JP2001338285A - Image processor and recording medium with image processing program recorded thereon - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generate a high quality interpolated image by reducing noises to be generated, in interpolation processing of a first color information added with a second color information. SOLUTION: For the image processor performing interpolation processing of the first color information in image data containing at least the first color information and the second color information, this device is provided with an interpolating means for determining the interpolation value of the first color information, corresponding to the first color information and/or the second color information around an interpolation object pixel and a structure deciding means for deciding an image structure in a local area including the interpolation object pixel, and changing the contribution rate of the second color information in the interpolation value determination of the interpolation object pixel corresponding to the decided result.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

TECHNICAL FIELD The present invention relates to a second color other than the first color.
The present invention relates to an image processing technique for obtaining an interpolation value of first color information in consideration of image structure information obtained from color information.

[0002]

2. Description of the Related Art Conventionally, there has been known an image processing apparatus that determines an interpolation value of first color information in consideration of second color information other than the first color (US Pat. No. 5,629,734). Hereinafter, an interpolation process performed by this type of image processing apparatus will be described. First, the image processing apparatus extracts the following 5-row, 5-column local region from the image data with the interpolation target pixel as the center. Here, G ○ means G color information, and Z ○ means other color information (R color or B color). .. G2... Z3G4Z5G6Z7... G8.... Z9... Next, the image processing apparatus calculates DH = | −Z3 + 2.Z5- based on the color information in this local area. Z7 | + | G4-G6 | ··· [Equation 1] DV = | -Z1 + 2 · Z5-Z9 | + | G2-G8 | ··· [Equation 2] G5H = (G4 + G6) / 2 + (-Z3 + 2 ・ Z5-Z7) / 4 ・ ・ ・ [Equation 3] G5V = (G2 + G8) / 2 + (-Z1 + 2 ・ Z5-Z9) / 4 ・ ・ ・ [Equation 4] G5A = (G2 + G4 + G6 + G8) / 2 + (-Z1-Z3 + 4 ・ Z5-Z7-Z9) / 8 [Equation 5] if DH <DV then G5 = G5H else if DV <DH then G5 = G5V else G5 = G5A A series of calculations consisting of [Equation 6] are executed to determine the value of the interpolation target pixel G5. The first term in the above [Equations 3 to 5] is a simple linear interpolation value based on the G color information. On the other hand, the second term is a curvature component (spatial second derivative) of other color information. In the conventional example, the detail component of the interpolated image is supplemented by adding the second term.

[0003]

However, in the above-described conventional example, dots such as black and white (hereinafter referred to as "dot noise") may occur in the interpolated image. Since the dot noise is noticeable in the interpolated image, it greatly impairs the image quality of the interpolated image. In view of the above, an object of the present invention is to generate a high-quality interpolated image.

[0004]

According to a first aspect of the present invention, there is provided an image processing apparatus for performing an interpolation process of first color information on image data including at least first color information and second color information. In the apparatus, an interpolation means for determining an interpolation value of the first color information according to the first color information and / or the second color information around the interpolation target pixel, and determining an image structure for a local region including the interpolation target pixel. And a structure judging means for changing the contribution ratio of the second color information in determining the interpolation value of the interpolation target pixel according to the judgment result.

[0005] Usually, at a place such as a color boundary in a screen,
Since the ratio of the color components changes spatially, the first color information and the second color information show different changes. When the conventional interpolation processing is performed at such a color boundary, different image structures of the second color information are mixed in the interpolated image of the first color information, and dot noise occurs. Thus, in the above configuration, the structure determining unit determines the image structure in a local region including the interpolation target pixel. The structure determining unit changes the contribution ratio of the second color information in determining the interpolation value according to the determination result. Therefore, using the above configuration, it is possible to temporarily reduce the contribution ratio according to the situation where different image structures of the second color information are mixed in the interpolation image. as a result,
The suppression of dot noise and the reproduction of detail can both be achieved in a well-balanced manner, and a high-quality interpolated image can be generated.

[0008] According to a second aspect of the present invention, in the image processing apparatus according to the first aspect, the structure determining means determines whether a peak or trap does not exist in a local area including the pixel to be interpolated. And means for reducing the contribution ratio of the second color information in determining the interpolation value of the interpolation target pixel.

In general, a portion where no peak or trap occurs in a local region indicates that the color is monotonically changing, and is more likely to be a color boundary than a detail. Therefore, when the peak and the trap of the second color information do not exist in the local region including the pixel to be interpolated, the structure determination unit temporarily reduces the contribution ratio of the second color information in determining the interpolation value. As a result, the occurrence of dot noise and the like are reliably reduced, and the dot noise can be reduced.

[0008] Even if the determination of the color boundary is incorrect, the detail components included in such a portion are originally small, so that there is little possibility that the detail reproducibility of the interpolated image will be impaired as a whole. Therefore, in the above configuration, the suppression of the dot noise and the reproduction of the detail are compatible with a good balance, and a high-quality interpolated image can be generated.

[0010] According to a third aspect of the present invention, in the image processing apparatus according to the first aspect, the structure determining means determines an image structure for a local area including the pixel to be interpolated, and determines the first color. When it is determined that the correlation between the image structure of the information and the image structure of the second color information is small, the means is a means for reducing the contribution ratio of the second color information in determining the interpolation value of the pixel to be interpolated.

In the above arrangement, the structure determining means determines the image structure in the local area including the pixel to be interpolated. As a result of this determination, when the correlation between the “image structure of the first color information” and the “image structure of the second color information” is small, the contribution rate of the second color information in determining the interpolation value is temporarily reduced. Therefore, the chance and amount of different image structures of the second color information entering the interpolated image of the first color information are reliably reduced, and the dot noise in the interpolated image can be reduced. on the other hand,
On the other hand, if the correlation is somewhat strong, the contribution of the second color information is not reduced. Therefore, the image structure of the first color information is appropriately supplemented with the image structure of the second color information, and it is possible to obtain an interpolated image with rich details.

According to a fourth aspect of the present invention, in the image processing apparatus according to any one of the first to third aspects, the similarity of each direction in a local region including a pixel to be interpolated is provided. , And the interpolation means calculates the interpolation value by increasing the contribution ratio of the color information positioned in the direction of high similarity determined by the direction determination means.

<Fifth aspect> According to a fifth aspect of the present invention, in the image processing apparatus according to the fourth aspect, the structure determining unit determines the image structure with respect to the “high similarity direction” determined by the direction determining unit. Is determined.

In the above arrangement, the structure determining means determines the image structure in the direction in which the interpolation operation is performed. Therefore, there is no need to determine the image structure for all directions,
It is possible to efficiently reduce the calculation amount of the determination process.

According to a sixth aspect of the present invention, there is provided an image processing apparatus for performing interpolation of first color information on image data including at least first color information and second color information. Interpolating means for determining an interpolation value of the first color information in accordance with at least the first color information and / or the second color information, and determining a saturation or a color difference with respect to a local area including a pixel to be interpolated. Alternatively, the image processing apparatus further includes a color determination unit configured to reduce a contribution rate of the second color information in determining an interpolation value of the interpolation target pixel when the color difference is determined to be large.

