JP2005260557A - Pseudo color reduction for color image - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide technology for appropriately reducing a pseudo color which is apt to occur in an edge pixel. <P>SOLUTION: The edge pixel existing in an edge in a color image and an edge direction in the edge pixel are first detected. A plurality of reference pixels which are confronted with each other across a virtual straight line along the edge direction and are adjacent to the edge pixel are detected. A color of the edge pixel is adjusted based on a relation between the colors of a plurality of the reference pixels and the color of the edge pixel. Thus, the pseudo color of the edge pixel is reduced. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a technique for reducing false colors in a color image.

  As a digital camera, a camera using a single-plate solid-state imaging device and a Bayer color filter is generally used. As is well known, in a Bayer color filter, only one color component filter is provided for each pixel. Original data (referred to as RAW data) obtained with such a single-plate solid-state image sensor is composed of only one color of data for each pixel. Usually, three color components RGB in each pixel are calculated by performing color interpolation (also referred to as “pixel interpolation”) on the RAW data.

  By the way, it is known that when such color interpolation is performed, a so-called false color is generated. In particular, since the pixel value of the RAW data fluctuates rapidly in the vicinity of the edge, a false color tends to occur. Therefore, various methods have been proposed in the past to reduce false colors.

JP 2002-209224 A JP 2002-77930 A JP 2002-232902 A

  However, the conventional method has a problem that the false color cannot be sufficiently reduced, or the sharpness is excessively lowered as the false color is reduced.

  An object of this invention is to provide the new technique for reducing a false color.

In order to achieve the above object, an apparatus of the present invention is an image processing apparatus for reducing false colors of a color image,
An edge detection unit that detects edge pixels existing at edges in a color image and an edge direction of the edge pixels;
A plurality of reference pixels that are opposed to and adjacent to the edge pixel across a virtual straight line along the edge direction are detected, and based on the relationship between the color of the plurality of reference pixels and the color of the edge pixel A false color reduction processing unit that reduces the false color of the edge pixel by adjusting the color of the edge pixel;
Is provided.

  According to this apparatus, since the color of the edge pixel is adjusted according to the color of the edge pixel and the color of the reference pixel on both sides of the edge, it is possible to appropriately reduce false colors that are likely to occur in the edge pixel.

  The false color reduction processing unit performs first color adjustment of the edge pixel so as to approach the hue of a color proximity reference pixel having a hue closer to the hue of the edge pixel among the plurality of reference pixels. A reduction mode may be provided.

  According to this configuration, when a false color is generated in the edge pixel, the false color can be reduced by bringing the hue of the edge pixel close to the hue of the color proximity reference pixel.

The false color reduction processing unit
(I) executing the first false color reduction mode when a difference between a hue of the color proximity reference pixel and a hue of the edge pixel is smaller than a predetermined first threshold;
(Ii) When the difference in hue is larger than the first threshold value, a second false color reduction mode for reducing the saturation of the edge pixel may be executed.

  According to this configuration, it is possible to reduce the false color by an appropriate method according to the magnitude of the difference between the hue of the color proximity reference pixel and the hue of the edge pixel. Specifically, when the difference in hue is relatively small, the original hue of the edge pixel can be approximated by bringing the hue of the edge pixel close to the hue of the color proximity reference pixel. In addition, when the hue difference is relatively large, the saturation of the edge pixel is reduced, so that the edge pixel can be made inconspicuous even if a false color is generated.

The first false color reduction mode is:
(I-1) When the hue difference is smaller than a second threshold smaller than the first threshold, the hue of the edge pixel is changed to the color proximity reference pixel while maintaining the saturation of the edge pixel. A first sub-mode for performing color adjustment of the edge pixel so as to approach the hue of
(I-2) When the hue difference is between the first threshold value and the second threshold value, the hue of the edge pixel is referred to the color proximity reference while reducing the saturation of the edge pixel. A second sub-mode that performs color adjustment of the edge pixel so as to approach the hue of the pixel may be included.

