TWI477155B - Color interpolation method and image demosaicking method using the same and computer program production thereof - Google Patents

Description

Color interpolation method and image demosaicing method using the color interpolation method and computer program product thereof

The present invention relates to a color interpolation method and an image demosaicing method using the color interpolation method, in particular, a color interpolation method applied to an image generated by a color filter and an image using the color interpolation method. Demosaicing method.

In recent years, Digital Still Cameras (DSCs) have become very popular consumer electronics products. A digital camera generally converts the sensed incident light into an electrical signal by means of a charge coupled device (CCD) or a complementary metal oxide semiconductor (CMOS) sensor, and then these analog signals are converted by analog/digital conversion. Become a digital image signal.

Please refer to FIG. 1 , which is a schematic structural diagram of a conventional color filter array (CFA). In order to save the size of the sampled color pixels and the sensor, a color filter array is generally used as the color sampling format, so a color filter array is usually plated on the CCD or CMOS sensor. A commonly used color filter array is a CFA published by Bryce E. Bayer, as shown in Fig. 1, wherein the FR system is a red filter, the FG system is a green filter, and the FB system is Blue filter. When using CFA, each sampling point has only one color, that is, one-third of the full-color system (ie, three colors of red, green, and blue). Since each filter in the CFA corresponds to a pixel and only allows a specific color component to pass, the color component lacking in each pixel must be reconstructed before the captured image is further processed or displayed. Each pixel contains all three color components.

In order to reconstruct a complete color image from the original sensing data, color interpolation is usually used to estimate the two color components that each pixel lacks. The so-called color interpolation method is based on the information provided by a part of the pixels of each target pixel, that is, when a CFA is known, the two color components lacking each pixel can be Reconstructed using information provided by adjacent pixels. The reconstruction of the overall image is a well-known demosaicking operation in the industry.

The current decoding demozaicking technology can only estimate and reconstruct the vertical and horizontal edges of the Bayer CFA color filter array (Bayer CFA). However, the other two may exist in the electronic image. There is no effective reconstruction method for the ° and 135° sides.

Accordingly, one aspect of the present invention is to provide a color interpolation method. This color interpolation method can be interpolated along the 135° side of the target pixel. The color interpolation method is applicable to a pixel array, the pixel array includes a plurality of pixels, the pixels include a plurality of first pixels and a plurality of second pixels, each of the first pixels having a corresponding one to a first color An intensity value ( G _i,j ), each second pixel having a second intensity value ( R _i,j ) corresponding to one of the second colors, wherein j represents the abscissa of each pixel, and i represents each pixel The ordinate. In this color interpolation method, an interpolation step is first performed for each second pixel to calculate an estimated intensity value ( P _i,j ) corresponding to one of the first colors for each second pixel. The color interpolation method performs an interpolation step for each second pixel to calculate an estimated intensity value ( P _i,j ) corresponding to one of the first colors for each second pixel. In this interpolation step, first, a high pass mask is provided according to the pixel array to detect the position of the second pixel (target pixel) relative to an edge of the pixel array, and provide detection Test result (f). Then, a first oblique intensity value calculation step is performed to calculate the first oblique intensity value ( ), wherein the first oblique intensity value calculation step is based on a formula: . Next, a second oblique intensity value calculation step is performed to calculate a second oblique intensity value ( ), wherein the second oblique intensity value calculation step is based on a formula: . Next, an estimation parameter calculation step is performed to calculate an estimation parameter ( ), wherein the estimation parameter calculation step is based on a formula:

Then, an estimated intensity value calculation step is performed to calculate the estimated intensity value ( P _i,j ), wherein the estimated intensity value calculation step is based on a formula:

Another aspect of the present invention is to provide a color interpolation method. This color interpolation method can be interpolated along the 45° side of the target pixel. The color interpolation method is applicable to a pixel array, the pixel array includes a plurality of pixels, the pixels include a plurality of first pixels and a plurality of second pixels, each of the first pixels having a corresponding one to a first color An intensity value ( G _i,j ), each second pixel having a second intensity value ( R _i,j ) corresponding to one of the second colors, wherein j represents the abscissa of each pixel, and i represents each pixel The ordinate. The color interpolation method performs an interpolation step for each second pixel to calculate an estimated intensity value ( P _{i , j} ) corresponding to one of the first colors of each second pixel. In this interpolation step, a high pass mask is first provided according to the pixel array to detect the position of the second pixel with respect to an edge of the pixel array, and the detection result (f) is provided. Then, performing a first oblique intensity value calculation step to calculate a first oblique intensity value ( ), wherein the first oblique intensity value calculation step is based on a formula: . Next, a second oblique intensity value calculation step is performed to calculate a second oblique intensity value ( ), wherein the second oblique intensity value calculation step is based on a formula: . Then, an estimation parameter calculation step is performed to calculate an estimation parameter ( ), wherein the estimation parameter calculation step is based on a formula: Next, an estimated intensity value calculation step is performed to calculate an estimated intensity value ( P _{i , j} ), wherein the estimated intensity value calculation step is based on a formula:

Yet another aspect of the present invention is to provide an image demosaicing method for processing a pixel array corresponding to a color filter array to reconstruct a color component lacking in the pixel array. In the image demosaicing method, the pixel array is first provided, the pixel array includes a plurality of first pixels and second pixels, wherein each of the first pixels has a first intensity value corresponding to one of the first colors ( G _{i , j} ), and each second pixel has a second intensity value ( R _{i , j} ) corresponding to one of the second colors. The second pixels comprise a plurality of estimated pixels and unestimated pixels, wherein the unestimated pixels comprise target pixels, and the remaining of the unestimated pixels and the estimated pixels are centered around the target pixels. Then, an intensity value estimation step is performed to calculate that the target pixel corresponds to one of the first colors and the second estimated intensity value. In this intensity value estimation step, an interpolation step is first performed to calculate a plurality of candidate estimated intensity values of the target pixel. Then, a decision step is performed to determine a second estimated intensity value of the target pixel based on the candidate estimated intensity value and the first estimated intensity value.

For convenience of explanation, in the following description, the "intensity value" will be used to represent the brightness of a certain color in the pixel, for example, the green intensity value represents the green brightness value of the pixel.

Referring to FIG. 2 and FIG. 3 simultaneously, FIG. 2 is a schematic structural diagram of a pixel array 100 according to an embodiment of the invention, and FIG. 3 is a schematic structural diagram of a processing region 102 according to an embodiment of the invention. . The pixel array 100 includes a plurality of first pixels R, second pixels G, and third pixels B, and the pixels may respectively correspond to different colors according to the structure of the color filter array. Taking the Bell filter array as an example, the first pixel R corresponds to red, the second pixel G corresponds to green, and the third pixel B corresponds to blue. In addition, in order to perform color interpolation, the user usually determines the size of the processing area 102 by using a decision window W, and then uses the pixels in the processing area 102 for interpolation. In the present embodiment, the size of the window is determined to be 5 by 5 pixels, but the present invention is not limited thereto.

A schematic diagram of the flow of the interpolation method 200. The color interpolation method 200 is used to estimate color components that are lacking in a target pixel of one of the processing regions 102. In general, since the human eye is sensitive to green, in the following description, the green intensity value p _{i in} the first pixel R (hereinafter referred to as the target pixel) whose coordinates are (i, j) is estimated _{, j is} an example to illustrate the color interpolation method 100.