In general, locations where the saturation or color difference is large are:
The signal levels of the first color information and the second color information are significantly different. When details of the image structure of the first information are supplemented with the image structure of the second information in such a portion, an error caused by the level difference is mixed in the interpolated image. Such errors may also cause dot noise. Therefore, in the above configuration, the color determination unit determines the saturation or the color difference in the local area including the interpolation target pixel. If it is determined that the saturation or the color difference is large,
The color determination means temporarily reduces the contribution rate of the second color information in determining the interpolation value. As a result, the chance and the amount of the error caused by the above-described level difference to be mixed into the interpolated image are surely reduced, and the dot noise can be reduced.

<Seventh Aspect> According to the seventh aspect, in the image processing apparatus according to any one of the first to sixth aspects, the contribution rate of the second color information is substantially reduced when the second color information is reduced. It is characterized by being set to zero.

As described above, the nonlinear processing for reducing the contribution ratio of the second information to substantially zero makes it possible to reliably prevent the generation of dot noise.

According to an eighth aspect of the present invention, there is provided an image processing apparatus for performing interpolation processing of first color information on image data including at least first color information and second color information. In a predetermined direction on the image passing through (the pixel position of Y (0) below), first color information X (-1) second color information Y (0) first color information X (+1) For a pixel array in which two-color information Y (+2) is arranged, the first color information X
(0) is calculated as X (0) = (1-a1) .X (-1) + a1.X (+1) + a2.Y (0) -a2.Y (+2) (Where a1 and a2 are asymmetrical interpolation means determined by a predetermined proportional coefficient).

The first and second terms of the above [Equation 7] are:
Using the surrounding first color information X (-1), X (+1), (1-a1) vs. a1
Is a term for which a linear interpolation value of X (0) is obtained by performing interpolation using the internal division ratio of. On the other hand, the third and fourth terms in [Equation 7] are the spatial first derivative (gradient) between the second color information Y (0) and Y (+2).
Corresponds to. This spatial first derivative is a component obtained by extracting a spatial change of the second color information from Y (0) toward Y (+2). In [Equation 7], the spatial first derivative is added to the linear interpolation value of X (0) by the weighting coefficient a2 to supplement the detail of the interpolated image.

Therefore, it is possible to generate an interpolated image rich in detail, which cannot be obtained only by the interpolation processing of the first color information. In particular, the present invention is characterized in that peripheral pixels of an interpolation symmetric pixel are used asymmetrically as compared with the above-described conventional example. Therefore, the number of peripheral pixels used is smaller than that of the symmetrical conventional example (5 → 4), and it is possible to reduce such an adverse effect that errors are mixed in the interpolation values across color boundaries and edges. It becomes possible. With such an operation, the chance of generating dot noise is reduced, and a high-quality interpolated image can be generated.

Further, the interpolation processing of the present invention can be applied if the above-mentioned pixel arrangement exists near the pixel to be interpolated. Therefore, the present invention is applicable not only to the Bayer array but also to image data of various color arrays. For example, even when such an arrangement exists only in a specific direction, the present invention is applicable to the specific direction. Also, for example, the present invention is applicable to a case where adjacent colors are different in the vertical direction and the horizontal direction as in a complementary color arrangement.

In a ninth aspect of the present invention, in the image processing apparatus according to the eighth aspect, in a local area including the pixel to be interpolated, a plurality of predetermined directions starting from the pixel to be interpolated are similar. Direction determining means for determining the sex, and the asymmetric interpolation means selects the predetermined direction determined to have high similarity by the direction determining means and calculates the first color information X (0). .

In the above arrangement, the direction determining means determines the similarity in a plurality of predetermined directions starting from the pixel to be interpolated. The asymmetric interpolation means executes an asymmetric interpolation operation in the predetermined direction having a high similarity. Therefore, the possibility that an asymmetric interpolation operation is performed across color boundaries and edges is further reduced, and the chance of generating dot noise is further reduced.

In a tenth aspect of the present invention, in the image processing apparatus according to the eighth aspect, in a local region including the pixel to be interpolated, the local region including the pixel to be interpolated is similar in a plurality of predetermined directions starting from the pixel to be interpolated. Direction determining means for determining the gender, and the asymmetric interpolation means weights and synthesizes the first color information X (0) calculated for each of the plurality of predetermined directions at a weight ratio according to the similarity determined by the direction determining means. It is characterized by the following.

With the above configuration, it is possible to reduce the weight ratio in the direction of low similarity. As a result, error mixing from a direction that straddles a color boundary or an edge is suppressed, and it is possible to reduce the occurrence opportunity and amount of dot noise. Therefore, it is possible to generate a high-quality interpolation image.

(11) An image processing program for causing a computer to function as the image processing apparatus according to any one of (1) to (10) is recorded on the recording medium according to claim 11. It is characterized by having.

By executing such an image processing program, the computer functions as each component according to any one of the first to tenth aspects. As a result, claim 1
The image processing device described in any one of the above items 10 to 10 can be realized on a computer.

[0028]

Embodiments of the present invention will be described below with reference to the drawings.

First Embodiment A first embodiment is an electronic camera according to the first, second, fourth, fifth, and seventh aspects.

[Schematic Configuration of Electronic Camera] FIG. 1 is a diagram showing a schematic configuration of the electronic camera 11. In FIG. 1, a photographing lens 12 is mounted on an electronic camera 11. The light receiving surface of the image sensor 13 is provided in the image space of the photographing lens 12. The image sensor 13 photoelectrically converts the light image on the light receiving surface to generate original image data in a Bayer array.
The original image data is digitized via a signal processing unit 14 and an A / D conversion unit 15 and then supplied to an image processing unit 16 including a signal processor and an image memory.

The image processing section 16 performs gradation conversion such as defective pixel correction, black level clamping, white balance processing, and γ correction, color interpolation processing, color space conversion, and spatial filter processing (edge processing) on the original image data. Image processing such as emphasis). The image data that has been subjected to the image processing by the image processing unit 16 is stored in a recording system (not shown) in the electronic camera 11.
Output to

In the configuration described above, the interpolation means of the present invention corresponds to the "function of performing G color interpolation processing" in the image processing section 16. The structure determining means corresponds to “the function of determining the image structure and changing the contribution ratio of the color information other than the G color in the G color interpolation according to the determination result” in the image processing unit 16. Further, the direction determining means corresponds to the “function of determining an interpolation direction by similarity determination” in the image processing unit 16.

[Outline of Color Interpolation Processing] FIG. 2 is a diagram for explaining the outline of the color interpolation processing in the image processing section 16. Hereinafter, before describing the features of the present invention, the overall operation of the color interpolation processing will be described. First, the image processing unit 16 performs white balance processing, gradation conversion, and the like on the original image data to generate image data before color interpolation. The image data before the color interpolation is G [X,
Y] and Z which is color information other than G color (R color or B color)
[X, Y]. Here, X is a coordinate value indicating a pixel position in the horizontal direction, and Y is a coordinate value indicating a pixel position in the vertical direction.

