TW201234866A - Pixel interpolation circuit, pixel interpolation method, and recording medium - Google Patents

Abstract

A correlation along a horizontal direction and a correlation along a vertical direction are calculated with respect to a neighboring region around an interpolation target pixel by using respective pixel values of R pixels, G pixels, and B pixels in Bayer data. The pixel value of the G pixel to be interpolated for the interpolation target pixel is determined based upon pixel values of pixels adjacent to the interpolation target pixel along a direction exhibiting a stronger correlation between the calculated correlations. The interpolation target pixel in the Bayer data is interpolated by using the G pixel of the pixel value determined.

Description

201234866 VI. Description of the Invention: [Technical Field] The present invention relates to a pixel interpolation power for an image pickup apparatus or the like for use, and a color image pickup method using: =. 1 prime interpolation method and s recording medium [previous technique] Bayer's image of the camera image of the image element arranged by the single-plate arrangement; the image sensor of the image is interpolated and arranged in the Bayer data. Like the k number, you need (the pixel value). In particular, because of the color information other than the G/α color filter, the information of the G pixel, the information affects the brightness information, here, the previous resolution and the sharpness of the corridor. R-pixels and interpolation target pixels that are not present in the information (G昼 is described in the Japanese Patent Publications | Μ Interpolation G-Pictures] has opened the 2nd 135564 bulletin In the data of the plug-in object, the prior art system considers the interpolation of the pixels in the interpolation object; the correlation of the pixels in the surrounding area is in the [invention]. The invention is intended to be solved. However, when these vouchers are interpolated with G 昼 = = = 插 对 对 对 对 对 对 对 对 Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ Γ When the G pixel is interpolated, the correlation of the pixels is referred to, and the problem of the G element in the vicinity of the interpolation target is determined. The method is correctly interpolated in the interpolation object. - 201234866 Therefore, the purpose of the λ 义 ♦ ♦ 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本The data, the correlation calculation unit, and the eigenvalues, calculate the interpolation target 佥I—, G pixel, and B昼Associated with each of the vertical direction of the drawing ^ horizontal 'direction of the peripheral area

The pixel of the pixel; and the book sound 'The interpolation object has no G # , $ prime value determination unit, and the i# calculation unit calculates the correlation between the correlation and the correlation. Related. Twins & + ^ The interpolated object pixel adjacent to the book Xin: "strong direction, based on the pre-interpolated object pixel; c, decided to interpolate the above-mentioned composition Painting: supplementing the pixel value determined by the above-mentioned pixel value determining unit. The invention JL #, 7 uses the Bayer data in the form of R, which includes: the first step, which is such that the calculated genus does not exist. f, G pixel and B pixel each pixel value, the horizontal direction of the domain, the pixel of the pixel of the pixel is in the surrounding area

The correlation between the correlation and the vertical square A is related to the above-mentioned first, the correlation; the second step ~ the direction of the long-term h M i calculated by the first step, and the dependence of each correlation in the correlation The strong value refers to the pixel value of the pixel adjacent to the pixel of the interpolation object; and the pixel interpolation of the G element of the interpolation object in the third step is (4) 2Λ according to the aforementioned Bayer data. The image signal, 其他 决定 决定 决定 决定 决定 决定 决定 疋 疋 疋 疋 其他 其他 其他 疋 其他 其他 疋 疋 疋 疋 其他 其他 其他 其他 其他 疋 其他 疋 其他 疋 其他 其他 其他 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素 素士 ' ° /, in the direction of the straw by the aforementioned i-th processing, according to the strong disc, according to the disc 5, + according to the interpolated object pixel