  According to this configuration, it is possible to reduce the false color by an appropriate method according to the magnitude of the difference between the hue of the color proximity reference pixel and the hue of the edge pixel.

The edge detection unit selects one of a plurality of predetermined directions set in advance at intervals of 45 degrees as the edge direction,
The false color reduction processing unit may select two of the eight neighboring pixels existing around the edge pixel as the plurality of reference pixels.

  According to this configuration, the reference pixel can be easily selected.

The image processing apparatus further includes a color interpolation processing unit that obtains all color components of each pixel by performing color interpolation processing on a color image before color interpolation in which each pixel has only some color components. And
The edge detection unit performs a process on the color image before the color interpolation process,
The false color reduction processing unit may execute processing on the color image after the color interpolation processing.

  According to this configuration, it is possible to execute appropriate false color reduction processing based on the color image before color interpolation.

  The present invention can be realized in various forms. For example, an image processing method and apparatus, a false color reduction method and apparatus, a computer program for realizing the functions of the method or apparatus, and the computer The present invention can be realized in the form of a recording medium recording the program, a data signal including the computer program and embodied in a carrier wave, and the like.

Next, embodiments of the present invention will be described in the following order.
A. Example:
B. Modified example

A. Example:
FIG. 1 is an explanatory diagram showing an image processing system as an embodiment of the present invention. This image processing system includes a digital camera 100 and a computer 200. The computer 200 includes an image processing unit 210 that creates bitmap data BMIG from RAW data of the digital camera 100. The image processing unit 210 may be provided inside the digital camera 100 or may be provided in another device (for example, a color printer).

  FIG. 2 shows an example of a color filter array (Bayer array) of the digital camera 100. Each square frame corresponds to one pixel. Each pixel is provided with a filter for only one color component of the three primary colors RGB. Accordingly, the pixel value of each pixel in the RAW data is composed of only one component of the R component, the G component, and the B component. As a color filter, a complementary color filter (YMCG filter) may be used instead of the primary color filter (RGB filter).

  The pixel interpolation unit 220 in FIG. 1 calculates all color components (for example, RGB and YCbCr) for each pixel by performing various processes on the RAW data. The pixel interpolation unit 220 includes an edge detection processing unit 222, a color interpolation processing unit 224, and a false color reduction processing unit 226. The processing contents of these units will be described later.

  The color adjustment unit 230 performs color adjustment such as white balance on color image data (for example, YCbCr data) after pixel interpolation, and the gradation correction unit 240 further performs gradation correction such as exposure correction and γ correction. To do. The bitmap image data BMIG generated in this way is stored in the image storage unit 250 provided in the hard disk of the computer 200. Note that the function of each unit of the image processing unit 210 is realized by a computer program.

  FIG. 3 is a flowchart showing a processing procedure of the pixel interpolation unit 220. In step S1, the edge detection processing unit 222 detects the edge strength and the edge direction of each pixel in the color image represented by the RAW data. FIG. 4A shows an example of a pixel array. For convenience of explanation, each pixel is assigned an R, G or B character indicating the color of the color filter and a two-digit number indicating the pixel position. The pixels provided with R, G, and B color filters are also referred to as “R pixels”, “G pixels”, and “B pixels”, respectively. In this example, the pixel R33 at the center is the target pixel for the edge detection process. FIGS. 4B and 4C show examples of edge filter coefficient arrays for detecting edge amounts in the 45-degree direction and the 135-degree direction, respectively. The edge amount Δd45 in the 45 degree direction and the edge amount Δd135 in the 135 degree direction in the processing target pixel Px are calculated by the following equations.

ここで、ａ〜ｈは所定の正の係数であり、Ｇ４１，Ｇ３２…は各画素の画素値（Ｇ成分値）である。なお、処理対象画素がＢ画素の場合も同じエッジフィルタを用いることができる。

Here, a to h are predetermined positive coefficients, and G41, G32... Are pixel values (G component values) of the respective pixels. The same edge filter can also be used when the processing target pixel is a B pixel.