The image processing section 16 performs a G color interpolation process on the image data before the color interpolation, and obtains a G plane FG [X,
Y] is generated. Next, the image processing unit 16 outputs Z [X,
Y] and FG [X, Y] to perform color difference conversion,
Two types of color difference planes DB [X, Y] and DR [X, Y] are generated. The image processing unit 16 calculates the color difference plane DB
The color difference interpolation is executed for [X, Y] and DR [X, Y], and the color difference planes FDB [X, Y] and FDR after the color difference interpolation are performed.
[X, Y] is generated.

Subsequently, the image processing section 16 controls the color difference plane FDB.
Based on [X, Y], FDR [X, Y] and G plane FG [X, Y], B plane FB [X, Y] and R plane FR [X,
Y] is restored. After a series of courses described above, RGB3
Interpolated images FR [X, Y], FG [X, Y], FB
[X, Y] is generated.

[Detailed Description of G Interpolation Processing] Next, the operation of the G interpolation processing which is a feature of the present invention will be described. FIG.
FIG. 9 is a diagram showing a G interpolation processing routine for an interpolation target pixel [i, j]. Hereinafter, the G interpolation processing will be described along the step numbers shown in FIG.

Step S1: The image processing unit 16 determines that Cv0 [i, j] = | G [i, j-1] -G [i, j + 1] | + (| G [i, j-1] -Z [i, j] | + | G [i, j + 1] -Z [i, j] |) / 2 [Equation 8] Ch0 [i, j] = | G [i-1, j] -G [i + 1, j] | + (| G [i-1, j] -Z [i, j] | + | G [i + 1, j]- Z [i, j] |) / 2 [Expression 9] is calculated, and the dissimilarities Cv0 and Ch0 are obtained.

Step S2: Next, the image processing unit 16
Is obtained by locally averaging the dissimilarities Cv0 and Ch0 according to the following equation,
The reference range of dissimilarity is expanded. Cv [i, j] = (4 ・ Cv0 [i, j] + Cv0 [i-1, j-1] + Cv0 [i + 1, j-1] + Cv0 [i-1, j + 1] + Cv0 [i + 1, j + 1]) / 8 [Equation 10] Ch [i, j] = (4 · Ch0 [i, j] + Ch0 [i-1, j-1] + Ch0 [ i + 1, j-1] + Ch0 [i-1, j + 1] + Ch0 [i + 1, j + 1]) / 8 [Equation 11]

Step S3: The image processing unit 16 compares the dissimilarities Cv and Ch with the reference range expanded to determine the direction of higher similarity. First, | Cv [i, j] -Ch [i, j] | ≦
If the threshold value is Th1, the image processing unit 16 determines that the similarity is substantially the same in both directions, and shifts the operation to step S10. Next, if Cv [i, j] <Ch [i, j], the image processing unit 16 determines that the similarity in the vertical direction is high, and proceeds to step S4.
Move the operation to. On the other hand, if Cv [i, j]> Ch [i, j],
The image processing unit 16 determines that the similarity in the horizontal direction is high, and shifts the operation to step S7.

Step S4: The image processing section 16 calculates G = (G [i, j-1] + G [i, j + 1]) / 2 in the vertical direction having high similarity [Equation 12]. To calculate the same color interpolation value G.

Step S5: The image processing unit 16 calculates Xr = | Z [i, j-2] -Z [i, j + 2] |-| 2 · Z [i, j for the vertical direction having high similarity. ] -Z [i, j-2] -Z [i, j + 2] |... [Equation 13] is calculated and set as the peak discrimination value Xr.

Step S6: The image processing section 16 sets the coefficient a used in step S15 to be described later to one. Thereafter, the image processing unit 16 shifts the operation to Step S13.

Step S7: The image processing section 16 calculates G = (G [i-1, j] + G [i + 1, j]) / 2 in the horizontal direction having high similarity [Equation 14]. To calculate the same color interpolation value G.

Step S8: The image processing section 16 determines that Xr = | Z [i-2, j] -Z [i + 2, j] |-| 2 · Z [i, j ] -Z [i-2, j] -Z [i + 2, j] |... [Equation 15] is calculated and set as the peak discrimination value Xr.

Step S9: The image processing section 16 sets the coefficient a used in step S15 described later to 0. Thereafter, the image processing unit 16 shifts the operation to Step S13.

Step S10: The image processing section 16 calculates G = (G [i, j-1] + G [i, j + 1] + G [i-1, j] + G [i] in both the vertical and horizontal directions. +1, j]) / 4 [Expression 16] is calculated, and the same color interpolation value G is obtained.

Step S11: The image processing section 16 determines that Xrv = | Z [i, j-2] -Z [i, j + 2] |-| 2 · Z [i, j] -Z [i, j- 2] -Z [i, j + 2] | [Equation 17] Xrh = | Z [i-2, j] -Z [i + 2, j] |-| 2 · Z [i, j] -Z [i-2, j] -Z [i + 2, j] | ... [Equation 18] Xr = min (Xrv, Xrh) ... [Equation 19] is calculated, and the peak determination value Xr is calculated. I do.

Step S12: The image processing section 16 sets the coefficient a used in step S15 to be described later to 1/2.
Set to. Thereafter, the image processing unit 16 proceeds to step S1
The operation moves to 3.

Step S13: Here, the image processing unit 1
Step 6 determines the value of the peak determination value Xr calculated in the above step. Here, when the peak discrimination value Xr is smaller than the threshold Th2, the image processing unit 16 determines that there is no peak and trap in the local area including the interpolation target pixel, and shifts the operation to step S14. On the other hand, the peak discrimination value Xr
Is greater than the threshold Th2, the image processing unit 16 determines that a peak or trap exists in the local area including the interpolation target pixel, and shifts the operation to step S15. The threshold Th2 may be set to, for example, zero. Further, in a plurality of test images, the value of the peak discrimination value Xr at the location where the dot noise occurs may be checked, and the threshold Th2 may be determined in consideration of the average value and the variance of the peak discrimination values Xr.

Step S14: The image processing section 16 determines that there is a high possibility that dot noise will occur in the pixel [i, j] to be interpolated, and temporarily reduces the contribution ratio of colors other than G in the interpolation value determination to substantially zero. And As a result, FG [i, j] = G (Equation 20) is used as the final interpolation value of the G color at the interpolation target pixel.

Step S15: The image processing section 16 determines that there is little possibility that dot noise will occur in the pixel [i, j] to be interpolated, and maintains the contribution ratios other than the G color in the interpolation value determination. As a result, FG [i, j] = G + [Z [i, j] / 2- [a · (Z [i, j-2] + Z [i, j + 2]) + (1-a) [(Z [i−2, j] + Z [i + 2, j])] / 4] [Expression 21] is used as the final interpolation value of the G color at the interpolation target pixel.

