JP3984252B2 - Image processing device - Google Patents

Description

  The present invention relates to an image processing apparatus that performs image processing using a spatial filter.

A conventional image creating apparatus includes an image input device realized by a scanner, an image processing device including a spatial filter, and an image output device. The image input device reads a document and generates image data. The image processing apparatus performs a process for improving image quality on image data using a spatial filter. The image output device outputs an image based on image data after image processing . One of image processing using the spatial filter is a process of emphasizing the edge in the image.

As a spatial filter for edge enhancement, a high-frequency enhancement filter shown in FIG. 41 is used. As shown in FIG. 42, the high frequency emphasis filter is an MTF (Modulation Transfer).
Function) has a characteristic that increases uniformly from a predetermined reference value. The graph of the characteristics of the high-frequency emphasis filter approximates a graph of a direct proportional function. When the high-frequency emphasis filter is realized by a 3 × 3 digital filter, the characteristics of the high-frequency spatial filter are determined by a matrix shown in FIG. In the matrix of FIG. 43, the filter coefficient at the center of the matrix is larger than the remaining filter coefficient, and the greater the distance between the filter coefficient at the center and the remaining filter coefficient, the greater the filter coefficient at the center and the remaining filter coefficient. And the difference is small.

  Edge enhancement processing using a spatial filter is performed only on portions corresponding to edges in image data. For this purpose, the image processing apparatus further includes an edge extracting unit that extracts an edge in the image.

  The edge extraction accuracy of the edge extraction unit is determined according to the input accuracy of the image input device. For example, when the input resolution of the image input device is 600 dpi (dot per inch), it is difficult to extract the edge of a line segment that is finer than 600 dpi. Even if an edge of a fine grayscale pattern whose spatial frequency is higher than a certain reference value is extracted by the edge extraction unit, the edge is not extracted. When there is a light and shade pattern with a spatial frequency less than the input resolution in the image, a noise component causing image color shift or turbidity is added to the image data in the image input device. As a result, the spatial frequency band that can be included in the image is divided into a band that includes only the spatial frequency of the grayscale pattern from which the edge extraction unit can extract the edge with a certain spatial frequency as a boundary, and the edge cannot be extracted. It is difficult to separate into a band including only the spatial frequency of the grayscale pattern. When the density pattern of the spatial frequency in the vicinity of the boundary of both bands and its vicinity exists in the image, the edge extraction unit may or may not extract the edge of the density pattern.

  In the image processing device, when the enhancement process is performed based on the extraction result of the edge extraction unit, the enhancement process may or may not be performed on the boundary between the two bands and the spatial frequency shading pattern in the vicinity thereof. To do. As a result, density unevenness occurs in a portion where the gray pattern in the image is present, so that the image quality of the image output based on the image data after the enhancement processing is lowered.

Patent Document 1 , Patent Document 2 and Patent Document 3 disclose edge enhancement processing using a spatial filter, respectively. The image processing apparatus of Patent Document 1 includes an edge amount detection unit, an edge enhancement filter, and a mixer for edge enhancement. The mixer mixes the image data before the edge enhancement filter processing and the image data after the edge enhancement filter processing at a mixing ratio corresponding to the edge amount detected by the detection means.

Prior to the filter processing for edge enhancement, an apparatus using the image processing method of Patent Document 2 divides an image to be processed into a plurality of blocks, determines the strength of the filter to be applied to each block, and The filter strength of each block is corrected so that the strengths of the filters of two adjacent blocks do not become extremely different from each other. After correcting the filter strength, the filter in the apparatus using the image processing method performs the filter processing on each block based on the corrected filter strength obtained for each block.

The MTF correction apparatus of Patent Document 3 includes a smoothing filter, an edge enhancement filter, and a filter that does not perform any processing for edge enhancement. Prior to the filtering process, the MTF correction device obtains an edge degree and a roughness degree, which are numerical values related to the image quality of an image to be processed, for each of all pixels constituting the image, and determines the edge degree and the roughness degree. One of the three filters is selected based on the magnitude relationship with a predetermined reference value. After the filter is selected, edge enhancement processing is performed on each of all the pixels using any one of the selected filters.

JP-A-5-145759 JP-A-5-344345 JP-A-10-28225

Further, as described above, the image processing apparatus of Patent Document 1 needs to detect edge amounts and mix image data in addition to filter processing for edge enhancement in an image. As a result, the processing of the entire apparatus becomes complicated and the structure of the apparatus becomes complicated. Prior to the filter processing, the apparatus using the image processing method of Patent Document 2 needs to detect the filter strength of each block and then detect the boundary surface to correct the filter strength with respect to the boundary surface. . As a result, the apparatus using the image processing method becomes complicated. The MTF correction apparatus of Patent Document 3 requires calculation of the edge degree and roughness and selection processing of three filters. As a result, the processing of the entire MTF correction apparatus becomes complicated and the structure of the apparatus becomes complicated. Further, since the MTF correction device switches between the smoothing filter and the edge enhancement filter based on the reference value, density unevenness may occur in a portion composed of pixels having the edge degree and the roughness degree close to the reference value.

  An object of the present invention is to provide an image processing apparatus capable of simplifying processing for edge enhancement and enhancing edges efficiently.

The present invention provides an edge extraction means for extracting an edge in an image based on image data to be processed;
A spatial filter means for edge enhancement that performs the enhancement processing of the extracted edge on a portion corresponding to the edge in the image data;
The characteristics of the spatial filter means for edge enhancement are determined by a matrix obtained by convolution calculation of a matrix that defines the characteristics of the enhancement filter and a matrix that defines the characteristics of the smoothing filter,
The characteristics of the smoothing filter are:
The reference MTF with a spatial frequency of 0 is 1.0,
Within the band where the spatial frequency is greater than 0 and less than the predetermined upper limit spatial frequency FFC, the MTF approaches 0 as the spatial frequency increases, and the MTF is 0 in the band above the upper limit spatial frequency FFC.
The upper limit spatial frequency FFC is a value exceeding the representative frequency f0.
It is a characteristic that smoothes the components of all spatial frequencies that can be included in the image,
The representative frequency f0 is a spatial frequency that is an average value of an upper limit value and a lower limit value in the erroneous determination frequency band WZ2, which is a band including a spatial frequency that causes an edge extraction error of the edge extraction unit,
The characteristics of the enhancement filter are:
The reference MTF where the spatial frequency is 0 is 1.0, and the MTF increases as the spatial frequency increases.
It is a characteristic that emphasizes the components of all spatial frequencies that can be included in the image,
Thus, the characteristics of the spatial filter means for edge enhancement are
In the erroneous determination frequency band WZ2, it is flat,
Wherein less than the lower limit value of the erroneous determination frequency band, in the spatial frequency band WZ1 edge extraction means comprises only reliably extractable spatial frequency edges, Ri characteristics der to emphasize the spatial frequency component,
It is a characteristic that attenuates the spatial frequency component in the non-extractable band WZ3 that includes only the spatial frequency that exceeds the upper limit value of the erroneous determination frequency band WZ2 and the edge extraction means cannot extract the edge at all.
The image is composed of pixels, the image data is composed of a plurality of color data obtained by color-separating the data of each pixel, and one of the plurality of color data is black data representing the density of the pixels ,
The enhancement filter used for the convolution calculation is an image processing device that performs enhancement processing only on black data .

  According to the present invention, the edge enhancement spatial filter means in the image processing apparatus keeps the spatial frequency component in the erroneous determination frequency band as it is, and converts it to a spatial frequency component less than the lower limit value of the erroneous determination frequency band. Emphasis on it. The misjudgment frequency band is a spatial frequency band corresponding to a spatial frequency boundary of a shading pattern from which an edge can be extracted by the edge detection means. The edge emphasizing spatial filter means can satisfactorily perform edge emphasis processing on a spatial frequency component that is likely to cause an erroneous determination of an edge. That is, the edge enhancement spatial filter means can perform edge enhancement processing uniformly and satisfactorily on the entire image. As a result, the spatial filter means for edge enhancement can prevent the image quality after the edge enhancement processing from being deteriorated due to an edge extraction error in the edge detection means.

  According to the present invention, since the characteristics of the edge enhancement spatial filter means in the image processing apparatus are determined by a single matrix, the edge enhancement spatial filter means is realized by a single digital filter. The spatial filter means for edge enhancement does not need to use a plurality of filters, and does not need to perform complicated processing such as changing the filter coefficient in accordance with the degree of edge enhancement. As a result, the edge emphasizing spatial filter means can easily maintain a flat characteristic in a spatial frequency band including a spatial frequency in which an edge extraction error is likely to occur, and at the same time, a spatial frequency outside the spatial frequency band can be accurately detected. In this case, it is possible to easily emphasize an edge of a grayscale pattern having a spatial frequency lower than the spatial frequency at which an edge extraction error may occur. Since the edge enhancement spatial filter means is realized by a single digital filter, the processing speed of the edge enhancement spatial filter means is improved, and the circuit configuration of the edge enhancement spatial filter means is simplified.

  According to the present invention, the spatial filter means for edge enhancement in the image processing apparatus includes image data composed of a plurality of color data of each pixel, and one of the plurality of color data is black data. In some cases, preferably only the black data for each pixel is processed. As a result, the edge enhancement spatial filter means performs edge enhancement processing only on black edges and does not perform enhancement processing on color edges. This is for the following reason. When the image subjected to the edge enhancement process is a character image, it is a black edge in the character image that the edge enhancement process is considered to have an appropriate effect. The color edge in the image to be subjected to the edge enhancement process is not limited to the edge of the color character, but may be the outline of the region existing in the image. Therefore, it is not always good to perform edge enhancement processing on color edges. Therefore, the spatial filter means for edge enhancement in the image processing apparatus performs processing only on black data representing the shading of pixels and does not perform processing on other color data representing pixel colors. As a result, the spatial filter means for edge enhancement can perform the edge enhancement processing more effectively, so that the quality of the image with the edge enhanced can be further improved.

The present invention also provides an edge extraction means for extracting an edge in an image based on image data to be processed;
A spatial filter means for edge enhancement that performs the enhancement processing of the extracted edge on a portion corresponding to the edge in the image data;
Characteristics of the spatial filter means for or falling edge of di emphasized, and a matrix defining the characteristics of the matrix and the smoothing filter for determining the characteristics of the enhancement filter is defined by a matrix obtained by convolution,
The characteristics of the smoothing filter are:
The reference MTF with a spatial frequency of 0 is 1.0,
Within the band where the spatial frequency is greater than 0 and less than the predetermined upper limit spatial frequency FFC, the MTF approaches 0 as the spatial frequency increases, and
MTF is 0 in the band above the upper limit spatial frequency FFC,
The upper limit spatial frequency FFC is a value exceeding the representative frequency f0.
It is a characteristic that smoothes the components of all spatial frequencies that can be included in the image,
The representative frequency f0 is a spatial frequency that is an average value of an upper limit value and a lower limit value in the erroneous determination frequency band WZ2, which is a band including a spatial frequency that causes an edge extraction error of the edge extraction unit,
The characteristics of the enhancement filter are:
A reference MTF with a spatial frequency of 0 is 1.0, and
The MTF increases as the spatial frequency increases,
It is a characteristic that emphasizes the components of all spatial frequencies that can be included in the image,
Thus, the characteristics of the spatial filter means for edge enhancement are
In the erroneous determination frequency band WZ2, it is flat,
The spatial frequency component is less than the lower limit value of the erroneous determination frequency band, and the spatial frequency component is emphasized in the spatial frequency band WZ1 including only the spatial frequency from which the edge extraction unit can reliably extract the edge,
It is a characteristic that attenuates the spatial frequency component in the non-extractable band WZ3 that includes only the spatial frequency that exceeds the upper limit value of the erroneous determination frequency band WZ2 and the edge extraction means cannot extract the edge at all.
When the image is constituted by pixels, and the pixel data is constituted by pixel luminance data and color difference data, the edge enhancement spatial filter means performs processing only on the luminance data of each pixel. a image picture processor you.