  FIG. 5A shows a pixel array centered on G pixels, and FIGS. 5B and 5C show edge filters for G pixels. The edge amounts Δd45 and Δd135 at this time are calculated by the following equations.

  By using the edge filters shown in FIGS. 4 and 5, the edge amounts in the 45 degree direction and the 135 degree direction in each pixel of the RAW data can be respectively calculated. As can be understood from Equations 1 and 2, in this embodiment, only the G component is used for edge detection when the processing target is any of R, B, and G pixels. This is because (1) G pixels are arranged every other pixel in a checkerboard pattern, and there are the most, and (2) G components have the greatest influence on luminance among RGB components. This is because of the two reasons that the influence on the occurrence of the edge is large. That is, the edge of each pixel can be accurately detected by using the G component.

  FIG. 6A shows an edge direction determination method using edge amounts Δd45 and Δd135 in the 45 degree direction and the 135 degree direction. Here, the edge direction θx and the edge strength Ax are calculated according to the following equations.

  As can be understood from this equation and FIG. 6A, the edge direction θx is an angle measured clockwise from the 135-degree direction of the pixel array. In this embodiment, as the value of the edge direction θx, a value closest to the edge angle obtained by the above Equation 3 is selected from four values (0 °, 45 °, 90 °, 135 °) every 45 °. Select one. Further, when the edge intensity Ax is equal to or greater than the predetermined threshold Ath, this pixel is determined to be an edge pixel, and when it is less than the threshold Ath, it is determined to be a non-edge pixel. Note that data indicating whether each pixel is an edge pixel (also referred to as “edge image data”) is temporarily stored in the memory of the computer 200. Or you may make it add the flag of whether it is an edge pixel to the pixel data of each pixel in color image data.

  In this embodiment, the size of the edge component for a plurality of edge directions is referred to as “edge amount”, and the size of the edge component obtained by combining the edge amounts in the plurality of edge directions is referred to as “edge strength”. It is out. However, both “edge amount” and “edge strength” can be called “edge amount”, or both can be called “edge strength”.

  FIG. 6B shows two reference pixels Pref1 and Pref2 in the case of four edge directions. Here, the “reference pixel” means a pixel used for color adjustment of the edge pixel Px in a false color reduction process described later. In the present embodiment, as the reference pixels Pref1 and Pref2, of the eight pixels P1 to P8 adjacent to the edge pixel Px, two pixels facing each other across a virtual straight line (indicated by an arrow) having the edge direction θx. Is selected. For example, when the edge direction θx is 0 °, the pixels P3 and P6 are selected as the reference pixels Pref1 and Pref2. This virtual straight line often coincides with the boundary of the edge in the image. It can be considered that the reference pixels Pref1 and Pref2 exist in regions on both sides of the edge boundary passing through the edge pixel Px. For example, when the first reference pixel Pref1 is in a high luminance region, the second reference pixel Pref2 is in a low luminance region. Therefore, the pixel values of the two reference pixels Pref1 and Pref2 are usually significantly different. The color of the edge pixel Px may generate a false color due to the influence of the adjacent pixels P1 to P8 when a color component is supplemented in a color interpolation process described later. Therefore, the false color can be reduced by adjusting the color of the edge pixel Px using the colors of the reference pixels Pref1 and Pref2. The selection of the reference pixel may be performed by the edge detection processing unit 222 or may be performed by the false color reduction processing unit 226.

  When the edge detection process is thus completed, the color interpolation processing unit 224 executes the color interpolation process in step S2 of FIG. This color interpolation process is a process of interpolating a color component lacking in each pixel of RAW data from surrounding pixels, and is a pixel interpolation process in a narrow sense. As this color interpolation processing, any processing such as simple interpolation can be used. As a result of this color interpolation processing, the pixel values of all the pixels become pixel values having three components of RGB. In this embodiment, the pixel value of the luminance / color difference color space (YCbCr color space) is calculated from the RGB values after color interpolation. This is because YCbCr data is used in the next false color reduction process.