[Effects of First Embodiment, etc.] FIG. 4 is an enlarged picture (an image obtained by dithering a halftone image on a display by dithering) to compare and explain the effects of the first embodiment. Data). Among them, FIG. 4A is a general interpolated image using only adjacent G color values. In FIG. 4A, in the interpolation processing, the oblique edge of the detail portion is not faithfully reproduced, and the edge shape is broken.

FIG. 4B shows an interpolated image in which color information other than the G color is taken into account, as in the above-described conventional example. In FIG. 4B, the oblique edge of the detail portion is faithfully reproduced and has a smooth edge shape. However, for color boundary points, dot noise (black spots)
Many have occurred.

FIG. 4C is an interpolated image processed according to the first embodiment. In FIG. 4C, the oblique edge of the detail portion is faithfully reproduced, and has a smooth edge shape. Further, the color boundary portion has a distinctive feature that the number of occurrences of dot noise is extremely small.

As described above, in the first embodiment, it is possible to generate a high-quality interpolated image with little dot noise and excellent detail reproducibility. Next, another embodiment will be described.

<< Second Embodiment >> A second embodiment is an electronic camera according to the first, second, fourth, fifth and seventh aspects. Note that the schematic configuration of the electronic camera according to the second embodiment is common to that of the first embodiment (FIG. 1).
The following description will be made using the reference numerals shown in FIG. 1 as they are.

Here, as for the correspondence between the constituent elements of the present invention and the second embodiment, the interpolation means uses the image processing unit 1
6 corresponds to the “function of performing G color interpolation processing”.
Further, the structure determination means corresponds to “the function of determining the image structure and changing the contribution ratio of colors other than G in the G color interpolation according to the determination result” in the image processing unit 16. Further, the direction determining means corresponds to the “function of determining an interpolation direction by similarity determination” in the image processing unit 16.

FIG. 5 is a diagram showing a G interpolation processing routine for the interpolation target pixel [i, j]. The difference in the operation of the second embodiment is that steps S13 to S15 in the first embodiment (FIG. 3) are replaced by step S2 shown in FIG.
1 to 22. Hereinafter, the operation of the difference will be described.

Step S21: The image processing section 16 determines the contribution ratio b according to the peak discrimination value Xr. Here, for example, a membership function b (Xr) as shown in FIG. 6 is used to determine the contribution rate b. This membership function b
(Xr) is a function that gradually reduces the contribution ratio b from the upper limit c to the lower limit zero as the peak discrimination value Xr increases.

Step S22: The image processing unit 16 uses the determined contribution ratio b to calculate FG [i, j] = G + b · [Z [i, j] / 2- [a · (Z [ i, j-2] + Z [i, j + 2]) + (1-a) · (Z [i-2, j] + Z [i + 2, j]) / 4] ・ ・ ・ [ Equation 22] is calculated and used as a G color interpolation value in the interpolation target pixel.

[Effects of Second Embodiment, etc.] In the second embodiment, when it is uncertain whether or not to generate dot noise, the contribution ratio b is set to an appropriate value between the upper limit value c and the lower limit value zero. Set to an intermediate value. Therefore, it is possible to appropriately control the suppression of the amplitude of the dot noise and the reproduction of the detail in an uncertain location. Next, another embodiment will be described.

Third Embodiment A third embodiment is an electronic camera according to the first, third, fourth, fifth, and seventh aspects. Note that the schematic configuration of the electronic camera according to the third embodiment is the same as that of the first embodiment (FIG. 1).
The following description will be made using the reference numerals shown in FIG. 1 as they are.

Here, as for the correspondence between the constituent elements of the present invention and the third embodiment, the interpolation means uses the image processing unit 1.
6 corresponds to the “function of performing G color interpolation processing”.
Further, the structure determination means corresponds to “the function of determining the image structure and changing the contribution ratio of colors other than G in the G color interpolation according to the determination result” in the image processing unit 16. Further, the direction determining means corresponds to the “function of determining an interpolation direction by similarity determination” in the image processing unit 16.

FIG. 7 is a diagram showing a G interpolation processing routine for the interpolation target pixel [i, j]. The difference in operation in the third embodiment is that steps S5, S8, S11, and S13 in the first embodiment (FIG. 3) are replaced with steps S23 to S26 shown in FIG. 7, respectively. . Hereinafter, the operation of the difference will be described.

Step S23: The image processing unit 16 calculates Xc = | Z [i, j-2] -Z [i, j + 2]-(G [i, j-1] for the vertical direction having high similarity. -G [i, j + 1]) · 2 | (Equation 23) is calculated as the correlation determination value Xc.

Step S24: The image processing section 16 calculates Xc = | Z [i-2, j] -Z [i + 2, j]-(G [i-1, j]) for the horizontal direction having a high similarity. -G [i + 1, j]) · 2 | (Equation 24) is calculated and set as the correlation determination value Xc.

Step S25: The image processing unit 16 determines that Xcv = | Z [i, j-2] -Z [i, j + 2]-(G [i, j-1] -G [i, j + 1) ]) · 2 | ... [Equation 25] Xch = | Z [i-2, j] -Z [i + 2, j]-(G [i-1, j] -G [i + 1, j ]] · 2 | ··· [Equation 26] Xc = max (Xcv, Xch) ··· [Equation 27] is used as the correlation determination value Xc.

Step S26: Here, the image processing unit 1
6 determines the value of the correlation determination value Xc calculated in the above step. Here, when the correlation determination value Xc exceeds the threshold Th3, the image processing unit 16 determines that the correlation between the image structure of G color and the image structure other than G color is low, and shifts the operation to step S14. On the other hand, when the correlation determination value Xc is smaller than the threshold Th3, the image processing unit 16 determines that the correlation between the image structure of G color and the image structure other than G color is high, and shifts the operation to step S15. It is preferable that the value of the threshold Th3 be determined by examining the value of the correlation discrimination value Xc at the place where the dot noise occurs in a plurality of test images and considering the average value and variance of these correlation discrimination values Xc.

[Effects of Third Embodiment, etc.] In the third embodiment, when the correlation between the image structure of G color and the image structure other than G color is small, the contribution ratio of the color other than G in determining the interpolation value is determined. Substantially zero. Therefore, there is no possibility that different image structures other than the G color are mixed into the G color interpolation image, and it is possible to reliably reduce the dot noise. Conversely, when the correlation is strong, the contribution rate is not reduced, the image structure of G color is appropriately supplemented with an image structure other than G color, and an interpolation image rich in detail is generated. Next, another embodiment will be described.

<< Fourth Embodiment >> A fourth embodiment is an electronic camera according to the sixth and seventh aspects. The schematic configuration of the electronic camera according to the fourth embodiment is as follows.
Since the second embodiment is common to the first embodiment (FIG. 1), the following description will be made using the reference numerals shown in FIG. 1 as they are.

