TWI455063B - Color interpolation device and method thereof - Google Patents

Description

Color interpolation device and method thereof

The present invention relates to a color interpolation device and method thereof, and more particularly to a color interpolation device and method for improving color interpolation accuracy by using an arrangement manner in which a specific pixel arrangement is regarded as a 45 degree or 135 degree interleaving manner .

In recent years, Digital Still Cameras have become a very popular consumer electronics product replacing traditional negative film cameras. Digital cameras use light-sensitive charge-coupled devices or complementary metal oxide semiconductors. (Complementary Metal-Oxide-Semiconductor) sensors replace the chemical sensitization of traditional camera negatives.

Among them, the color expression of the traditional negative film camera is through layer-by-layer color sensing of silver salt particles sensitive to different color wavelengths, so that white sunlight can automatically perform color separation on the film layer.

However, digital cameras are limited by technology and cost, and only one layer of photosensitive substrate. In order to simulate the color separation of traditional negative film cameras, the Color Filter Array comes into being. The principle is to divide the color filters into One-quarter of the red and blue, and one-half of the green, and then through the digital image processing engine for color interpolation (Interpolation) to unrecorded three-quarters of red and blue, and One-half of the green color is added to the image file by means of data calculation to form a complete three-primary color file, which is then combined into a digital photo.

For example, the Republic of China Patent Publication No. I274908 discloses a method of image color interpolation of a color filter array. The interpolation method of the present invention utilizes a pre-estimated minimum square error (MSE) to interpolate the color filter array after capturing an image by a color filter array (CFA). Color image. In order to estimate the color pixel values lost by the color filter array, weights (eights) adjacent to the color pixel values to be used are first calculated, and the weight values are obtained by matrix operations. In the course of the operation, a color model is also used to make the restored color closer to the human visual response.

However, the color interpolation method described above can only estimate and reconstruct the vertical and horizontal edge directions for the color filter array, but the other two may also exist in the electronic image at 45° and 135°. In the edge direction, there is no effective reconstruction method, and the image boundary of the image is easy to produce distortion such as zipper shape, color overlap, and complementary color error.

Therefore, how to design a color interpolation device that can estimate and reconstruct the color filter array at the 45° and 135° edge directions, thereby effectively avoiding distortion such as zipper shape, color overlap and complementary color error at the image boundary of the image. And its methods have become the goal of the joint efforts of relevant manufacturers and related research and development personnel.

The present inventors have in view of the fact that the conventional color interpolation apparatus and method thereof cannot estimate and reconstruct the color filter array in the 45° and 135° edge directions, and it is easy to produce zipper shape, color overlap, and complementary color error at the image boundary of the image. The lack of equal distortion is an active development, in order to improve the above-mentioned shortcomings, and through continuous experimentation and efforts, finally developed the present invention.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a color interpolation that can estimate and reconstruct a color filter array at 45° and 135° edge directions while avoiding distortion such as zipper, color overlap, and complementary color errors at the image boundary of the image. Apparatus and method therefor.

In order to achieve the above object, the color interpolation device of the present invention is applicable to a pixel array including a plurality of first pixels, a plurality of second pixels, and a plurality of third pixels, each of the first pixels having a first color The intensity value, each second pixel has a second color intensity value, and each third pixel has a third color intensity value, the color interpolation device includes: an edge feature detecting unit, respectively The arrangement of the plurality of first pixels of the pixel array is regarded as an arrangement of 45 degrees and 135 degrees of interleaving, and then detecting edge features of the pixel array in a direction along 45 degrees and 135 degrees respectively; an interpolation unit Linking with the edge feature detecting unit, the first pixels are interpolated according to the edge features to generate at least two interpolation results, so that the second pixels and the third pixel systems respectively generate the An intensity value of the first color, and spectrally interpolating the pixel array, such that the first pixel and the second pixel systems each generate an intensity value of the third color, and the first The pixel and the third pixel respectively generate intensity values of the second color; a weighted median filter is respectively connected to the edge feature detecting unit and the interpolation unit, and according to the edge feature, The interpolation result assigns different weight values, and performs Directional Weighted Median Filtering to determine the best intensity value of the first color, and the optimal intensity value of the first color and The second color generated by the spectral interpolation and the intensity value of the third color are subjected to Center Weighted Median Filtering; and a post processing unit is connected to the weighted median filter, and And generating the first output of the final output according to the optimal intensity value of the first color after performing the intermediate weighted median filtering, and the intensity value of the second color and the third color generated after performing spectral interpolation The color, the second color, and the intensity value of the third color.