  Since there is no direct relationship between the color interpolation process in step S2 and the edge detection process in step S1, the color interpolation process can be executed before the edge detection process.

  In step S3, the false color reduction processing unit 226 executes a false color reduction process. FIG. 7 is a flowchart showing the detailed procedure of step S3. The routine in FIG. 7 is executed for each pixel in the color image.

  In step T1, it is determined whether the processing target pixel is an edge pixel. In the case of a non-edge pixel, saturation adjustment for the non-edge pixel is executed in step T2. FIG. 8 is a graph showing the content of saturation adjustment for non-edge pixels. In FIG. 8A, the horizontal axis indicates the value of luminance Y, and the vertical axis indicates the value of coefficient α used for saturation adjustment. The Cb ′ and Cr ′ components after saturation adjustment are obtained by multiplying the original Cb and Cr components by a coefficient α. As shown in FIG. 8B, the saturation S of the target pixel color Px is given by the distance from the origin on the CbCr plane (FIG. 8B) to the target pixel color Px. The hue H is given by an angle from the Cr axis to the target pixel color Px. As shown in FIG. 8A, when the luminance Y is in the range from 0 to a predetermined value (200 in FIG. 8A), the coefficient α is 1, and the saturation is substantially maintained. When the luminance Y is 200 or more, the coefficient α gradually decreases from 1.0 to 0. In this example, the decrease in coefficient α is linear. FIG. 8B shows a state where the saturation is reduced when α is less than 1. That is, it can be understood that the target pixel color Px ′ (indicated by a black square) after the saturation adjustment has the same hue H as that before the adjustment and the saturation S is lowered.

  As described above, the reason why the saturation is reduced in the bright non-edge pixels whose luminance Y is larger than the predetermined value is that false colors tend to occur even in the non-edge pixels in a bright region. That is, even if a false color is generated in a bright non-edge pixel, the saturation is reduced by the processing of FIG. 8, so that the false color can be made inconspicuous.

  If the target pixel is an edge pixel in step T1 in FIG. 7, steps T3 to T5 are executed. In step T3, the hues of the edge pixel and the reference pixel are calculated. FIG. 9A shows an example of the hue H (x) of the edge pixel Px on the CbCr plane and the hues H (ref1) and H (ref2) of the reference pixels Pref1 and Pref2. Hue H is calculated by the following equation.

  In step T4, the hue differences ΔH1 and ΔH2 between the edge pixel Px and the two reference pixels Pref1 and Pref2, and the minimum value ΔHmin are calculated by the following equations.

  In step T5, the color of the edge pixel is adjusted according to the value of the minimum hue difference ΔHmin. FIG. 9B shows a color adjustment mode for edge pixels. In this example, the minimum hue difference ΔHmin is divided into three ranges by two thresholds TH1 and TH2, and different color adjustment modes are executed in each range. At this time, as described in FIG. 9A, the presence / absence of the target hue Htg for color adjustment is determined according to whether or not the minimum hue value ΔHmin is equal to or less than the first threshold value TH1. That is, when ΔHmin ≦ TH1, the hue of the reference pixel that gives the minimum hue value ΔHmin is used as the target hue Htg. In the example of FIG. 9A, since the hue difference ΔH1 with respect to the first reference pixel Pref1 is smaller, the hue H (pref1) becomes the target hue Htg. On the other hand, if TH1 <ΔHmin, it is determined that there is no target hue Htg. As shown in FIG. 9B, when the target hue Htg exists, a color adjustment mode (hue rotation mode or hue / saturation adjustment mode) using the target hue Htg is executed. On the other hand, when the target hue Htg does not exist, the saturation reduction mode that does not use the target hue Htg is executed.

  FIG. 10A shows the content of the hue rotation mode. This mode is executed when the difference between the hue of the edge pixel Px and the target hue Htg is sufficiently small (when ΔHmin ≦ TH2). In the hue rotation mode, the CbCr component of the edge pixel Px is adjusted so that the hue of the edge pixel Px matches the target hue Htg while keeping the saturation constant. In FIG. 10A, the color of the edge pixel Px ′ after color adjustment is indicated by a black square. Note that the luminance Y is preferably maintained constant. The same applies to other modes.