であり、
強調フィルタの特性を定めるマトリクスＭＣ２は、

であり、
平滑フィルタの特性を定めるマトリクスＭＣ１と強調フィルタの特性を定めるマトリクスＭＣ２とのコンボリューション演算して得られるマトリクスＭＣ３は、

であることを特徴とする。 According to the present invention, the edge enhancement spatial filter means in the image processing apparatus performs processing only on the luminance data when the image data is constituted by the luminance data and color difference data of the pixels. This is for the following reason. The edge-enhanced spatial filter means mainly aims to improve the quality of the image, and converts the characteristics of the spatial frequency of the image to be processed to a desired characteristic. If the appearance of the color of the image changes as a result of converting the spatial frequency characteristics of the image to be processed, the quality of the image after conversion is lower than the quality of the image before conversion. Therefore, the spatial filter means for edge enhancement preferably converts the spatial frequency characteristics of the processing target image so that the appearance of the color of the processing target image is maintained. The edge emphasizing spatial filter means in the image processing apparatus performs processing only on the luminance data representing the density of the pixel, and not on the color difference data representing the color of the pixel. As a result, the edge enhancement spatial filter can satisfactorily perform edge enhancement processing on image data without degrading image quality.
In the present invention, the representative frequency f0 is 6.5 lines / mm,
The matrix MC1 that defines the characteristics of the smoothing filter is:

And
The matrix MC2 that defines the characteristics of the enhancement filter is:

And
A matrix MC3 obtained by convolution calculation of the matrix MC1 that determines the characteristics of the smoothing filter and the matrix MC2 that determines the characteristics of the enhancement filter is:

It is characterized by being.

  As described above, according to the present invention, the characteristic of the spatial filter means for edge enhancement is flat in the erroneous determination frequency band, and the spatial frequency component is reduced in the spatial frequency below the lower limit value of the erroneous determination frequency band. It is a characteristic to emphasize. As a result, the edge enhancement spatial filter means can perform edge enhancement processing uniformly and satisfactorily on the image as a whole, thereby preventing degradation in image quality after edge enhancement processing. it can.

  According to the invention, the matrix that determines the characteristics of the spatial filter means for edge enhancement is obtained by convolution calculation of a matrix for determining the characteristics of the enhancement filter and a matrix for determining the characteristics of the smoothing filter. It is done. As a result, the spatial filter means for edge enhancement is realized by a single digital filter, so that the processing speed is improved and the circuit configuration is simplified.

  Further, according to the present invention, the edge emphasizing spatial filter means comprises a plurality of color data and image data, and one of the plurality of color data is black data. Process only the data. This prevents a change in the color of the image, so that the edge enhancement spatial filter means can perform the edge enhancement processing more effectively.

  Furthermore, according to the present invention, the edge enhancement spatial filter means performs processing only on luminance data when image data is constituted by pixel luminance data and color difference data. As a result, the spatial enhancement filter for edge enhancement prevents changes in the color of the image, so that the edge enhancement processing can be applied to the image data without degrading the image quality.

  FIG. 1 is a schematic front sectional view showing a configuration of an image forming apparatus 11 provided with an image processing apparatus as a reference example of the present invention. The image forming apparatus 11 is realized by a digital color copying machine. Inside the main body of the image forming apparatus 1, an image input device 13 as an image input unit, an image processing device 14, and an image output device 15 as an image output unit are provided. A transparent document table 311 and an operation device 16 are provided on the upper surface of the main body of the image forming apparatus 11.

On the upper surface of the document table 311, a double-sided automatic document feeder (RADF: Reversing Automatic
Document Feeder) 312 is attached. The double-sided automatic document feeder 312 is supported in an openable / closable state with respect to the document table 311 and has a predetermined positional relationship with the surface of the document table 311. The double-sided automatic document feeder 312 first transports a document so that one side of the document faces the image input device 13 at a predetermined position on the document table 311, and after reading of the one side is finished, The document is conveyed toward the document table 311 with the front and back sides thereof reversed so that the other side of the document is opposed to the image input device 13 at a predetermined position of the document table 311. The document conveying operation and the front / back reversing operation in the duplex automatic document feeder 312 are controlled in relation to the entire operation of the image forming apparatus 11.

  The image input device 13 is disposed below the document table 311 in order to read an image of the document conveyed on the document table 311 by the double-sided automatic document feeder 312. The image input device 13 includes a document scanning body that reciprocates in parallel with the lower surface of the document table 311, an optical lens 315, and a CCD (Charge Coupled Device) line sensor 316 that is a photoelectric conversion element.

  The document scanning body includes a first scanning unit 313 and a second scanning unit 314. The first scanning unit 313 includes an exposure lamp that exposes the surface of the document, and a first mirror that deflects a reflected light image from the document in a predetermined first direction. The first scanning unit 313 reciprocates in parallel with the lower surface of the document table 311 while maintaining a predetermined distance with respect to the lower surface of the document table 311. The second scanning unit 314 includes a second mirror and a second mirror in order to further deflect the reflected light image from the document deflected by the first mirror of the first scanning unit 313 in a predetermined second direction. Has 3 mirrors. The second scanning unit 314 reciprocates in parallel with the lower surface of the document table 311 while maintaining a predetermined speed relationship with the first scanning unit 313.

  The optical lens 315 reduces the reflected light image deflected by the third mirror of the second scanning unit, and forms the reduced reflected light image at a predetermined position on the CCD line sensor 316. The CCD line sensor 316 sequentially photoelectrically converts the formed reflected light image and outputs an analog image signal which is an electrical signal. Specifically, the CCD line sensor 316 provided in the image forming apparatus 11 is a three-line color CCD line sensor. The three-line color CCD line sensor reads a black and white image or a color image, and separates the reflected light image into images of color components of red (R), green (G), and blue (B). A reflectance signal composed of a signal corresponding to the component image is output. The reflectance signal generated by the CCD line sensor 316 is given to the image processing device 14. In the following description, the three colors of red, green, and blue are collectively referred to as “RGB”.

  The image processing device 14 performs predetermined processing described later on the given RGB reflectance signal. As a result, the RGB reflectance signals are digital image signals (hereinafter referred to as “image data”) composed of signals corresponding to cyan (C), magenta (M), yellow (Y), and black (K) color component images, respectively. "). In the following description, the four colors of cyan, magenta, yellow, and black are collectively referred to as “CMYK”.

  Below the image output device 15, a paper feed mechanism 211 having a paper tray is provided. In the image forming apparatus 11 of this reference example, a cut sheet-like paper that is a recording medium is used as the paper P. In general, the paper feed mechanism 221 separates the paper P in the paper tray one by one and supplies it to the image output device 15. When duplex printing is performed, the paper feed mechanism 211 transports the paper P on which an image is formed on one side so that the paper P is re-supplied to the image output device 15 in accordance with the image formation timing of the image output device 15. . In the paper feed mechanism 221, when the paper P delivered from the paper tray is supplied into the guide of the paper feed transport path in the paper feed mechanism 221, a sensor provided in the paper feed transport path is used to detect the paper P. Detecting the tip part of the sensor and outputting a detection signal.

  A pair of registration rollers 212 is disposed in front of the image output device 15. The registration roller 212 temporarily stops the transported paper P based on a detection signal from a sensor in the paper feed transport path. The separately supplied paper P is conveyed to the image output device 15 at a supply timing controlled by the registration roller 212.

  In the image output device 15, the first image forming portion Pa, the second image forming portion Pb, the third image forming portion Pc, the fourth image forming portion Pd, the transfer conveyance belt mechanism 213, and the fixing device 217 are included. Are provided. The transfer / conveying belt mechanism 213 is disposed in a lower portion in the image output device 15. The transfer / conveying belt mechanism 213 includes a driving roller 214, a driven roller 215, and a transfer / conveying belt 216 stretched between these rollers 214 and 215 so as to extend substantially in parallel. A pattern image detection unit 232 is provided in the vicinity of the lower side of the transfer conveyance belt 216.

  The first image forming unit Pa, the second image forming unit Pb, the third image forming unit Pc, and the fourth image forming unit Pd are located above the transfer conveyance belt 216 in the image output device 15 and close to the transfer conveyance belt 216. Then, they are arranged in order from the upstream side of the sheet conveyance path. The fixing device 217 is disposed on the downstream side of the transfer conveyance belt mechanism 213 in the sheet conveyance path. A paper suction charger 228 is provided between the first image forming portion Pa and the paper feed mechanism 21. A static eliminator 229 for separating the paper is provided between the fourth image forming unit Pd and the fixing device 217 and at a position almost directly above the driving roller 214.

  The transfer / conveying belt mechanism 213 is generally configured to convey the paper P supplied by the registration rollers 212 while electrostatically attracting the sheet P to the transfer / conveying belt 216, and specifically operates as follows. To do. The sheet suction charger 228 charges the surface of the transfer conveyance belt 216. The transfer conveyance belt 216 is driven in the direction indicated by the arrow Z in FIG. Therefore, the transfer / conveying belt 216 holds the paper P fed through the paper feeding mechanism 211, and the first image forming unit Pa, the second image forming unit Pb, the third image forming unit Pc, and the fourth image forming unit. The paper P is sequentially conveyed to Pd. Since the surface of the transfer / conveying belt 216 is charged by the sheet adsorbing charger 228, the transfer / conveying belt mechanism 213 securely adsorbs the sheet P supplied from the paper feeding mechanism 211 to the transfer / conveying belt 216. The paper P can be stably conveyed from the first image forming portion Pa to the fourth image forming portion Pd so as not to cause a positional shift. In order to separate the electrostatically adsorbed paper P from the transfer conveyance belt 216, an alternating current is applied to the static eliminator 229 for paper separation.

  The first image forming unit Pa, the second image forming unit Pb, the third image forming unit Pc, and the fourth image forming unit Pd have substantially the same configuration. Each image forming apparatus Pa. Pb, Pc, and Pd include photosensitive drums 222a, 222b, 222c, and 222d that are driven to rotate in the direction indicated by arrow F in FIG. Around the photosensitive drums 222a to 222d, chargers 223a, 223b, 223c, 223d, developing devices 224a, 224b, 224c, 224d, transfer members 225a, 225b, 225c, 225d, and cleaning devices 226a, 226b, 226c and 226d are sequentially arranged along the rotation direction F of the photosensitive drums 222a to 222d.

  Laser beam scanner units 227a, 227b, 227c, and 227d are provided above the photosensitive drums 222a to 222d, respectively. The laser beam scanner units 227a to 227d include semiconductor laser elements, polygon mirrors 240a, 240b, 240c, and 240d that are deflection devices, fθ lenses 241a, 241b, 241c, and 241d, and two mirrors 242a, 242b, 242c, and 242d. ; 243a, 243b, 243c, 243d, respectively. In FIG. 1, the semiconductor laser element is not shown.

  A signal corresponding to the black color component image of the color document image in the image data from the image processing device 14 is input to the laser beam scanner unit 227a of the first image forming unit Pa. A signal corresponding to the cyan color component image of the color document image in the image data is input to the laser beam scanner unit 227b of the second image forming unit Pb. A signal corresponding to the magenta color component image of the color original image in the image data is input to the laser beam scanner unit 227c of the third image forming unit Pc. A signal corresponding to the yellow color component image of the color original image in the image data is input to the laser beam scanner unit 227d of the fourth image forming unit Pd.

  The chargers 223a to 223d uniformly charge the photosensitive drums 222a to 222d, respectively. The semiconductor laser elements of the laser beam scanner units 227a to 227d emit laser beams that are modulated in accordance with signals given to the units. Polygon mirrors 240a to 240d deflect laser light from the semiconductor laser element in a predetermined main scanning direction. The fθ lenses 241a to 241d and the mirrors 242a to 242d; 243a to 243d image the laser beams deflected by the polygon mirrors 240a to 240d on the surfaces of the charged photosensitive drums 222a to 222d. As a result, electrostatic latent images corresponding to the four color components of the color original image are formed on the photosensitive drums 222a to 222d.

  The developing device 224a of the first image forming unit Pa contains black toner. The developing device 224b of the second image forming unit Pb contains cyan toner. The developing device 224c of the third image forming unit Pc contains magenta toner. The developing device 224d of the fourth image forming unit Pd contains yellow toner. The developing devices 224a to 224d develop the electrostatic latent images on the photosensitive drums 222a to 222d with toner stored therein. As a result, the image output device 15 reproduces the original image as black, cyan, magenta, and yellow toner images.

  The transfer members 225a to 225d transfer the toner images on the photosensitive drums 222a to 222d onto one surface of the paper P conveyed by the transfer conveyance belt 216. The cleaning devices 226a to 226d remove toner remaining on the photosensitive drums 222a to 222d after the transfer, respectively.

  After the transfer of the toner image is completed in the fourth image forming unit Pd, the sheet P is sequentially peeled from the transfer conveyance belt 219 from the leading end portion by the static eliminator 229 and guided to the fixing device 217. The fixing device 217 fixes the toner image transferred onto the paper P onto the paper P.