In order to achieve the above object, the color interpolation method of the present invention is applicable to a pixel array including a plurality of first pixels, a plurality of second pixels, and a plurality of third pixels, each of the first pixels having a first color The intensity value, each second pixel has a second color intensity value, and each third pixel has a third color intensity value, the color interpolation method comprising the steps of: multiplexing the first pixel of the pixel array The arrangement manner is respectively regarded as an arrangement of 45 degrees or 135 degrees of interleaving, and then detecting edge features of the pixel array in the direction of 45 degrees or 135 degrees respectively; performing the first pixels according to the edge features Initially interpolating to generate at least two interpolation results, after the preliminary interpolation, the second pixels and the third pixel systems each generate an intensity value of the first color; according to the edge features, the The interpolation result assigns different weight values and performs directional weighted median filtering; determining the best value according to the intensity value of the first color of the interpolation result after performing the directional weighted median filtering An intensity value of the first color; spectrally interpolating the pixel array, wherein the first pixel and the second pixel are each generating an intensity value of the third color, and the first pixel and the first The three pixel systems each generate an intensity value of the second color; and intermediately weighting the optimal intensity value of the first color and the intensity value of the second color and the third color generated after performing spectral interpolation Median filtering to determine the intensity value of the first color, the second color, and the third color that are finally output.

Through the above apparatus and method, the present invention can detect and reconstruct the edge features of the color filter array along the direction of 45 degrees or 135 degrees, thereby avoiding zipper shape, color overlap, and the image boundary of the image. Distortion such as complementary color error.

The preferred embodiments of the present invention are described in detail below with reference to the drawings.

Please refer to the first to fifth c-pictures, the color interpolation device (1) of the present invention is applied to a pixel array (2), the pixel array (2) includes a plurality of first pixels (20), and a plurality of second a pixel (21) and a plurality of third pixels (22), each of the first pixels (20) having a first color intensity value (not shown), and each second pixel (21) having a second color intensity a value (not shown), each third pixel (22) has a third color intensity value (not shown), and the color interpolation device (1) includes: an edge feature detecting unit (10), The arrangement of the plurality of first pixels (20) of the pixel array (2) is regarded as an arrangement of 45 degrees and 135 degrees, and the pixel array is detected from 45 degrees and 135 degrees respectively. (2) edge features in the complex direction; an interpolation unit (11) is coupled to the edge feature detecting unit (10), and is generated by interpolating the first pixels (20) according to the edge features And at least two interpolating results, wherein the second pixels (21) and the third pixels (22) respectively generate intensity values of the first color, and perform spectrum on the pixel array (2) Inserting, the first pixel (20) and the second pixels (21) each generating an intensity value of the third color, and the first pixel (20) and the third pixel (22) are Generating an intensity value of the second color; a weighted median filter (12) is respectively connected to the edge feature detecting unit (10) and the interpolation unit (11), and according to the edge feature, The interpolation result assigns different weight values, and performs Directional Weighted Median Filtering to determine the optimal intensity value of the first color, and the optimal intensity value of the first color. And performing the intermediate weighted median filtering on the second color generated by the spectral interpolation and the intensity value of the third color; and a post processing unit (13), and the weighted median filtering The device (12) is connected, and according to the optimal intensity value of the first color after performing intermediate weighted median filtering, and the intensity value of the second color and the third color generated after performing spectral interpolation, Producing the first color of the final output, A second color and the third color intensity value.

