JP2008078830A - Image processing apparatus, and image processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing apparatus capable of suppressing interference moire caused by dot structure that an original has and halftone processing. <P>SOLUTION: In the image processing apparatus 100, an original optically read by an image input means 101 such as a scanner is converted into digital image signals of rgb each having 8 bits and the digital image signal is output. The output image signal is input to a scanner γ correction means 102 and a reflection factor linear rgb signal is converted into a density-linear RGB signal by an LUT (lookup table) or the like. Simultaneously, the rgb signal is input to an image area separation means 103 and the image attribute of the input image is decided in each pixel and the decided result is output as an image area separation result. Further, the g signal is input to a first edge quantity calculation means 104 and the edge degree of the input image is calculated as edge quantity e1 and output. A filter processing means 105 adaptively executes edge emphasis processing or smoothing processing on the basis of the decision result s1 decided by the image area separation means 103 and the edge quantity e1 calculated by first edge quantity calculation means 104. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an MF copying machine (multifunction copying machine), a facsimile, a scanner, which processes image data and copies an image or transfers a file to an image input from a scanner or the like that reads a document or an external device. The present invention relates to an image processing apparatus such as a printer, and more particularly to an image processing apparatus that includes an image area separation unit and a dither halftone processing unit, and is intended to reduce the occurrence of interference moiré at a halftone dot portion. is there.

As a conventional technique, Patent Document 1 discloses an image forming apparatus that inputs an image signal to form an image on a recording medium, calculates a power spectrum for each RGB included in the image signal, and calculates the spatial frequency of the spectrum. An image forming method is disclosed in which an angle is obtained, whereby a screen pattern for each YMC in a printer unit is determined, and an image forming signal that is pulse-width modulated by the screen pattern is generated to form an image.
Further, Patent Document 2 discloses a network that analyzes a halftone dot structure of an input image signal in an image processing apparatus that executes and outputs spatial frequency processing that emphasizes / de-emphasizes the sharpness of the input image signal. An image processing apparatus is disclosed that includes point structure analysis means and filter processing means for controlling two-dimensional directional characteristics in spatial frequency processing according to the analysis result by the halftone structure analysis means.
JP 2001-45306 A JP 2002-344743 A

However, the prior art disclosed in Patent Document 1 detects the screen angle and the number of screen lines of a halftone image and controls halftone processing, and does not perform image area separation. By suppressing the moire, there is a problem that the image at the character portion deteriorates.
The prior art disclosed in Patent Document 2 detects the screen angle of a halftone image and changes the two-dimensional direction characteristics of the filter processing. There is a possibility that the effect of suppressing moire is reduced by the characteristics of the filter. is there.
The present invention has been made in view of the above problems, and the remaining degree of halftone dot undulation by the adaptive filter is caused by the number of halftone dots, so the line number information is retained, and interference is performed based on the line number information. An object of the present invention is to provide an image processing apparatus capable of suppressing interference moiré due to a halftone dot structure of a document and halftone processing by applying halftone processing parameters in which moiré is unlikely to occur.

In order to solve such a problem, the present invention provides a color image input unit that reads a color image, an image attribute for each area of an image signal read from the color image input unit, and at least a halftone dot. Image area separation means having line number detection means for detecting the number of halftone dots in the area, filter processing means for correcting spatial frequency characteristics according to the separation result of the image area separation means, and processing performed by the filter processing means Halftone processing comprising edge detection means for performing edge detection on a processed image, and threshold selection means for selecting a threshold according to the detection result of the edge detection means and the number of halftone lines detected by the line number detection means And the halftone processing unit controls generation of periodic or aperiodic dots by the threshold selection unit.
According to a second aspect of the present invention, the threshold selection unit selects a threshold so as to generate a dot having a weak periodic structure in the low line number halftone dot region where the line number information is lower than a predetermined threshold, and the low line In a region other than the halftone dot region, a threshold value is selected so as to generate a dot having a strong periodic structure.

According to a third aspect of the present invention, the threshold selection condition of the threshold selection unit is changed in accordance with the spatial frequency characteristic corrected by the filter processing unit.
According to a fourth aspect of the present invention, when the spatial frequency characteristic corrected by the filter processing means is a characteristic that emphasizes higher frequency components, the threshold selection condition of the threshold selection means is periodic even if the number of lines is high. Control is performed so as to generate a dot having a weak structure.
The fifth image processing mode includes a first image processing mode in which character images are emphasized and a second image processing mode in which image images are emphasized, and the second image processing mode in the first image processing mode. The threshold selection condition by the threshold selection means is controlled so as to generate dots with a weak periodic structure even with a high number of lines.