Here, as for the correspondence between the constituent features of the present invention and the fourth embodiment, the interpolation means uses the image processing unit 1.
6 corresponds to the “function of performing G color interpolation processing”.
In addition, the structure determination means corresponds to the “function of determining a color difference and changing a contribution ratio other than the G color in the G interpolation according to the determination result” in the image processing unit 16.

FIG. 8 is a diagram showing a G interpolation processing routine for the interpolation target pixel [i, j]. The operation difference in the fourth embodiment is that steps S5, S8, S11, and S13 in the first embodiment (FIG. 3) are replaced with steps S27 to S30 shown in FIG. 8, respectively. . Hereinafter, the operation of the difference will be described.

Step S27: The image processing section 16 calculates Xs = | Z [i, j-2] + Z [i, j] -G [i, j-1] · 2 | + | Z [i, j + 2] + Z [i, j] -G [i, j + 1] · 2 | (Equation 28) is calculated and set as the color difference determination value Xs.

Step S28: The image processing section 16 determines that Xs = | Z [i-2, j] + Z [i, j] -G [i-1, j] · 2 | + | Z [i + 2, j] + Z [i, j] -G [i + 1, j] · 2 | (Equation 29) is calculated and used as the color difference determination value Xs.

Step S29: The image processing unit 16 determines that Xsv = | Z [i, j-2] + Z [i, j] -G [i, j-1] · 2 | + | Z [i, j + 2] + Z [i, j] -G [i, j + 1] · 2 |... [Equation 28a] Xsh = | Z [i−2, j] + Z [i, j] -G [i −1, j] · 2 | + | Z [i + 2, j] + Z [i, j] −G [i + 1, j] · 2 | (Equation 29a) Xs = max (Xsv, Xsh)... [Equation 30] is calculated as the color difference determination value Xs.

Step S30: Here, the image processing unit 1
6 determines the value of the color difference determination value Xs calculated in the above step. Here, if the color difference determination value Xs exceeds the threshold Th4, the image processing unit 16 determines that the color difference (saturation) of the local region is large, and shifts the operation to step S14. On the other hand, when the color difference determination value Xs is smaller than the threshold Th4, the image processing unit 16
Determines that the color difference in the local region is small, and determines in step S1
Operation is shifted to 5. It is preferable that the value of the threshold Th4 be determined by examining the value of the color difference discrimination value Xs at the place where the dot noise occurs in a plurality of test images, and considering the average value and variance of these color difference discrimination values Xs.

[Effects of the Fourth Embodiment] In the fourth embodiment, when the color difference (saturation) of the G color is large, the contribution ratios other than the G color in the determination of the interpolation value are set to substantially zero. Therefore, it is possible to reliably reduce the dot noise that is likely to be generated in a high chroma portion. Conversely, when the saturation is small, the contribution rate is not reduced, and the details of the G interpolation pixels are appropriately enhanced with an image structure other than the G color. next,
Another embodiment will be described.

Fifth Embodiment A fifth embodiment is an electronic camera according to the eighth and ninth aspects. The schematic configuration of the electronic camera according to the fifth embodiment is as follows.
Since the second embodiment is common to the first embodiment (FIG. 1), the following description will be made using the reference numerals shown in FIG. 1 as they are.

Here, as for the correspondence between the constituent elements of the present invention and the fifth embodiment, the direction determining means determines whether or not the image processing section 16 determines “similarity determination for a plurality of predetermined directions starting from the pixel to be interpolated. Function to perform Further, the asymmetric interpolation means corresponds to the “function of obtaining an asymmetric interpolation value of G color in a predetermined direction with high similarity” in the image processing unit 16.

FIG. 9 is a diagram showing a G interpolation processing routine according to the fifth embodiment. Hereinafter, the operation of the G interpolation processing will be described along the step numbers shown in FIG.

Step S41: The image processing section 16 performs similarity determination in the vertical and horizontal directions in the local area including the pixel to be interpolated, and determines a direction having a high similarity as the interpolation direction.

Step S42: The image processing section 16 arranges in the direction of interpolation around the pixel to be interpolated (the pixel position of Y (0) below). Second color information Y (-2) First color information X (- 1) 2nd color information Y (0) 1st color information X (+1) 2nd color information Y (+2) Yth = | Y (-2) -Y (0) | -| Y (+2) -Y (0) |... [Equation 31] is calculated to determine the similarity determination value Yth. The similarity determination value Yth here is a result of comparing similarities in two directions starting from the pixel to be interpolated (ie, asymmetry).

Step S43: The image processing section 16 determines the value of the similarity determination value Yth. Here, the similarity determination value Y
When the absolute value of th is equal to or smaller than the threshold th5, the image processing unit 16
It is determined that there is no clear difference in similarity, and the operation moves to step S44. If the similarity determination value Yth exceeds the threshold th5, the image processing unit 16 determines that the similarity on the Y (+2) side is high, and shifts the operation to step S45. On the other hand, when the similarity determination value Yth is smaller than the threshold value -th5, the image processing unit 1
6 judges that the similarity on the Y (-2) side is high, and determines in step S4
The operation moves to 6. It is preferable that the threshold value Th5 be determined on a trial basis while evaluating the frequency of occurrence of dot noise in a plurality of test images.

Step S44: Since there is no clear difference in similarity, the image processing section 16 executes an interpolation operation symmetrical in the interpolation direction to determine an interpolation value X (0) of the interpolation symmetric pixel.

Step S45: Since the similarity on the Y (+2) side is high, the image processing unit 16 executes an asymmetric interpolation operation using the Y (+2) side according to the following equation. X (0) = a1 · X (-1) + (1-a1) · X (+1) + a2 · [Y (0) −Y (+2)] (32) where the coefficient About a1, for example, about 1/2 is preferable. Also, to increase the weighting ratio in the direction of higher similarity,
The coefficient a1 may be set to a value slightly smaller than 1/2. Further, it is preferable that the coefficient a2 is determined on a trial basis by subjecting a plurality of test images to subjective evaluation of the amount of detail enhancement and the like.

Step S46: Since the similarity on the Y (-2) side is high, the image processing unit 16 executes an asymmetric interpolation operation using the Y (-2) side according to the following equation. X (0) = a1 · X (+1) + (1-a1) · X (−1) + a2 · [Y (0) −Y (−2)] (33) where the coefficient About a1, for example, about 1/2 is preferable. Also, to increase the weighting ratio in the direction of higher similarity,
The coefficient a1 may be set to a value slightly smaller than 1/2. Further, it is preferable that the coefficient a2 is determined on a trial basis by subjecting a plurality of test images to subjective evaluation of the amount of detail enhancement and the like.

[Effects of Fifth Embodiment] As described above, in the fifth embodiment, an asymmetric interpolation operation is executed after selecting a direction having a high similarity. Therefore, there is little possibility that an error is mixed in the interpolated image over the color boundary or the edge, and a high-quality interpolated image with little dot noise can be generated.