, the value of the denier, the imputation of the G value of the imputed object pixel and the third treatment, the vertical will be auspicious 2" >, you are based on the monthly Bayer data number 'interpolation by The effect of the invention of the pixel value determined by the second processing described above is based on the present invention, and the implementation method of the present invention can be explained by correctly interpolating the data of the Bayer data. First, the reference is made to the present invention. The overall configuration of the digital camera 1 。 The imaging lens 62 includes a focusing lens and a zoom lens, and a lens driving block 63. The lens driving block 63 is driven by the focusing lens and the zoom lens respectively. The parallel focus motor and the zoom motor are configured to be driven by a focus drive driver that drives the focus motor and the zoom motor from the CPU No. 70. The aperture shutter 64 includes a drive circuit (not shown), and the path is controlled by the CPU 70. The signal is used to make the aperture fast. In addition, the aperture shutter 64 functions as a color filter provided on the light receiving surface in the CCD 50 of the imaging element after the aperture and the shutter. The CCD 50 passes through the imaging lens 62 and The light of the scene projected by the circle 64 is converted into an electric signal, and the converted number is output as an image pickup signal to the image processing unit 20. Further, the image of the image of the picture of the pixel of the pixel in the first correlation is G. Fig. 1 shows the control actuator connected to the optical axis of the unillustrated and changes the driving of the electric actuator. Yes, Bayer arranges the telecom after the shutter> CCD50 201234866 according to TG66 The drive is driven in accordance with a timing signal of a specific frequency generated by a control signal from the CPU 70. The memory card 8 is detachably connected to the card I/F 76. The memory 69 stores a program required to control the respective portions of the digital camera 1 by the CpU 70. The DRAM 68 is used as a buffer memory for temporarily storing image data of the CPU 70 after being captured by the CCD 5 ,, and is also used as a working memory of the CPU 7. The image display unit 7 is composed of a liquid crystal display and its driving circuit. When the digital camera 100 is in the shooting standby state, the scene captured by the CCD 50 is displayed as a passing image on the liquid crystal display. The image display unit 7 is in a digital (four) 1〇〇 reproduction map. When the image is 'displayed, the image read from the memory card 8〇 is displayed on the liquid crystal display. The input button 72 includes: a shutter button, a mode button, a button, a cross button: a plurality of operation keys such as a shooting mode selection button, and will be used. The operation signal of the key operation is output to the CPU 70. The CPU 70 controls the entirety of the digital camera 1 to perform various actions by still collaborating with the purchase of the spotted memory from the memory 69. The structure of the image processing unit 2A will be described with reference to Fig. 2. The image processing unit 20, i, ι 20 is balanced by the analog processing unit (CDS, AGC, A/d) 21, ..β_ Ρ 22 The adjustment unit 23, the γ correction unit 24, the pixel interpolation unit 30, the yuv #towing decoration plus conversion processing unit 27, and the JPEG conversion unit 28 are configured. The analog 'correlated double sampling circuit (CDS) output from the CCD 50 in the image processing unit 20 reduces noise included in the image pickup signal of the signal. Since 201234866, the dynamic gain control circuit (AGC) adjusts the gain of the image signal after the noise is reduced. The A/D converter (A/D) converts the gain-adjusted imaging signal into a digital signal, that is, into a Bayer array of color filters (displaying red information pixel data), G (display) Bayer data for the pixel data of the green information) and B (the blue information of the blue information). The black level alignment unit 22 sandwiches the Bayer data into a specific black level. The white balance adjusting unit 23 performs gain adjustment for each of R, G, and B on the Bayer data after the sandwiching, and performs adjustment of the white balance. The correction unit 24 performs the γ characteristic on the Bayer data after the white balance is corrected ( Gray scale characteristic) correction. The pixel interpolation unit 30 generates an image signal for the information of the Bayer data interpolated pixel value after the 丫 characteristic correction. The detailed content of the pixel interpolation unit 3 产生 is generated by the luminance signal. (Υ) and the color difference signal (u, i '5 children signal. The JPEG conversion unit 28 is formed by an image code of jpEG^ shape. The compression-encoded image 彳_ < signal is compressed and compiled. It has been recorded on the memory card 80. The book, ηι图 'details the detailed description of the pixel interpolation unit 3°. ==:=::=;:, the second electric one. The first circuit part 3. & Broadband 6 加 加 加 36 36 Μ Μ Μ ... 播 播 播 播 播 播 播 播 播 播 播 播 播 播 播 播 播 播 播 播 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 When the data is generated as a color map, when the patent is the same, at the B-value, the pixel is interpolated in the pixel. When the image letter is still frequent, although the sharpness is due to Bayer's 3:6, the signal is transmitted from the γ correction unit 24 for the reason, and then the circuit is 第 σ | The image signal of the image signal generated in the denier interpolation unit 30 is generated by the signal processing of the information of the missing element. The signal of the first φ circuit part 3 0 a The processing is also performed by the signal that is disclosed in the Japanese Patent Publication No. JP-A No. Hei. In the R pixel of the information of the prime, and the element of the G pixel (hereinafter referred to as "the subject of the target"), only the pixel value of the G pixel around the pixel of the broadcast object is referred to. When the pixel value of the G pixel is interpolated in the pixel, the reference value afl decreases. As a result, the correlation between the pixel values of the interpolation pair and the pixel value of the interpolation cannot be correctly determined, and the pixel of the interpolation target cannot be accurately obtained. The correct pixel value of the pixel. The unsuitable G pixel becomes an isolated point, sometimes in the image signal. Therefore, '& The component removes the image signal (noise component, etc.) by passing the number through the LPF 26 of the low-pass chopper. However, the image signal is passed [Μ% to remove the noise component, and The image in the image signal is lowered. The reproduced image according to the image signal becomes blurred. Therefore, the second circuit portion 3〇b, the material sent from the γ correction unit 24 is separately extracted. Then, in the adder, the image of the image signal passing through the LPF 26 is added to the image of the pixel, and the information of the image is affected by the sharpness of the reconstructed image _,