  The reason for performing such a hue rotation mode is as follows. As described above, the reference pixels Pref1 and Pref2 are considered to exist in the regions (high luminance region and low luminance region) on both sides of the edge boundary. When the difference between the hue of the edge pixel Px and the target hue Htg is sufficiently small, the color of the edge pixel Px is not significantly deviated from the original color of the subject by the color interpolation processing in step S3, and the original hue of the edge pixel Px. Is likely to be close to the target hue Htg. Therefore, in this case, the false color can be reduced by matching the hue of the edge pixel Px with the hue Htg of the reference pixel Pref1 (also referred to as “color proximity reference pixel”) having a hue close thereto. .

  FIG. 10B shows the contents of the hue / saturation adjustment mode. This mode is executed when the difference between the hue of the edge pixel Px and the target hue Htg is in an intermediate range (when TH2 <ΔHmin ≦ TH1). In the hue / saturation adjustment mode, the color component CbCr of the edge pixel is adjusted so that the hue of the edge pixel Px matches the target hue Htg while reducing the saturation. In FIG. 10B, hue and saturation are simultaneously adjusted by reducing the Cb component while maintaining the Cr component of the edge pixel Px. In the first quadrant, when the hue of the edge pixel color Px is on the right side of the hue of the reference pixel Pref1, contrary to FIG. 10B, Cr is maintained while maintaining the Cb component of the edge pixel Px. Reduce ingredients. However, it is also possible to match the hue of the edge pixel Px with the target hue Htg while reducing the saturation by changing both the Cb component and the Cr component.

  In this hue / saturation adjustment mode, the saturation of the edge pixel Px decreases, so that false colors can be made inconspicuous. Further, since the original hue of the edge pixel Px is considered to be close to the target hue Htg, the false color can be further effectively reduced by making this coincide with the target hue Htg.

  The two modes shown in FIGS. 10A and 10B are common in that the hue of the edge pixel Px matches the target hue Htg. In these two modes, the hue of the edge pixel Px is made to match the target hue Htg, but it is not necessary to make it completely match. Generally, the hue of the edge pixel Px may be made closer to the target hue Htg. .

  FIG. 10C shows the content of the saturation reduction mode. This mode is executed when the difference between the hue of the edge pixel Px and the target hue Htg is large (when TH1 <ΔHmin). In the saturation reduction mode, the saturation is reduced while maintaining the hue of the edge pixel Px. In FIG. 10B, the Cb and Cr components of the edge pixel Px are both set to 0. Therefore, the color of the edge pixel Px ′ after color adjustment is gray. Instead, the saturation may be reduced by multiplying the Cb and Cr components by a common coefficient less than 1. In this way, it is possible to reduce only the saturation without changing the hue.

  In this saturation reduction mode, the saturation of the edge pixel Px is lowered, so that the false color can be made inconspicuous. In particular, when the original hue of the edge pixel Px is far away from the hues of the two reference pixels Pref1 and Pref2, if the hue of the edge pixel Px matches the hue of the reference pixels Pref1 and Pref2, a false color is displayed. There is a possibility of strengthening. Therefore, there is a high possibility that false colors can be reduced more effectively by reducing only the saturation without changing the hue of the edge pixel Px.

  When the false color reduction processing for each pixel is completed in this way, the processed YCbCr data is supplied from the pixel interpolation unit 220 to the color adjustment unit 230 (FIG. 1). The color adjustment unit 230 performs color adjustment such as white balance on the YCbCr data. The gradation correction unit 240 also performs gradation correction such as exposure correction and γ correction. As a result, bitmap image data BMIG in which the color and contrast are appropriately corrected is generated.