According to a sixth aspect of the present invention, the image area separating unit stores representative line number information of the number of halftone lines detected by the line number detecting unit as line number information for the page, and a threshold value based on the line number information. The threshold selection means is controlled to change the selection condition.
According to a seventh aspect of the present invention, the image area separating unit stores halftone dot information having the lowest line number among the halftone line numbers detected by the line number detecting unit as line number information for the page, and the line number information is stored in the line number information. The threshold selection unit is controlled to change the threshold selection condition based on the threshold selection condition.
8. A color image input step for reading a color image, and an image attribute for each region of an image signal read from the color image input step, and a line number detection for detecting the number of halftone dots in at least a halftone dot region An image region separation step having a step, a filter processing step for correcting a spatial frequency characteristic according to a separation result of the image region separation step, and edge detection for performing edge detection on the processed image processed by the filter processing step A halftone processing step including a threshold selection step for selecting a threshold according to a detection result of the edge detection step and halftone line number information detected by the line number detection step. Controls generation of periodic or aperiodic dots by the threshold selection step. To.

  According to the present invention, a color image input unit that reads a color image, and a line number detection that determines an image attribute for each region of an image signal read from the color image input unit and detects the number of halftone dots in at least a halftone dot region An image area separation means having a means, a filter processing means for correcting a spatial frequency characteristic according to a separation result of the image area separation means, and an edge detection means for performing edge detection on the processed image processed by the filter processing means, A halftone processing unit having a threshold selection unit that selects a threshold according to the detection result of the edge detection unit and the halftone line number information detected by the line number detection unit. The halftone processing unit includes a threshold selection unit. Controls the generation of periodic or non-periodic dots, so it is possible to maintain half-line processing parameters and to prevent interference moire based on line number information. Can apply a result, it is possible to suppress interference moire due halftone structure and halftoning with the document.

Hereinafter, the present invention will be described in detail with reference to embodiments shown in the drawings. However, the components, types, combinations, shapes, relative arrangements, and the like described in this embodiment are merely illustrative examples and not intended to limit the scope of the present invention only unless otherwise specified. .
FIG. 1 is a block diagram of an image processing apparatus according to an embodiment of the present invention. In the image processing apparatus 100, a document optically read by an image input means 101 such as a scanner is converted into an 8-bit digital image signal of rgb and output. The output image signal is input to the scanner γ correction means 102, and the linear reflectance rgb signal is converted into a density linear RGB signal by an LUT (Look Up Table) or the like. At the same time, the rgb signal is input to the image area separation means 103, and the input image is subjected to image attribute determination for each pixel and output as an image area separation result. Further, the g signal is input to the first edge amount calculation means 104, and the degree of edge of the input image is calculated and output as the edge amount e1. The filter processing unit 105 adaptively performs edge enhancement processing or smoothing processing based on the determination result s1 from the image area separation unit 103 and the edge amount e1 from the first edge amount calculation unit 104.

The image signal after the filter processing is input to the compression unit 106 and compressed, and further stored in the storage unit 107. The compressed image signal once stored is again decompressed by the decompressing means 109, and subsequent image processing is started. When there are a plurality of copy requests from the user by an operation unit (not shown) or the like, the compressed image signal stored in the storage means 107 is read a plurality of times for the number of copies, subjected to subsequent image processing and output. Further, the image area separation signal s1 from the image area separation means 103 is accumulated in the separation signal accumulation means 108.
The image signal expanded by the expansion unit 109 is output to the color correction unit 110. The color conversion means 106 converts the RGB signal into a CMY signal suitable for a printer color material by a masking operation or the like. Further, it is converted into a C′M′Y′K ′ signal by the under color removal / black generation unit 111 and output to the printer γ correction unit 112 and the second edge amount calculation unit 113. In the second edge amount calculation means 113, the edge amounts e2 to e5 are calculated by a predetermined edge amount calculation filter for each CMYK plate and output to the halftone processing means 114. In the halftone processing unit 114, the edge amounts e2 to e5 and the separation signal s2 from the separation signal accumulating unit 108 are used so as to achieve both the sharpness of the character image and the halftone image and the graininess and gradation of the continuous tone image. Control dither adaptively. The output image signal processed by the halftone processing unit 114 is output to the image output unit 115 such as a laser printer and printed on a storage medium such as a recording sheet (not shown).