In the color interpolation method 200, a mask providing step 210 is first performed to provide a preset high pass mask, such as a Laplacian mask or a Gaussian Laplacian mask. (Laplacian of Gaussian Mask), to detect the position of the target pixel relative to the edge of the pixel array, and output a detection result f, wherein the edge refers to a place where the gray scale changes greatly in the pixel array. The size of the mask is determined according to the decision window. Since the decision window used in this embodiment is 5 by 5 pixels, the size of the mask is also 5 by 5 pixels. Next, an oblique intensity value calculation step 220 is performed. In the first oblique intensity value calculation step 220, a formula is used to calculate the oblique intensity value This formula is expressed as follows:

among them The system represents the oblique intensity value estimated from the north to the west centered on the target pixel, and Gi _-1,j represents the green intensity value of the second pixel G with coordinates (i-1,j), G _{i,j- The 1} series represents the green intensity value of the second pixel G whose coordinates are (i, j-1). Then, a second oblique intensity value calculation step 230 is performed. In the second oblique intensity value calculation step 230, a formula is used to calculate the oblique intensity value This formula is expressed as follows:

among them, The system represents the oblique intensity value estimated from the east to the south centered on the target pixel, and G _i+1,j represents the green intensity value of the second pixel G with coordinates (i+1,j), G _{i,j The +1} line represents the green intensity value of the second pixel G whose coordinates are (i, j+1). Next, an estimated parameter calculation step 240 is performed to calculate the estimated parameters . In the filter parameter calculation step 240, the obtained result f is detected by a high-pass mask and the estimated parameter L _i,j is calculated according to a formula, which is expressed as follows:

It can be seen from the above equation that when the result f is greater than or equal to zero, the estimated parameter L _i,j is with The largest of them; when the result f is less than zero, the estimated parameter L _i,j is with The smallest of them. Then, an intensity value calculation step 250 is performed. In the estimated intensity value calculation step 250, the estimated intensity value P _i,j is calculated according to a formula, which is expressed as follows:

Where R _i,j represents the red intensity value of the target pixel, and R _i-2,j+2 represents the red intensity value of the first pixel R with coordinates (i-2,j+2), R _{i+2,j -2} represents the red intensity value of the first pixel R whose coordinates are (i+2, j-2).

The color interpolation method 200 of the present embodiment can estimate the missing color along the 135° side of the target pixel. For pixel arrays having edges of 135[deg.] sides, the color interpolation method 200 can improve the accuracy of the estimation.

Please refer to FIG. 5, which is a flow chart of a color interpolation method 300 according to an embodiment of the invention. The color interpolation method 300 of the present embodiment is similar to the color interpolation method 200, but differs in that the color interpolation method 300 is used to estimate the missing color along the 45° side of the target pixel.

In the color interpolation method 300, a mask providing step 310 is first performed to provide a preset high pass mask to detect the position of the target pixel relative to the edge of the pixel array and to provide a detection result f. Next, a diagonal intensity value calculation step 320 is performed. In the oblique intensity value calculation step 320, a formula is used to calculate the oblique intensity value This formula is expressed as follows:

among them The system represents the oblique intensity value estimated from north to east centering on the target pixel, and G _i-1, j represents the green intensity value of the second pixel G with coordinates (i-1, j), G _{i,j The +1} line represents the green intensity value of the second pixel G whose coordinates are (i, j+1). Then, an oblique intensity value calculation step 330 is performed. In the oblique intensity value calculation step 330, a formula is used to calculate the oblique intensity value This formula is expressed as follows:

among them, The system represents the oblique intensity value estimated from the west to the south centered on the target pixel, and G _i+1,j represents the green intensity value of the second pixel G with coordinates (i+1,j), G _{i,j The -1} series represents the green intensity value of the second pixel G of coordinates (i, j-1). Next, an estimated parameter calculation step 340 is performed to calculate the estimated parameters. . In the estimated parameter calculation step 340, the obtained result f is detected by the high-pass mask and the estimated parameter is calculated according to a formula. This formula is expressed as follows:

It can be seen from the above equation that when the result f is greater than or equal to zero, the estimated parameter L _i,j is with The largest of them; when the result f is less than zero, the estimated parameter L _i,j is with The smallest of them. Then, an estimated intensity value calculation step 350 is performed. In the estimated intensity value calculation step 350, the estimated intensity value P _i,j is calculated according to a formula, which is expressed as follows:

Where R _i,j represents the red intensity value of the target pixel, and R _i-2,j-2 represents the red intensity value of the first pixel R with coordinates (i-2,j-2), R _{i+2,j +2} represents the red intensity value of the first pixel R whose coordinates are (i+2, j+2).