  Thus, in the above embodiment, the color of the edge pixel is adjusted based on the relationship between the color of the edge pixel and the colors of a plurality of reference pixels adjacent to the color of the edge pixel. Can be appropriately reduced. In particular, as shown in FIG. 9B, since the color adjustment mode is switched according to the range of the minimum hue difference ΔHmin between the edge pixel and the reference pixel, the color adjustment mode is changed according to the relationship between the edge pixel and the surrounding color. It is possible to reduce false colors by an appropriate method.

B. Variation:
The present invention is not limited to the above-described examples and embodiments, and can be implemented in various modes without departing from the gist thereof. For example, the following modifications are possible.

B1. Modification 1:
In the above embodiment, the color of the edge pixel is adjusted using two reference pixels, but it is possible to use three or more reference pixels. However, in this case as well, it is preferable to select a pixel group that faces and sandwiches a virtual straight line along the edge direction and is adjacent to the edge pixel. Here, the “pixel group adjacent to the edge pixel” does not have to be in contact with the edge pixel, and it is sufficient that the pixel closest to the edge pixel is in contact with the edge pixel. For example, it is possible to select a pixel outside the eight adjacent pixels shown in FIG. 6B as a part of the reference pixel group. Even when three or more reference pixels are used, it is preferable to adjust the color of the edge pixel using the hue of the reference pixel having the hue closest to the edge pixel, as in the above embodiment.

B2. Modification 2:
As a method for adjusting the color of the edge pixel, various methods other than the method employed in the above embodiment can be employed. For example, in the above embodiment, the color adjustment of the edge pixel is executed in accordance with the relationship between the hue of the edge pixel and the hue of the reference pixel. However, the color adjustment of the edge pixel is performed in consideration of the luminance of the edge pixel and the reference pixel. It is also possible to execute. For example, edge pixels with low luminance tend to be inconspicuous even if false colors are generated. Therefore, the false color reduction process may not be performed on edge pixels with low luminance.

  Further, the false color reduction process (FIG. 8) for non-edge pixels may be omitted. However, as in the above-described embodiment, it is preferable that the false color reduction process for non-edge pixels and the false color reduction process for edge pixels are executed by different methods. By doing so, it is possible to perform false color reduction processing suitable for edge pixels and non-edge pixels, respectively.

  In the above embodiment, the case where the minimum value ΔHmin of the color difference between the edge pixel and the reference pixel is equal to or smaller than the first threshold value TH1 is divided into two submodes (hue rotation mode and hue / saturation adjustment mode). These can be combined into one mode, or divided into three or more submodes. In either case, in each mode of ΔHmin ≦ TH1, it is preferable to perform color adjustment so that the hue of the edge pixel approaches the target hue Htg.

  Note that the mode in the case of TH1 <ΔHmin may also be divided into two or more submodes.

B3. Modification 3:
As a method for detecting the edge direction, various methods other than the methods shown in FIGS. 4 to 6 can be used. For example, other edge filters such as a Sobel filter can be used.

B4. Modification 4:
In the above embodiment, the YCbCr color space is used in the false color reduction process, but any other color space can be used.

B5. Modification 5:
In the above embodiment, the false color reduction process is performed on the RAW data generated by the digital camera 100. However, the false color reduction process of the present invention can be applied to various color image data other than the RAW data. is there. For example, in bitmap data (YCbCr data or RGB data) obtained by decompressing JPEG image data, an edge may exist at the boundary of the pixel block, and a false color may be generated at the edge pixel. As described above, it is possible to reduce the false color in the edge pixel by executing the false color reduction process of the present invention even for the image data in which each pixel has all the color components.

B6. Modification 6:
In the above embodiment, a part of the configuration realized by hardware may be replaced with software, and conversely, a part of the configuration realized by software may be replaced by hardware.