Hereinafter, the operation of the core block constituting the image processing apparatus according to the present invention will be described more specifically. First, the image area separation unit 103 will be described.
FIG. 2 is a configuration diagram of the image area separating means 103. As shown in FIG. 2, the image area separation means 103 analyzes the input image signal and outputs two signals, s1 signal and s2 signal. The s1 signal is a signal indicating a 2-bit image attribute, and indicates any one of a black character area, a color character area, and a picture area (the picture area is an area other than the black character area and the color character area). Yes, it is information obtained in units of pixels. The s2 signal is a signal indicating the number of lines in the halftone dot image region, and is a 2-bit signal in this embodiment.
The s1 signal is obtained from the determination unit 1311 that detects the feature of the image by the edge detection unit 1301, the halftone detection unit 1307, and the color detection unit 1304, and further comprehensively determines the detection results by these three detection units. Specific operations will be described below.
A general operation of the edge detection unit 1301 will be described with reference to FIG. As shown in FIG. 2, the edge detection unit 1301 is configured to input g data of rgb data, perform pattern matching 1302 using a matrix of a predetermined size, and detect an edge portion of a character / line drawing. In the present embodiment, an example is shown in which detection is performed using a 3 × 3 matrix. Detection patterns are prepared as shown in FIGS. 3-1 to 3-4, and edges in each direction can be detected. For example, when detecting a vertical edge as shown in FIG. 3A, when a pixel marked with a white circle is a white pixel and a pixel marked with a black circle is a black pixel, the target pixel (marked with *) is regarded as an edge candidate To do. The white pixel and the black pixel are obtained by performing threshold processing on an input image signal with two predetermined threshold values, and trinating the input image signal into a white pixel, a black pixel, and a gray pixel. Dilation processing 1303 is performed on the edge candidate pixels thus obtained, and the obtained area is output to the determination unit 1311 as an edge portion.

Next, the general operation of the color pixel detection means 1304 will be described with reference to FIG.
The color pixel detection unit 1304 first performs a color difference determination 1305. When rgb data is input and the difference between the rgb color values in the target pixel is larger than a predetermined threshold, this is set as a color pixel candidate. Further, expansion processing 1305 is performed on the pixels detected as color pixel candidates, and the obtained region is output to the determination unit 1311 as a chromatic color region.
Next, the operations of the halftone dot detection means and the line number detection means 1307 will be described with reference to FIG. First, the halftone dot detection means and the line number detection means 1307 input g data of rgb data, and the center pixel is a peak pixel in a matrix of a predetermined size (3 × 3 in this embodiment, see FIG. 4). Whether or not there is a peak detection 1308 is performed. When the pixel of interest (marked with *) has a value larger than all the surrounding pixels (a to h) and the absolute value of the difference between the average value of a to h and the pixel of interest is larger than a predetermined threshold value, Let the pixel be a halftone dot candidate (mountain peak). Similarly, the target pixel (marked with *) is smaller in value than all the surrounding pixels (a to h), and the absolute value of the difference between the average value of a to h and the target pixel value is larger than a predetermined threshold value. At this time, the pixel of interest is a halftone dot candidate (valley peak). This is because the peak pixels of the halftone dot composing the halftone image detect features whose density is sufficiently higher or lower than the surrounding pixels. For the obtained peak pixel, the density determination 1309 counts how many peak pixels exist in the predetermined region. If the count value is equal to or greater than a predetermined threshold value, it is determined that the halftone dot candidate pixel is a halftone dot pixel, and the area is expanded by the subsequent expansion process 1310, and the obtained area is set as a halftone dot area.

By the way, the dilation processes 1303, 1306, and 1310 are applied to pixels detected as edge pixels, color pixels, and halftone pixels, and are filled in to obtain a region. In this embodiment, the expansion in the expansion processes 1303 and 1306 is 3 × 3 (expands one pixel at the periphery). The expansion processing 1310 for halftone dot detection is 7 × 7 (expands by 3 pixels in the vicinity), and is a size that can fill a halftone dot image of about 120 lines (lpi).
Each detection result obtained as described above is comprehensively determined on the basis of the truth table shown in FIG. 23 by the determination means 1311 and separated into three image areas, a black character area, a color character area, and a design area. Is output as the s1 signal.
Further, the count values obtained by the density determination 1309 in the halftone dot detection unit and the line number detection unit 1307 are classified into four levels by the four leveling unit 1312 and sent to the AND circuit 1313 as a 2-bit signal. The AND circuit 1313 performs an AND process on the signal determined as the halftone dot region by the expansion processing 1310, and generates and outputs an s2 signal as peak pixel density information in the halftone dot region. The s2 signal is a signal indicating the number of halftone dots, and is determined using the feature that a halftone dot region having a high number of lines if the number of peak pixels is large and a halftone dot region having a low number of lines if the number of peak pixels is small. It is what you are doing.