In this way, it is emphasized that the reproduced image based on the Bayer data is not blurred. W ~ 9 - 201234866 ", - person, in the second circuit part 30b. In the second circuit unit 3 〇b, the number processing is detailed in response to the interpolation object book in the Bayer data (the ruler is used for extraction). The information of the interpolation element has no circuit. As shown in Fig. 4, the reference image generating unit 41 and the epoch-the circuit 40 of the 33 "quality value" are defined by the horizontal direction cliff 41, the vertical direction reference direction correlation calculating unit 43, and the vertical 邛42. The level value determining unit 45 and the book extracting unit A. The visibility calculation unit 44 and the pixel weight I capsule output unit 46 are configured. The horizontal direction reference image generating unit 41: the binary value of the pixel value of the pixel connected to the horizontal direction = the use of the reference image. The horizontal direction reference map = the direction in which the horizontally interpolated object pixel is obtained and used in the direction of the Bayer data, and the correlation is strong. In particular, the horizontal direction reference = each interpolation object The paintings around the pixels:::: er data to any, rational, ^ = (A1) is in the horizontal direction adjacent to the pixel of the g-pixel pixel value minus the level - ', pixel, The processing of the average value of the pixel values of the element is adjacent to the G picture (A2) of the R pixel, and is adjacent to the horizontal direction of the r in the horizontal direction adjacent to the G picture, and subtracts the G from The average value of the pixel value of the pixel of the pixel of the pixel, (A3) is adjacent to the horizontal direction: square = ... pixel of the picture element = the pixel value of the G pixel is subtracted Processing + mean, -10- 201234866