1 is an explanatory diagram showing an image processing system as one embodiment of the present invention. FIG. 3 is an explanatory diagram illustrating an example of a color filter array of a digital camera. The flowchart which shows the process sequence of a pixel interpolation part. Explanatory drawing which shows the edge filter for R pixel and B pixel. Explanatory drawing which shows the edge filter for G pixels. Explanatory drawing which shows the determination method of edge direction, and the reference pixel according to edge direction. The flowchart which shows the detailed procedure of step S3. The graph which shows the content of the saturation adjustment with respect to a non-edge pixel. Explanatory drawing which shows the color adjustment mode with respect to the relationship between the hue of an edge pixel and a reference pixel, and an edge pixel. Explanatory drawing which shows the processing content of three color adjustment modes.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 ... Digital camera 200 ... Computer 210 ... Image processing part 220 ... Pixel interpolation part 222 ... Edge detection processing part 224 ... Color interpolation processing part 226 ... False color reduction processing part 230 ... Color adjustment part 240 ... Tone correction part 250 ... Image Storage

Claims (8)

An image processing apparatus for reducing false colors in a color image,
An edge detection unit that detects edge pixels existing at edges in a color image and an edge direction of the edge pixels;
A plurality of reference pixels that are opposed to and adjacent to the edge pixel across a virtual straight line along the edge direction are detected, and based on the relationship between the color of the plurality of reference pixels and the color of the edge pixel A false color reduction processing unit that reduces the false color of the edge pixel by adjusting the color of the edge pixel;
An image processing apparatus comprising: The image processing apparatus according to claim 1,
The false color reduction processing unit performs first color adjustment of the edge pixel so as to approach the hue of a color proximity reference pixel having a hue closer to the hue of the edge pixel among the plurality of reference pixels. An image processing apparatus having a reduction mode. The image processing apparatus according to claim 1 or 2,
The false color reduction processing unit
(I) executing the first false color reduction mode when a difference between a hue of the color proximity reference pixel and a hue of the edge pixel is smaller than a predetermined first threshold;
(Ii) An image processing device that executes a second false color reduction mode for reducing the saturation of the edge pixel when the difference in hue is larger than the first threshold. The image processing apparatus according to claim 3,
The first false color reduction mode is:
(I-1) When the hue difference is smaller than a second threshold smaller than the first threshold, the hue of the edge pixel is changed to the color proximity reference pixel while maintaining the saturation of the edge pixel. A first sub-mode for performing color adjustment of the edge pixel so as to approach the hue of
(I-2) When the hue difference is between the first threshold value and the second threshold value, the hue of the edge pixel is referred to the color proximity reference while reducing the saturation of the edge pixel. And a second sub mode for performing color adjustment of the edge pixel so as to approach the hue of the pixel. The image processing apparatus according to any one of claims 1 to 4,
The edge detection unit selects one of a plurality of predetermined directions set in advance at intervals of 45 degrees as the edge direction,
The false color reduction processing unit is an image processing apparatus that selects two of the eight adjacent pixels existing around the edge pixel as the plurality of reference pixels. The image processing apparatus according to claim 1, further comprising:
A color interpolation processing unit for obtaining all color components of each pixel by performing color interpolation processing on a color image before color interpolation in which each pixel has only a part of color components,
The edge detection unit performs a process on the color image before the color interpolation process,
The false color reduction processing unit is an image processing device that executes processing on the color image after the color interpolation processing. A method for reducing false colors in a color image,
(A) detecting an edge pixel existing at an edge in a color image and an edge direction of the edge pixel;
(B) detecting a plurality of reference pixels facing each other across the virtual line along the edge direction and adjacent to the edge pixel;
(C) reducing the false color of the edge pixel by adjusting the color of the edge pixel based on the relationship between the color of the plurality of reference pixels and the color of the edge pixel;
A method comprising: A computer program for reducing false colors in a color image,
An edge detection function for detecting an edge pixel existing at an edge in a color image and an edge direction of the edge pixel;
A plurality of reference pixels that are opposed to and adjacent to the edge pixel across a virtual straight line along the edge direction are detected, and based on the relationship between the color of the plurality of reference pixels and the color of the edge pixel A false color reduction processing function for reducing the false color of the edge pixel by adjusting the color of the edge pixel;
A computer program that causes a computer to realize

Images