FIG. 5 is a diagram showing the relationship between the number of halftone lines and the number of peak pixels. It can be seen that the higher the number of lines, the greater the number of peak pixels. Conversely, the number of lines can be specified to some extent by the number of peak pixels. In this embodiment, the density determination 1309 counts the number in the 20 × 9 pixel block and outputs it to the four-level conversion unit 1312. In 1312, a low line number halftone dot (100 lpi or less), a medium line by two threshold values. The number of halftone dots (higher than 100 lpi and less than or equal to 133 lpi) and high line number halftone dots (higher than 133 lpi) are classified and determined. Further, the number of lines for the halftone dot region is output by AND processing. That is, as shown in FIG. 6, the 2-bit s2 signal is output as 00 with a low number of lines, 01 with a medium number of lines, 10 with a high number of lines, and 11 with a non-halftone dot.
In this embodiment, the number of halftone lines is determined based on the number of peak pixels. However, the number of halftone lines may be detected using an autocorrelation function of the pixel region of interest such as Japanese Patent Publication No. 8-31952. Alternatively, a method of determining a halftone dot period and detecting the number of lines by performing a Fourier transform and calculating a spatial frequency characteristic as disclosed in JP-A-7-220072 may be used.

Next, the operation of the first edge amount calculation means 104 will be described with reference to FIG. The first edge amount calculation unit 104 inputs the g signal read by the image input unit 101 and calculates the edge amount. The input g signal is input to four edge amount detection filters 1 to 4 (numbers in FIG. 7 corresponding to 181 to 184), and the edge amount is detected. Each of the four edge amount detection filters is a primary differential filter having a 7 × 7 matrix size shown in FIGS. 8A to 8D. The vertical edge, the horizontal edge, the left diagonal edge, and the right diagonal edge in this order. It is for detecting. The calculation results obtained by these edge amount detection filters are input to the subsequent absolute value conversion circuits 185 to 188 and converted into absolute values. Further, the maximum value among these four absolute values is selected by the maximum value selection circuit 189 and output as the edge amount signal e1.
Here, an example using a primary differential filter is shown, but a secondary differential filter may be used. Since the secondary differential filter calculates a high edge amount at the center of the line, it may be advantageous for edge enhancement processing. Of course, the primary differentiation and the secondary differentiation may be combined or the matrix size may be changed according to the purpose.

Next, the operation of the filter processing means 105 will be described with reference to FIG. The filter processing unit 105 inputs the image signal after the scanner γ correction processing 102, and based on the separation signal s 1 (character / picture) from the image area separation unit 103 and the edge amount e 1 from the first edge amount calculation unit 104. Perform characteristic conversion. Here, the R signal of RGB will be described.
The upper path in FIG. 9 is a process applied to the character region, and a uniform edge emphasis process 1042 is performed by the edge emphasis filter after the through buffer 1041 and is output to the selector 1045. The filter coefficient of the edge enhancement filter is as shown in FIG. 11, and is designed with emphasis on the sharpness of characters. The lower pass is a process applied to the picture area. After smoothing by the smoothing processing means 1043 uniformly with a filter as shown in FIG. 12, adaptive edge enhancement processing 1044 is performed, Adaptive edge enhancement based on the calculated edge amount e1 is performed.
The adaptive edge enhancement processing 1044 will be described in detail with reference to FIG. The edge amount e1 obtained by the first edge amount calculating means 104 is input to the edge amount conversion table 1048 and converted into a corrected edge amount e1 ′ by a predetermined conversion table. The image signal after the smoothing process 1043 is input to the Laplacian filter 1046, the Laplacian component is extracted by the filter process shown in FIG. 10, and then multiplied by the above-described corrected edge amount e1 ′ by the multiplier 1049, and further the original image The signal is added by the adder 1047 and output.

For example, as shown in FIG. 14, the edge amount conversion table 1048 has a characteristic that the corrected edge amount e1 ′ is smaller as the edge amount e1 is smaller, and the edge amount e1 such as a halftone dot photo area or a halftone portion of a character background. Is corrected so that e1 ′ is further reduced by correction so that weak edge enhancement is applied. On the contrary, a region having a relatively large edge amount, such as characters in a halftone image, is corrected so that e1 ′ is further increased by correction so that strong edge enhancement is applied.
The image signal processed in the upper and lower two passes as described above (returning to FIG. 9) is selectively output by the separation signal s1 from the image area separation means 103. That is, a signal from the upper path is selected and output for a pixel having s1 color character or black character attribute, and a signal from the lower path is selected and output for a pixel having a picture area attribute. . By doing so, it is possible to perform edge emphasis on the character on the halftone dot without emphasizing the halftone dot portion, and it is possible to perform processing that achieves both the graininess of the pattern portion and the character sharpness.
The compression unit 106 performs image compression on the filtered image signal, but an irreversible compression method such as JPEG may be used, or a reversible compression method may be used. Thereafter, the data is stored in the storage unit 107 such as an HDD (hard disk drive), and is output to the color correction unit 110 which is restored to an image signal not compressed by the decompression unit 109 again.