In a preferred embodiment of the present invention, the edge feature detecting unit (10) first considers the arrangement of the plurality of first pixels (20) of the pixel array (2) as a 135-degree interlaced arrangement, The space of the first pixel (20) represents the second pixel (21) or the third pixel (22), i represents the ordinate of the pixel array (2), and j represents the abscissa of the pixel array (2). Since the arrangement of the plurality of first pixels (20) of the pixel array (2) is regarded as a 45-degree interlaced arrangement in a symmetrical relationship with the 135-degree interleaved arrangement, no further description is provided herein.

The edge feature detecting unit (10) detects edge features of the pixel array (2) in five directions along a direction of 135 degrees, which are a first direction, a second direction, a third direction, and a first a fourth direction and a fifth direction, the first direction, the second direction, the third direction, the fourth direction, and the fifth direction are respectively from a plurality of diamond squares from the pixel array (2) origin (i , j) extending to the two ends, each of the diamond-shaped squares inscribed a second pixel (21) or the third pixel (22), and each of the four corners of each of the diamond-shaped squares corresponds to a first pixel ( 20).

If the origin (i, j) is the second pixel (21), the intensity values of the second color of the first pixels (20) are first simulated, and the first direction and the second direction are respectively obtained. In the third direction, the fourth direction, and the fifth direction, the first point corresponding to the closest angle between the diamond square of the origin (i, j) and the diamond square closest to the two ends thereof is first a color difference between the first color intensity value of the pixel (20) and the second color intensity value, and then averaging the color difference between the first color intensity value and the second color intensity value, the average value plus the origin (i, j) The second color intensity value is the edge feature of the edge feature detecting unit (10) detecting the origin (i, j) of the pixel array (2) along the direction of 135 degrees, but The invention is not limited thereto. If the origin (i, j) is the third pixel (21), the intensity value of the second color is replaced by the intensity value of the third color.

Then, the weighted median filter (12) is based on the average values obtained in the first direction, the second direction, the third direction, the fourth direction, and the fifth direction, and the first direction, The second direction, the third direction, the fourth direction, and the fifth direction perform directional weighted median filtering, and the first direction, the second direction, the third direction, the fourth direction, and the The average value obtained by the fifth direction is assigned different weights to change the edge feature detecting unit (10) from the first direction, the second direction, the third direction, the fourth direction, and the fifth direction. The number of times the average.

The edge feature detected by the edge feature detecting unit (10) further includes a horizontal width and a vertical width of the image content, and the interpolation unit (11) further generates a width according to the horizontal width and the vertical width of the image content. As a result of the interpolation, the second pixels (21) and the third pixels (22) each generate an intensity value of the first color.

In an embodiment of the present invention, if the origin (i, j) is a second pixel (21), when the horizontal width of the image content is greater than the vertical width, the width interpolation result is from the image content. The color difference between the first color intensity value of the first pixel in the vertical direction and the color intensity value of the simulated second color intensity, and the average of the second color intensity values of the second pixel symmetrically up and down in the vertical direction of the image content, when the image content level If the width is smaller than the vertical width, the width interpolation result is a color difference between the first color intensity value of the first pixel and the simulated second color intensity value in the horizontal direction of the image content, and the horizontal direction of the image content is bilaterally symmetric. The second color intensity values of the second pixels are averaged, and the edge features in the directional weighted median filtering performed by the weighted median filter (12) include the width interpolation results.

The spectral interpolation performed by the interpolation unit (11), such that the first pixel (20) and the third pixel (22) each generate an intensity value of the second color, wherein the central pixel (i, j) is the first pixel (20), and the color difference between the first color intensity value and the second color intensity value is simulated by the second pixel (21) adjacent to the center pixel (i, j), respectively. Inserting a second color intensity value of the central pixel (i, j), wherein the second pixel (21) adjacent to the central pixel (i, j) is likely to be above the central pixel (i, j) Position or left and right position, if the central pixel (i, j) is the third pixel (22), the simulated first color intensity value of the central pixel (i, j), and the periphery of the central pixel (i, j) The simulated first color intensity value and the second color intensity value of the second pixel (21) in the four corner directions interpolate the intensity value of the second color.