  The sheet P that has passed through the nip between the fixing rollers provided in the fixing device 217 passes through the conveyance direction switching gate 218. The transport direction switching gate 218 passes the transport path of the paper P after the toner image is fixed, the first path for discharging the paper P out of the main body of the image forming apparatus 11, and the paper P toward the image output apparatus 15. Selectively switch between a second path for resupply. When the transport path is switched to the first path by the transport direction switching gate 218, the paper P is discharged onto the paper discharge tray 220 attached to the outer wall of the main body of the image forming apparatus 11 by the discharge roller. When the transport path is switched to the second path by the transport direction switching gate 218, the paper P is transported to the switchback transport path 221, the front and back are reversed by the switchback transport path 221, and then passed through the paper feed mechanism 211. Then, it is supplied again to the image output device 15.

  In the above description, optical writing is performed on the photoconductors 222a to 222d by scanning and exposing the laser beams with the laser beam scanner units 227a to 227d. Instead of the laser beam scanner units 227a to 227d, an LED (Light Emitting Diode) head which is a writing optical system including a light emitting diode array and an imaging lens array may be used. The LED head is silent compared to the laser beam scanner unit and is silent because there are no moving parts. Therefore, an LED head is preferably used in a tandem digital color image forming apparatus that requires a plurality of optical writing units.

  In the above description, since the CMYK image data is output from the image processing device 14, the image output device 15 is realized by an electrophotographic printer. The image output device 15 is not limited to an electrophotographic printer, and may be realized by another device, for example, an ink jet printer.

The image processing apparatus 14 may output not only CMYK image data but also other image data, for example, RGB image data or L ^* a ^* b ^* image data. In this case, the configuration of the image output device 15 and the configuration of the image forming device 11 are in accordance with the configuration of image data output from the image output device. Hereinafter, the description of FIGS. 2 to 22 is an example of a configuration in which RGB data is output from the image processing apparatus 14. When RGB data is output, the image output device 15 is realized by a display device, for example.

  FIG. 2 is a block diagram illustrating a configuration of an image forming apparatus 11 including an image processing apparatus 14 that is a reference example of the present invention. In addition to the image processing device 14, the image forming device 11 includes an image input device 13, an image output device 15, and an operation device 16. The image input device 13 inputs an image to be processed. The image processing device 14 performs a predetermined process on the input image. The image output device 15 outputs the image processed by the image processing device 14. The operation device 16 is used when a user of the image forming apparatus 11 gives an instruction to the image forming apparatus 11. The input resolution of the image in the image input device 13 and the output resolution of the image in the image output device 15 are determined in advance. In this reference example, the image input device 13 will be described as a scanner device capable of reading a color image. The scanner device reads the surface of a document to be read with a predetermined input resolution, generates an RGB reflectance signal as an analog image signal indicating an image on the document, and provides the image processing device 14 with the RGB reflectance signal.

  The image processing apparatus 14 includes an analog / digital (hereinafter abbreviated as “A / D”) conversion unit 21, a shading correction unit 22, an input tone correction unit 23, a color correction unit 24, an image area separation processing unit 25, and filter control. Section 26, spatial filter processing section 27, and halftone output gradation processing section 28.

  The A / D converter 21 converts the RGB color system reflectance signal given from the image input device 13 into a digital signal. The shading correction unit performs a shading correction process on the A / D converted reflectance signal. The shading correction process is performed to remove various distortions that occur in the image signal due to the configuration of the illumination system, the imaging system, and the imaging system of the image input device 13. The input tone correction processing unit 23 performs input tone processing on the reflectance signal that has been subjected to the shading correction processing. The input tone correction process is a process for converting the reflectance signal into a signal that can be easily handled by the image processing apparatus 14 such as a density signal. The input tone correction unit 23 may further perform color balance processing on the reflectance signal. The color correction unit 24 performs a color correction process on the density signal output from the input tone correction unit 23 in order to realize faithful color reproduction in the image output device 15. In the following description, A / D conversion, shading correction processing, input tone correction processing, and color correction processing are collectively referred to as “pre-processing”. The density signal of the preprocessed image is a digital signal composed of red, green and blue (hereinafter referred to as “RGB”) data of each pixel constituting the image, and hereinafter referred to as “image data”. Called. The RGB data of the pixel is obtained by separating the pixel into three colors of RGB. The image data is given to the image area separation processing unit 25 and the spatial filter processing unit 27.

  The image area separation processing unit 25 performs area separation processing based on image data. The area separation process is a process for discriminating characters, photographs and halftone dots and separating them into character areas, photograph areas and halftone dot areas. The image area separation processing unit 25 corresponds to halftone dot area extraction means for extracting a halftone dot area in an image. The halftone dot area corresponds to a halftone displayed portion using halftone dots in an image on a document. There are a plurality of types of halftone dots having different linear densities. In order to determine whether or not an arbitrary region in the image is a halftone dot region, it is only necessary to determine whether or not the shading pattern in the arbitrary region has periodicity using, for example, a Fourier transform method. . When the spatial frequency distribution of an area is measured using the Fourier transform method, if the distribution is biased to a specific band, the shading pattern in the area has periodicity, so that the area is a halftone dot area. To be judged. The processing result of the image area separation processing unit 25 is given to the filter control unit 26 and the halftone output gradation processing unit 28. The processing result of the image area separation processing unit 25 may be further given to the image output device 15.

  The filter control unit 26 adjusts the characteristics of the spatial filter processing unit 27 for each region in the image based on the processing result of the image region separation processing unit 25 and the image data after the preprocessing. The spatial filter processing unit 27 performs spatial filter processing on the pre-processed image data based on the adjusted characteristics to limit the spatial frequency component of the region in the image. Details of the spatial filter processing unit 27 will be described later.

  The halftone output tone processing unit 28 performs tone correction processing and halftone generation processing on the image data output from the spatial filter processing unit 27. The halftone generation process is a process for dividing an image into a plurality of pixels so that the gradation of each pixel can be reproduced. The halftone output gradation processing unit 28 may perform processing for converting the density value of the image data into a halftone dot area ratio that is a characteristic value of the image output device 15. In the following description, tone correction processing and halftone generation processing are collectively referred to as “post-processing”. Data of the post-processed image is given to the image output device 15.

  FIG. 3 is a diagram for explaining a process from image input to image output in the image forming apparatus 11 of FIG. When the user operates the operation device 16 to instruct to copy the document, the image input device 13 reads the document and gives the read result to the image processing device 14 as processing in step S1. As processing of step S2, the part from the A / D conversion unit 21 to the color correction unit 24 of the image processing apparatus 14 performs preprocessing. As the processing in step S3, the image area separation processing unit 25 performs area separation processing.

  The spatial filter processing unit 27 performs the moire of the halftone dot region on the portion corresponding to the halftone dot region of the preprocessed image data (hereinafter referred to as “halftone dot portion”) as the process of step S4. A spatial filter process is performed for removal. Spatial filter processing is performed individually for each piece of RGB data. The characteristics of the spatial filter processing unit 27 when processing RGB data are equal to each other. The spatial filter processing unit 27 may output a portion other than the halftone dot portion of the image data without performing any spatial filter processing, and a spatial filter corresponding to the spatial frequency characteristics of the region other than the halftone dot region of the image. You may output, after processing. The halftone output tone processing unit 28 performs post-processing on the image data on which the spatial filter processing has been performed as the processing in step S <b> 5, and provides the processed data to the image output device 15. The image output device 15 outputs an image based on the post-processed image data as the process of step S6.

  The image processing apparatus 14 uses the spatial filter processing unit 27 for the purpose of suppressing moire generated in the halftone area. A main cause of generation of moire in the halftone dot region is due to period interference associated with the combination of the input resolution of the image input device 13 and the output resolution of the image output device 15. When the input resolution and the output resolution are different from each other, moire is formed because the halftone dot formation period interferes. In the image forming apparatus 11, the input resolution and the output resolution are determined in advance. Therefore, the spatial frequency at which moiré is most likely to occur can be obtained in advance by a model based on a combination of a halftone dot formation period in the image input device 13 and a halftone dot formation period in the image output device 15.

  FIG. 4 is a graph schematically showing the spatial frequency characteristics of a halftone dot region using 65-line halftone dots (hereinafter abbreviated as “half-tone dot region of 65-line halftone dot”). FIG. 5 is a graph schematically showing a spatial frequency characteristic of a halftone dot region using 133-line halftone dots (hereinafter abbreviated as “halftone dot region of 133-line halftone dot”). 4 and 5 illustrate an example in which the input resolution is 600 dpi, the output resolution is 600 dpi, and the image is represented by a dither halftone display method using a 3 × 3 dither matrix. . 4 and 5, “main scanning spatial frequency” indicates the spatial frequency of the halftone dot in the same direction as the scanning direction of the image input device 13, and “sub-scanning spatial frequency” indicates the scanning of the image input device 13. The spatial frequency of a halftone dot in a direction orthogonal to the direction is shown. A plane defined by the main scanning and sub-scanning spatial frequency axes is defined as a spatial frequency plane. The unit of the spatial frequency is the number of lines per 1 mm [lines / mm], that is, the linear density [lpi]. The vertical axis of the graph is MTF (Modulation Transfer Function).

  In the halftone dot region of the 65-line halftone dot in FIG. 4, the MTF of the point PF1 where the main scanning and sub-scanning spatial frequencies in the spatial frequency plane are each 4 lines / mm is the maximum, and the point PF1 is the center. It can be seen that the MTFs at the points in the band WF1 for ± 1 line / mm are mountain-shaped. Further, in the halftone dot region of the 133-line halftone dot in FIG. 5, the MTF of the point PF2 where the main scanning and sub-scanning spatial frequencies in the spatial frequency plane are each 8 lines / mm is the maximum, and the point PF2 is the center. As can be seen, the MTFs at points in the band WF2 for ± 1 line / mm are mountain-shaped. In FIGS. 4 and 5, portions in the spatial frequency plane corresponding to the bands WF1 and WF2 are indicated by hatching. In the following description, the spatial frequency characteristic graph of the halftone dot region shown in FIGS. 4 and 5 is simplified by replacing the spatial frequency plane with a one-dimensional spatial frequency axis as shown in FIGS. Shown in the form.

  The image processing apparatus 14 of the present reference example is characterized by the characteristics of the spatial filter processing unit 27 when performing spatial filter processing for moire removal. Hereinafter, the characteristics of the spatial filter processing unit 27 at the time of moire removal will be described in detail by taking as an example a halftone dot area of 65 line halftone dots and a halftone dot area of 133 line halftone dots. In the present specification, a spatial frequency band including all spatial frequencies that can be included in an image to be processed is referred to as an “allowable spatial frequency band”. The minimum spatial frequency that an image can contain is zero.

  The characteristics of the spatial filter processing unit 27 when performing the spatial filter processing for removing moiré are the characteristics of a predetermined spatial frequency (hereinafter referred to as “moire frequency”) that attenuates all spatial frequency components that can be included in the image and can cause moire. ")") Is further attenuated or removed. The spatial filter processing unit 27 in a state where such characteristics are set corresponds to the spatial filter means for removing moire. In the halftone dot region, not only the moire frequency component but also the spatial frequency component in the vicinity of the moire frequency can be a cause of moire. A spatial frequency band that is composed of a moire frequency and a spatial frequency in the vicinity of the moire frequency and is likely to generate moire is abbreviated as a “moire frequency region”.

  FIG. 8 is a graph of the ideal characteristic GA1 for the 65-line halftone dot of the spatial filter processing unit 27. FIG. 9 is a graph of the ideal characteristic GB1 for the 133-line halftone dot of the spatial filter processing unit 27. In FIG. 8, “FMA” indicates the moire frequency at the 65-line halftone dot, and in FIG. 9, “FMB” indicates the moire frequency at the 133-line halftone dot. In the graph of the filter characteristics of the present specification, the horizontal axis is obtained by replacing the spatial frequency plane with a one-dimensional spatial frequency axis, and the vertical axis is MTF. When the MTF is less than 1.0, the filter characteristics tend to attenuate. When the MTF exceeds 1.0, the filter characteristics tend to be emphasized.

  The characteristics GA1 and GB1 of the ideal spatial filter processing unit 27 are basically characteristics of an attenuation tendency, and are characteristics in which a spatial frequency band in which moire is particularly likely to occur is suppressed. For this reason, the MTFs in the characteristics GA1 and GB1 of the ideal spatial filter processing unit 27 are attenuated more than the MTF in the case where the spatial frequency is 0 (hereinafter referred to as “reference MTF”) over the entire allowable spatial frequency band, and moire. It is a minimum value at the frequencies FMA and FMB. The reference MTF is 1.0.

  The spatial filter processing unit 27 is preferably realized by a digital filter. The characteristics of the spatial filter processing unit 27 composed of a digital filter are determined by a matrix composed of a plurality of filter coefficients. The matrix is used when performing calculations for spatial filtering on the data of the pixels constituting the block set in the image based on any one pixel (hereinafter referred to as “target pixel”). . The calculation for the spatial filter processing is performed a plurality of times while changing the target pixel. The filter coefficient is a weighting coefficient for pixel data.