Next, the operation of the color correction unit 110 will be described. Various methods can be considered for the color correction processing. Here, it is assumed that the following masking operation is performed.
[Formula 1]
C = α11 × R + α12 × G + α13 × B + β1
M = α21 × R + α22 × G + α23 × B + β2
Y = α31 × R + α32 × G + α33 × B + β3
However, α11 to α33 and β1 to β3 are predetermined color correction coefficients, and the output CMY is also an 8-bit (0 to 255) signal. The above coefficients convert RGB image signals depending on the device of the image input means 101 (scanner) into CMY signals suitable for the color material of the image output means 115.

Next, the under color removal / black generation means 111 generates a K signal as a black component and performs under color removal (UCR) from the CMY signal. The generation of the K signal and the removal of the under color from the CMY signal are performed as follows.
[Formula 2]
K = Min (C, M, Y) × β4
C ′ = C−K × β5
M ′ = M−K × β5
Y ′ = Y−K × β5
However, Min (C, M, Y) represents the minimum of the CMY signals.
However, β4 and β5 are 8-bit signals with predetermined coefficients.
The C′M′Y′K ′ signal output from the under color removal / black generation unit 111 is input to the printer γ correction unit 112 and the second edge amount calculation unit 113.

In the second edge amount calculation unit 113, the edge amount is calculated again by a method equivalent to that of the first edge amount calculation unit 104 described above. Although the edge amount is calculated for the g signal earlier, the second edge amount calculating unit 111 independently calculates the edge amount for each of the image signals C ′, M ′, Y ′, and K ′. It is configured to calculate. The edge amounts e2 to e5 obtained in this way are input to the halftone processing means 114.
The halftone processing unit 114 is configured so as to realize high-quality image reproduction by continuously controlling characteristics and processing according to the obtained edge amount.

  Another example of halftone processing is disclosed in Japanese Patent Application Laid-Open No. 2001-128004 “Image processing method, image processing apparatus and storage medium”. In Japanese Patent Laid-Open No. 2001-128004, the quantization threshold for error diffusion processing is made dithered, and the dither amplitude is controlled by the edge amount. The sharpness of the character and the pattern are applied by applying a dither threshold value with a large amplitude to the pattern part with a small edge amount, and applying a dither threshold value with a small amplitude or a threshold value with zero amplitude to the outline part of the character with a large edge amount The graininess of the parts is compatible. In the embodiment, there is disclosed a method for controlling the amplitude level by multiplying the basic dither amplitude by the magnification obtained based on the edge amount. The edge amounts e2 to e5 obtained by the second edge amount calculating unit 113 are input to the halftone processing unit 114 that is continuously controlled based on the edge amount as described above, and the halftone process is performed to thereby perform the defect. High-quality image reproduction without any image can be performed.

The halftone processing means of the present invention performs image diffusion that reproduces dots with excellent tone characteristics such that error diffusion is performed using a threshold having an amplitude of such a dither shape, and error diffusion using a threshold having a flat shape. This is a process of selectively switching between image areas that perform dot reproduction with excellent sharpness. A more detailed description of the halftone processing means 114 will be described with reference to FIG. The halftone processing means in this embodiment is performed for each of the four plates C ′, M ′, Y ′, and K ′. FIG. 15 shows an example of processing for the cyan (C ′) plate. However, the same applies to the other versions.
The image signal C ′ from the printer γ correction unit 112 is input to the adder 1140. Adder 1140 adds a signal from error diffusion calculation section 1145 (to be described later) and input signal C ′ to obtain mi, and outputs the result to comparator 1142. The comparator 1142 compares the thresholds thm and mi set by the threshold selection unit 1141 and performs quantization according to the threshold thm. That is, if it is larger than thm, out = 1 (ON dot) is output, and if it is smaller, out = 0 (OFF dot) is output. The quantization result is simultaneously sent to the weight generation unit 1146. The weight generation unit generates a weight of 255 when out = 1 and 0 when out = 0, and outputs the weight to the subtractor 1143. The subtractor 1143 subtracts the weight wt from the progress mi from the adder 1140 and outputs the error value err to the error diffusion calculation unit 1145. The error diffusion calculation unit 1145 uses the error value err based on a predetermined diffusion matrix 1144. Is diffused to surrounding pixels, and a mer signal is output to the adder 1140. The above description is the basic configuration operation of error diffusion, and it is possible to realize halftone processing that enables appropriate density reproduction for an input image.