The color interpolation method 300 of the present embodiment can estimate the missing color along the 45° side of the target pixel, which can increase the accuracy of image reconstruction.

Please refer to FIG. 6 and FIG. 7 simultaneously. FIG. 6 is a schematic structural diagram of a pixel array 400 according to an embodiment of the present invention, and FIG. 7 is a schematic structural diagram of a processing region 402 according to an embodiment of the invention. . The pixel array 400 is similar to the pixel array 100, and the first pixel R includes the estimated pixels 111-116 and the unestimated pixels 117-125, wherein the estimated color values of the pixels 111-116 corresponding to the green color have been The official estimate shows that the unmeasured pixels 117~125 corresponding to the green color intensity values have not been formally estimated. Further, in order to perform the image demosaicing method, the user usually determines the size of the processing area 402 by using the determination window W, and then performs de-mosaic using the pixels in the processing area 402. In the present embodiment, the size of the window is determined to be 5 by 5 pixels, but the present invention is not limited thereto.

Please refer to FIG. 8 , which is a flow chart of an image demosaicing method 500 according to an embodiment of the invention. The demosaicing method 500 is based on a soft decision structure, and the estimated pixel is used to estimate the color intensity value lost by the target pixel, so that the reconstructed image obtained can have higher precision. In the following description, the demosaicing method 500 will be described by taking the green component missing from the unestimated pixel 117 whose coordinates are (i, j) in the estimation processing region 402 as an example, in which the pixels 111 to 116 have been estimated. The green intensity value has been estimated, and the green intensity values of the unestimated pixels 117-125 have not been estimated.

In the demosaicing method 500, an interpolation step 510 is first performed to calculate a plurality of candidate estimated intensity values for pixels 117-125 that correspond to green. The interpolation step 510 can generate a plurality of candidate estimated intensity values for the pixels 117-125 using conventional horizontal interpolation, vertical interpolation, and 45° edge interpolation and 135° edge interpolation of the present invention. These candidate estimated intensity values may be the interpolated intensity values calculated by the interpolation method, or the intensity values obtained by further processing the interpolated intensity values by some calculation method (for example, the average calculation method). Next, a decision step 520 is performed to calculate the green estimated intensity value of the pixel 117 based on the candidate estimated intensity values of the pixels 117-125 and the green intensity values of the estimated pixels 111, 112, 113, and 116.

In decision step 520, a color difference calculation step 520a is first performed to calculate a plurality of color difference values within the window W region centered on the pixel 117. with , wherein the color difference calculation step 520a is calculated using the following formula:

Where i, j represents the coordinate value of the pixel, and R represents the red intensity value of the pixel. Represents the estimated green intensity value of the pixel, Representing the level of the pixel to estimate the green intensity value, Represents the vertical estimated green intensity value of the pixel, representing the estimated green intensity value of the pixel at 45° , The estimated green intensity value of the pixel on the 135° side. Next, an indicator calculation step 520b is performed to calculate an indicator set Ω _i,j ={Δ H _i,j ,Δ V _i,j ,Δ45° _i,j ,Δ135° _{i , j} }, wherein the indicator factor calculates the difference in color difference variation along the vertical, horizontal, 45°, and 135° directions in the window W region, and the present embodiment is calculated by the following formula, but the present invention is not limited thereto.

Then, a selection step 520c is performed to select the estimated green intensity value of the pixel 117 from the candidate estimated intensity values according to the indicator factor group. The selection step 520c selects the estimated intensity value corresponding to the minimum direction of the color difference change difference, and the present embodiment is calculated by the following formula, but the invention is not limited thereto.