(A4) is in the horizontal direction adjacent to the GR when the sound is estimated to subtract the pixel value of the G pixel adjacent to the element in the horizontal direction from the pixel value of the pixel. Processing of the average value Fig. 5 shows the position of the pixel of the pixel of the Bayer data, which is marked with B, and is located in the Bayer data. G1, G3, · · .G47, etc., where G is noted, is placed in the Bayer data. R8, R10, · · · R4 — Not 诅 Han 4〇 # The position of the R note is the position where the R element is placed in the Bayer data. Further, in Fig. 5, b 〇, R 、, Τ β2, · · · Β 48 show the pixel values of the respective pixels. ^ G1 G3' · · · G47 displays the pixel values of each G pixel. R8, Ri〇, · · . R4 shows the prime value of each R pixel. ‘The horizontal direction reference 40H is shown in each of the 6 figures in the Bayer data. The pixel of the position of the horizontal direction reference image (at the time of the interpolation target 40H, attention is paid to the peripheral region where the target pixel is interpolated when the horizontal reference image is generated in the fifth. The horizontal reference of the example of the first image The image 40H is the horizontal reference image of the "R24 pixel" shown in Fig. 5, such as the 7 χ 7 pixels of the peripheral region of the pixel at the position R24 where the horizontal reference image is not displayed. B〇, G1+: B2, · · · The location of each element of the Β48 position. In addition, the surrounding area may not be an area composed of 7x7 pixels. Next, an example of each component in the horizontal direction reference image will be described for each component in the horizontal direction reference image 4A. 201234866 The component "h8" in the horizontal direction reference image 40H is obtained by the processing of (昼1;) of the pixel values of G7, R8, and G9 shown in Fig. 5, and becomes the following value. H8 = R8 - (G7 + G9) / 2 The component r h9" in the horizontal direction reference image 40H, by using the processing of the pixel values of R8, G9, and R1 shown in Fig. 5 (A2), And become the following value. H9 = (R8 + R10)/2 - G9 The component "hl5" in the horizontal direction reference image 40H is processed by using the pixel values of B14, G15, and B16 shown in Fig. 5 (A3). Become the following value. Hl5 = (B14 + B16) / 2- G15 The component "hl6" in the horizontal reference image 40Η, by using the processing of the pixel values of G15, B16, and G17 shown in Fig. 5 (A4) Become the following value. Hl6 = B16- (G15 + G17) Similarly, the other components in the horizontal direction reference image 40H can be obtained by performing the processes of (A1) to (A4). In addition to the processing of (A1) to (A4), the pixels surrounding the pixel of the Bayer data can be processed by any of the following (A5) to (A8). A horizontal reference circle image is generated. (A5) The R pixel adjacent to the G pixel in the horizontal direction is subtracted from the average value of the pixel value of the G pixel adjacent to the R pixel in the horizontal direction, and the pixel value of the R element is subtracted. Handling 201234866 (A 6 ) ί especially in the horizontal direction Hao η G G is asked by the R pixel (} pixel, subtracted from the pixel value of the pixel in the ten direction of the horizontal date The treatment of the average value of the denier value of the R picture of G 昼 ( (A 7 ) is adjacent in the horizontal direction; J straight · ^ η ▲ .j. The door is adjacent to the G pixel of B pixel, from the element The processing of the prime value minus the average value of the pixel values of the Β 邻接 邻接 你 你 你 你 你 你 你 你 你 你 你 你 Α Α Α Α Α Α Α Α Α Α Α Α Α Α Α Α Α Α Α Α The process of subtracting the pixel value of the B element from the average value of the pixel values of the G pixels adjacent to the pixel in the horizontal direction, and the vertical reference image generating unit 42 The Bayer data is used to perform a vertical direction reference image using an operation of a pixel value adjacent to a pixel in the vertical direction. The vertical reference image is used for each interpolation target pixel in the Bayer data. - v q κ 1 冢 之 Λ Λ Λ 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 ' 5 ' ' ' ' ' ' ' ' ' ' ' ' ' ' 5 5 5 5 5 5 5 5 5 5 (Β1) to (Β4) any _ processing, and a vertical direction reference image is produced. (Β1) The rule element adjacent to the G element in the vertical direction is subtracted from the pixel value of the R element in the vertical direction. The process (B2) in which the average value of the pixel values of the g-crystals adjacent to the R element is adjacent to the G element of the R pixel in the vertical direction, and the R drawn adjacent to the G element in the vertical direction The average value of the prime value of the element, the process of subtracting the pixel value of the G element (B3) is adjacent to the g element of the β pixel in the vertical direction, and is adjacent to the G element in the vertical direction. The average value of the pixel value of the halogen, the process of subtracting the pixel value of the G element from -13 to 201234866 (B4) is adjacent to the b pixel of the g element in the vertical direction, the pixel value of the B pixel The process of subtracting the average value of the pixel values adjacent to the B element in the vertical direction is shown in FIG. 7 'displaying the vertical direction reference image 40V as the vertical reference image Here, each component of the vertical direction reference image 4〇v will be described. The component "V8" in the vertical direction reference image 40V is obtained by making the pixel values of G1, R8, and G15 shown in FIG. The processing of B1) becomes the following value: v8 = R8 - (Gl+Gl 5)/ 2 The component "vl5" in the vertical reference image 40V, and the pixels of R8, G15, and R22 shown in Fig. 5 The value of (β2) becomes the following value: V15 = (R8 + R22) / 2 - G15 The component "v9" in the vertical reference image 40V, by making B2, G9, B16 shown in Fig. 5 The processing of the prime value (B3) becomes the following value. V9 = (B2 + B16) / 2- G9 The component rvl6" in the vertical reference image 40V is represented by the following (Β4) of the pixel values of G9, Β16, and G23 shown in Fig. 5 value. V16 = B16- (G9 + G23) / 2 Similarly, other processing in the vertical direction reference image 40V can be obtained by performing the processing of (B1) to (B4). From the use of the G 昼 straight, and the use of the use, the use of the first, and the use of the rules, points -14-201234866 In addition to the processing of (B1) to (B4), the Bayer data in the vicinity of the interpolation object Each of the elements can also produce a vertical beauty image by performing any of the processing of (B5) to (B8) in sequence. The bauxite, (B5) is adjacent to the R element of the G element in the vertical direction, and subtracts the R element from the average of the pixel values of the G pixel adjacent to the R pixel in the vertical direction. The processing of the prime value (B6) is adjacent to the G pixel of the R pixel in the vertical direction, and subtracts the pixel value of the ruler of the g pixel adjacent to the g pixel from the pixel value of the G pixel. The processing of the average value (B7) is adjacent to the pixel element of the B pixel in the vertical direction, and subtracts the pixel value of the B pixel adjacent to the g pixel in the vertical direction from the pixel value of the G pixel. The processing of the average value (B8) is adjacent to the b pixel of the G pixel in the vertical direction, and subtracts the B pixel from the average value of the pixel values of the G pixel adjacent to the B pixel in the vertical direction. The processing of the pixel value: the direction correlation degree calculation unit 43 calculates the correlation degree 1 in the horizontal direction and the second correlation degree Hr from the horizontal direction reference image. The horizontal direction correlation degree calculation unit 43 calculates the horizontal correlation degree 藉 by executing the processing of - people (C1) to (C3). In the dry (C1), the difference value between each component in the horizontal direction reference image is obtained (C2)'. The absolute value of each difference value obtained in (C1) is calculated. (3) Calculate the (four) value obtained in (C2) by multiplying the specific coefficient t and 'as the horizontal correlation degree Hr. -15- 201234866 When the processing examples of (C1) to (C3) are listed, regarding the fifth figure The horizontal correlation Hr of the pixel at the "R24" position is as follows.