Here, the threshold selection unit 1141 operates to select from the threshold set 1147 based on the signal s2 signal indicating the number of halftone lines and the e2 signal from the second edge amount calculation unit 113. The threshold set 1147 includes a threshold having a dither shape with excellent gradation characteristics (strong dither), a threshold having a flat shape excellent in sharpness (flat), and a dither threshold having an intermediate shape between the two (a dither threshold). Dither weak) is stored, and the optimum threshold is used by e2 and s2. More specifically, it is as follows.
First, an edge determination threshold set is selected for each pixel based on the dotted line number signal s2. As shown in FIG. 16, different threshold sets are selected according to the number of halftone lines. In the case of a low line number halftone dot, a relatively low threshold value is selected for both th1 and th2, and a higher threshold value is selected as the line number becomes higher or as the non-halftone point is reached. Here, th1 and th2 are determination threshold values for the edge amount e2, and are threshold values for classifying into three areas of a strong edge area, a weak edge area, and a non-edge area. According to the values of th1 and th2 of the selected edge determination threshold set, the area is classified into three areas as shown in FIG. 17, and an area having an edge amount smaller than th1 is determined as a non-edge area. Is selected. A region having an edge amount equal to or greater than th1 and smaller than th2 is determined to be a weak edge region, and a dither weak threshold is selected. Further, a region having an edge amount equal to or greater than th2 is determined as a strong edge region, and a flat threshold value is selected.
In this way, by changing the edge determination threshold according to the dot line number information and controlling the amplitude of the threshold to be applied, high-quality image reproduction that does not interfere with the tone of the halftone dot of the original document can be achieved. An image processing apparatus that can be provided can be provided.

  By the way, the performance of the adaptive filter described above can provide a sufficient smoothing effect for a halftone dot with a high number of lines, but cannot perform smoothing sufficiently for a halftone dot with a low number of lines. There is a feature that the halftone dot structure of the is left. Furthermore, since an adaptive filter based on the edge amount is performed, a smoothing effect is obtained for halftone dots with a specific number of coins or more, but a smoothing effect is sharply applied to halftone images with a number of specific numbers or less. Has the property of getting worse. The actual filter effect will be described with reference to FIG. FIG. 18 is a diagram illustrating the processing effect of the filter processing unit on the halftone image. The solid line shows the undulation (peak to peak) of the halftone dot of the document. In the example of FIG. 18, it is a plot in a halftone dot image with a 30% area ratio. The dotted line indicates the undulation (peak-to-peak) of the halftone image signal after the filter processing means. The horizontal axis is the number of lines. Although the original document also has a lower undulation as the number of lines increases, it can be seen that sufficient smoothing is performed at 133 lpi or more in the example of FIG. On the other hand, it can be seen that smoothing is not sufficiently performed in the region where the number of lines is lower than 133 lpi. This is caused by the original operation of the adaptive filter, and smoothing is not performed on parts where strong edges such as characters exist, but rather sharpness of the characters on the halftone dots is maintained by emphasis. Therefore, it is unavoidable that a high edge is taken for a halftone dot portion having a low number of lines. As described above, there are a line number region that can be sufficiently smoothed by the process using the adaptive filter, and a line number region that cannot be smoothed.

When halftone processing having a dither amplitude is performed on a region where such halftone undulations remain, the halftone of the document interferes with the dither tone by the halftone processing, resulting in a moire-like abnormal image. Therefore, the method of determining the intensity of the dither amplitude based on the information on how much halftone dots remain by the adaptive filter in the previous stage is effective for image reproduction without interference moire, and the present invention realizes this. To do.
That is, by selectively applying the amplitude of the dither threshold of the halftone processing unit based on the halftone line number information in the image area separating unit and the edge amount in the second edge amount calculating unit, a high level of interference moire is not generated. Quality image reproduction is possible.