As can be seen from the above description, in the present embodiment, the pixels 111, 112, 113, 116 and the un-estimated pixels 118, 121, 122, and 123 are surrounded by the target pixel 117, and thus can be Formally estimating the estimated intensity values of the pixels 111, 112, 113, 116 and the candidate intensity values of the un-formally estimated pixels 118, 121, 122, 123 corresponding to the plurality of interpolation methods ( The one of the candidate intensity values is determined as the estimated green intensity value of the target pixel 117, or the candidate green estimated intensity value is further calculated (eg, the average calculation method) to calculate the estimated target pixel 117. Measure the green intensity value.

In addition, when the green intensity value of the target pixel 117 is estimated, the decision window W determines the next processing area of 5 by 5 pixels, for example, the processing area centered on the unestimated pixel 118. At this time, since the green intensity value of the target pixel 117 has been calculated, the green estimated intensity value of the unrecognized pixel 118 may be based on the officially estimated pixels 112, 113, 114, 117 and the un-estimated pixel 119. , 122, 123, 124 to calculate.

Please refer to FIG. 9 , FIG. 10 and FIG. 11 , FIG. 9 is a schematic structural diagram of a pixel array 400 according to an embodiment of the invention, and FIG. 10 is a diagram showing a processing area according to an embodiment of the invention. FIG. 11 is a schematic flow chart of an image demosaicing method 700 according to an embodiment of the invention. In the present embodiment, the image demosaicing method 700 uses the pixel 117 whose green intensity value has been officially estimated to estimate the green intensity value for the next un-estimated pixel 118. In the image demosaicing method 700, the demosaicing method 500 is first performed to obtain the green component lost by the target estimation pixel 117. When the green intensity value of the target pixel 117 is officially estimated, the decision window W moves to the processing region 602 centered on the un-estimated pixel 118. Next, an interpolation step 710 is performed to calculate a plurality of candidate estimated intensity values corresponding to green that are not formally estimated pixels 118, 119, 122, 123, 124. Interpolation step 710 can generate a plurality of pixels 118, 119, 122, 123, 124 using conventional horizontal interpolation, vertical interpolation, and 45° degree edge interpolation and 135° edge interpolation of the present invention. Candidate estimated intensity value. These candidate estimated intensity values may be the interpolated intensity values calculated by the interpolation method, or the intensity values obtained by further processing the interpolated intensity values by some calculation method (for example, the average calculation method). Next, a decision step 720 is performed to calculate the green estimated intensity value of the pixel 118 based on the candidate estimated intensity value and the green intensity values of the officially estimated pixels 112, 113, 114, and 117.

In decision step 720, a color difference calculation step 720a is first performed to calculate a plurality of color difference values within the window W region centered on the pixel 118. , wherein the color difference calculation step 720a is calculated using the following formula:

Where i, j represents the coordinate value of the pixel, and R represents the red intensity value of the pixel. Represents the estimated green intensity value of the pixel, Representing the level of the pixel to estimate the green intensity value, Represents the vertical estimated green intensity value of the pixel, representing the estimated green intensity value of the pixel at 45° , The estimated green intensity value of the pixel on the 135° side. Next, an indicator calculation step 720b is performed to calculate an indicator set Ω _i,j ={Δ H _i,j ,Δ V _i,j ,Δ45° _i,j ,Δ135° _{i , j} }, wherein the indicator factor calculates the difference in color difference variation along the vertical, horizontal, 45°, and 135° directions in the window W region, and the present embodiment is calculated by the following formula, but the present invention is not limited thereto.

Then, a selection step 720c is performed to select the estimated green intensity value of the pixel 118 from the candidate estimated intensity values according to the indicator factor group. The selecting step 720c selects the estimated difference value of the smallest difference direction of the color difference change degree, and the present embodiment is calculated by the following formula, but the present invention is not limited thereto:

In addition, it should be noted that the color interpolation method 200, 300 and the image demosaicing method 500, 700 of the present invention can be applied to a computer program product, and when the computer program product is loaded on a computer, the computer can perform color. Interpolation methods 200, 300 and image demosaicing method 500.