  The matrix that defines the characteristics of the spatial filter processing unit 27 for removing moire is a matrix that defines the characteristics of a filter (hereinafter referred to as “first filter”) having a characteristic of attenuating or removing the spatial frequency component of the moire frequency; It is obtained by convolution calculation with a matrix that defines the characteristics of the smoothing filter. In the following description, a matrix that defines the characteristics of the spatial filter processing unit 27 for removing moire is referred to as a “spatial filter matrix”, and a matrix that defines the characteristics of the first filter is referred to as a “first filter matrix”. The matrix that defines the characteristics is referred to as a “smoothing filter matrix”.

  FIG. 10 is a graph showing an ideal characteristic GA3 of the first filter used for defining the characteristic of the ideal spatial filter processing unit 27 for the 65-line halftone dot shown in FIG. FIG. 11 is a graph showing an ideal characteristic GB3 of the first filter used for defining the characteristic of the ideal spatial filter processing unit 27 for the 133-line halftone dot shown in FIG. In the ideal characteristics GA3 and GB of the first filter, the MTFs of the moire frequencies FMA and FMB are minimum values, and the MTF of the moire frequency region is less than 1.0.

  The first filter is preferably realized by a band cut filter in which the MTF has a minimum value at the moire frequency. The cut-off frequency of such a band cut filter matches the moire frequency. The characteristics of the first filter are not limited to the characteristics of the band cut filter and may be realized by characteristics of other filters as long as the characteristics of the filter have characteristics of attenuating or removing the moire frequency component.

  The characteristic of the smoothing filter is to attenuate all spatial frequency components that can be included in the image. FIG. 12 is a graph showing the ideal characteristic GA4 of the smoothing filter used for defining the characteristic of the ideal spatial filter processing unit 27 for the 65-line halftone dot shown in FIG. FIG. 13 is a graph showing the ideal characteristic GB4 of the smoothing filter used for defining the characteristic of the ideal spatial filter processing unit 27 for the 133-line halftone dot shown in FIG. In ideal smoothing filter characteristics GA4 and GB5, the MTFs of all spatial frequencies in the allowable spatial frequency band are 1.0 or less. As long as the MTF of the smoothing filter is always 1.0 or less, it may be changed in any manner as the spatial frequency changes.

  The characteristics of the spatial filter processing unit 27 are the characteristics described with reference to FIGS. 8 and 9 for the following reason. When a 65-line halftone dot is used in the halftone dot region, the 65-line halftone dot spatial frequency characteristic as shown in FIG. 6 is obtained by obtaining the moire frequency of the 65-line halftone dot in advance. In the obtained spatial frequency characteristics of the 65-line halftone dot, since the band in which the MTF is emphasized in a mountain shape corresponds to the moire frequency region, the spatial filter processing unit 27 attenuates the spatial frequency component in the moire frequency region. Like that. For attenuation of the spatial frequency component in the moire frequency region, a first filter that attenuates or removes the spatial frequency component in the moire frequency region as shown in FIG. 10 is used.

  In the halftone dot region, moire due to a spatial frequency in a band other than the moire frequency region occurs in addition to the moire due to the spatial frequency in the moire frequency region. Moire caused by the spatial frequency in other bands is less visually noticeable than moire caused by the spatial frequency in the moire frequency range. When only the first filter of FIG. 10 is used for the attenuation processing of the spatial frequency component in the halftone dot region of the 65-line halftone dot, only the occurrence of moire due to the spatial frequency in the moire frequency region is suppressed, and Moire caused by the spatial frequency is generated as it is. As a result, the moire suppression state of the entire halftone dot area after the attenuation process is not unified, and the quality of the halftone dot area is degraded. Therefore, in order to suppress moire caused by the spatial frequency in the other band, it is necessary to perform a smoothing process on the component of the spatial frequency in the other band. For smoothing the spatial frequency components in other bands, as shown in FIG. 12, a smoothing filter that attenuates the spatial frequency components in the entire allowable spatial frequency band is used.

  Based on the reason described above, the characteristics of the ideal spatial filter processing unit 27 for the 65-line halftone dot are the combination of the characteristics of the first filter of FIG. 10 and the characteristics of the smoothing filter of FIG. 12, as shown in FIG. It has the characteristics of a bandpass filter obtained together. As a result, the spatial filter processing unit 27 can perform a smoothing process for attenuating all spatial frequency components in the remaining bands other than the moire frequency band while performing a process for particularly attenuating the spatial frequency components in the moire frequency band. . When the spatial filter processing unit 27 having the characteristics shown in FIG. 8 is used, the moire is suppressed to a level that is not sufficiently conspicuous over the entire allowable spatial frequency band. As a result, the spatial filter processing unit 27 can further improve the quality of the image after the spatial filter processing, as compared with the band cut filter that performs attenuation processing only on the spatial frequency components in the moire frequency region.

  When the halftone dot used in the halftone dot region is a 133-line halftone dot, as in the case where the halftone dot is a 65-line halftone dot, first, based on the input resolution and the output resolution, as shown in FIG. Moire frequency range is required. Next, the characteristics of the filter that attenuates the spatial frequency components in the entire allowable spatial frequency band and further attenuates or removes the obtained spatial frequency components in the moire frequency band are shown in FIG. And the characteristics of the smoothing filter. If the spatial filter processing is performed using the spatial filter processing unit 27 having the characteristics shown in FIG. 9, the generation factor of moire in the halftone dot region is visually reduced over the entire wavelength region. Thereby, the quality of the image after the spatial filter processing using the spatial filter processing unit 27 is further improved than the quality of the image after the spatial filter processing using only the band cut filter.

  Based on the reason described above, the setting of the spatial filter matrix is performed based on the moire frequency range obtained according to the combination of the input resolution and the output resolution. The filter coefficient of the spatial filter matrix is set by a convolution operation between the first filter matrix in which the filter coefficient is set according to the moire frequency region and the smoothing filter matrix. As a result, it is possible to provide a filter that is particularly effective for the moire frequency band and has a moderate effect for the remaining spatial frequency band.

  In the spatial filter processing in step S4 in FIG. 3, the spatial filter processing target of the present reference example is a halftone dot region, and the characteristics of the spatial filter processing unit 27 change according to the type of halftone dot. Therefore, prior to the execution of the spatial filter processing, a spatial filter matrix corresponding to the type of halftone dot is determined in advance. The determined spatial filter matrix is stored in a storage unit provided in the spatial filter processing unit 27 or the filter control unit 26. When two or more spatial filter matrices are prepared, it is preferable that the filter control unit 26 appropriately switches the spatial filter matrix used by the spatial filter processing unit 27 according to the type of halftone dot.

  FIG. 14 is a diagram for explaining processing for controlling the characteristics of the spatial filter unit 27 in the filter control unit 26. FIG. 14 shows an example in which a 65-line halftone dot spatial filter matrix and a 133-point spatial filter matrix are already set. When the user operates the operation device 16 to instruct document copying, the filter control unit 26 starts processing after the image input device 13 reads the document. In step S11, the filter control unit 26 provides the halftone dot portion of the image data to the spatial filter processing unit 27, and among the 65-line halftone dot spatial filter matrix and the 133-point halftone dot spatial filter matrix. One of the above is given to the spatial filter processing unit 27. Which of the two spatial filter matrices is given is preferably selected depending on whether the halftone dot used in the halftone dot region is a 65-line halftone dot or a 133-line halftone dot. In step S12, the filter control unit 26 causes the spatial filter processing unit 27 to start spatial filter processing.

  As a result of the processing in FIG. 14, since the spatial filter processing unit 27 is given a single spatial filter matrix and a halftone dot portion, the spatial filter processing unit 27 performs spatial filter processing for moire removal. In the example of FIG. 14, two spatial filter matrices are prepared. However, the present invention is not limited to this, and three or more spatial filter matrices may be prepared. Further, only one spatial filter matrix may be prepared, and the spatial filter processing unit 27 may always use the same spatial filter matrix.

  The actual characteristics of the spatial filter processing unit 27 change as the characteristics of the image input device 13 and the image output device 15 change. The actual characteristics of the spatial filter processing unit 27 are optimized by the following method. First, a test original is copied to the image forming apparatus 11 having the configuration shown in FIG. 2, and an experiment is performed to visually confirm, for example, the presence or absence of moire in a halftone dot region in the image output as a result. The experiment is performed a plurality of times while mutually changing the characteristics of the spatial filter processing unit 27. The characteristic of the spatial filter processing unit 27 when an image with the least occurrence of moire is output as a result of a plurality of experiments may be set to an optimum characteristic. The actual characteristics of the spatial filter processing unit 27 are restricted by the configuration of the digital filter, and may not match the ideal characteristics. In this case, a characteristic that is most similar to the ideal characteristic may be selected from among a plurality of characteristics used in the optimization experiment.

  In the characteristics of the spatial filter processing unit 27, the spatial frequency at which the MTF becomes a minimum value can be changed by changing the weighting for the central pixel of the first filter used for the convolution calculation of the spatial filter matrix. The MTF of the entire characteristic of the spatial filter processing unit 27 can be changed by changing the weighting for the central pixel of the smoothing filter used for the convolution calculation of the spatial filter matrix. The greater the weighting corresponding to the center pixel of the smoothing filter, the stronger the MTF of the entire characteristic of the spatial filter processing unit 27.

  The optimum spatial filter matrix will be described below by taking 65-line halftone dots and 133-line halftone dots as examples. In the following description, the first filter matrix and the smoothing filter matrix are both realized by a matrix of 3 rows and 3 columns (3 × 3), and the spatial filter matrix is realized by a matrix of 5 rows and 5 columns (5 × 5). Take the case as an example. In the 3 × 3 matrix and the 5 × 5 matrix, the target pixel is in the center of the block, and the filter coefficient corresponding to the target pixel is in the center of the matrix. Among all the pixels in the block, the remaining pixels other than the target pixel are referred to as “peripheral pixels”. The smoothing filter matrix used to obtain the spatial filter matrix is preferably realized with a smoothing filter matrix in which all the filter coefficients are equal to each other. The smoothing filter matrix used for obtaining the spatial filter matrix is more preferably realized by a smoothing filter matrix in which the filter coefficient of the pixel of interest is larger than the filter coefficients of the surrounding pixels.

  First, an optimal spatial filter matrix that is determined using smoothing filter matrices in which all the filter coefficients are equal to each other will be described with reference to FIGS. In this reference example, all the filter coefficients in the smoothing filter matrix are 1. The calculation of the spatial filter processing using the smoothing filter matrix in which all the filter coefficients are equal to each other corresponds to the calculation for calculating the arithmetic average of the data of all the pixels in the block.

  FIG. 15 is a diagram showing a determinant for determining an optimum spatial filter matrix for a 65-line halftone dot. The first term on the left side of the determinant in FIG. 15 is the optimum first filter matrix MA1 for 65-line halftone dots, and the second term on the left side of the determinant in FIG. 15 is the optimum smoothing filter for 65-line halftone dots. This is the matrix MA2. The right side of the determinant in FIG. 15 is an optimal spatial filter matrix MA3 for 65 halftone dots, and an optimal first filter matrix MA1 for 65 line halftone dots and an optimal smoothing filter matrix MA2 for 65 line halftone dots. It is obtained by the convolution calculation.

  FIG. 16A is a graph of the characteristic GA5 of the optimum first filter for the 65-line halftone dot. The optimum first filter characteristic GA5 is determined by the optimum first filter matrix MA1 of FIG. The optimum first filter characteristic is that of a band cut filter having a cutoff frequency of 6.5 lines / mm.

  FIG. 16B is a graph of the characteristic GA6 of the optimum smoothing filter for the 65-line halftone dot. The optimum smoothing filter characteristic GA6 is determined by the optimum smoothing filter matrix MA2 in FIG. In the characteristic GA6 of the optimum smoothing filter, the reference MTF is 1.0, the MTF approaches 0 as the spatial frequency increases within a band greater than 0 and less than the predetermined upper limit spatial frequency FFA, and the upper limit spatial frequency FFA. The MTF of the above spatial frequency is zero. The upper limit spatial frequency FFA is a spatial frequency equal to or higher than the moire frequency FMA, and is 11 lines / mm in the example of FIG.

  FIG. 16C is a graph of the characteristic GA7 of the optimum spatial filter processing unit 27 for 65 halftone dots. The characteristic GA7 of the optimum spatial filter processing unit 27 is determined by the optimum spatial filter matrix MA3 in FIG. The optimum characteristic GA7 of the spatial filter processing unit 27 is that the MTF with the spatial frequency of 6.75 lines / mm is a minimum value instead of the MTF of the moire frequency FMA of the 65-line halftone dot, and 11 lines / mm or more. The MTF of the spatial frequency component is 0, and the other configuration is equal to the characteristic GA1 of the ideal spatial filter processing unit 27 for the 65-line halftone dot in FIG.

