TW201034470A - Color interpolation method and image demosaicking method using the same and computer program production thereof - Google Patents

Abstract

Disclosed are a color interpolation method and an image demosaicking method using the same and a computer program production thereof for reconstructing missing color components of the pixels of a pixel array, wherein the pixel array is corresponding to a color filter array and includes a plurality of estimated pixels and a target pixel. The color interpolation method is used to estimate the strength of a target pixel in a 45 DEG diagonal or a 135 DEG diagonal direction. In the image demosaicking method, the color interpolation method is firstly performed to obtain a plurality of candidate estimated strengths of the target pixel. Then, one of the candidate estimated strength is determined to be the estimated strength of the target pixel in accordance with the strengths of the estimated pixels.

Description

201034470 _ VI. 明明: ............ [Technical Field of the Invention] The present invention relates to a color interpolation method and an image demosaicing method using the color interpolation method, in particular It is a color interpolation method applied to processing an image generated by a color filter and an image demosaicing method using the color interpolation method. > [Prior Art] In recent years, digital still cameras (Digital Still Cameras; DSCs) have become very popular consumer electronics products. Digital cameras typically use a Charge Couple Device (CCD) or a Complementary Metal Oxide Semiconductor (CMOS) sensor to convert the sensed incident light into an electrical signal, which is then analog/digital converted. , become a digital image signal. Please refer to FIG. 1 for a schematic diagram showing the structure of a conventional color filter array (CFA). In order to save the size of the sampled color pixels and sensors, color filter arrays are most commonly used as color sampling formats, so a color filter array is usually plated on a CCD or CMOS sensor. A commonly used color filter array is a CFA published by Bryce E. Bayer, as shown in Fig. 1, wherein fr is a red calender, FG is a green filter, and fb is Blue fishing light film. When using CFA, each sample point has only one color, which is one-third of the full-color system (ie, 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 201034470 color components. - In order to reconstruct a complete color image from the original sensed data, color interpolation is usually used to estimate the two color defects that are missing from each pixel. The so-called color interpolation method is based on the information provided by a part of each target pixel, that is, when a CFA is known, the two color components lacking in one pixel can be Reconstructed by using information provided by neighboring pixels. The reconstruction of the overall image is the industry's known demosaicking operation. The purpose of decoding the demosaicking technology for Bayer's color filter array (Bayer CFA) can only achieve the vertical and horizontal side of the sum and reconstruction, but for the other two may also exist in the electronic image 45 . And 135. There is no effective reconstruction method in the side direction. SUMMARY OF THE INVENTION Accordingly, one aspect of the present invention is to provide a color interpolation method. (4) The interpolation method can be along 135 of the target pixel. The color interpolation method is applicable to a pixel array, the pixel array includes a plurality of pixels, and the pixels include a plurality of first pixels and a plurality of second pixels, each of the first pixels having a corresponding one a first intensity value of the first color (GJ, each second pixel having a second = degree value (A) corresponding to one of the second colors, where j represents the abscissa of each pixel, and 丨 represents each pixel In the color interpolation method, an interpolation step is first performed on each of the second pixels to calculate an estimated intensity value (4) corresponding to one of the first colors per second pixel. The interpolation method performs an interpolation step for each second pixel to calculate an estimated intensity value (6) corresponding to the first color 201034470 every 1 2 pixels. In this interpolation step, first, the root array is Providing a high pass mask, then measuring the position of the (target pixel) relative to one of the edges of the pixel array, ^, ,, (f>. Then 'performing the first-to-slant intensity value meter Step = intensity value (譬 this first oblique intensity value According to a formula ··. Then the oblique intensity value calculation step is calculated to calculate the -筮- 弟一 grab intensity value calculation step according to the formula: ιΐτ丨+Γγ ° Then, the _estimation parameter calculation (4) , by the formula: "number of turns (4)) 'where the estimated parameters are calculated according to the - public

