next, as an initial setting in the main scanning direction, the maximum pixel position of interest is set to the leftmost pixel (P (1) ) (step 35).

Next, a processing interval is set between the maximum pixel position of interest and, for example, the next maximum pixel position that advances in the main scanning direction, and the interval intersection from the maximum pixel position of interest to the next maximum pixel position is determined (step 38). The interval statement obtained here is the number of pixels that read the halftone period in the main scanning direction of the input halftone dot image (the number of pixels between the maximum value of interest and the next maximum value of 11 in the horizontal scanning direction, (a positive integer).

Attached Table 1 shows the screen angles of 45 degrees and 90 degrees, which are commonly used, and halftone dots of 65 lines/inch to 150 lines/inch were read at reading pixel densities of 8 wood/mm and 16 wood/mm, respectively. 2 is a table showing the halftone dot period in the 1-mo scanning direction in terms of the number of pixels.

Note that FIGS. 4(A) and 4(B) are diagrams for explaining halftone dots. i1i written in angle da1 of attached table 1! 90
The halftone dot means the halftone dot shown in FIG. 4(A), and the halftone dot period of this halftone dot in the scanning direction is #gl, which is indicated by L+ in the figure. Further, the 45 degree halftone dot means the halftone dot shown in FIG. 4(B), and the halftone dot period of this halftone dot with respect to the scanning direction is the distance indicated by L2 in the figure.

In this embodiment, the reading pixel density is 16 pixels/mm, so the halftone dots are less than 100 lines/inch and 45
It can be seen that when the diagonal halftone dots are used, the halftone dot period is the longest, and the period becomes the angler's pixel. Note that for halftone dots with 100 lines/inch or less, a better image can be obtained by binarizing them with hard W threshold 1 rather than performing pseudo halftoning. This point will be explained excluding the objects to be processed by the image processing method of the present invention.

Next, the occurrence of moiré due to interference between the halftone dot period and a signal with a period comparable to the halftone period included in the input image signal is prevented.
In order to avoid this problem, a process is performed to remove a signal having a cycle comparable to the halftone dot cycle from the input image signal using f stages of attenuation.

Incidentally, the term "period comparable to the halftone dot period" includes not only a case where the period is almost the same as the halftone dot period, but also a case where the period is the same as the halftone dot period.

In this case, as the first embodiment, the attenuation stage is configured by software of the microprocessor. This attenuation stage is configured so that the frequency characteristic of the filter can be switched to one type h based on the interval (sometimes referred to as period) of the processing section described above.

Therefore, the frequency component (sometimes referred to as halftone frequency) that has the IW period V equal to the periodic intersection can be removed from the image signal corresponding to the processing section.

Specifically, if the density information from which the halftone frequency of the first pixel has been removed is C''(i), then C°゛(i) is expressed as (
3) Determined from the formula.

C” (i)=AC(i)+B (C(i+1)×C(
+-1) ) +C(C(i+2) +C(+-2) l +D(C(i+3)+C
(i-3)j...(3) However, A-D indicates the coefficients of the digital filter. Here, for the periodic sentence, the coefficient A-
If D is set to a value as shown in Attached Table 2, the frequency characteristics of the Tr filters will be as shown in FIGS. 5 to 9. still,
The frequency characteristic diagrams shown in FIGS. 5 to 9 are graphs in which the horizontal axis represents the frequency, the vertical axis represents the attenuation level (dB), and the attenuation product is plotted against the frequency.

Incidentally, FIG. 5 shows a characteristic curve diagram of the filter when attenuating a frequency component exhibiting a period of 3 pixels (5.3 pixels/mm). FIG. 6 shows a characteristic curve diagram of a filter when attenuating a frequency component having a period of 4 pixels (4,0 pixels/mm). FIG. 7 shows a characteristic curve diagram of a filter when attenuating a frequency component having a period of 5 pixels (3,2 pixels/mm). FIG. 8 shows a characteristic curve diagram of a filter in which a frequency component having a period of 6 pixels (2,7/mm) can be attenuated. FIG. 9 shows a characteristic curve diagram of a filter when attenuating a frequency component exhibiting a period of 7 pixels (2,3/mm).

Therefore, it is possible to set g-attenuation characteristics that can attenuate periodic fundamental waves and harmonics from the image signal in accordance with the periodic text.