Although the present invention has been disclosed in the above embodiments, it is not intended to limit the present invention, and the present invention can be modified and retouched without departing from the spirit and scope of the present invention. The scope is subject to the definition of the scope of the patent application attached.

100. . . Pixel array

102. . . Processing area

111~116. . . Estimated pixel

117~125. . . Unestimated pixel

200. . . Color interpolation method

210. . . Mask providing step

220. . . Oblique intensity value calculation step

230. . . Oblique intensity value calculation step

240. . . Estimated parameter calculation step

250. . . Estimated intensity value calculation step

300. . . Color interpolation method

310. . . Mask providing step

320. . . Oblique intensity value calculation step

330. . . Oblique intensity value calculation step

340. . . Estimated parameter calculation step

350. . . Estimated intensity value calculation step

400. . . Pixel array

402. . . Processing area

500. . . Mosaic method

510. . . Interpolation step

520. . . Decision step

520a. . . Color difference calculation step

520b. . . Indicator calculation step

520c. . . Selection step

602. . . Processing area

700. . . Image demosaicing method

710. . . Interpolation step

720. . . Decision step

720a. . . Color difference calculation step

720b. . . Indicator calculation step

720c. . . Selection step

FG. . . Green filter

FR. . . Red filter

G. . . Second pixel

FB. . . Blue filter

W. . . Decision window

R. . . First pixel

i. . . Coordinate value

B. . . Third pixel

j. . . Coordinate value

The above and other objects, features, and advantages of the present invention will become more apparent and understood.

FIG. 1 is a schematic view showing the structure of a conventional color filter array (CFA).

2 is a schematic structural view of a pixel array according to an embodiment of the invention.

3 is a schematic structural view of a processing region according to an embodiment of the present invention.

4 is a flow chart showing a color interpolation method according to an embodiment of the present invention.

FIG. 5 is a flow chart showing a color interpolation method according to an embodiment of the present invention.

FIG. 6 is a schematic structural view of a pixel array according to an embodiment of the invention.

Figure 7 is a block diagram showing the structure of a processing area according to an embodiment of the present invention.

FIG. 8 is a flow chart showing an image demosaicing method according to an embodiment of the invention.

FIG. 9 is a schematic structural view of a pixel array according to an embodiment of the invention.

Figure 10 is a block diagram showing the structure of a processing area according to an embodiment of the present invention.

11 is a flow chart showing an image demosaicing method according to an embodiment of the invention.

500. . . Mosaic method

510. . . Interpolation step

520. . . Decision step

520a. . . Color difference calculation step

520b. . . Indicator calculation step

520c. . . Selection step

Claims (14)