L 135 max min Guanwei, catch SE, otherwise Green intensity value Formula V Wu Ruru, rWW2) Color interpolation method for color... Color interpolation (4) Suitable for interpolating in the direction of (4). For this pixel, the pixels comprise a plurality of:: column II array comprises a complex (four) each pixel has a pair of pixels and a plurality of second pixels, (~), each - second pixel 1::: one of the colors first Discussion

^ t J 彳冢 杈 杈 coordinates, i represents the 201034470 ordinate of each pixel. The color interpolation method performs an interpolation step for each second pixel to calculate an estimated intensity value (heart) corresponding to one of the first colors per 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 to provide a detection result (f). Then, a first oblique intensity value calculation step is performed to calculate a -first oblique intensity value (four) sticker, wherein the first oblique intensity value calculation step is based on a common ^ 8 +1. Connect * 'to - the second oblique intensity value calculation step, two calculate a second oblique intensity value (〇 'where the second oblique strong calculation step is based on - formula: ~. Then, enter 彳 ^# test Parameter calculation (4), to calculate - estimate the parameter calculation (4) in the system

L

45 U maxfsiV^^^ ^ l, j Uj j her {catch Ί V hj Li if /> 〇 otherwise Next, enter the 估 an estimated intensity value calculation step (m towel... / "丨故" 十算-estimate Intensity value, wherein the estimated intensity value calculation step is

P

i,, ΐ (〇h2, 广- U ^ invention again - the state is provided - 郷 image 2 processing corresponding - color filter, mirror array - like 3 = two pixels, 4 2 1 ΓΓ contains a plurality of - The first pixel like the ^^^^ mother has a corresponding value to a first face $+, an intensity value (G.), and each -m _ pushes one of the colors of the Β山弟| all, j) to the mother The pixel has a corresponding intensity value corresponding to a first-shoaling uv). The seventh person, the first color - the first pixel contains a plurality of estimated transmissions 201034470 .^ ' and these estimated pixels contain the target pixel, the remaining ten of the estimated pixels and the estimated The pixel is centered around the target pixel. Narrow, = an intensity value estimation step to calculate a second estimated intensity value for one of the corresponding colors of the target pixel. In the intensity value estimation step, the step 'to calculate a plurality of candidate estimated intensity lines of the target pixel-determining step to estimate the second target of the mosquito target pixel according to the candidate estimated intensity value and the first estimated intensity value Measure the intensity value. [Embodiment] For the sake of convenience of explanation, in the following description, the intensity value of the image color, for example, the green intensity value represents the color of the y image according to the processing region of the present invention. The T element can be divided into two according to the structure of the color filter array. The first pixel R corresponds to red and the other corresponds to green. The third pixel 8 corresponds to blue. Window color interpolation 'Users usually use - decide to view eClslon wmd〇w) w to determine the size of the processing area, and then use the pixels in 3 = to interpolate. In this embodiment, the size of the chimney is 5 times. s main 丨* This is not 5 pixels in size, but the present invention is not limited to the 〇月参..., Fig. 4' is a schematic flow diagram of the coloring method 2010100470 interpolation method 200 according to the present invention. Color interpolation. Method 2 is used to estimate the color component lacking in the target pixel of one of the processing regions 102. Generally, since the human eye is sensitive to green, in the following description, the coordinates will be estimated. The first pixel R of (i, j) (hereinafter referred to as In the color interpolation method 200, the mask providing step 21 is first performed to provide a preset high pass mask (high pass mask). ), for example: Laplacian mask or Lapladan 〇f Gaussian Mask to detect the position of the target pixel relative to the edge of the pixel array and output a Excerpt result f, where the edge refers to the pixel array = where the gray scale changes greatly. The size of the mask is determined according to the mosquito window, since the decision window used in this embodiment is 5 by 5 pixels,