By the method described above, the coefficient A-D corresponding to the circumference of the processing section of interest is selected, and by equation (3), the maximum value pixel that appears next in the main scanning direction from the maximum value pixel of interest within this processing section is selected. The image signals up to one before are filtered to remove halftone dot frequencies (steps 37 and 38).

Next, the image signal from which the halftone frequency has been removed is converted into pseudo-halftone by the systematic teaser method (step 39).

Next, the 11 maximum pixel positions of interest are set as the next maximum value pixel position to appear in the main scanning direction (step 40). continue,
Main scanning direction - Determine whether processing for a line is completed (
Step 41). - If the processing for the rows has not been completed, return to step 36 and perform the processing in sequence. Also, if the process of striking the bell is completed, move the line of interest one line to line F (in the sub-scanning direction) (step 42). Next, in the sub-scanning direction, for example, it is determined whether or not the processing for -pages is P:rI (step 43), and if the processing is not completed, then
Return to step 32 and complete all processing in the sub-scanning direction.
Processing is performed sequentially until

FIG. 1O is a configuration diagram showing each means for realizing the processing order shown in FIG. 1 and the relationship between each stage.

In FIG. 10, reference numeral 51 indicates a reading means for reading an image at a pixel density of, for example, 16 pixels/mm. Further, 53 indicates an input memory that stores image information of an input image. 55 is a microprocessor (sometimes called a microcomputer)
shows. The microcomputer 55 reads image information from the manual memory 53 in accordance with the processing order shown in FIG. 1, processes it using the work memory 57, and outputs a pseudo-halftone image signal to the output memory 59. The image signal read out from the output memory 59 is recorded in the 111 output f stage 6].

Further, FIG. 11 is a block diagram showing a second embodiment of the present invention, and is a block diagram showing a case where the processing procedure of FIG. 1 is implemented by hardware.

In FIG. 11, reference numeral 71 indicates a reading device, and an image signal 73 read by this reading device 71 is stored in an image memory 75 and also input to an extreme position detection circuit 77. This extreme value position detection circuit 77 detects the pixel position exhibiting the maximum value of the image signal 73 and outputs this pixel position to the maximum value position memory 78.

Further, 81 indicates a digital filter as an attenuation f stage, which removes a desired frequency component from the image signal. Reference numeral 83 indicates a dither processing circuit for performing pseudo-halftone conversion, and 85 indicates a recording device for recording the pseudo-halftone image signal. Further, 87 indicates a microprocessor that controls each device such as the reading device 71 and the memory to perform desired image processing.

FIG. 12 is a block diagram showing an embodiment of the extreme position detection circuit shown in FIG. 11. In FIG. 12, 101 and 102 represent one-pixel delay circuits, 103 and 104 represent power n multipliers, 105 represents a multiplier with coefficient A, and 106 represents a multiplier with coefficient B. In this case, the coefficients A and B are A=0.5, B=o, and 25. These one pixel delay circuits 101, 102,! :, adder 1
03 and 104 and multipliers 105 and 1013 constitute a noise removal stage (corresponding to step 32 in FIG. 1). Also, 10? 1 represents a one-pixel delay circuit, and 108 represents a subtracter, and this one-pixel delay circuit 107 and g calculator 108 constitute means for determining a difference (corresponding to step 33 in FIG. 1).

lO8 indicates a difference value memory that stores the determined difference. Further, +10 indicates a microprocessor, which detects the pixel position where one extreme value is located from the obtained difference (corresponding to step 34 in FIG. 1).

Figure 13 shows the digital filter 8 shown in figure wrJl1.
1. FIG.

In this case, one pixel delay circuits 121 to 126, adders 131 to 138, multipliers 141 to 144 having coefficients A, B, and C, and coefficients A to D of multiplier 14]-144.
A one-dimensional non-recursive form (F I R: Finite Is +
pu1geResponse) filter (corresponding to step 38 in FIG. 1). A switching signal corresponding to the periodic sentence is sent from the processor 87 shown in FIG. 11 to the coefficient RO.
When outputting to M151, a coefficient value corresponding to the periodicity is selected from the coefficient values stored in coefficient ROM 151, and multiplier 1
These coefficients are set in 41 to 144.