In the embodiment described above, the threshold value for the edge amount e2 is selectively changed in accordance with the halftone line number information s2, and change control is performed for the areas to which the dither strong, dither weak, and flat threshold values are applied. However, similar effects can be obtained in the following embodiments.
As shown in FIG. 19, the edge amount e2 is threshold-processed according to predetermined threshold values (th1, th2) set in advance, and is classified into a non-edge region, a weak edge region, and a strong edge region, and these classified three regions and lines The dither strong, dither weak, and flat threshold values are selectively used in combination with the numerical information s2. As shown in FIG. 19, in the high line number and non-halftone dot regions, the dither strong to dither weak to flat threshold values are applied as the non-edge to weak edge to strong edge change. However, the dither strength is not used, but the dither weak to flat threshold is applied, and only the flat threshold is applied in the low line number region. Even with this configuration, the effect of reducing interference moiré can be obtained as a flat threshold for a halftone dot region having a small number of lines.
As described above, in a region where a high edge amount is detected or a region where the number of lines is low, a halftone process such as a flat threshold value that generates a dot with a weak periodic structure is applied. Interference moiré with halftone dots can be suppressed.

  As described above, it has been explained that the filter processing means in the previous stage has many halftone dots remaining or is sufficiently smoothed depending on the number of halftone dots. Apart from this, even with halftone dots having the same number of lines, changing the image quality mode changes the remaining degree of undulation of the halftone dots. For example, in a character mode that emphasizes the sharpness of characters, it is common to use an image design that emphasizes edge enhancement compared to the default character photo mode. In other words, the filter processing means performs correction for enhancing the enhancement degree with respect to the detected edge amount. With this control, the remaining halftone dot undulation changes as shown in FIG. It can be seen that halftone dot undulations remain in the long dotted line plot after filtering (character mode) compared to the short dotted line after filtering (character photo mode). This also has a tendency, which means that the undulations in the midline number region are more likely to remain rather than the overall undulations remaining. As shown in FIG. 19, in the halftone dot area of 150 lpi or more, the size of the halftone dot undulation is almost the same in both the character photograph mode and the character mode. The character photograph mode can sufficiently smooth halftone dots of about 133 lpi, whereas the character mode results in undulations remaining. This phenomenon is also caused by an adaptive filter that controls the amount of enhancement based on the edge amount. In a region where the edge amount is sufficiently small (a high line number region such as 150 lpi or more), the filter setting is the same as that in the character photograph mode. The degree is for controlling the degree of emphasis higher than that in the character photo mode. When the characteristics of such filter processing means change, it is desirable to change the threshold setting in the halftone processing section in the subsequent stage.

The state of the control will be described with reference to FIG. FIG. 21 is a table for setting a threshold value for the edge amount e2 in accordance with the number of halftone dots s2 as in FIG. In the filter processing mode in which halftone undulations remain, th1 and th2 of the number of low to medium lines are set to be lower than in FIG. With this configuration, a dither weak dither amplitude or a flat threshold is applied to a smaller edge amount, and dot generation with a weak periodic structure is generated than dither setting in the character photo mode. This makes it easier to select the image, which makes it difficult to interfere with the halftone dot structure of the document, and enables high-quality image reproduction without moire.
The embodiment described above is configured to hold halftone line number information for each pixel. In such a configuration, information on which area of an image is an image having a number of lines of lpi is maintained, so that an optimum halftone process is performed for each area even for an image having a mixed number of halftone lines. Parameters can be applied, and high-quality images can be obtained. However, since the information is stored in units of pixels, the information storage memory in the separated signal storage means is large and has a demerit that increases the cost. In another invention, an apparatus is provided that can reduce the cost by minimizing the information to be stored.

FIG. 22 shows another embodiment. FIG. 22 is basically the same as FIG. 1, but does not store the halftone dot number result s2 by the image area separation unit 103, but performs the minimum line number determination 116 from s2. Of the detected number of lines, the value of the minimum number of lines is stored in the temporary buffer, and when the result of s2 is obtained for the entire image, the information of the lowest number of lines is set to s3 and the line number information holding means 117 is stored. To accumulate. This is the only value such as 133 lpi or 175 lpi, and is the only value held for the page.
The halftone processing unit 114 selects an edge determination threshold set for thresholding e2 based on the minimum line number information s3 (the same operation as in FIG. 16), and based on the obtained edge determination thresholds th1 and th2. It is configured to classify into a non-edge region, a weak edge region, and a strong edge region. In general, halftone lines are rarely mixed in a printed document, and it is often sufficient to hold only single line number information. Also, even if halftone images with multiple lines are mixed in the page, if the minimum number of lines is determined and interference moiré does not occur for the minimum number of lines, then However, interference moiré does not occur with respect to a high number of lines, and a device that does not generate moiré can be provided.