SK = The size of the cover is also 5 by 5 pixels. Next, an oblique intensity value calculation step 220 is performed. In the first oblique intensity value calculation step, the oblique strength value ^ is calculated by using the - formula, which is expressed as follows: : Gw + Gu - i nrrrNW ^ _ where the representative represents the target pixel as the center, and the intensity value The value of the 健 健 为 ( ( 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估 估The value then proceeds to the first oblique intensity value calculation step to the intensity value calculation step 230, using a formula ~ at the first oblique value milk, this formula is expressed as follows: The leaf calculates the oblique intensity where the 'ah y system represents The target pixel is centered, the intensity value, the Gi+1J system represents the southward estimate of the (i+1,j) intensity value, and the GiJ+1 system represents the coordinate _ one pixel G to the pixel G of the green 201034470 Then, an estimated parameter calculation step 240 is performed to calculate the estimated parameter β. In the filter parameter calculation step 240, 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: f NW SE) max srv , SIV 135_ l Uj 'NW SE~\ Min SIV ,SIV Uj) if /^〇 otherwise

It can be seen from the above equation that when the result f is greater than or equal to zero, the estimated parameter 1,,; is the largest of them; when the result f is less than zero, the estimated parameter is the smallest of the minds. Then, an intensity value calculation step 250 is performed. In the estimated intensity value calculation step 250, the estimated intensity value A is calculated according to a formula, which is expressed as follows: where Ri, j represents the red intensity value of the target pixel, and 艮-2, 3+2 represents the coordinate The red intensity value of the first pixel R of (i-2, j+2), Ri+2, and 2 represent 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 a pixel array having an edge of 135° edge direction, 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, except 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, 201034470 is first performed. A preset high pass mask is provided to detect the position of the target pixel relative to the edge of the pixel array, and the detection result f is provided. Next, the oblique intensity value meter is calculated to calculate step 320. In the oblique intensity value calculation step 320, a formula is used to calculate the oblique intensity value y; f, the formula is expressed as follows: sidj+Gu+' where ah, the 7 series represents the target pixel as the center, and is estimated from the north to the east. The measured oblique intensity value, GM, j represents the green intensity value of the second pixel G of the coordinate, and Gi, ^ represents the green intensity φ value of the second pixel G whose coordinates are (ij + Ι). 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, which is expressed as follows: siv = = gmj + g,, b

Among them, the representative represents the oblique intensity value centered on the target pixel from west to south, and Gi+1,j represents the green intensity value of the second pixel G with coordinates (i+l,j), Gy" Represents the green intensity value of the second pixel G of coordinates (i, j_l). Next, an estimated parameter calculation step 340 is performed to calculate the estimated parameter C. In the estimated parameter calculation step 340, the result f obtained by the high-pass mask detection is used and the estimated parameter β is calculated according to a formula, which is expressed as follows:

L 45 hj max min r NE SW^ SIV , SIV , i, ju) NE SW\ SIV , SIV i, j hj ) if otherwise It can be seen from the above equation that when the result f is greater than or equal to zero, the estimated parameter & To see the biggest of 匕7 and ahτ; when the result f is less than zero, the estimated parameter /^ is the smallest of 邶7 and Γ. Then, an estimated intensity value calculation step 350 is performed. In the estimated intensity value calculation step 350, the estimated intensity value I is calculated according to a formula 11 201034470. This formula is expressed as follows: . P z· , J (Ζ5 ) + 士(2, factory-2 J -2 - 尺 +2 J+2 ) where Ri, j represents the red intensity value of the target pixel, and Ri-2, j-2 represents the red intensity of the first pixel R of the coordinates (i-2, j-2) The value, Ri+2J+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° direction 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 A formal estimate reveals 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 decision window W, and then uses the pixels in the processing area 402 to perform demosaicing. 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 the softness 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 accuracy. In the following description, the green component lost by the unestimated pixel 117 having coordinates (i, j) in the estimation processing region 402 will be taken as an example to illustrate the demosaicing 12 201034470, in which the pixels 111 to 116 have been estimated. The green intensity value has been estimated • The green intensity value of the σ' unestimated pixels 117~125 has not been estimated. In the demosaicing method 500, 'interpolation step 510 is first performed, and the plurality of candidate estimation strength interpolation steps 510 corresponding to the green pixels 117-125 corresponding to green can utilize the conventional horizontal interpolation method and vertical interpolation method. 45 of the present invention. Edge interpolation and 135. Edge interpolation produces a plurality of candidate estimated intensity values for pixel 125. These candidate estimated intensity values can be: the interpolated intensity value calculated by the interpolation method or the higher the interpolated intensity value is obtained by some calculation method J such as the average calculation method. :value. Next, a decision step 52 is performed to calculate the green estimated intensity value of the pixel 117 based on the estimated intensity value and the green intensity values of the estimated pixels (1), U2, 113, and 116 according to the pixels ι 7 to ΐ25. In the decision step 52, the color difference calculation step 520a is first performed, with a plurality of color difference values, meanings, C and Cf in the window w region centered on the pixel 117, wherein the color difference calculation step 52 is performed Formula calculation: i^, y+n~ '0JJ+n , 々·+«, if « = if « = :-2 = 0,2 [Ri+n,j- ' ^i+n,j , if / 7 = =-2 λ^η,Γ -3⁄4 ”J, if « = = 0,2 if n =-2 ^Ri+n,j- n~GilnJ-n^ if « = 0,2 < ^+ n9J+n ^i+nj+n , if« =~2 ^nJ^G}lnj+n^ if n -0,2 & U is the coordinate value of the pixel, R is the red intensity of the pixel, G Represents the estimated green intensity value of the pixel, 仏 represents the pixel of the pixel ,, /) 0,·/) = d) 13 201034470 - The estimated green intensity value, the square represents the vertical estimated green intensity value of the pixel, The representative pixel is at 45. The estimated green intensity value ό 45., ό 135. represents the estimated green intensity value of the pixel on the 135° side. Next, an indicator calculation step 520b is performed to calculate the indicator group (indicator set) ω,;={δ//,7,δρ^,Μ5〇,;,δι35..}, where the indicator is calculated The variation of the color difference in the W region along the vertical, horizontal, 45, and 135° directions is calculated by the following formula, but the present invention is not limited thereto. C (four) force - "suppressing + 丨 L L = |C;f (Bu 2J-2)-e(U.)HC(W)-C(i+2J+2). Then, a selection step 520c' is performed to estimate the intensity value from the candidate according to the indicator group The estimated green intensity value & amp of the selected pixel 117, wherein the selecting step 520c selects the estimated intensity value corresponding to the smallest direction of the color difference change difference direction, and the embodiment is calculated by the following formula, but the invention is not limited Here I(Gj + G[j), if any two of the indicators = min (Ω( y), if ΜΓ, ν = min (Ω, -y), if Δ. = = min(QIJ) /45. Gi,j, if Δ45〇•J = min (Ω/>7) if Δ135° tj =min(Q,y)