Hr=abs(h8-h9)+2xabs(h9-hl0)+2xabs(hl0-hll)+abs(hll-hl2)+2x{abs(hl5-hl6)+2xabs(hl6-hl7)+2xabs(hl7- h 18) + abs(hl8-hl9)} + 2χ {abs(h22-h2 3) + 2xabs(h2 3-h24) + 2 xabs(h24-h25)+abs(h25-h26)}+2x{abs( H29-h30)+2xabs(h 30-h31)+2xabs(h31-h32)+abs(h32-h33)}+abs(h36-h37)+2 xabs(h37-h38)+2xabs(h38-h39)+ Abs(h39-h40) where abs() represents the operation for finding the absolute value. The smaller the value of the horizontal correlation Hr, the stronger the correlation in the horizontal direction. The vertical direction correlation degree calculation unit 44 calculates the vertical correlation degree Vd of the correlation degree in the vertical direction from the vertical direction reference image with respect to each interpolation target element in the Bayer data. The vertical direction correlation degree calculation unit 44 calculates the vertical correlation degree Vd by performing the processes of (D1) to (D3). (D 1 ) The difference value between the components in the vertical reference image is obtained. (D2) The absolute value of each difference value obtained in (D 1 ) is obtained. (D3) Calculate the sum of the absolute values obtained in (D2) multiplied by the specific coefficient as the vertical correlation Vd. When the processing examples of (D1) to (D3) are listed, the vertical correlation Vd of the pixels in the position "R24" shown in Fig. 5 is as follows.