  FIG. 17 is a diagram showing a determinant for determining an optimum spatial filter matrix for a 133-line halftone dot. The right side of the determinant in FIG. 17 is an optimal spatial filter matrix MB3 for 133 halftone dots, and the optimal first filter matrix MB1 that is the first term on the left side, and the smoothing filter matrix MB2 that is the second term on the left side. Is obtained by convolution calculation. In the second term on the left side of FIG. 17, the same matrix as the optimum smoothing filter matrix MA2 for the 65-line halftone dot in FIG. 15 is used.

  FIG. 18A is a graph of the characteristic GB5 of the optimum first filter for the 133-line halftone dot. The optimum first filter characteristic GB5 is determined by the first filter matrix MB1 of FIG. The optimum characteristic GB5 of the first filter is that of a band cut filter having a cutoff frequency of 9.2 lines / mm. FIG. 18B is a graph of the characteristic GB6 of the optimum smoothing filter for the 133-line halftone dot. The optimum smoothing filter characteristic GB6 is determined by the smoothing filter matrix MB2 of FIG. The optimum smoothing filter characteristic GB6 is equal to the smoothing filter characteristic GA6 of FIG.

  FIG. 18C is a graph of the characteristic GB7 of the optimum spatial filter processing unit 27 for the 133-line halftone dot. The characteristic GB7 of the optimum spatial filter processing unit 27 is determined by the optimum spatial filter matrix MB3 in FIG. The optimum characteristic GB7 of the spatial filter processing unit 27 is that the MTF having a spatial frequency of 9.2 / mm is a minimum value instead of the MTF of the moire frequency FMB of the 133-line halftone dot, and 22 / mm or more. The MTF of the spatial frequency of the band is zero, and the other configuration is equal to the characteristic GB1 of the ideal spatial filter processing unit 27 for the 133-line halftone dot in FIG.

  As described above with reference to FIGS. 15 to 18, when the spatial filter matrix is determined using the smoothing filter matrix in which all the filter coefficients are equal to each other, the characteristics of the spatial filter processing unit 27 are changed to the spatial filter for moire removal. The characteristics can be made more suitable for processing.

  Next, an optimum spatial filter matrix determined using a smoothing filter matrix in which the filter coefficient of the pixel of interest is larger than the filter coefficients of the surrounding pixels will be described with reference to FIGS. The matrix operation using matrices with different filter coefficients corresponds to an operation for obtaining a weighted average of data of all the pixels in the block including the target pixel.

  FIG. 19 is a diagram showing a determinant for determining an optimum spatial filter matrix MA5 for a 65-line halftone dot. The right side of the determinant in FIG. 19 is an optimum spatial filter matrix MA5 for 65 halftone dots, the optimum first filter matrix MA1 being the first term on the left side and the optimum smoothing filter matrix being the second term on the left side. It is obtained by convolution calculation with MA4. In the first term on the left side in the determinant of FIG. 19, the optimum first filter matrix MA1 for the 65 dot halftone dot of FIG. 15 is used. In the optimum smoothing filter matrix MA4 in FIG. 19, the filter coefficient of the pixel of interest is 4. In addition, the filter coefficient of the pixels located above, below, left, and right of the target pixel is 2. The filter coefficients of the remaining four neighboring pixels on the diagonal line of the target pixel are set to the minimum value of all the filter coefficients, that is, 1 because the peripheral pixel is located farthest from the target pixel in the block.

  FIG. 20A is a graph of the characteristic GA5 of the optimal first filter for the 65-line halftone dot, which is the same as FIG. FIG. 20B is a graph of the characteristic GA8 of the optimum smoothing filter for the 65-line halftone dot. The optimum smoothing filter characteristic GB8 is determined by the optimum smoothing filter matrix in FIG. The optimum smoothing filter characteristic GA8 has an upper limit spatial frequency FFA of 16 lines / mm, and the other configuration is the same as the optimum smoothing filter characteristic GA4 of FIG.

  FIG. 20C is a graph of the characteristic GA9 of the optimum spatial filter processing unit 27 for a 65-line halftone dot. The characteristic GA9 of the optimum spatial filter processing unit 27 is determined by the spatial filter matrix MA5 in FIG. The optimum characteristic GA9 of the spatial filter processing unit 27 is that the MTF having a spatial frequency of 6.5 lines / mm is a minimum value instead of the MTF of the moire frequency FMA of the 65-line halftone dot, and 16 lines / mm or more. The MTF of the frequency component of the spatial frequency of the band is 0, and the other configuration is equal to the characteristic GA1 of the ideal spatial filter processing unit 27 for the 65-line halftone dot in FIG.

  FIG. 21 is a diagram showing a determinant for determining the optimum spatial filter matrix MB5 for the 133-line halftone dot. The right side of the determinant in FIG. 21 is an optimum spatial filter matrix MB5 for 133 halftone dots, the optimum first filter matrix MB1 being the first term on the left side and the optimum smoothing filter matrix being the second term on the left side. MB4 is obtained by convolution calculation. As the first term on the left side in the determinant of FIG. 21, the optimal first filter matrix MB1 for the 133 dot halftone dot of FIG. 17 is used. The optimum smoothing filter matrix MB4 for 133 halftone dots in FIG. 21 is equal to the optimum smoothing filter matrix MB4 for 65 dot halftones in FIG.

  FIG. 22A is a graph of the characteristic GB5 of the optimal first filter for the 133-line halftone dot, which is the same as FIG. FIG. 22B is a graph of the characteristic GB8 of the optimum smoothing filter for the 133-line halftone dot. The optimum smoothing filter characteristic GB8 is determined by the smoothing filter matrix MB4 of FIG. The optimum smoothing filter characteristic GB8 is equal to the smoothing filter characteristic GA8 of FIG.

  FIG. 22C is a graph of the characteristic GB9 of the optimum spatial filter processing unit 27 for the 133-line halftone dot. The characteristic GB9 of the optimum spatial filter processing unit 27 is determined by the spatial filter matrix MB5 in FIG. The characteristic GB9 of the optimum spatial filter processing unit 27 is that the MTF having a spatial frequency of 9.2 lines / mm is a minimum value instead of the MTF of the moire frequency FMB of the 133-line halftone dot, and 16 lines / mm or more. The MTF of the frequency component of the spatial frequency in this band is 0, and the other configuration is equal to the characteristic GB1 of the ideal spatial filter processing unit 27 for 133 points in FIG.

  As described above, when the spatial filter matrix is determined using the smoothing filter matrix in which the filter coefficient of the target pixel is larger than the filter coefficients of the surrounding pixels, the characteristics of the spatial filter processing unit 27 are more suitable for the spatial filter processing. Can be characteristic. The reason why it is more preferable to use a smoothing filter matrix in which the filter coefficient of the pixel of interest is larger than the filter coefficient of the surrounding pixels is as follows. When removing moiré from an image, it is ideal that the moiré is removed while retaining the features of the pixel of interest in the image. The feature of the target pixel is, for example, the difference between the color of the target pixel and the colors of pixels around the target pixel. When the filter coefficient of the target pixel is larger than the filter coefficients of the surrounding pixels, the smoothing filter can leave more features of the target pixel. Therefore, in the smoothing filter matrix, the filter coefficient of the target pixel is larger than the filter coefficients of the surrounding pixels. In a smoothing filter matrix in which the filter coefficient of the target pixel is larger than the filter coefficient of the peripheral pixel, preferably, the greater the distance between the filter coefficient of the peripheral pixel and the filter coefficient of the target pixel, the smaller the filter coefficient of the peripheral pixel. Yes. As a result, the smoothing filter can leave more features of the target pixel.

  4 to 22, the case where there is only one moire frequency region in the halftone dot region extracted by the image region separation processing unit 25 is described as an example. Actually, the moire frequency in the halftone dot region is not limited to one, and there may be a plurality of moire frequencies. In the case where there are a plurality of moire frequencies in the halftone dot region, the first filter matrix preferably defines the characteristics of a filter (hereinafter referred to as “second filter”) having a characteristic of attenuating or removing each moire frequency component. (Hereinafter referred to as “second filter matrix”) is a matrix obtained by mutual convolution calculations. The second filter is preferably realized by a band cut filter in which the MTF of the moire frequency is a minimum value.

  A characteristic GX1 of the spatial filter processing unit 27 when there are a plurality of moire frequencies in the halftone dot region will be described with reference to FIG. In the following description, as shown in FIG. 23A, the case where the spatial frequency characteristic of the halftone dot region has three moire frequency regions A1, A2, A3 is taken as an example. As shown in FIG. 23B, the characteristic GX1 of the spatial filter processing unit 27 shows a characteristic that basically has a tendency to attenuate and that the moire frequency regions A1, A2, and A3 are suppressed. . Therefore, the reference MTF is 1.0, the MTFs of the moire frequencies FM1, FM2, and FM3 of the three moire frequency regions A1, A2, and A3 are minimum values, and the MTFs of all spatial frequencies in the allowable spatial frequency band are It is less than 1.0.

  The matrix that defines the characteristic GX1 of the ideal spatial filter processing unit 27 is obtained by convolution calculation of a matrix that defines the characteristics GX2, GX3, and GX4 of the three second filters and a matrix that defines the characteristics of the smoothing filter. FIG. 23C, FIG. 23D, and FIG. 23E are graphs showing the characteristics GX2, GX3, and GX4 of the second filter, respectively. Regarding the characteristics GX2, GX3, and GX4 of the second filter, the reference MTF is 1.0, and the MTFs of the moire frequencies FM1, FM2, and FM3 are minimum values. The spatial frequencies at which the MTFs become minimum values in the three second filters are different from each other. FIG. 23F is a graph illustrating characteristics of the smoothing filter characteristic GX5. A smoothing filter characteristic GX5 in FIG. 23F is equal to the smoothing filter characteristic used in the description of FIGS.

  When there are a plurality of moire frequencies in the halftone dot region, the spatial filter processing unit 27 performs a spatial filter process using a spatial filter matrix for defining the characteristics shown in FIG. Accordingly, the spatial filter processing unit 27 can perform a suitable spatial filter process on the halftone dot region. In FIG. 23, there are three moire frequency regions in the halftone dot region. However, the present invention is not limited to this, and it is sufficient that there is at least one moire frequency region.

When the image to be processed by the spatial filter processing unit 27 is a color image, the image data may be composed of luminance data of each pixel and color difference data of each pixel. Examples of the image data in which the luminance data and the color difference data are separated include L ^* a ^* b ^* color system image data. The L ^* a ^* b ^* color system is a color system standardized by the International Commission on Illumination (CIE) in 1976, and is a standard color space in the printing industry. The data of the L ^* a ^* b ^* color system image represents a pixel by a combination of a lightness index L ^* and two types of perceptual chromaticities a ^* and b ^* . The lightness index L ^* corresponds to luminance data, and the two types of perceptual chromaticity a ^* and b ^* correspond to color difference data. In order to provide the spatial filter processing unit 27 with the L ^* a ^* b ^* color system image data in the image processing apparatus 14, the color correction unit 24 converts the RGB color system image data into L ^* a ^* b ^*. It suffices to have a color conversion unit for converting to color system image data.

If the L ^* a ^* b ^* display system image data is given, the spatial filter processing unit 27 processes only the luminance data L ^* and does not process the color difference data a ^* and b ^*. It is preferable. The reason for this is as follows. In this reference example, the spatial filter process converts the spatial frequency characteristics of the image to be processed into a desired spatial frequency characteristic for the main purpose of improving the quality of the image. When the image data is composed of pixel luminance data and pixel color difference data, if the color difference data is subjected to spatial filter processing, the image before the spatial filter processing and the image after the spatial filter processing are obtained. The color appearance changes. Therefore, it is not preferable to perform spatial filtering on the color difference data because it is difficult to improve image quality. When the L ^* a ^* b ^* color system image data is given, the spatial filter processing unit 27 performs processing only on the luminance data L ^* representing the shade of the pixel, and the color difference data a representing the color of the pixel a ^{* And} b ^* are not processed. As a result, the spatial filter processing unit 27 can perform good spatial filter processing without causing a reduction in image quality associated with the spatial filter processing.