Interpolating the spectral interpolation performed by the interpolation unit (11) such that the first pixel (20) and the second pixels (21) each generate an intensity value of the third color, wherein if the central pixel ( i, j) is the first pixel (20), and the color difference between the first color intensity value and the third color intensity value is simulated by the third pixel (22) adjacent to the central pixel (i, j), respectively. Interpolating a third color intensity value of the central pixel (i, j), wherein a third pixel (22) adjacent to the central pixel (i, j) is likely to be above and below the central pixel (i, j) Or the left and right positions, if the central pixel (i, j) is the second pixel (21), the simulated first color intensity value of the central pixel (i, j) is used, and the central pixel (i, j) is surrounded by four The simulated first color intensity value and the third color intensity value of the third pixel (21) in the corner direction interpolate the intensity value of the third color.

The intermediate weighted median filtering performed by the weighted median filter (12) first utilizes the optimal intensity value of the first color and the second color generated after spectral interpolation and the first The intensity value of the three colors is used to obtain the color difference between the first color intensity value and the second color intensity value of the pixel array (2), and the color difference between the first color intensity value and the third color intensity value, and then the precision needs to be enhanced. The pixels at the position are weighted to increase the color difference between the first color intensity value and the second color intensity value of the specific pixel, and the number of color differences between the first color intensity value and the third color intensity value, and finally the post processing The unit (13) reuses the color difference between the first color intensity value and the second color intensity value, and the color difference between the first color intensity value and the third color intensity value, and according to the optimal intensity value of the first color, and The second color generated by the spectral interpolation and the intensity value of the third color produce a final output of the first color, the second color, and the intensity value of the third color.

The preliminary interpolation performed by the interpolation unit (11) and the spectral interpolation are achieved by bilinear interpolation or bicubic interpolation.

In a preferred embodiment of the present invention, the first color is green, the second color is red, and the third color is blue, but the invention is not limited thereto, and another implementation of the present invention. In the example, the first color is green, the second color is blue, and the third color is red.

Referring to the second to sixth figures, the color interpolation method (3) of the present invention is applicable to a pixel array (2) including a plurality of first pixels (20) and a plurality of second pixels. (21) and a plurality of third pixels (22), each first pixel (20) has a first color intensity value, and each second pixel (21) has a second color intensity value, each third The pixel (22) has a third color intensity value, and the color interpolation method (3) comprises the steps of:

Step 300: The arrangement of the plurality of first pixels (20) of the pixel array (2) is regarded as an arrangement of 45 degrees and 135 degrees, respectively, and then the pixels are detected along the directions of 45 degrees and 135 degrees, respectively. Array (2) edge features in the complex direction;

Step 301: Perform preliminary interpolation on the first pixels (20) according to the edge features to generate at least two interpolation results, and after performing preliminary interpolation, the second pixels (21) and the third pixels. (22) each generating an intensity value of the first color;

Step 302: assign different weight values to the interpolation results according to the edge features, and perform directional weighted median filtering;

Step 303: Determine an optimal intensity value of the first color according to the intensity value of the first color of the interpolation result after performing the directional weighted median filtering;

Step 304: performing spectral interpolation on the pixel array (2), so that the first pixels (20) and the second pixels (21) respectively generate intensity values of the third color, and the first pixels (20) and the third pixels (22) each generating an intensity value of the second color;

Step 305: Perform intermediate weighted median filtering on the intensity value of the first color and the intensity value of the second color and the third color generated after spectral interpolation to determine the final output. The intensity value of a color, the second color, and the third color.

The step 300 detects the edge features of the five directions of the pixel array (2) along a direction of 135 degrees, which are a first direction, a second direction, a third direction, a fourth direction, and a a fifth direction, the first direction, the second direction, the third direction, the fourth direction, and the fifth direction are respectively from a plurality of diamond squares from an origin (i, j) of the pixel array (2) The two ends extend, and each of the diamond-shaped squares inscribes a second pixel (21) or the third pixel (22), and each of the four corners of each of the diamond-shaped squares corresponds to a first pixel (20).