Vd = abs(v8-vl5) + 2xabs(vl5-v22) + 2xabs(v22-v29) + a bs(v29-v36) + 2X {abs(v9-v16) + 2xabs(vl6-v23) + 2xabs(v23 -v30)+abs(v30-v37)}+2x{abs(vl0-vl7)+2xabs(vl7-v24)+ 2xabs(v24-v31)+abs(v31-v38))+2x(abs(vll-vl8) )+2xabs( -16- 201234866

Vl8-v25)+2xabs(v25-v32)+abs(v32-v39)}+abs(vl2-vl9)+ 2xabs(vl9-v25)+2xabs(v26-v33)+abs(v33-v40) where 'abs 〇 denotes the operation of finding the absolute value. The smaller the value of the vertical correlation Vd, the stronger the correlation in the vertical direction. The pixel value determining unit 45 determines the pixel value of the G pixel of the interpolation of the interpolation target pixel based on the horizontal correlation degree Hr and the vertical correlation degree Vd. When the horizontal correlation degree Hr is smaller than the vertical correlation degree Vd, in other words, when the correlation in the horizontal direction is stronger than the vertical direction, the pixel value determining unit C determines the neighboring pixels in the horizontal direction of the interpolation target pixel. ^ The average value of the value is used as the pixel of the Gt element of the interpolation target. In this case, the pixel value determining unit 45 displays the image shown in Fig. 5, ϋ 4 t (interpolated pixel of interest). "The pixel value of the G element of the interpolation" is the pixel value of the G pixel adjacent to the "R24" position in the horizontal direction ^G23", and "G25, the m value is determined." G25", The following pixels G24 = (G23 + G25) / 2: The outer 'vertical correlation Vd is lower than the horizontal correlation Hr... When the correlation between the value and the straight direction is stronger than the horizontal direction, 佥: 昼The average value of the prime value of the prime = the pixel value of the element adjacent to the vertical direction.乍 is the position 2 of the interpolation target pixel interpolation: The pixel value determination unit 45 performs the interpolation of the "R24: set pixel (interpolation target pixel) interpolation shown in Fig. 5" The following pixel values are determined by using the book 201234866 y, which is adjacent to the "R24" position in the vertical direction, and the pixel values "〇17" and "G31" of the prime. G24 = (Gi7 + G3 1) / 2 The pixel extracting unit 46 extracts the pixel determined by the pixel value determining unit as the pixel value of the book of the interpolation target pixel interpolation in the Bayer data. A > G pixel extraction circuit 40 except for the pixel at the "R24" position: all the interpolation elements in the Bayer data (the R pixel of the G pixel information is not in the R pixel, " The process of the above description is carried out, and the pixel value of the G gas which is to be interpolated for each interpolation object is extracted. That is, the G-cell extraction circuit 4〇 is all inserted into the (four) symbolic elements, and the horizontal direction reference image and the vertical direction reference image are generated, and the horizontal direction reference image and the vertical direction reference map are used. In the image, the correlation between the pixel value in the horizontal direction and the pixel value in the vertical direction is calculated, and the pixel value to be interpolated for each interpolation object is extracted. This processing, and the date of the ^, _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ In the second circuit portion 3b, the G pixel is extracted from the G image signal through the HpF37 of the high-pass filter, and the low-frequency component of the G image signal is removed. For performing signal processing in the first circuit portion, 诵埚T ^ 2 & through the color image signal of the LPF (Low Pass Filter) 26, removing the low frequency Λ8 + ^ Batu. & G component of the frequency component Signal, that is, the implementation of interpolation. " Go to the description of 3 i, when using this embodiment, except in the interpolation pair ", Outside the G pixel, also in reference to -18-201234866 surrounding the interpolation target pixel day