As can be seen from the above description, in the present embodiment, the officially estimated pixels 14 201034470 - 111 , 112 , 113 , 116 and the un-estimated pixels 118 , 121 , 122 , • 123 are surrounded by the target pixel 117 . Therefore, the estimated intensity values (3⁄4, sentences) obtained by the estimated thickness values of the pixels 1U, 112, 113, and 116 and the non-formally estimated pixels 118, 121, 122, and 123 corresponding to the plurality of interpolation methods may be used. To determine the candidate intensity values as the estimated green intensity value of the target pixel 117, and then calculate the target pixel by some calculation (such as the average calculation method) of the candidate green estimated intensity value. Estimated green intensity value of 117. 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, t: 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, and 124 are calculated. _Please refer to FIG. 9 , FIG. 10 and FIG. 9 at the same time, and the ninth embodiment of the present invention is not in accordance with an embodiment of the present invention. FIG. FIG. 11 is a schematic view of the flow (fourth) of the method 70G according to the present invention. In this implementation, the image is used by the Marseille method, and the green intensity value has been officially estimated to be the pixel n? to estimate the green intensity value of the measured pixel m. In the image demosaicing method 7GG, first, the solution #5 method 5()() is performed, and the target estimates the green component lost by the pixel n7. After the target pixel 117 ^ 4 color intensity value is officially estimated, the decision window w moves to the processing area 6 〇 2 centered on the official estimated pixel U8. Next, proceed to = 15 201034470 to insert the step 71G' to calculate the number of candidate estimated strength values corresponding to the green of the prime m, 119, 122, 123, and 124. The interpolation step 710 can utilize the conventional level (four) method, the vertical (four) method, and the shirt of the present invention. The degree interpolation method and the 135 side interpolation method generate pixels 118, 119, η, (2), and a plurality of candidate estimated intensity values. These candidate estimated intensity values may be interpolated to calculate the _plug strength value, or the interpolation strength value may be further advanced by some calculation method (eg, averaging + calculation method). 'Determining step 72' to determine the green estimate intensity value of the pixel 118 based on the candidate estimated intensity value and the green intensity values of the pixels 112, 113, Μ, and m. In the decision step 72G, first, the color difference calculation step 7 is performed to calculate the plurality of color difference values ^ in the window W region centered on the second image f118, wherein the color difference calculation step 720a utilizes the following / ^J'+n [^J+n if « = -2 if w = 0,2 (^+»J' ~^J+n,j , if n = =-2 ~ ^ϊ+nj > if w = :0,2 ^i+n,jn~^i+nj-n , if n =—2 , Ri+n,j- -n-GUn,jn^ if n = 0,2 1 ^i+n,j+n~ ^ i+n,j+n , if n =-2 \Ri+n,j+n~^ilnJ+n^ if n = 0,2 where · ·, value ~, J represents the coordinate value of the pixel, R system The red intensity of the representative pixel is 'G' represents the estimated green intensity value of the pixel, 仏 represents the water representative of the pixel, i, and '彔 intensity value, & represents the vertical estimated green intensity value of the pixel, pixel at 45 The estimated green intensity value of the edge is 75. The representative pixel is 〇, . -> 135° 16 201034470 . The estimated green intensity value of the 135° edge. Next, an indicator calculation step 720b is performed to Calculate the indicator set ω,;={δ//,7, Δγ",Δ45°ί7.,Δ135.,"}, where the indicator factor is calculated The difference in chromatic aberration variation along the vertical, horizontal, 45°, and 135° directions in the W region is calculated by the following formula, but the present invention is not limited thereto. ^ =| ^g(U-2)- c^g(ij) |+| 〇,7)-ς^(ΰ·+2)|

AVU =| I +| ^g(ij)-^g(i+zj) I Δ135;. =|^f(/-2,y-2)-^f(w)|+|^J(/; 7VC (/+2, y > 2). Then, a selection step 720c is performed to select an estimated green intensity value (^, wherein the selection step 720c is selected from the candidate estimated intensity values according to the indicator group. The chromatic aberration change degree difference direction is the corresponding estimated intensity value, and the present embodiment is calculated by the following formula, but the present invention is not limited thereto:

If any two of the indicators = min (Ω,. y)

Gj, if Δ//, ν=πιΐη(Ω,ν)