A color interpolation method, applicable to a pixel array, the pixel array comprising a plurality of pixels, the pixels comprising a plurality of first pixels and a plurality of second pixels, each of the first pixels having a corresponding first color a first intensity value ( G _{i , j} ), each of the second pixels having a second intensity value ( R _{i , j} ) corresponding to one of the second colors, wherein j represents a horizontal of each of the pixels a coordinate, i represents the ordinate of each of the pixels, and the color interpolation method comprises: performing an interpolation step on each of the second pixels by using a processor of a computer to calculate each of the The two pixels correspond to an estimated intensity value ( P _{i , j} ) of the first color, wherein the interpolating step comprises: providing a high pass mask according to the pixel array to detect the second pixel a position of the pixel relative to an edge of the pixel array, and outputting a detection result (f); performing a first oblique intensity value calculation step to calculate a first oblique intensity value ( ), wherein the first oblique intensity value calculation step is based on a formula: Performing a second oblique intensity value calculation step to calculate a second oblique intensity value ( ), wherein the second oblique intensity value calculation step is based on a formula: Perform an estimation parameter calculation step to calculate an estimation parameter ( ), wherein the estimation parameter calculation step is based on a formula: Performing an estimated intensity value calculation step to calculate the estimated intensity value ( P _{i , j} ), wherein the estimated intensity value calculation step is based on a formula: And using the processor of the computer to reconstruct the color information of each of the second pixels according to the estimated intensity values corresponding to the first color of each of the second pixels. The color interpolation method of claim 1, wherein the high-pass mask is a Lapalacian Mask or a Laplacian of Gaussian Mask. The color interpolation method of claim 1, wherein the first color is green and the second color is red or blue. A computer program product, after loading and executing the program via a computer, the computer can perform the color interpolation method as described in claim 1 of the patent application. A color interpolation method is applicable to a pixel array, wherein the pixel array includes a plurality of pixels, the pixels include a plurality of first pixels and a plurality of second pixels, each of the first pixels having a corresponding first One of the first color intensity values ( G _{i , j} ), each of the second pixels having a second color intensity value ( R _{i , j} ) corresponding to one of the second colors, wherein j represents each of the The horizontal coordinate of the pixel, i represents the ordinate of each of the pixels, and the color interpolation method comprises: performing an interpolation step on each of the second pixels by using a processor of a computer to calculate each The second pixels correspond to one of the first color estimated intensity values ( P _{i , j} ), wherein the interpolating step comprises: providing a high pass mask according to the pixel array to detect Measure a position of the second pixel relative to an edge of the pixel array, and output a detection result (f); perform a first oblique intensity value calculation step to calculate a first oblique intensity value ( ), wherein the first oblique intensity value calculation step is based on a formula: Performing a second oblique intensity value calculation step to calculate a second oblique intensity value ( ), wherein the second oblique intensity value calculation step is based on a formula: Perform an estimation parameter calculation step to calculate an estimation parameter ( ), wherein the estimation parameter calculation step is based on a formula: Performing an estimated intensity value calculation step to calculate the estimated intensity value ( P _{i , j} ), wherein the estimated intensity value calculation step is based on a formula: And using the processor of the computer to reconstruct the color information of each of the second pixels according to the estimated intensity values corresponding to the first color of each of the second pixels. The color interpolation method of claim 5, wherein the high pass mask is a Laplacian mask or a Gaussian Laplacian mask. The color interpolation method of claim 5, wherein the first color is green and the second color is red or blue. A computer program product, after loading and executing the program via a computer, the computer can perform the color interpolation method as described in claim 5 of the patent application. An image demosaicing method for processing a pixel array corresponding to a color filter array to reconstruct a color component lacking the pixel array, the image demosaicing method comprising: providing the pixel array, wherein the pixel array comprises: a plurality First pixels, wherein each of the first pixels has a first intensity value ( G _{i , j} ) corresponding to one of the first colors; and a plurality of second pixels, wherein each of the second pixels has a corresponding a second intensity value ( R _{i , j} ) to a second color, the second pixels comprising: a plurality of estimated pixels, wherein the estimated pixels correspond to a plurality of first estimated intensity values The first estimated intensity values correspond to the first color; and the plurality of unestimated pixels, wherein the unestimated pixels comprise a target pixel, and the rest of the unestimated pixels and the Estimating the pixel is centered around the target pixel; using a processor of a computer to perform an intensity value estimation step to calculate a second estimated intensity value corresponding to the first color of the target pixel, The strength The value estimation step includes: performing an interpolation step to calculate a plurality of candidate estimated intensity