In the above-described embodiment, the asymmetric interpolation operation is executed by selecting a direction having a high similarity. However, the present invention is not limited to this. More generally, claim 1
As described in “0”, the asymmetric interpolation values in a plurality of directions may be weighted and synthesized at a weight ratio according to the similarity. For example, b1 = | Y (+2) -Y (0) | (Expression 34a) b2 = | Y (-2) -Y (0) | (Expression 34b) Weighted synthesis may be performed as described above. X (0) = ((a1 ・ X (-1) + (1-a1) ・ X (+1) + a2 ・ (Y (0) -Y (+2))) ・ b2 + (a1 ・ X ( +1) + (1-a1) · X (-1) + a2 · (Y (0) −Y (-2))) · b1) / (b1 + b2) (Equation 35) The interpolation calculation can be executed at a weighting ratio that accurately reflects the degree of similarity. Next, another embodiment will be described.

<< Sixth Embodiment >> A sixth embodiment is an electronic camera according to claims 1, 4, 5, 8, and 10. Since the schematic configuration of the electronic camera according to the sixth embodiment is common to that of the first embodiment (FIG. 1), the following description will be made using the reference numerals shown in FIG. 1 as they are.

Here, regarding the correspondence between the constitutional requirements of claims 1, 4 and 5 and the sixth embodiment, the interpolating means corresponds to the "function of performing G color interpolation processing" in the image processing section 16. I do. Further, the structure determination means corresponds to “the function of changing the contribution ratio of the G interpolation other than the G color according to the determination result of the image structure” in the image processing unit 16. On the other hand, with regard to the correspondence between the constitutional requirements of claims 8 and 10 and the sixth embodiment, the direction determining means performs the “ Function to perform ”. Also, the asymmetric interpolation means corresponds to the “function of determining an asymmetric interpolation value of G color in a plurality of predetermined directions and determining a G color interpolation value by weight-synthesizing these asymmetric interpolation values” in the image processing unit 16. .

FIG. 10 is a diagram showing a G interpolation processing routine according to the sixth embodiment. Hereinafter, the operation of the G interpolation processing will be described along the step numbers shown in FIG.

Step S51: The image processing section 16 determines the similarity in the vertical and horizontal directions in the local area including the pixel to be interpolated, and determines the direction having the higher similarity as the interpolation direction.
If the vertical direction is determined as the interpolation direction, the image processing unit 1
6 shifts the operation to step S52. When the horizontal direction is determined to be the interpolation direction, the image processing unit 16 determines in step S
The operation shifts to 55. On the other hand, when there is no clear difference between the vertical and horizontal similarities, the image processing unit 16 sets the interpolation direction to the vertical and horizontal directions and shifts the operation to step S58.

Step S52: Here, the image processing unit 1
6 is bt1 = | Z [i, j] -Z [i, j-2] | +1 ... [Expression 36] bt2 = | Z [i, j] -Z [ i, j + 2] | +1 (Expression 37) mbt = bt1 + bt2 (Expression 38) Note that the addition term “+1” in [Equations 36 to 37]
Is a minimum number for avoiding a division by zero error in step S61 described later, and can be omitted when error processing is separately performed.

Step S53: The image processing section 16 calculates G = (G [i, j-1] + G [i, j + 1]) / 2 in the vertical direction as the interpolation direction [Equation 39]. To calculate the same color interpolation value G.

Step S54: The image processing section 16 sets the coefficient a used in step S61 described later to one. Thereafter, the image processing unit 16 shifts the operation to Step S61.

Step S55: Here, the image processing unit 1
6 is by1 = | Z [i, j] -Z [i-2, j] | +1... [Equation 40] in the horizontal direction which is the interpolation direction by2 = | Z [i, j] -Z [ i + 2, j] | +1 (Expression 41) mby = by1 + by2 (Expression 42) is calculated. Note that the addition term “+1” in [Equations 40 to 41]
Is a minimum number for avoiding a division by zero error in step S61 described later, and can be omitted when error processing is separately performed.

Step S56: The image processing section 16 calculates G = (G [i-1, j] + G [i + 1, j]) / 2 in the horizontal direction which is the interpolation direction [Equation 43]. To calculate the same color interpolation value G.

Step S57: The image processing section 16 sets the coefficient a used in step S61 described later to zero. Thereafter, the image processing unit 16 proceeds to step S61.
Move the operation to.

Step S58: Here, the image processing unit 1
6 is bt1 = | Z [i, j] -Z [i, j-2] | +1... [Equation 44] bt2 = | Z [i, j] -Z [ i, j + 2] | +1 (Equation 45) mbt = bt1 + bt2 (Equation 46) by1 = | Z [i, j] -Z [i-2, j] | +1 .. [Equation 47] by2 = | Z [i, j] -Z [i + 2, j] | +1 (Equation 48) mby = by1 + by2 (Equation 49)
Is calculated.

Step S59: The image processing section 16 calculates G = (G [i, j-1] + G [i, j + 1] + G [i-1, j] + G [i + 1, j]) / 4 [Equation 50] is calculated to obtain the same color interpolation value G.

Step S60: The image processing section 16 sets the coefficient a used in step S61 to be described later to 1/2.
Set to. Thereafter, the image processing unit 16 proceeds to step S6
The operation is shifted to 1.

Step S61: The image processing section 16 calculates FG [i, j] = G + a. [(Z [i, j] -Z [i, j-2]). Bt2 + (Z [i, j] -Z [i, j + 2]) · bt1] / (2 · mbt) + (1-a) · [(Z [i, j] -Z [i-2, j]) · by2 + (Z [i , j] -Z [i + 2, j]) · by1] / (2 · mby) [Equation 51] is used to determine the final interpolation value of the G color at the interpolation target pixel.

[Effects of Sixth Embodiment] As described above, in the sixth embodiment, the asymmetrical interpolation operation is performed, and at the same time, the contribution ratios other than the G color in the interpolation value determination are changed. Therefore, it is possible to obtain the effects described in claims 1 and 8 synergistically, and to generate a high-quality interpolated image.

The above-described sixth embodiment may be modified as follows. First, the image processing unit 16 compares the vertical and horizontal similarities of the local area including the interpolation target position [i, j], and determines the direction of higher similarity as the interpolation direction. here,
When the vertical direction is set as the interpolation direction, the image processing unit 16 calculates b1 = abs (Z [i, j] -Z [i, j-2]) + 1... [Expression 52] b2 = abs (Z [ i, j] -Z [i, j + 2]) + 1 [Equation 53] mb = b1 + b2 [Equation 54] G1 = (3 * G [i, j-1] + G [i, j + 1]) / 4+ (Z [i, j] -Z [i, j-2]) / 2 [Equation 55] G2 = (3 * G [i, j + 1] + G [i, j-1]) / 4+ (Z [i, j] -Z [i, j + 2]) / 2 ... [Equation 56] FG [i, j] = (b2 * G1 + b1 * G2) / mba [Expression 57] is calculated, and FG [i, j] is used as a G interpolation value.