FIG. 24 illustrates a process from image input to image output in the image forming apparatus 11 of FIG. 2 when the spatial filter processing unit 27 processes L ^* a ^* b ^* color system image data. It is a figure for demonstrating. In FIG. 24, steps that perform the same processes as those in FIG. 3 are given the same reference numerals, and descriptions thereof are omitted. After the preprocessing in step S2, the color conversion unit in the color correction unit 24 of the image processing apparatus 14 converts the preprocessed RGB color system image data to L ^* as the processing in step S21 ^. Convert to a ^* b ^* color system image data. L ^* a ^* b ^* color system image data and RGB color system image data can be converted into each other by calculation.

The spatial filter processing unit 27 of the image processing apparatus 14 performs spatial filter processing using a spatial filter matrix only on the luminance data L ^* in the data of the L ^* a ^* b ^* color system image as the process of step S24. Apply. The spatial filter matrix used in step S24 is obtained by convolution calculation of the first filter matrix and the smoothing filter matrix ^* as in the description of FIGS. The color difference data a ^* and b ^* in the L ^* a ^* b ^* color system image data are supplied to the halftone output gradation processing unit 28 without being subjected to the spatial filter processing.

When the spatial filter processing is performed only on the luminance data L ^* , there is a difference in the spatial filter matrix used in the spatial filter processing unit 27 depending on the type of halftone dot. Therefore, when the spatial filter process is performed only on the luminance data L ^* , the spatial filter matrix for the 65-point halftone dot and the spatial filter matrix for the 133-point halftone dot may be switched appropriately. The spatial filter matrix switching process is the same as the process described in FIG.

  When the image to be processed by the spatial filter processing unit 27 is a color image, the image data includes a plurality of color data representing the color of each pixel and black data representing the density of each pixel. It may be. The image data includes CMYK image data. The CMYK image data is constituted by data obtained by color-separating each pixel data into four colors of cyan (C), magenta (M), yellow (Y), and black (K).

  FIG. 25 is a diagram illustrating a configuration of the image processing apparatus 14 in a case where the image data to be processed is realized by CMYK image data. In the example of FIG. 25, the color correction unit 24 includes a color conversion unit 41 and a black generation and under color removal unit 42. The color conversion unit 41 converts RGB color system image data into CMY color system image data. The black generation under color removal unit 42 generates black data representing pixel shades based on CMY color system image data, and further performs under color removal processing on the CMY color system image data. . As a result, the RGB color system image data is converted into CMYK image data and is provided to the spatial filter processing unit 27. When the CMYK image data is given, the spatial filter processing unit 27 performs spatial filter processing on the CMYK color data. The spatial filter matrix used for the spatial filter processing for each piece of CMYK data is obtained by convolution calculation of the first filter matrix and the smoothing filter matrix, as in the description of FIGS. When the CMYK image data is given, the spatial filter processing unit 27 performs the spatial filter processing only on the black data representing the shade of the pixel, and performs the spatial filter processing on each CMY data representing the color of the image. It may be configured to output without any. As a result, the spatial filter processing unit 27 can prevent a change in the color of the image while sufficiently removing moire from the halftone dot region.

  The image processing apparatus 14 of the reference example may be used alone, may be combined only with the image input apparatus 13, or may be combined only with the image output apparatus 15. The image processing apparatus 14 of the reference example may include at least the spatial filter processing unit 27 and the image area separation processing unit 25, and other processing units may be omitted as appropriate. The image output device 15 is realized by, for example, a display device using a cathode ray tube or a display device using a liquid crystal display element when the image processing device 14 outputs RGB color system image data. When the image processing device 14 outputs CMYK image data, for example, it is realized by a printing device using an electrophotographic method or a printing device using an ink jet method. When the image processing apparatus 14 outputs CMYK image data, the image forming apparatus 11 is realized by a copying machine. The image forming apparatus 1 in FIG. 1 is an example of the image forming apparatus of the present invention, and the specific configuration of the image forming apparatus is not limited to the configuration in FIG. In the reference example, two types of 65-dot halftone dot and 133-line halftone dot are given as the types of halftone dots for which the spatial filter characteristics are set. However, the present invention is not limited to this. A spatial filter characteristic corresponding to some other halftone dot may be set.

  FIG. 26 is a block diagram illustrating a configuration of an image forming apparatus 101 including the image processing apparatus 103 according to an embodiment of the present invention. The image forming apparatus 101 according to the embodiment is different from the image forming apparatus 11 of the reference example only in the configuration of parts described below, and the configuration of parts not described is the same. In the description of the image forming apparatus 101 according to the embodiment, the same reference numerals are given to the same components as those of the image forming apparatus 11 of the reference example, and the description thereof is omitted.

  The image forming apparatus 101 includes an image input device 13, an image output device 15, and an operation device 16 in addition to the image processing device 103. The image processing apparatus 103 includes an A / D conversion unit 21, a shading correction unit 22, an input tone correction unit 23, a color correction unit 24, an image area separation processing unit 106, a spatial filter processing unit 108, and halftone output tone processing. Part 28 is included.

  The image area separation processing unit 106 extracts an edge in the image based on the preprocessed image data. The edge extraction process is realized by a process using a Sobel filter. In the following description, pixels constituting an edge are referred to as edge pixels, and an area composed only of edge pixels in an image is referred to as an edge area. The edge extraction result is given to the spatial filter processing unit 108 and the halftone output gradation processing unit 28. That is, the image area separation processing unit 106 corresponds to an edge extraction unit that extracts an edge of an image. The spatial filter processing unit 108 performs spatial filter processing on the pre-processed image data. A detailed description of the spatial filter processing unit 108 will be described later.

  FIG. 27 is a diagram for explaining a process from image input to image output in the image forming apparatus 101 of FIG. When the user operates the operation device 16 to instruct to copy the document, the image input device 13 reads the document and supplies the read result to the image processing device 103 as a process of step S101 as a process of step S1. The portions from the A / D conversion unit 21 to the color correction unit 24 of the image processing apparatus 103 perform preprocessing such as A / D conversion processing, shading correction processing, and input tone correction processing as processing in step S102. The image area separation processing unit 106 performs edge extraction processing as processing in step S103.

  The spatial filter processing unit 108 performs spatial filter processing for edge enhancement on the portion corresponding to the edge region in the preprocessed image data (hereinafter abbreviated as the edge portion) as the processing of step S104. Apply. Further, the spatial filter processing unit 108 may output the portion other than the edge region in the image data as it is without performing any spatial filter processing, and according to the spatial frequency characteristics of the region other than the edge region in the image. After performing the spatial filter processing, it may be output. Spatial filter processing is performed individually for each of the RGB data of the pixels constituting the image data. The halftone output tone processing unit 28 performs post-processing on the image data on which the spatial filter processing has been performed as the processing of step S <b> 105, and provides the processed data to the image output device 15. The image output device 15 outputs an image based on the post-processed image data as the process of step S106.

  The image processing apparatus 103 according to the embodiment is characterized in the spatial filter processing for edge enhancement in the spatial filter processing unit 108. The characteristics of the spatial filter processing unit 108 of the image processing apparatus 103 will be described in detail below. In the following description, it is assumed that the characteristics of the spatial filter processing unit 108 when performing spatial filter processing on each RGB data are equal to each other.

  The characteristic of the spatial filter processing unit 108 during the spatial filter processing for edge enhancement is a spatial frequency band including a spatial frequency that causes an edge extraction error of the image area separation processing unit 106 (hereinafter referred to as “error determination frequency band”). In the spatial frequency band that is flat and less than the lower limit value of the erroneous determination frequency band, the spatial frequency component is emphasized. The spatial filter processing unit 108 in which such characteristics are set corresponds to edge enhancement spatial filter means.

In the following description, the allowable spatial frequency band including all the spatial frequencies that can be included in the image is divided into three areas of an extractable band WZ1, an erroneous determination frequency band WZ2, and an unextractable band WZ3. The erroneous determination frequency band WZ2 includes a spatial frequency of a shading pattern in which the edge extraction unit of the image region separation processing unit 106 is likely to make an error in edge extraction. The extractable band WZ1 is a band that is less than the lower limit value of the erroneous determination frequency band WZ2, and includes only the spatial frequency of the light and shade pattern in which the edge extraction unit of the image area separation processing unit 106 can reliably extract the edge. The unextractable band WZ3 is a band that exceeds the upper limit value of the erroneous determination frequency band WZ2, and includes only the spatial frequency of the grayscale pattern in which the edge extraction unit of the image area separation processing unit 106 cannot extract the edge at all. In the extractable band WZ1, the erroneous determination frequency band WZ2, and the non-extractable band WZ3, the spatial frequencies in the bands WZ1 to WZ3 are sequentially increased in this order.

FIG. 28 is a graph of the ideal characteristic GC2 of the spatial filter processing unit. The ideal characteristic GC2 of the spatial filter processing unit 108 is that the spatial frequency MTF in the extractable band WZ1 is larger than 1.0 and the spatial frequency MTF in the erroneous determination region WZ2 is almost 1.0. . The characteristics of the spatial filter processing unit 108 tend to attenuate when the MTF is less than 1.0, flat when the MTF is approximately 1.0, and when the MTF exceeds 1.0. It tends to be emphasized. When the characteristics of the spatial filter processing unit 108 are flat, the spatial filter processing unit 108 does not emphasize or attenuate the spatial frequency component, so that the spatial frequency component is maintained almost as it is. Ideal characteristics GC2 of the spatial filter processing section 108, in the extraction allowed band WZ3, that has become a property of attenuating components of the spatial frequency.

When the spatial filter processing unit 108 is realized by a digital filter, the characteristics of the spatial filter processing unit 108 are determined by a matrix. The matrix that defines the characteristics of the spatial filter processing unit 108 for edge enhancement is obtained by convolution calculation of a matrix that defines the characteristics of the enhancement filter and a matrix that defines the characteristics of the smoothing filter. The characteristic of the ideal enhancement filter is a characteristic that amplifies the components of all the spatial frequencies within the allowable spatial frequency band. Characteristics of an ideal flat Namerafu filter is adapted to the characteristics of attenuating components of the whole spatial frequency within the allowable frequency range. In the following description, the matrix that defines the characteristics of the spatial filter processing unit 108 for edge enhancement is referred to as “spatial filter matrix”, the matrix that defines the characteristics of the enhancement filter is referred to as “enhancement filter matrix”, and the characteristics of the smoothing filter are defined. The defined matrix is referred to as a “smoothing filter matrix”.

A matrix obtained by convolution calculation of the enhancement filter matrix and the smoothing filter matrix is a matrix that defines the characteristics of the bandpass filter. In the present embodiment, as shown in FIG. 29A, the smoothing filter matrix and the enhancement filter matrix are each realized by a 3 × 3 matrix, and as shown in FIG. 29B, the spatial filter matrix. Is realized by a band-pass filter whose characteristics are determined by a 5 × 5 matrix. When the characteristic GC2 of the spatial filter processing unit 108 is realized by the characteristics of the bandpass filter, in the following description, in the spatial frequency plane, the spatial frequency band in which the component is emphasized and the spatial frequency band in which the component is attenuated Is defined as “representative frequency f0”. The representative frequency f0 is any one spatial frequency within the erroneous determination frequency band WZ2. Note characteristic of the spatial filter processing section 108, an erroneous determination in the frequency band is flat, and Ru are implemented in to emphasize the spatial frequency components less than the lower limit value of the erroneous determination frequency band resolver Ndopasufiru data.

  The actual characteristics of the edge enhancement spatial filter processing unit 108 change as the input resolution of the image input device 13 and the extraction accuracy of the edge extraction unit of the image area separation processing unit 106 change. Therefore, the actual characteristics of the spatial filter processing unit 108 are optimized based on the procedure experiment described in the reference example. The characteristic optimization procedure of the spatial filter processing unit 108 in the embodiment is based on density unevenness caused by edge enhancement instead of moire as an evaluation criterion, compared with the characteristic optimization procedure described in the reference example. The other configurations are the same as the reference example. In addition, the procedure for optimizing the characteristics when the actual characteristics of the filter do not match the ideal characteristics is the same as the procedure described in the reference example.

FIG. 30 is a diagram illustrating a determinant for determining the optimum spatial filter matrix MC3. Note that the determinant in FIG. 30 is an example in which the representative frequency f0 is 6.5 lines / mm. Determinant left first term of FIG. 30, the optimum is a smoothing filter matrix MC1, has the same matrix as the optimal smoothing filter matrix MA4 of FIG. The second term on the left side of the determinant in FIG. 30 is the optimum enhancement filter matrix MC2. The right side of the determinant in FIG. 30 is the optimum spatial filter matrix MC3, which is determined by the convolution operation of the optimum smoothing filter matrix MC1 and the optimum enhancement filter matrix MC2.