R pixel and B佥% I

A prime, the more I value of the G book Xin interpolated with Bayer data. In this way, the stalking 丨 丨 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一value. In the present embodiment, the pixel values of the illustrator, the R pixel, and the B element are used, and the quasi-image and the vertical direction reference image are used, and the horizontal and vertical images are used. The component in the direction reference image is calculated as the correlation between the correlation between the pixel value and the pixel value in the vertical direction. When the correlation with the pixel value is different, the ^, ° of the reference pixel As a result, the possibility of misjudging the direction of the correlation of the pixel values can be reduced. Further, in the present embodiment, the signal processing is performed in the first circuit portion 3a, and the color image k number of the LPF (Low Pass Filter)% is interpolated in the second circuit portion 3b. The correct information of the G-salt extracted in the middle. Since the information of the G pixel affects the sharpness of the outline in the reproduced image, the image quality of the reproduced image can be improved by emphasizing the rim of the reproduced image based on the Bayer data. <Modified Example> This is a modified example of the present embodiment. Just in the case of interpolating the elements around the pixel, there is a small difference in the correlation between the correlation in the horizontal direction and the vertical direction. That is to say, in the vicinity of the imputation object, sometimes the gray level is moderated. At this time, according to the processing of the first embodiment, when the pixel of the interpolation object of Bayer's feed is interpolated with G element, the value of the interpolation object forms a distinct value (isolated point), and According to the problem of deterioration of the quality of the reproduced image of the image signal, the present modification corresponds to the problem. -19- 201234866 In this modified example, the difference from the above embodiment is as shown in Fig. 8 m叱_ The mixture ratio determination unit 47 is provided only in the output circuit 40. The other configuration and processing system are the same as the above-described embodiment. The mixture ratio determination unit 4 transmits the τ τ mountain, The horizontal correlation degree Hr calculated from the horizontal direction correlation degree calculation unit 43 is determined from the vertical correlation degree Vd calculated by the verticals correlation coefficient calculation unit 44 to determine the mixture ratio. The horizontal correlation degree Hr vl·· + Dt t , ^ 4 H is smaller than the vertical correlation Vd, that is, the correlation in the horizontal direction is greater than the vertical square Θ4 day. When the correlation is strong, the mixture ratio determining unit 47 determines the value of Hr/Vd as the Λ丨 ~ horizontal phase, for mixing & ratio, in addition, vertical correlation Vd is more than '^10 Hr is small, and when the correlation between the county and the ancient times is strong, the mixture is two: the correlation ratio is higher than the horizontal direction. The ratio of H to the 疋 determines the value of Vd/Hr as the correlation between the flat direction and the vertical direction. In the case of strong time, the pixel and the fixed portion 45 determine the mean value of the pixel value of the 佥/in water + direction-pixel of the G 昼 邻接 邻接 依据 依据 插 插 插 插 插 插 插 插In the 7-direction and the vertical direction, the pixel values of the G-pixels of the four pixels of the interpolation target pixel (the mean value of the 5 gold tones 〇#音_^^) are adjacent to the pixel of the interpolation object. For example, the pixel value "G24" of the G element of the pixel interpolation at the R24j position shown in Fig. 5 becomes the following pixel value. 024 = (1.0 - (Hr/ Vd )) X(G23 + 025)/2 + rw / Vd)X (Gl7+G31 + G23+G25)/4 When the correlation between the vertical direction is stronger than the horizontal direction, check and 疋. P 45 decides to add the average value of the pixel values adjacent to the pixel of the interpolation target according to the ratio of the mixture ratio: -20- 201234866 The pixel value adjacent to the average value of the interpolation in the horizontal direction and the vertical direction G pixel, the pixel value of G pixel. The four G pixels of the complement pixel are the pixels of the R24" position of the interpolation target pixel. The following pixel values. + G31) / 2 + (Vd plug

At this time, as shown in Fig. 5, the pixel value "〇24" of the G pixel is G24 = (1.0 - (Vd/ Hr)) x (Gl7