G:j, if A^.=min(Q,.y) if Δ45; .=ιηΐη(Ω,ν) , if Δ135°. =min(Q,.y). In addition, it is worth noting that the present invention The color interpolation method 200, 300 and the image demosaicing method 500, 700 can be applied to a computer program product. When the computer program product is loaded into a computer, the computer can perform the color interpolation method 200, 300 and image solution. Mosaic method 500. The present invention has been disclosed in the above embodiments, and is not intended to limit the invention, and various modifications and refinements may be made without departing from the spirit and scope of the invention. The scope of the invention is defined by the scope of the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS 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 view showing the structure of a pixel array according to an embodiment of the present invention. Figure 3 is a block diagram showing the structure of a processing area in accordance with an embodiment of the present invention. Fig. 4 is a schematic diagram showing the flow of the color interpolation method according to an embodiment of the present invention. Figure 5 is a flow chart showing a color interpolation method according to an embodiment of the present invention. Figure 6 is a block diagram showing the structure of a pixel array in accordance with an embodiment of the present invention. Figure 7 is a block diagram showing the structure of a processing area in accordance with an embodiment of the present invention. Figure 8 is a schematic diagram showing the flow of an image demosaicing method according to an embodiment of the present invention. 18 201034470. Fig. 9 is a block diagram showing the structure of a pixel array according to an embodiment of the present invention. Fig. 10 is a view showing the structure of a processing area according to an embodiment of the present invention. Figure 11 is a flow chart showing an image demosaicing method according to an embodiment of the present invention. [Main component symbol description] 100: pixel array 102: processing regions 111 to 116: estimated pixels 117 to 125: unestimated pixel 200: color interpolation method 210: mask providing step 220: oblique intensity value calculation step 230: Oblique Strength Value Calculation Step 240: Estimation Parameter Calculation Step 250: Estimated Intensity Value Calculation Step 300: Color Interpolation Method 310: Mask Providing Step 320: Oblique Strength Value Calculation Step 330: Oblique Strength Value Calculation Step 340: Estimation parameter calculation step 350: Estimation intensity value calculation step 400: Pixel array 402: Processing area 500: Demosaicing method 510: Interpolation step 520: Decision step 520a: Color difference calculation step 520b: Indication factor 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 Factor Calculation Step 720c: Selection Step FG: Green Filter 19 201034470, FR: Red Filter Light sheet FB: blue filter «R: first pixel B: third pixel G: 'second pixel W: decision window i: coordinate value j: coordinate value

❿ 20

Claims (1)