If the origin (i, j) is the second pixel (21), the intensity values of the second color of the first pixels (20) are first simulated, and the first direction and the second direction are respectively obtained. In the third direction, the fourth direction, and the fifth direction, the first point corresponding to the closest angle between the diamond square of the origin (i, j) and the diamond square closest to the two ends thereof is first a color difference between the first color intensity value of the pixel (20) and the second color intensity value, and then averaging the color difference between the first color intensity value and the second color intensity value, the average value plus the origin (i, j) The second color intensity value is the edge feature on the origin (i, j) of the pixel array (2) detected along the direction of 135 degrees, but the invention is not limited thereto. When the origin (i, j) is the third pixel (21), the intensity value of the second color is replaced by the intensity value of the third color.

The step 302 is based on the average values obtained in the first direction, the second direction, the third direction, the fourth direction, and the fifth direction, and the first direction, the second direction, and the third The direction, the fourth direction, and the fifth direction perform directional weighted median filtering, and the average values of the first direction, the second direction, the third direction, the fourth direction, and the fifth direction are averaged Different weights are used to change the number of times the average value is obtained from the first direction, the second direction, the third direction, the fourth direction, and the fifth direction.

The spectral interpolation performed by the step 304 is such that the first pixel (20) and the third pixel (22) each generate an intensity value of the second color, wherein if the central pixel (i, j) for the first pixel (20), respectively, using the second pixel (21) adjacent to the central pixel (i, j) to simulate the color difference between the first color intensity value and the second color intensity value, interpolation a second color intensity value of the central pixel (i, j), wherein the second pixel (21) adjacent to the central pixel (i, j) is likely to be above or below the central pixel (i, j) Position, if the central pixel (i, j) is the third pixel (22), the simulated first color intensity value of the central pixel (i, j) and the four corners surrounding the central pixel (i, j) The simulated first color intensity value and the second color intensity value of the second pixel (21) of the direction interpolate the intensity value of the second color.

The spectral interpolation performed in the step 304 is such that the first pixel (20) and the second pixels (21) respectively generate an intensity value of the third color, wherein if the central pixel (i, j) for the first pixel (20), respectively, using the third pixel (22) adjacent to the central pixel (i, j) to simulate the color difference between the first color intensity value and the third color intensity value, interpolation a third color intensity value of the central pixel (i, j), wherein the third pixel (22) adjacent to the central pixel (i, j) is likely to be above or below the central pixel (i, j) Position, if the central pixel (i, j) is the second pixel (21), the simulated first color intensity value of the central pixel (i, j) and the four corners surrounding the central pixel (i, j) The simulated first color intensity value and the third color intensity value of the third pixel (21) of the direction interpolate the intensity value of the third color.

The intermediate weighted median filtering performed in step 305 is performed by first using the optimal intensity value of the first color and the intensity value of the second color and the third color generated after spectral interpolation. a color difference between the first color intensity value and the second color intensity value of the pixel array (2), and a color difference between the first color intensity value and the third color intensity value, and then weighting the pixel at a position where the precision is required to be Increasing a color difference between the first color intensity value and the second color intensity value of the specific pixel, and a color difference between the first color intensity value and the third color intensity value, and finally using the first color intensity value and the second color a color difference of the intensity value, and a color difference between the first color intensity value and the third color intensity value, and according to the optimal intensity value of the first color, and the second color generated after spectral interpolation and the first The intensity values of the three colors produce a final output of the first color, the second color, and the intensity value of the third color.

The edge features further include a horizontal width and a vertical width of the image content, and preliminary interpolating the first pixels (20) according to the horizontal width and the vertical width of the image content to generate at least one width interpolation. As a result, the second pixels (21) and the third pixels (22) each generate an intensity value of the first color, and the interpolation result in the step 302 includes the width interpolation results.