1 is a block diagram of an image processing apparatus according to an embodiment of the present invention. 3 is a configuration diagram of an image area separation unit 103. FIG. It is a figure of a detection pattern. It is a figure of the matrix of a predetermined size. It is a figure which shows the relationship between the number of halftone lines, and the number of peak pixels. It is a figure which shows the relationship between s2 and the number of lines. FIG. 6 is a diagram illustrating an operation of the first edge amount calculation unit 104. It is a figure which shows the 1st differential filter of 7x7 matrix size. FIG. 6 is a diagram showing the operation of the filter processing means 105. It is a figure which shows the filter coefficient of an edge emphasis filter. It is a figure which shows the filter coefficient of an edge emphasis filter. It is a figure which shows the filter coefficient of an edge emphasis filter. It is a figure explaining the adaptive edge emphasis process 1044. FIG. It is a figure which shows an edge amount conversion table. It is a block diagram of a halftone processing means. It is a figure which shows the relationship between the number of halftone lines, and an edge determination threshold value. It is the figure which classified into 3 edge amounts. It is a figure which shows the processing effect of the filter process means with respect to a halftone image. It is a figure which shows the relationship between the number of halftone lines, and edge amount. It is a figure which shows the relationship between the number of halftone dots, and halftone dot relief. It is a figure which shows the relationship between the number of halftone lines, and an edge determination threshold value. It is a block diagram of the image processing apparatus which concerns on other embodiment of this invention. It is a figure showing a truth value.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 100 Image processing apparatus 101 Image input means 102 Scanner gamma correction means 103 Image area separation means 104 First edge amount calculation means 105 Filter processing means 106 Compression means 107 Storage means 108 Separation signal storage means 109 expansion means, 110 color correction means, 111 under color removal / black generation means, 112 printer gamma correction means, 113 second edge amount calculation means, 114 halftone processing means, 115 image output means

Claims (8)

An image area having color image input means for reading a color image, and a line number detection means for determining an image attribute for each area of the image signal read from the color image input means and detecting at least the number of halftone dots in the halftone area Separating means; filter processing means for correcting a spatial frequency characteristic according to a separation result of the image area separating means; edge detecting means for performing edge detection on a processed image processed by the filter processing means; Halftone processing means having threshold selection means for selecting a threshold according to the detection result of the detection means and the dotted line number information detected by the line number detection means,
The image processing apparatus, wherein the halftone processing unit controls generation of a periodic or aperiodic dot by the threshold selection unit.   The threshold selection means selects a threshold value so as to generate a dot having a weak periodic structure in a low line number halftone dot area where the line number information is lower than a predetermined threshold, and other than the low line number halftone dot area The image processing apparatus according to claim 1, wherein a threshold value is selected so as to generate a dot having a strong periodic structure in the region.   The image processing apparatus according to claim 1, wherein a threshold selection condition of the threshold selection unit is changed according to a spatial frequency characteristic corrected by the filter processing unit.   When the spatial frequency characteristic corrected by the filter processing means is a characteristic that emphasizes a higher frequency component, even if the threshold selection condition of the threshold selection means is a high number of lines, weak dot generation with a periodic structure is generated. The image processing apparatus according to claim 3, wherein control is performed so that   A first image processing mode in which character images are emphasized; and a second image processing mode in which image images are emphasized. In the first image processing mode, the second image processing mode is compared with the second image processing mode. The image processing apparatus according to claim 2, wherein the threshold selection condition by the threshold selection unit is controlled to generate a dot having a weak periodic structure even when the number of lines is high.   The image area separating unit stores representative line number information of the number of halftone dots detected by the line number detecting unit as line number information for the page, and changes a threshold selection condition based on the line number information. The image processing apparatus according to claim 1, wherein the threshold selection unit is controlled as described above.   The image area separating unit stores halftone dot information having the lowest line number among the halftone line numbers detected by the line number detecting unit as line number information for the page, and a threshold selection condition based on the line number information The image processing apparatus according to claim 6, wherein the threshold value selection unit is controlled to change the threshold value. An image area having a color image input step for reading a color image, and a line number detection step for determining an image attribute for each area of the image signal read from the color image input step and detecting at least the number of halftone dots in the halftone area A separation step, a filter processing step for correcting a spatial frequency characteristic according to a separation result of the image region separation step, an edge detection step for performing edge detection on the processed image processed by the filter processing step, and the edge A halftone processing step having a threshold selection step of selecting a threshold according to the detection result of the detection step and the halftone line number information detected by the line number detection step,
In the halftone processing step, generation of periodic or aperiodic dots is controlled by the threshold selection step.

Images