On the other hand, if the horizontal direction is the interpolation direction, the image processing section 16 calculates b1 = abs (Z [i, j] -Z [i-2, j]) + 1 (Equation 58) b2 = abs (Z [i, j] -Z [i + 2, j]) + 1 [Equation 59] mb = b1 + b2 [Equation 60] G1 = (3 * G [i-1 , j] + G [i + 1, j]) / 4+ (Z [i, j] -Z [i-2, j]) / 2 [Equation 61] G2 = (3 * G [i + 1, j] + G [i-1, j]) / 4+ (Z [i, j] -Z [i + 2, j]) / 2 [Equation 62] FG [i, j] = (b2 * G1 + b1 * G2) / mb (Equation 63) is calculated, and FG [i, j] is used as a G interpolation value. Also in this case, it is possible to obtain the respective effects described in claims 1 and 8 synergistically, and to generate a high-quality interpolation image.

<< Supplementary Items of Embodiment >> In the above-described embodiment, the case where the interpolation processing is executed on the electronic camera 11 has been described, but the present invention is not limited to this. For example, it is of course possible to execute an interpolation processing routine as shown in FIGS. 3, 5, and 7 to 10 on a computer. Further, as described in claim 11, a recording medium on which an image processing program for executing such interpolation processing is recorded is also an embodiment of the present invention. Note that the act of manufacturing such a recording medium includes an act of transferring program data from a remote location via a communication medium such as the Internet and recording the program data in a memory or a hard disk in a partner computer.

Further, in the above-described embodiment, a case has been described where the G-color vacancy position in the Bayer array is interpolated, but the present invention is not limited to this. The present invention is a technique applicable to general interpolation processing for image data including color information of two or more colors. Therefore,
The present invention can be applied to interpolation processing of a color array other than the Bayer array (for example, a complementary color array or a honeycomb array), and can also be applied to interpolation processing other than the G color (for example, color difference interpolation or interpolation in another color system). Applicable. Further, the present invention is not limited to the interpolation processing of the vacant lattice position, and is of course applicable to the interpolation processing when converting the number of pixels of the image data. Further, when there is a process of determining an interpolation value of a specific color by adding two or more colors other than the specific color, the interpolation value of the specific color (final interpolation value) is added by adding one color other than the specific color. The present invention is not limited to this, and the present invention can be applied to a process of determining an intermediate value in a process of deriving an interpolation value.

Although the second embodiment uses the second color information of the pixel to be interpolated, the present invention is not limited to this. For example, the detail of the interpolation image of the first color information may be enhanced by using only the surrounding second color information without directly using the second color information of the interpolation target pixel. In this case, the high frequency component of the spatial frequency does not increase unnecessarily, and it is possible to obtain the two effects of reducing the image noise and reducing the amount of compression code.

In the above-described embodiment, the interpolation processing is performed on the image data after the white balance adjustment. In this case, since the extreme color bias is corrected in advance, the places where the dot noise is likely to occur are reduced in advance. Therefore, by performing the correction processing of the present invention after white balance adjustment, an excellent advantage that dot noise is synergistically reduced and detail reproducibility is synergistically enhanced is obtained. However, the present invention is not limited to this, and a sufficient effect can be obtained even if executed before white balance adjustment.

In the above-described embodiment, the interpolation processing is performed on the image data after the gradation conversion. In this case, since the extreme grayscale bias is corrected in accordance with the visual characteristics, the places where the dot noise is likely to occur are reduced in advance. Therefore, by performing the correction processing of the present invention after gradation conversion, an excellent advantage of synergistically reducing dot noise and synergistically improving detail reproducibility can be obtained. However, the present invention is not limited to this, and it is possible to obtain a sufficient effect even if it is executed before gradation conversion.

In the above-described embodiment, the case where the image processing unit 16 performs the interpolation processing by software processing has been described, but the present invention is not limited to this. Of course, the image processing unit 16 may be realized by a hardware configuration such as an ASIC. Note that, in the above-described embodiment, in order to disclose the invention as much as possible, an explanation is given using an expression that specifically considers the range of the local region. These equations have an excellent advantage that a good interpolation image can be generated for a wide variety of image data in general. However, in practice, the range of the local area is appropriately optimized according to the number of pixels of the image, the size on the display screen, the size at the time of printing, and the like (in a situation where it is not necessary to visually see small parts, Local area may be temporarily enlarged). Therefore, the present invention is not narrowly limited by the equation specifically considering the local region.

In the first and second embodiments, it is determined whether or not a peak or trap exists near the pixel to be interpolated using the above-described peak discrimination value Xr.
This determination method has an excellent advantage that determination can be made in a short time with a small amount of calculation. However, the present invention may be any method for determining whether a peak or trap exists in a local region. For example, a minimum value and a maximum value are obtained for the second color information (or first color information) around the interpolation target pixel, and the second color information (or interpolation) of the interpolation target pixel is determined from the range between the minimum value and the maximum value. It may be determined whether or not a peak or trap exists by determining whether or not the first color information adjacent to the target pixel) protrudes.

In the third embodiment, the G color space 1
The correlation between image structures is determined by comparing the second derivative and the first derivative of the space other than the G color. This determination method has an excellent advantage that the correlation of the image structure can be determined in a short time with a small amount of calculation. However, the present invention may be any method that determines the correlation between image structures. For example, various correlation determination methods such as pattern matching and a method for obtaining a cross-correlation coefficient may be used.

In the fourth embodiment, the color difference is determined. This determination method has an excellent advantage that determination can be made in a short time with a small amount of calculation compared to the determination of saturation. However, the present invention is not limited to this determination method. For example, a method of determining saturation may be used.

Further, in the fourth embodiment, the color difference is determined in a relatively wide local area. This has the advantage of being less susceptible to image noise. However, the present invention is not limited to this.
For example, the contribution ratio of the second information may be changed according to the color difference (or saturation) obtained from the “linear interpolation value of the interpolation target pixel” and the “second color information of the interpolation target pixel”.

[0116]

According to the first aspect of the present invention, an image structure is determined for a local region including an interpolation symmetric pixel, and "the contribution ratio of the second color information in determining the interpolation value" is changed according to the determination result. . By using this configuration, it is possible to flexibly cope with an image structure in which the degree of occurrence of dot noise is high and to reduce the contribution ratio. As a result, it is possible to generate a high-quality interpolation image with little dot noise.

According to the second aspect of the invention, when there is no peak or trap in the local area, the contribution of the second color information is reduced. As a result, the contribution rate is reduced in a place where dot noise is highly likely to be generated and does not significantly contribute to the enhancement of the detail, and the dot noise can be reduced.

According to the third aspect of the invention, when the correlation between the "image structure of the first color information" and the "image structure of the second color information" is small, the contribution ratio of the second color information is reduced. Therefore, the chance that different image structures of the second color information are mixed in the interpolated image is reduced, and the dot noise can be reliably reduced.