FIG. 31 is a graph showing an optimum smoothing filter characteristic GC3. The optimum smoothing filter characteristic GC3 of FIG. 31 is determined by the optimum smoothing filter matrix MC1 of FIG. In the characteristic GC3 of the optimum smoothing filter, the reference MTF is 1.0, the MTF approaches 0 as the spatial frequency increases within a band where the spatial frequency is greater than 0 and less than the predetermined upper limit spatial frequency FFC, and The MTF is 0 in the band above the upper limit spatial frequency FFC. The upper limit spatial frequency FFC is a value exceeding the representative frequency f0.

  FIG. 32 is a graph showing the optimum enhancement filter characteristic GC4. The optimum enhancement filter characteristic GC4 of FIG. 32 is determined by the optimum enhancement filter matrix MC2 of FIG. In the optimum enhancement filter characteristic GC4, the reference MTF is 1.0, and the MTF increases as the spatial frequency increases.

As shown in FIG. 28, the graph of the characteristics of the spatial filter processing unit 108 defined by the optimal spatial filter matrix MC3 in FIG. 30 has an MTF of a spatial frequency less than the representative frequency f0 greater than 1.0, The MTF is equal to 1.0, and the MTF having a spatial frequency greater than the representative frequency f0 is less than 1.0. In this case, the representative frequency is 6.5 lines / mm. As a result the spatial filter processing section 108, in the edge region of the image, extractable by sufficiently increasing the component of the spatial frequency band WZ1, substantially keeping unchanged the components of the spatial frequencies in the erroneous determination frequency band WZ2, extracted Call in the band WZ3 Ru attenuates the components of the spatial frequency to the charge amount. As a result, the occurrence of density unevenness in the image after filtering is more sufficiently suppressed.

  The reason why the characteristics of the spatial filter processing unit 108 are set as shown in FIG. 28 will be described below with reference to the case where the image forming apparatus 101 processes the image shown in FIG.

  The edge extraction accuracy of the edge extraction unit of the image area separation processing unit 106 is determined according to the input resolution of the image input device 13. The allowable spatial frequency band is divided into two parts: a band including a spatial frequency of a grayscale pattern that can be clearly read by the image input apparatus 13 and a band including a spatial frequency of a grayscale pattern that cannot be read by the image input apparatus 13 at all. It ’s difficult. Also, the input resolution and the density pattern of the spatial frequency in the vicinity thereof may or may not be read clearly. As a result, a noise component that causes a color shift or turbidity of the image is included in the image data.

  The above phenomenon will be described by taking as an example an experiment using a manuscript on which an image of a test gray pattern shown in FIG. 33 is drawn. The test shading pattern is a pattern in which a plurality of lines having different widths are arranged in parallel to each other. Further, in the test light and shade pattern, the width of the plurality of lines is gradually reduced from the left to the right in the drawing. The width of each of the lines at the left end 113 is the reciprocal of the spatial frequency in the extractable band WZ1, and the width of each of the lines at the center 112 is the reciprocal of the spatial frequency in the erroneous determination frequency band WZ2. In FIG. 33, the area to which the diagonal line rising to the right inside each line is actually painted in the same color as the outline of the line.

  In the experiment using the image of FIG. 33, first, the image input device 13 is caused to read the document of FIG. 33, and the edge extraction unit of the image area separation processing unit 106 stores the image in the image based on the image data as a reading result. Let the edges be extracted. FIG. 34 is a diagram showing the distribution in the image of the edge pixels extracted by the edge extraction unit of the image area separation processing unit 106 as a result of the experiment. In FIG. 34, a frame line 111 indicates a region in which a line is drawn in FIG. 33, and a region with a diagonally downward slanting line indicates an edge region 114. A comparison between FIG. 33 and FIG. 34 shows that the pixels constituting the line of the central portion 112 of the test pattern are not determined as edge pixels. That is, in this experiment, it can be seen that the edge extraction unit of the image area separation processing unit 106 has erroneously extracted the edge of the line of the central portion 112.

It is assumed that the edge enhancement processing using the conventional edge enhancement spatial filter is performed on the image data of FIG. 33 based on the edge extraction result of FIG. FIG. 35 is a diagram showing an image that has been subjected to the edge enhancement processing of the prior art. Note that in FIG. 35, a portion with a diagonally downward slanting line is a portion that has been emphasized as a result of the edge emphasis processing, and a portion with a diagonally upward sloping line is a portion that has not been emphasized. As shown in FIG. 35, when using a prior art edge enhancement processing, the portion having the components of the spatial frequencies in the erroneous determination frequency band WZ2 in the image, that is, the line of the central portion 112, that is locally stressed I understand that. As a result, density unevenness occurs in the central line, so that the quality of the image in FIG. 35 is lower than the quality of the image in FIG.

  It is assumed that the edge enhancement processing of the spatial filter processing unit 108 of the present embodiment is performed on the image data of FIG. 33 based on the extraction result of FIG. FIG. 36 is a diagram illustrating an image on which the edge enhancement processing according to the present embodiment has been performed. In FIG. 36, a portion with a diagonally downward slanting line is a portion that has been emphasized as a result of the edge emphasis processing, and a portion with a diagonally upward sloping line is a portion that has not been emphasized. As shown in FIG. 36, when the spatial filter processing unit 108 of the present embodiment is used, only the part having the spatial frequency of the extractable band WZ1 in the image, that is, the line of the left end part 113 is emphasized, and the central part 112 The whole line is not highlighted.

  The spatial filter processing unit 108 according to the present embodiment prevents emphasis of the edge of the spatial frequency shading pattern in the erroneous determination frequency band WZ2, and reliably ensures only the edge of the spatial frequency shading pattern in the extractable band WZ1. Can be emphasized. As a result, the occurrence of density unevenness due to the erroneous extraction of the edge can be suppressed, so that the image quality of the image can be prevented from deteriorating. Based on the above reason, it is preferable to set the characteristic of the spatial filter processing unit 108 to the characteristic GC2 in FIG. The spatial filter processing unit 108 does not need to perform complicated processing such as appropriately switching characteristics according to changes in spatial frequency components, and can perform good edge enhancement processing at a time on an image.

  As described above, the image processing apparatus 103 uses, for edge enhancement, a single spatial filter that has smooth characteristics in the erroneous determination frequency band WZ2 and has enhancement characteristics in the extractable region WZ1. As a result, processing in the spatial filter processing unit 108 is facilitated, and the processing time of the entire image forming apparatus 101 is shortened. In addition, since the spatial filter processing unit 108 has been adjusted so that the characteristic in the erroneous determination frequency band WZ2 is flat from the beginning, it is possible to perform a better edge enhancement process.

The erroneous determination frequency band WZ2 and the representative frequency f0 are obtained by an experiment using the manuscript of FIG. First, the original shown in FIG. 33 is copied by the image forming apparatus 101 in which the edge enhancement characteristics of the prior art are set in the spatial filter processing unit 108. Next, in the output image, a portion where density unevenness due to an erroneous determination of an edge occurs is confirmed, and the spatial frequency band including the spatial frequency of the light and shade pattern in the confirmed portion is the erroneous determination frequency band WZ2. As determined. Furthermore, one spatial frequency within the determined erroneous determination frequency band WZ2 is selected as the representative frequency f0. The representative frequency f0 is an average value of the upper limit spatial frequency, that is, the upper limit value and the lower limit spatial frequency, that is, the lower limit value of the erroneous determination frequency band WZ2. That is, when the erroneous determination frequency band WZ2 is a band of 12 lines / mm or more and 14 lines / mm or less, the spatial frequency at the middle position of the band, that is, 13 lines / mm may be set as the representative frequency f0.

In the reference example, the spatial filter processing unit 108 individually performs edge enhancement processing using a spatial filter matrix on each color data of RGB. The image obtained as a result of such edge enhancement processing has a high quality because the ratio of density unevenness is uniform throughout and the edges are well enhanced.

  In order to further improve the edge enhancement effect, a spatial filter matrix in which the degree of enhancement is set in accordance with the characteristics of the data of each color is used for the processing of the data of each color. Since the spatial filter matrix is obtained by performing a convolution operation on the smoothing filter matrix and the enhancement filter matrix, the enhancement degree of the enhancement filter matrix may be changed in order to change the enhancement degree of the spatial filter matrix. If a convolution calculation is performed on the enhancement filter matrix and the smoothing filter matrix in which the enhancement degree is set according to the characteristics of the RGB color data, the spatial filter matrix finally obtained is a filter corresponding to the RGB data. It becomes a matrix.

  FIG. 37 is a diagram for explaining a process from image input to image output in the image forming apparatus 101 when the spatial filter matrix used for the spatial filter processing for each color data is individually set. . In the following description, in the figure for explaining the process of the image forming apparatus, the steps for performing the same processing as in FIG. In the process of FIG. 37, a spatial filter matrix for red (R), a spatial filter matrix for blue (B), and a spatial filter matrix for green (G) are preset. After the preprocessing in step S102 is performed on the image data, the spatial filter processing unit 108 performs spatial filtering on red data using a red spatial filter matrix, and green space as processing in step S114. Spatial filter processing for green data using a filter matrix and spatial filter processing for blue data using a blue spatial filter matrix are performed. Image data composed of RGB data after the spatial filter processing in step S114 is provided to the halftone output tone correction unit. As a result of the above processing, the spatial filter processing unit 108 can perform even better edge enhancement processing on the image, and can avoid the occurrence of problems such as image non-uniformity and image quality degradation.

  In edge enhancement, it may be desired to enhance only one of red, green, and blue. When the spatial filter matrix used for the spatial filter processing for each color data is individually set, it is only necessary to change the degree of enhancement of the spatial filter matrix used for the spatial filter processing for any one color data. As a result, a better edge enhancement process can be performed on the image.

In one embodiment of the present invention, when the image to be processed in the spatial filter processing section 108 is a color image, data of images, Ru is constituted by the data of the color difference of each pixel and the luminance data of each pixel. The description will proceed by taking as an example the case where the image data to be processed is, for example, L ^* a ^* b ^* color system image data. In order to give L ^* a ^* b ^* color system image data to the spatial filter processing unit 108 in the image processing apparatus 103, the color correction unit 24 converts the RGB color system image data to L ^* a ^* b ^*. It suffices to have a color conversion unit for converting to color system image data. If the L ^* a ^* b ^* color system image data is given, the spatial filter processing unit 108 processes only the luminance data L ^* and processes the color difference data a ^* and b ^*. It has such. This reason is equal to the reason described in FIG. 23 of the reference example.

FIG. 38 illustrates a process from image input to image output in the image forming apparatus 101 in FIG. 26 when the spatial filter processing unit 108 processes data of an L ^* a ^* b ^* color system image. It is a figure for demonstrating. After the preprocessing in step S102 is completed, the color conversion unit in the color correction unit 24 of the image processing apparatus 103 converts the preprocessed RGB color system image data to L ^* as the processing in step S121 ^. Convert to a ^* b ^* color system image data.

The spatial filter processing unit 108 performs a spatial filter process using a spatial filter matrix only on the luminance data L ^* in the L ^* a ^* b ^* color system image data as the process of step S124. The spatial filter matrix used for the spatial filter processing for the luminance data L ^* is obtained by convolution calculation of the enhancement filter matrix and the smoothing filter matrix, as in the description of FIGS. The color difference data a ^* and b ^* are given to the halftone output gradation processing unit 28 without being subjected to the spatial filter processing. Through the above processing, the edge enhancement spatial filter processing unit 108 can prevent the occurrence of density unevenness due to edge extraction errors and prevent the image quality of the image subjected to the spatial filter processing from being deteriorated. it can.

In another embodiment of the present invention, when the image to be processed by the spatial filter processing unit 108 is a color image, the image data includes data representing the color of each pixel and black representing the shade of each pixel. Ru is constituted by the data. FIG. 39 is a diagram illustrating a configuration of the image processing apparatus 103 in a case where the processing target image data is realized by CMYK image data. In the example of FIG. 39, as in the example of FIG. 25, the color correction unit 24 includes a color conversion unit 41 and a black generation under color removal unit 42. As a result, CMYK image data is provided to the spatial filter processing unit 108.

  When the CMYK image data is given, the spatial filter processing unit 108 preferably performs spatial filter processing only on black data (K) in the image data. FIG. 40 is a diagram for explaining a process from image input to image output in the image forming apparatus 101 of FIG. 26 when the spatial filter processing unit 108 processes CMYK image data. After the preprocessing in step S102 is completed, as the processing in step S131, the color conversion unit 41 in the color correction unit 24 of the image processing apparatus 103 converts the data of the RGB color system image on which the preprocessing has been performed into a CMY table. Further, the black generation under color removal unit 42 in the color correction unit 24 of the image processing apparatus 103 generates the black data of the image, converts the data into color system image data, and the CMY color system image data and CMYK image data is created from the black data.