values of the target pixel, wherein the interpolating step comprises: providing a high-pass mask according to the pixel array to detect the target Positioning the pixel at an edge of the pixel array and outputting a detection result (f); performing a first oblique intensity value calculation step to calculate a first oblique intensity value ( ), wherein the first oblique intensity value calculation step is based on a formula: Performing a second oblique intensity value calculation step to calculate a second oblique intensity value ( ), wherein the second oblique intensity value calculation step is based on a formula: Performing an estimation parameter calculation step to calculate an estimation parameter L _{i , j} , wherein the estimation parameter calculation step is based on a formula: And performing a candidate estimated intensity value calculation step to calculate a value of one of the candidate estimated intensity values ( P _{i , j} ), wherein the candidate estimated intensity value calculation step is based on a formula: And performing a determining step of determining the second estimated intensity value according to the candidate estimated intensity values and the first estimated intensity values, wherein the determining step comprises: performing a color difference calculating step to calculate Taking the target pixel as the center, a plurality of color difference values in a window region; performing an indicator calculation step to calculate a plurality of color difference variations in the window region along the vertical, horizontal, 45°, and 135° directions And performing a selection step of selecting the second estimated intensity value according to the difference in the color difference changes, wherein the second estimated intensity value is a direction corresponding to a minimum of the differences in the color difference changes Estimating the intensity value; and utilizing the processor of the computer to reconstruct the color information of the target pixel according to the estimated intensity value corresponding to the first color of the target pixel. An image demosaicing method for processing a pixel array corresponding to a color filter array to reconstruct a color component lacking the pixel array, the image demosaicing method comprising: providing the pixel array, wherein the pixel array comprises: a plurality First pixels, wherein each of the first pixels has a first intensity value ( G _{i , j} ) corresponding to one of the first colors; and a plurality of second pixels, wherein each of the second pixels has a corresponding a second intensity value ( R _{i , j} ) to a second color, the second pixels comprising: a plurality of estimated pixels, wherein the estimated pixels correspond to a plurality of first estimated intensity values The first estimated intensity values correspond to the first color; and the plurality of unestimated pixels, wherein the unestimated pixels comprise a target pixel, and the rest of the unestimated pixels and the Estimating the pixel is centered around the target pixel; using a processor of a computer to perform an intensity value estimation step to calculate a second estimated intensity value corresponding to the first color of the target pixel, The strength The value estimation step includes: performing an interpolation step to calculate a plurality of candidate estimated intensity values of the target pixel, wherein the interpolating step comprises: providing a high-pass mask according to the pixel array to detect the target Positioning the pixel at an edge of the pixel array and outputting a detection result (f); performing a first oblique intensity value calculation step to calculate a first oblique intensity value ( ), wherein the first oblique intensity value calculation step is based on a formula: Performing a second oblique intensity value calculation step to calculate a second oblique intensity value ( ), wherein the second oblique intensity value calculation step is based on a formula: Performing an estimation parameter calculation step to calculate an estimation parameter L _{i , j} , wherein the estimation parameter calculation step is based on a formula: And performing a candidate estimated intensity value calculation step to calculate a value of one of the candidate estimated intensity values ( P _{i , j} ), wherein the candidate estimated intensity value calculation step is based on a formula: And performing a determining step of determining the second estimated intensity value according to the candidate estimated intensity values and the first estimated intensity values, wherein the determining step comprises: performing a color difference calculating step to calculate Taking the target pixel as the center, a plurality of color difference values in a window region; performing an indicator calculation step to calculate a plurality of color difference variations in the window region along the vertical, horizontal, 45°, and 135° directions And performing a selection step of selecting the second estimated intensity value according to the difference in the color difference changes, wherein the second estimated intensity value is a direction corresponding to a minimum of the differences in the color difference changes Estimating the intensity value; and utilizing the processor of the computer to reconstruct the color information of the target pixel according to the estimated intensity value corresponding to the first color of the target pixel. The image demosaicing method according to claim 9 or claim 10, wherein the determining step selects one of the candidate estimated intensity values as the second estimated intensity value. The image demosaicing method of claim 9 or 10, wherein the high pass mask is a Laplacian mask or a Gaussian Laplacian mask. The image demosaicing method of claim 9 or 10, wherein the first color is green and the second color is red or blue. A computer program product, after loading and executing the program via a computer, the computer can perform an image demosaicing method as described in claim 9 or 10.