In an embodiment of the present invention, if the origin (i, j) is a second pixel (21), when the horizontal width of the image content is greater than the vertical width, the width interpolation result is from the image content. The color difference between the first color intensity value of the first pixel in the vertical direction and the color intensity value of the simulated second color intensity, and the average of the second color intensity values of the second pixel symmetrically up and down in the vertical direction of the image content, when the image content level If the width is smaller than the vertical width, the width interpolation result is a color difference between the first color intensity value of the first pixel and the simulated second color intensity value in the horizontal direction of the image content, and the horizontal direction of the image content is bilaterally symmetric. The average of the second color intensity values of the second pixels is obtained. The edge features in the directional weighted median filtering performed in step 302 include the width interpolation results.

The preliminary interpolation of the step 301 and the spectral interpolation of the step 304 are achieved by bilinear interpolation or bicubic interpolation.

In a preferred embodiment of the present invention, the first color is green, the second color is red, and the third color is blue, but the invention is not limited thereto, and another implementation of the present invention. In the example, the first color is green, the second color is blue, and the third color is red.

Since the present invention detects and reconstructs edge features of the color filter array in a direction of 45 degrees or 135 degrees, it is possible to avoid distortion such as zipper shape, color overlap, and complementary color error at the image boundary of the image; Moreover, the structure and the method of the steps are not easily reached by those skilled in the art, and are novel and progressive.

Through the above detailed description, it can fully demonstrate that the object and effect of the present invention are both progressive in implementation, highly industrially usable, and are new inventions not previously seen on the market, and fully comply with the invention patent requirements. , 提出 apply in accordance with the law. The above is only the preferred embodiment of the present invention, and is not intended to limit the scope of the invention. All changes and modifications made in accordance with the scope of the invention shall fall within the scope covered by the patent of the invention. I would like to ask your review committee to give a clear explanation and pray for it.

(1). . . Color interpolation device

(10). . . Edge feature detection unit

(11). . . Interpolation unit

(12). . . Weighted median filter

(13). . . Post processing unit

(2). . . Pixel array

(20). . . First pixel

(twenty one). . . Second pixel

(twenty two). . . Third pixel

(3). . . Color interpolation method

300. . . step

301. . . step

302. . . step

303. . . step

304. . . step

305. . . step

The first figure is an architectural diagram of the color interpolation device of the present invention;

The second figure is a schematic diagram of the arrangement of the plurality of first pixels of a pixel array as a 135 degree interleaving arrangement;

The third a diagram is a schematic diagram of detecting edge features of the first direction of the pixel array along a direction of 135 degrees;

The third b diagram is a schematic diagram of detecting edge features of the second direction of the pixel array along a direction of 135 degrees;

The third c-picture is a schematic diagram of detecting edge features of the third direction of the pixel array along a direction of 135 degrees;

The third d-picture is a schematic diagram of detecting edge features of the fourth direction of the pixel array along a direction of 135 degrees;

The third e-picture is a schematic diagram of detecting edge features of the fifth direction of the pixel array along a direction of 135 degrees;

The fourth a diagram is a schematic diagram of the horizontal width edge feature of an image content;

The fourth b diagram is a schematic diagram of the vertical width edge features of the image content;

The fifth embodiment is a first embodiment in which a first pixel generates a second color intensity value;

The fifth b is a second embodiment in which the first pixel generates an intensity value of the second color;

The fifth c-picture is an embodiment in which a third pixel generates an intensity value of the second color;

The sixth figure is a flow chart of the method of the color interpolation method of the present invention.