In addition, the extreme value position detection 111 explained in the embodiment described in
The circuit, attenuation means, etc. are not limited to this embodiment, and can be changed depending on the content of image processing required.

In addition, the embodiment described above detects the pixel position where the image signal level exhibits a maximum value from the input image signal, and detects the pixel position of the maximum value of interest in the input image signal and the next pixel proceeding from this pixel in the main scanning direction. An example has been described in which the processing interval is defined as the area between the pixel position and the maximum value pixel position that appears.

However, the present invention is not limited to this embodiment, and the processing interval is defined as a processing interval between a pixel position exhibiting a minimum value and a pixel position exhibiting a minimum value 1 that appears next in one scanning direction from this pixel. Alternatively, the processing interval may be set between the pixel position at which the maximum value is reached and the pixel position that is 11 feet away from the next minimum value i that appears in the scanning direction from this pixel. Furthermore, note [1 pole 4rl+ (maximum i〆1
Or, between the local minimum pixel position (1) and other extreme pixel positions, for example, the maximum pixel position of interest and n 1] local maximum ((i pixel position) The processing time may be set as 1%.

(Effects of the Invention) As is clear from the above description, according to the image processing method of the present invention, the halftone dot 1 interlock period of the halftone dot image is determined by detecting the extreme "1" of both signal levels. Therefore, the number of arithmetic units can be reduced compared to the conventional case of using 1-1 autocorrelation, and the period can be easily determined.

Also, following this period, frequency components having a period comparable to this halftone period are removed from the halftone dot image.

Therefore, moiré caused by interference between the halftone dot period and a signal having a period comparable to the halftone dot period included in the halftone dot image is prevented. Furthermore, since only the frequency components necessary to prevent moiré can be removed, image blur can be significantly reduced.

Therefore, even if an image with a case f-pattern or a periodicity is processed using the teaser method, the occurrence of moiré can be prevented.

Therefore, it is possible to provide an image processing method that can obtain an image with excellent gradation, less blur, and less noticeable moiré even when using the teaser method.

Attached Table 2 However, the sentence indicates the interval between processing sections, and A-D indicates the coefficient of equation (3).

[Brief explanation of drawings]