  The spatial filter processing unit 108 performs the spatial filter processing using the spatial filter matrix only on the black data in the CMYK image data as the processing in step S134. The spatial filter matrix used for the spatial filter processing for the black data is obtained by convolution calculation of the enhancement filter matrix and the smoothing filter matrix, as in the description of FIGS. The cyan (C) data, the magenta (M) data, and the yellow (Y) data are supplied to the halftone output gradation processing unit 28 without being subjected to the spatial filter processing.

  As a result of the processing described with reference to FIG. 40, the spatial filter processing unit 108 for edge enhancement emphasizes only the black edge in the image and does not enhance the color edge in the image. This is for the following reason. When the image subjected to the edge enhancement process is a character image, it is a black edge in the character image that the edge enhancement process is considered to have an appropriate effect. When there is a color edge in the image to be subjected to edge enhancement processing, the color edge is not necessarily the edge of a color character, but may be the outline of a region existing in the image. When the edge enhancement processing is performed on the color edge, the data regarding the pixel color is changed in the portion of the image where the edge enhancement processing is performed. Therefore, the portion subjected to the edge enhancement processing is changed in color or turbid in color compared to the portion not subjected to the edge enhancement processing. The spatial filter processing unit 108 described with reference to FIG. 40 performs processing only on black data and does not perform processing on other data. As a result, the spatial filter processing unit 108 can perform the edge enhancement process more effectively, so that the quality of the image with the edge enhanced can be further improved.

  The image processing apparatus 103 according to the embodiment described above may be used alone, may be combined only with the image input apparatus 13, or may be combined only with the image output apparatus 15. The image processing apparatus 103 according to the embodiment only needs to include at least the spatial filter processing unit 108 and the image area separation processing unit 106, and other processing units may be omitted as appropriate.

  The image processing apparatuses 14 and 103 according to the reference example and the embodiment described above are examples of the image processing apparatus of the present invention, and can be implemented in various other forms as long as the main operations are equal. In particular, the detailed operation of each processing unit in the image processing apparatus is not limited to this and may be realized by other operations as long as the same processing result is obtained.

In the image processing apparatuses 14 and 103 according to the reference example and the embodiment, the image to be processed is not limited to a color image but may be a monochrome image. When the processing target image is a color image, the processing target image data is limited to RGB color system image data, L ^* a ^* b ^* color system image data, and CMYK image data. Instead, each pixel may be configured by data of components necessary for outputting a color pixel. Image data composed of data representing color and black data for each pixel is not limited to CMYK image data, and may be data of other configurations. Image data composed of luminance data and color difference data for each pixel is not limited to L ^* a ^* b ^* color system image data, but other types of data such as so-called L ^* u ^* v. ^* Color image data or YIQ color image data may be used.

  The image area separation processing unit 25 of the reference example, the spatial filter processing unit 27 and the filter control unit 26 for removing moire, the image area separation processing unit 106 of the embodiment, and the spatial filter processing unit 108 for edge enhancement are: It may be provided in a single image processing apparatus. In this case, the quality of the image processed by the image processing apparatus is further improved because the occurrence of moire and the occurrence of density unevenness due to edge extraction errors are prevented.

1 is a front cross-sectional view illustrating a configuration of an image forming apparatus including an image processing apparatus according to a reference example of the present invention. 1 is a block diagram illustrating a configuration of an image forming apparatus including an image processing apparatus that is a reference example of the present invention. FIG. 3 is a diagram for explaining a process of the image forming apparatus in FIG. 2 when processing data of an RGB color system image. It is a graph which shows the spatial frequency characteristic of the halftone dot area | region of a 65 line halftone dot. It is a graph which shows the spatial frequency characteristic of the halftone dot area | region of a 133 line | wire halftone dot. It is a graph which shows the spatial frequency characteristic of the halftone dot area | region of a 65 line halftone dot. It is a graph which shows the spatial frequency characteristic of the halftone dot area | region of a 133 line | wire halftone dot. It is a graph of the characteristic of the spatial filter process part for an ideal 65 line halftone dot. It is a graph of the characteristic of the spatial filter process part for an ideal 133 line | wire halftone dot. It is a graph which shows the ideal characteristic of the 1st filter for determining the characteristic of the spatial filter process part for ideal 65 line halftone dots. It is a graph which shows the ideal characteristic of the 1st filter for determining the characteristic of the spatial filter process part for ideal 133 line | wire halftone dots. It is a graph which shows the ideal characteristic of the smoothing filter for determining the characteristic of the spatial filter process part for ideal 65 line halftone dots. It is a graph which shows the ideal characteristic of the smoothing filter for determining the characteristic of the spatial filter process part for ideal 133 line | wire halftone dots. It is a figure for demonstrating the process which controls the characteristic of a spatial filter part. It is a figure which shows the determinant for determining the optimal spatial filter matrix for 65 line | wire halftone dots using the smooth filter matrix in which all the filter coefficients are mutually equal. It is a graph which shows the characteristic defined by the matrix of the determinant of FIG. It is a figure which shows the determinant for determining the optimal spatial filter matrix for 133 line | wire halftone dots using the smoothing filter matrix in which all the filter coefficients are mutually equal. It is a graph which shows the characteristic defined by the matrix of the determinant of FIG. It is a figure which shows the determinant for determining the optimal spatial filter matrix for 65 line | wire halftone dots using the smooth filter matrix with a large central filter coefficient. It is a graph which shows the characteristic defined by the matrix of the determinant of FIG.

It is a figure which shows the determinant for determining the optimal spatial filter matrix for 133 line | wire halftone dots using the smooth filter matrix with a large filter coefficient of the center. It is a graph which shows the characteristic defined by the matrix of the determinant of FIG. It is a graph for demonstrating the characteristic of a spatial filter process part in case there exist multiple spatial frequencies which a moire can produce in a halftone dot area | region. FIG. 3 is a diagram for explaining a process of the image forming apparatus in FIG. 2 when processing data of an L ^* a ^* b ^* color system image. FIG. 3 is a diagram illustrating the configuration of the image processing apparatus in FIG. 2 when processing CMYK image data. 1 is a block diagram illustrating a configuration of an image forming apparatus including an image processing apparatus according to an embodiment of the present invention. FIG. 27 is a diagram for describing a process of the image forming apparatus in FIG. 26 in a case where RGB color system image data is a processing target. It is a graph of the characteristic of the ideal spatial filter process part for edge emphasis. It is a figure which shows the structure of the spatial filter process part for edge emphasis. It is a figure which shows the determinant for defining the optimal spatial filter matrix for edge emphasis. It is a graph which shows the characteristic of the smoothing filter for determining the characteristic of the optimal spatial filter process part for edge emphasis. It is a graph which shows the characteristic of the emphasis filter for determining the characteristic of the spatial filter process part for optimal edge emphasis. It is a figure which shows the image for a test for the emphasis process of an edge. It is a figure which shows the result of the edge extraction process with respect to the image of FIG. It is a figure which shows the result of the edge enhancement process of the prior art with respect to the image of FIG. It is a figure which shows the result of the edge emphasis process of one Embodiment with respect to the image of FIG. FIG. 27 is a diagram for describing a process of the image forming apparatus in FIG. 26 in a case where a spatial filter matrix for each data of RGB of pixels is individually set. FIG. 27 is a diagram for describing a process of the image forming apparatus in FIG. 26 in a case where L ^* a ^* b ^* color system image data is a processing target. FIG. 27 is a diagram illustrating a configuration of the image processing apparatus in FIG. 26 in a case where CMYK image data is a processing target. FIG. 27 is a diagram for describing a process of the image forming apparatus in FIG. 26 in a case where CMYK image data is a processing target.

It is a figure which shows the high-pass spatial filter used for the edge enhancement process of a prior art. It is a graph which shows the characteristic of the high-pass spatial filter of FIG. It is a figure which shows the matrix which defines the characteristic of the high-pass spatial filter of FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 13 Image input device 14,103 Image processing device 15 Image output device 25,106 Image area separation process part 26 Filter control part 27,108 Spatial filter process part

Claims (3)

Edge extraction means for extracting edges in the image based on the image data to be processed;
A spatial filter means for edge enhancement that performs the enhancement processing of the extracted edge on a portion corresponding to the edge in the image data;
The characteristics of the spatial filter means for edge enhancement are determined by a matrix obtained by convolution calculation of a matrix that defines the characteristics of the enhancement filter and a matrix that defines the characteristics of the smoothing filter,
The characteristics of the smoothing filter are:
The reference MTF with a spatial frequency of 0 is 1.0,
Within the band where the spatial frequency is greater than 0 and less than the predetermined upper limit spatial frequency FFC, the MTF approaches 0 as the spatial frequency increases, and the MTF is 0 in the band above the upper limit spatial frequency FFC.
The upper limit spatial frequency FFC is a value exceeding the representative frequency f0.
It is a characteristic that smoothes the components of all spatial frequencies that can be included in the image,
The representative frequency f0 is a spatial frequency that is an average value of an upper limit value and a lower limit value in the erroneous determination frequency band WZ2, which is a band including a spatial frequency that causes an edge extraction error of the edge extraction unit,
The characteristics of the enhancement filter are:
The reference MTF where the spatial frequency is 0 is 1.0, and the MTF increases as the spatial frequency increases.
It is a characteristic that emphasizes the components of all spatial frequencies that can be included in the image,
Thus, the characteristics of the spatial filter means for edge enhancement are
In the erroneous determination frequency band WZ2, it is flat,
Wherein less than the lower limit value of the erroneous determination frequency band, in the spatial frequency band WZ1 edge extraction means comprises only reliably extractable spatial frequency edges, Ri characteristics der to emphasize the spatial frequency component,
It is a characteristic that attenuates the spatial frequency component in the non-extractable band WZ3 that includes only the spatial frequency that exceeds the upper limit value of the erroneous determination frequency band WZ2 and the edge extraction means cannot extract the edge at all.
The image is composed of pixels, the image data is composed of a plurality of color data obtained by color-separating the data of each pixel, and one of the plurality of color data is black data representing the density of the pixels ,
An image processing apparatus , wherein an enhancement filter used for a convolution operation performs enhancement processing only on black data . Edge extraction means for extracting edges in the image based on the image data to be processed;
A spatial filter means for edge enhancement that performs the enhancement processing of the extracted edge on a portion corresponding to the edge in the image data;
Characteristics of the spatial filter means for or falling edge of di emphasized, and a matrix defining the characteristics of the matrix and the smoothing filter for determining the characteristics of the enhancement filter is defined by a matrix obtained by convolution,
The characteristics of the smoothing filter are:
The reference MTF with a spatial frequency of 0 is 1.0,
Within the band where the spatial frequency is greater than 0 and less than the predetermined upper limit spatial frequency FFC, the MTF approaches 0 as the spatial frequency increases, and
MTF is 0 in the band above the upper limit spatial frequency FFC,
The upper limit spatial frequency FFC is a value exceeding the representative frequency f0.
It is a characteristic that smoothes the components of all spatial frequencies that can be included in the image,
The representative frequency f0 is a spatial frequency that is an average value of an upper limit value and a lower limit value in the erroneous determination frequency band WZ2, which is a band including a spatial frequency that causes an edge extraction error of the edge extraction unit,
The characteristics of the enhancement filter are:
A reference MTF with a spatial frequency of 0 is 1.0, and
The MTF increases as the spatial frequency increases,
It is a characteristic that emphasizes the components of all spatial frequencies that can be included in the image,
Thus, the characteristics of the spatial filter means for edge enhancement are
In the erroneous determination frequency band WZ2, it is flat,
The spatial frequency component is less than the lower limit value of the erroneous determination frequency band, and the spatial frequency component is emphasized in the spatial frequency band WZ1 including only the spatial frequency from which the edge extraction unit can reliably extract the edge,
It is a characteristic that attenuates the spatial frequency component in the non-extractable band WZ3 that includes only the spatial frequency that exceeds the upper limit value of the erroneous determination frequency band WZ2 and the edge extraction means cannot extract the edge at all.
When the image is constituted by pixels, and the pixel data is constituted by pixel luminance data and color difference data, the edge enhancement spatial filter means performs processing only on the luminance data of each pixel. It is that images processing apparatus. The representative frequency f0 is 6.5 lines / mm,
The matrix MC1 that defines the characteristics of the smoothing filter is:

And
The matrix MC2 that defines the characteristics of the enhancement filter is:

And
A matrix MC3 obtained by convolution calculation of the matrix MC1 that determines the characteristics of the smoothing filter and the matrix MC2 that determines the characteristics of the enhancement filter is:

The image processing apparatus according to claim 1 or 2, wherein the at.

Images