(1). . . Color interpolation device

(10). . . Edge feature detection unit

(11). . . Interpolation unit

(12). . . Weighted median filter

(13). . . Post processing unit

Claims (10)

A color interpolation device is applicable to a pixel array, the pixel array includes a plurality of first pixels, a plurality of second pixels, and a plurality of third pixels, each of the first pixels having a first color intensity value, each second The pixel has a second color intensity value, and each of the third pixels has a third color intensity value. The color interpolation device includes: an edge feature detecting unit, wherein the plurality of first pixels of the pixel array are respectively The arrangement is regarded as an arrangement of 45 degrees and 135 degrees of interleaving, and then detects edge features of the pixel array in a direction along 45 degrees and 135 degrees respectively; an interpolation unit, and the edge feature detection a unit connection, wherein the first pixels are interpolated according to the edge features to generate at least two interpolation results, so that the second pixels and the third pixel systems respectively generate the intensity values of the first color, and Performing spectral interpolation on the pixel array, such that the first pixel and the second pixel system respectively generate intensity values of the third color, and the first pixels and the third pixel systems are respectively Self-generating the intensity value of the second color; a weighted median filter is respectively connected to the edge feature detecting unit and the interpolation unit, and according to the edge feature, assigning different weight values to the interpolation results, And performing Directional Weighted Median Filtering to determine the best intensity value of the first color, and the optimal intensity value of the first color and the spectral interpolation The second color and the intensity value of the third color are subjected to Center Weighted Median Filtering; and a post-processing unit is connected to the weighted median filter and is subjected to intermediate weighted median filtering. The best intensity value of the first color, and the second color generated by the spectral interpolation and the intensity value of the third color, the last color of the first color, the second color, and the The intensity value of the third color. The color interpolation device of claim 1, wherein the edge feature detected by the edge feature detecting unit further includes a horizontal width and a vertical width of the image content, and the interpolation unit is further based on the image content. The horizontal width and the vertical width generate a width interpolation result, so that the second pixel and the third pixel system each generate an intensity value of the first color, and the weighted median filter performs a directional weighted median The interpolation result in the filtering includes the width interpolation results. The color interpolation device according to claim 1, wherein the preliminary interpolation performed by the interpolation unit and the spectral interpolation are performed by bilinear interpolation or bicubic interpolation. (bicubic interpolation) is achieved. The color interpolation device of claim 1, wherein the first color is green, the second color is red, and the third color is blue. The color interpolation device of claim 1, wherein the first color is green, the second color is blue, and the third color is red. A color interpolation method is applicable to a pixel array, the pixel array includes a plurality of first pixels, a plurality of second pixels, and a plurality of third pixels, each of the first pixels having a first color intensity value, each second The pixel has a second color intensity value, and each third pixel has a third color intensity value. The color interpolation method includes the steps of: respectively arranging the plurality of first pixels of the pixel array as 45 degrees and 135 degree interlaced arrangement, and then detecting edge features of the pixel array in the direction of 45 degrees and 135 degrees respectively; preliminary interpolating the first pixels according to the edge features to generate at least two Interpolating results, after performing preliminary interpolation, the second pixels and the third pixel systems respectively generate intensity values of the first color; according to the edge features, assigning different weight values to the interpolation results, and Performing directional weighted median filtering; determining an optimal intensity value of the first color according to the intensity value of the first color of the interpolation result after performing the directional weighted median filtering; Performing spectral interpolation on the pixel array, such that the first pixel and the second pixel systems respectively generate intensity values of the third color, and the first pixels and the third pixel systems each generate the second The intensity value of the color; and the intermediate weighted median filtering of the optimal intensity value of the first color and the intensity value of the second color and the third color generated after spectral interpolation to determine the final output The intensity values of the first color, the second color, and the third color. The color interpolation method according to claim 6, wherein the edge features further comprise a horizontal width and a vertical width of the image content, and the first pixels are according to the horizontal width and the vertical width of the image content. Performing preliminary interpolation to generate at least one width interpolation result, such that the second pixels and the third pixel systems each generate an intensity value of the first color, and the interpolation in the directional weighted median filtering step The result includes the width interpolation results. The color interpolation method of claim 6, wherein the preliminary interpolation and the spectral interpolation are achieved by bilinear interpolation or bicubic interpolation. The color interpolation method of claim 6, wherein the first color is green, the second color is red, and the third color is blue. The color interpolation method of claim 6, wherein the first color is green, the second color is blue, and the third color is red.