According to the fourth aspect of the present invention, the direction of interpolation is controlled using the result of determination of the direction similarity. Therefore,
It is possible to sufficiently maintain the image structure (for example, vertical stripes and horizontal stripes) of the original image.

In the invention according to claim 5, it is determined whether or not the contribution ratio of the second color information is reduced in the interpolation direction. Therefore, it is possible to determine realistically whether to add the second color information in the interpolation direction. Further, it is not necessary to perform determination in all directions, and the amount of calculation in the determination process can be efficiently reduced.

According to the sixth aspect of the invention, when the saturation or the color difference is large, the contribution of the second color information is reduced. Therefore, there is less possibility that an error based on the difference in the color component ratio will be mixed into the interpolated image, and it is possible to reduce dot noise.

According to the seventh aspect of the present invention, since the contribution ratio of the second information is reduced to substantially zero, it is possible to reliably prevent the generation of dot noise.

According to the eighth aspect of the present invention, details of the interpolated image are supplemented by adding the spatial first derivative of the second color information to the interpolated image of the first information. Therefore, unlike the conventional example using the spatial second derivative (Laplacian), it becomes possible to select the peripheral pixels used for the interpolation processing in an asymmetrical manner. Such an asymmetric shape reduces the probability that peripheral pixels cross color boundaries and edges, and reduces the possibility that errors will be mixed in the interpolated image. As a result, it is possible to reduce dot noise and generate a high-quality interpolated image.

According to the ninth aspect of the present invention, an asymmetric interpolation operation is performed in a predetermined direction having a high similarity. Therefore, there is little possibility that an asymmetric interpolation operation is performed across color boundaries or edges, and a high-quality interpolated image with little dot noise can be generated.

According to the tenth aspect, the weight ratio is adjusted according to the similarity determination. With this configuration, it is possible to reduce the weighting ratio in a predetermined direction with low similarity. As a result, it is possible to reduce a possibility that an error is mixed in the interpolated image over a color boundary or an edge, and to generate a high-quality interpolated image with little dot noise.

By using the recording medium according to the eleventh aspect, a computer can be used.
It is possible to function as the image processing device according to any one of the above.

[Brief description of the drawings]

FIG. 1 is a diagram showing a schematic configuration of an electronic camera 11;

FIG. 2 is a diagram illustrating an outline of a color interpolation process performed by an image processing unit 16;

FIG. 3 is a diagram illustrating a G interpolation processing routine according to the first embodiment.

FIG. 4 is an enlarged screen photograph (a binarized image obtained by dithering a halftone image on a display) for comparatively explaining the effect of the first embodiment.

FIG. 5 is a diagram illustrating a G interpolation processing routine according to the second embodiment.

FIG. 6 is a diagram illustrating an example of a membership function b (Xr).

FIG. 7 is a diagram illustrating a G interpolation processing routine according to a third embodiment.

FIG. 8 is a diagram illustrating a G interpolation processing routine according to a fourth embodiment.

FIG. 9 is a diagram illustrating a G interpolation processing routine according to a fifth embodiment.

FIG. 10 is a diagram illustrating a G interpolation processing routine according to a sixth embodiment.

[Explanation of symbols]

 Reference Signs List 11 electronic camera 12 photographing lens 13 image sensor 14 signal processing unit 15 A / D conversion unit 16 image processing unit

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5B057 BA02 CA01 CB01 CE02 CE14 CE17 5C077 MP07 MP08 NN08 PP02 PP32 PP34 PP37 PP47 PP68 PQ08 PQ18 RR19 TT09 5C079 HB01 HB06 LA02 LA28 MA11 NA02

Claims (11)

[Claims] An image processing apparatus for performing an interpolation process of first color information on image data including at least first color information and second color information, the first color information and / or the second color information around an interpolation target pixel. Interpolating means for determining an interpolation value of the first color information according to the color information; determining an image structure for a local region including the interpolation target pixel; determining an interpolation value of the interpolation target pixel according to the determination result And a structure determining unit for changing a contribution ratio of the second color information. 2. The image processing apparatus according to claim 1, wherein the structure determining unit determines the interpolation value of the interpolation target pixel when a peak and a trap do not exist in a local area including the interpolation target pixel. An image processing apparatus, characterized in that it is means for reducing the contribution ratio of the second color information. 3. The image processing apparatus according to claim 1, wherein the structure determining unit determines an image structure of a local region including the interpolation target pixel, and determines an image structure of the first color information and a second color information. If it is determined that the correlation with the image structure is small,
An image processing apparatus, comprising: means for reducing a contribution rate of the second color information in determining an interpolation value of the interpolation target pixel. 4. The image processing apparatus according to claim 1, further comprising: a direction determination unit configured to determine a similarity for each direction in a local region including the interpolation target pixel; The image processing apparatus calculates the interpolation value by increasing the contribution ratio of the color information positioned in the “high similarity direction” determined by the direction determination unit. 5. The image processing apparatus according to claim 4, wherein the structure determination unit determines an image structure with respect to the “high similarity direction” determined by the direction determination unit. apparatus. 6. An image processing apparatus for performing an interpolation process of first color information on image data including at least first color information and second color information, wherein the first color information and / or the second color information around an interpolation target pixel are provided. An interpolation means for determining an interpolation value of the first color information according to the color information; and determining a saturation or a color difference for a local area including the interpolation target pixel, and determining that the saturation or the color difference is large.
An image processing apparatus comprising: a color determination unit configured to reduce a contribution ratio of the second color information in determining an interpolation value of the interpolation target pixel. 7. The image processing apparatus according to claim 1, wherein the contribution rate of the second color information is set to substantially zero when the second color information is reduced. Processing equipment. 8. An image processing apparatus for performing an interpolation process on first color information for image data including at least first color information and second color information, the image data passing through an interpolation target pixel (pixel position of Y (0) below). A pixel array in which a first color information X (-1), a second color information Y (0), a first color information X (+1), and a second color information Y (+2) are arranged in a predetermined direction on the image. , The first color information X (0) of the pixel to be interpolated is expressed as X (0) = (1-a1). -a2 · Y (+2) (where a1 and a2 are predetermined proportionality coefficients). 9. The image processing apparatus according to claim 8, further comprising: a direction determining unit that determines similarity in a plurality of predetermined directions starting from the interpolation target pixel in a local region including the interpolation target pixel, The asymmetric interpolation means selects a predetermined direction determined to have a high similarity by the direction determination means, and selects the first color information.
An image processing apparatus for calculating X (0). 10. The image processing apparatus according to claim 8, further comprising: a direction determining unit that determines similarity in a plurality of predetermined directions starting from the interpolation target pixel in a local region including the interpolation target pixel, The image processing, wherein the asymmetric interpolation means weights and synthesizes the first color information X (0) calculated for each of the plurality of predetermined directions at a weight ratio according to the similarity determined by the direction determination means. apparatus. 11. A machine-readable recording medium in which an image processing program for causing a computer to function as the image processing apparatus according to claim 1 is recorded.