201034470 VII. The scope of the patent application includes a plural number of inserts: ί like: [::: prime array 'this pixel, - the first intensity value (CU, each - the color of the - color - the second intensity ship pixel (four) Should be a second standard, i for each - the secret image, the vertical two represents each - the charm of the horizontal seat pair - each of the first color interpolation method includes: the second pixel corresponds to the ί! Γ color: = The interpolation step is calculated to include: one of the color estimated intensity values (6), the face = pixel array to provide - the high pass mask _ pass 第 display - the position of the pixel relative to one (four) pixel array edge 'And output -_, result (f);, line - the first oblique intensity value calculation step to calculate a first system according to = (")' where the first - oblique intensity value calculation step ... grasp, : proceed a second oblique intensity value calculation step to calculate a two-direction oblique intensity value (see ^7), wherein the second oblique intensity value calculation step is based on a formula: SIVS = GUXJ + GIJ + X; Estimating a parameter calculation step to calculate an estimated parameter ($), wherein the estimating parameter calculating step is based on a formula 21 201034470 NW SE') max siv , srv < \ hj i,j) min two NW SE, SIV . . , SIV ·. .\ , l, J hj) if otherwise ; and perform an estimated intensity calculation step To calculate the estimated intensity value (4) 'where the estimated intensity value calculation step is based on a formula: PU wide goods], +\(2Ri, factory hd2, j-2, . 2. 2. If the patent application scope The color interpolation method according to Item 1, wherein the same mask is a Lapalacian Mask or a Laplacian of Gaussian Mask. The color interpolation method according to item 1, wherein the first color is green, and the second color is red or blue. ^ 4· A computer program product, which is loaded into the program via a computer and executed Lu = the computer can Performing a color interpolation as described in the scope of the patent application, the color interpolation method is applicable to a pixel array, wherein the pixel array includes subtracted pixels, and the pixels include a plurality of pixels and a plurality of pixels Two pixels 'each-the first-pixels have a corresponding to - the first color - the first a color intensity value (GJ, each - the second pixels have a corresponding = second, color - second color intensity value (4), wherein - · represents the abscissa of each of the i pixels, i represents each of the pixels The ordinate includes: the color interpolation method includes: 22 201034470 performing an interpolation step for each of the second pixels to calculate a second pixel corresponding to the color - the interpolation step in the estimation The method includes: providing a high pass mask according to the pixel array to position a second pixel relative to an edge of the pixel array, and outputting a detection result (1); a first oblique intensity value calculation step to calculate a first oblique intensity value, wherein the first oblique intensity is based on a formula: grab d, j+Gij+', enter a second oblique intensity value Steps to calculate a first, oblique intensity value (〇, wherein the second oblique (four) degree value calculation step is based on a formula: mNE, grab sw, ij ^(sjvne>Sivsw) v Uj i,J Measuring parameter calculation steps to calculate an estimated parameter 'where the estimate The parameter calculation steps are based on a formula · (-_ NR L 45. Hj if />〇; and otherwise W ^ ^ ancient measured intensity value calculation step to calculate the estimated intensity (chemical) 'where the estimated intensity value calculation step is based on a formula: Ύ ") 4 (2 ά 2, / The method of color interpolation according to claim 5, wherein the first color is green and the second color is red or blue. ^ 8. A computer program product that loads the program and executes it via a computer. The computer can perform the color interpolation method as described in item 5 of the patent application. 9. An image demosaicing method for processing a pixel corresponding to a color filter array to reconstruct a color component lacking the pixel matrix, the image demosaicing method comprising: providing the pixel array, wherein the pixel array The method includes: a plurality of first-pixels, wherein each of the first pixels has a first intensity value (D; and a second pixel) of the first to the first color, wherein each of the second pixels has a pair Γ.—one of the second intensity values (heart), the second pixel pack : a plurality of estimated pixels, wherein the estimated pixels correspond to measured intensity values, the first estimated intensity values are intended to be the first color; and the first unestimated pixels are selected The unestimated pixel includes ^the target====_the estimated step is calculated to calculate the target pixel corresponding. The performing-inner intensity value' the intensity value estimating step includes a step to calculate the The number of target pixels 1 24 201034470 Candidate estimated material value, n These = candidate estimated intensity value and estimated intensity value method 1 ^. As stated in the patent application, the decision step is from The image demosaicing side is selected as one of the second estimated (four) candidate estimated intensity values. • If the patent application scope ninth method, the interpolation step includes: % like the mosaic mosaic party a high-pass mask to detect the position of the target image ί: ΓΓ edge and output - _ result (7); enter the 仃 - first - oblique intensity value calculation step, = 4 value (alarm where the first oblique intensity value The calculation step is based on saJ+Gl, H ·, and a second oblique step is performed according to one The second oblique intensity value calculation step, the intensity value ("ST^f", wherein the second oblique intensity value formula: SIVu=Gmj+Gij+] is used to perform an estimation parameter calculation step to calculate It makes the estimation parameter calculation step according to a formula: L hj (NW SE\ max SIV ., 57F. I, if /> 〇V l, J hj) (NW SE\ min SIV . . , SIV L otherwise K l, J hj and a candidate for calculating the intensity value calculation step to calculate the value of the calculation step for the candidate evaluations 25 201034470. The candidate value is +2Ί2. The scope of the patent application is η. Wherein the high-pass material is a Laplace material or a Gaussian; ^拉=克罩方> f rb, such as the image demosaicing method described in Item 9 of the U.S., wherein the (four) (four) comprises: providing a pixel according to the pixel array a high-pass mask, which is located at the edge of one of the pixel arrays, i to the position of the image edge, and outputs a detection result (7); a first oblique intensity value calculation step of the lamp to calculate - first oblique = degree value «), wherein the first - oblique intensity value calculation step is based on a SIV, ~ + G Performing a second oblique intensity value calculation step to calculate a second oblique intensity value «)', wherein the second oblique intensity value calculation step performs an estimation according to a formula: SIVu=Gi+lJ+Gl a parameter calculation step to calculate an estimated parameter, wherein the estimated parameter calculating step is based on a formula: L hj min NE SW^ SIV , SIV κ hj i, j and otherwise performing a candidate estimated intensity value calculating step Calculate the candidate estimates 26 201034470 • Measure the intensity value - go to ", ^ . According to the value of a ^ ^ (6) 'where the candidate estimated intensity value is calculated step Π, ,, hj-广RM, j+2 ). 4. The image demosaicing method of claim 13 wherein the high pass mask is a Laplacian mask α__ηΜ_). φ. The image demosaicing method of claim 9, wherein the first color is green and the second color is red or blue. 16. A computer program product, after loading and executing the program via a computer, the computer can perform the image demosaicing method as described in claim 9 of the patent application.