Fig. 1 is an operation flowchart showing the processing f order of the image processing method of this invention, Figs. 2 and 3 are diagrams explaining the image processing power V of this invention, and Fig. 4 explains halftone dots. Figures 5 to 9 are diagrams showing the frequency characteristics of the filter used in the present invention. Fig. 1O is a block diagram showing an example of means for realizing the present invention. Fig. 11 is a block diagram showing an embodiment of the present invention, and is a block diagram showing an example for realizing the present invention with hardware. , FIG. 12 is a block diagram showing an example of an extremum position detection circuit suitable for use with the image processing power D of the present invention, and FIG. 13 is a block diagram showing an example of a digital circuit suitable for use with the image processing power D of the present invention. A bronch diagram showing an example of a filter. FIG. 14 is a diagram used to explain the conventional technique (1 to i). 21... Local maximum value due to halftone period 23... 4! Straight 51.71...
・Reading means (reading device) 53...Manual memory 55.87.110...Microprocessor 57...
・Work memory, 59...Output memory 81.8
5... Recording means (recording device) 73... Image signal,
75... Image memory 77... Pole (fi position detection circuit, 79... Extreme value position memory 81... Attenuation means (-dimensional non-recursive (FIR) filter) 83... Teaser processing circuit 101 , 102, !07.121-12B... One pixel delay circuit 103.104.131-136... Adder 105, 108.141-144... Multiplier 109.
・・Difference value memory, 151 °°° coefficient ROM + 5
3...Switching signal. - Seroko J ¥ 1 > Counternumber (book/i,) The promise of this invention (; I scale 6 track diagram Figure 3 Diagram: The line section of this invention that shares the same line with the Somei 1 ni Itinai that shares the 1 ni of Kono Tora and the Ming Dynasty 2θ 4
θ 5θ 2θ frequency (book/m) Diagram corresponding to the explanation of this invention Fig. 7 Round number (T/m==) Diagram corresponding to the explanation of this invention Fig. S (Book/--) /θ mefθ2.107: 1 pixel 4 extensions - lower each lθ
3. IO4: , Calo゛Ro91, 7 Figure 7 This $ clear use () Itchichiku 1 Pro of the digital filter left over from home, Figure 77 Explanation diagram of conventional limb technique Figure 14 Procedure Neichi 1 F'Th! -1' January 7, 1985 [-1 Commissioner of the Patent Office Kuro 111 Kaii Tono 1 case display Showa ti011 = Patent application 221147
+- No. 2 Name of the invention Image processing method 3 Relationship with the person making the amendment Patent applicant address (〒-105) 1-1-17-12 Toranomon, Minato-ku, Tokyo Name (029)
Oki Electric Power Co., Ltd. Representative Nankai Hashimoto 4 Agent 170 6 (!188) 5563 Address 1-20-5 Higashiikebukuro, Toyoshima-ku, Tokyo No. 905, Ikebukuro White House Building Column (1) of the detailed description of the invention in the specification and the scope of claims in the specification are amended as follows. lr2, Claims (1) An image processing method that performs pseudo-halftoning after attenuating the image signal level of an image signal obtained by reading a halftone image pixel by pixel with a reading means. In this step, a pixel position exhibiting an extreme value is detected from the image signal level of each pixel, and a processing interval is defined between the pixel position where the extreme value is shaded and another pixel position exhibiting an extreme value other than the pixel position. , Image processing characterized by attenuating the image signal level of an image signal having a period comparable to the halftone period from an image corresponding to the processing period based on the halftone period determined from the length of the processing period. Method. (2) The image processing method according to claim 1, wherein the detection of a pixel position exhibiting an extreme value is performed by changing the sign of a difference between image signal levels of adjacent pixels. (3) The image processing method according to claim 1, wherein the image signal level is attenuated using a FIR filter in the L scanning direction. (4) The image processing method according to claim 1, characterized in that, after noise contained in the image signal is removed, a pixel I' device exhibiting an extreme value is detected. J(2), 2nd page of the specification [■ Line 14 [-1, 2nd page, line 15 [1, 2nd page, 18th line 1, 5th page, 12th line 11, 7t'] No. 1st line, page 11, line 19] 1, 14th Tei, line 15, and page 15, line 18] "pseudo" is corrected to "pseudo J." (3), 6th line of the specification [1, line 15 [Correct 1 [halftone dots in this processing section] to "halftone period". (4), 6th [[17th line [1] ``Therefore, the halftone period'' is corrected to ``Therefore, the pseudo halftoning period,'' 1. (5), The specification 61i, line 18, line 11, [interference with (a) having a period comparable to the period] is replaced by "interference with r period"
I am corrected. (6), "-uchibu" on page 9, line 16 of the specification is 1-
Correct line part 1. (7) In the fourth line of page 12 of the specification, ``Next, the halftone dot period,'' is corrected to read ``first order'' and ``pseudo halftoning period'' as j. (8), "Interference with a signal with a period comparable to the period" on page 12, line 5 of the specification is corrected as "interference with the period j." (9), page 12 of the specification, Correct "fundamental wave and harmonics" in the 6th line to 1 fundamental wave and 1 harmonic thereof. (In), on page 19, line 18 of the specification, ``Therefore, the halftone period,'' is corrected to ``1, therefore, the period of pseudo-halftoning.'' (II), page 19, line 19 of the specification, ``interference with No. 18 having a period comparable to the period'' is defined as ``interference with r period''.
I am corrected.

Claims (4)

[Claims] (1) In an image processing method that performs pseudo halftoning after attenuating the image signal level of an image signal obtained by reading a halftone image pixel by pixel with a reading means, the image of each pixel is A pixel position exhibiting an extreme value is detected from the signal level, a processing interval is defined between the pixel position exhibiting an extreme value and another pixel position exhibiting an extreme value other than the pixel position, and the length of the processing interval is determined. An image processing method characterized by attenuating the image signal level of an image signal having a period comparable to the halftone period from an image corresponding to the processing section based on the halftone period determined from the above. (2) The image processing method according to claim 1, wherein the detection of a pixel position exhibiting an extreme value is performed by changing the sign of a difference between image signal levels of adjacent pixels. (3) The image processing method according to claim 1, wherein the image signal level is attenuated using a one-dimensional non-recursive (FIR) filter in the main scanning direction. (4) The image processing method according to claim 1, further comprising detecting a pixel position exhibiting an extreme value after removing noise contained in the image signal.