Hr)x (G17+G31+G23+G25)/4 The pixel extracting unit 46 extracts the pixel value determined by the pixel value determining unit as a G picture to be interpolated by the pixel of the interpolation target in the Bayer data. Prime prime value. Thus, an image signal (G image signal) composed only of the information of the correct pupil extracted from the Bayer data is obtained. Then, returning to Fig. 3, in the adder 36, the signal processing is performed in the first circuit portion 30a, and the color image signal of the LpF (low pass filter) 26 is added with the G image signal, that is, Implement interpolation. According to the modified example described above, when the pixel value of the G pixel of the Bayer data interpolation is determined, the average value of the two pixels adjacent to the interpolation target pixel is in the direction of strong correlation, The average value of the four pixels adjacent to the interpolation target pixel in the horizontal direction and the vertical direction is also used. Thus, even if the correlation between the correlation between the horizontal direction and the vertical direction is different, the pixel value of the interpolation target pixel does not become a significant value (isolated point), and the reproduction map according to the image signal can be prevented. The temperament of the image deteriorates. The present invention is not limited to the above-described embodiments and modifications, and includes design changes that do not depart from the gist of the present invention. -2 1- 201234866 [Simple description of the drawing] Fig. 1 shows the overall structure of the digital camera 100. Fig. 2 is a view showing the configuration of the image processing unit 20. Fig. 3 is a view showing the structure of the pixel interpolation unit 30. Fig. 4 is a structural diagram of the G gas extraction circuit 40. Figure 5 is a diagram showing an example of a Bayer arrangement. Figure 6 is a diagram showing an example of a horizontal reference image. Figure 7 is a diagram showing an example of a vertical reference image. Fig. 8 is a structural diagram of a G-halogen extraction circuit 40 in a modified example. [Description of main component symbols] 20 Image processing unit 21 Analog processing unit 22 Black level adjustment unit 23 White balance adjustment unit 24 γ correction unit 26 LPF ( Low-pass filter) 27 YUV conversion processing unit 28 JPEG conversion unit 30 Prime interpolation unit 30a First circuit unit 30b Second circuit unit 31 Broadband LPF 32 G interpolation circuit 33 R interpolation circuit 34 B interpolation circuit - 22- 201234866 36 37 38 40 40 41 42 43 44 45 46 47 50 62 63 64 66 68 69 70 71 72 80 100

R, G, B adder

HPF subtractor

G-dimensional extraction circuit, pixel interpolation circuit, horizontal direction reference image generation unit, vertical direction reference image generation unit, horizontal direction correlation calculation unit, vertical direction correlation calculation unit, pixel value determination unit, pixel extraction unit, mixture ratio determination unit, CCD

Camera lens lens drive block aperture shutter TG

DRAM memory

CPU image display part key input part memory card digital camera element, signal -23- 201234866 76 card I/F 40H horizontal direction reference image Hr horizontal correlation Vd vertical correlation degree Y brightness signal U, V color difference signal CDS correlation double Sampling Circuit AGC White Gain Control Circuit A/DA/D Converter-24-

Claims (1)

With: 201234866 VII, the scope of application for patents: 1 · A morpheme interpolation electric pixel interpolating part, its Zuo # A m area ... Su Zhishu: use only an interpolation pair "Cha Xin's G ^ value, and Deciding to interpolate the former Hungarian value, and interpolating the determined G-value by the interpolation target pixel, so that the pixel value in the interpolation object is not present in the interpolation object pixel The correlation calculation unit is a peripheral value or a value of the peripheral region of the above-mentioned pixel of interest, which is the value of the pixel of the left-eye pixel, the G pixel, and the B element. The calculation is performed to calculate the correlation between the interpolation in the horizontal direction of the region, the vertical/in particular, and the upper pixel, and the pixel value of the peripheral element is different from the above-mentioned target sales. The pixel value of the peripheral pixels; the pixel value determining unit determines the pixel value of the pixel adjacent to the object element in the direction in which the correlation in each correlation calculated by the correlation is strong. The second G interpolation imputed value of the G 昼 插 ; ; ; ;; The image acquisition unit' is configured to acquire a G image. The ten image signal system is composed only of G pixels having the G interpolation pixel value by the pixel value determining unit 2; the high pass filter is removed. The low-frequency component of the aforementioned G image signal obtained by the above-mentioned Ο image letter; and the 图像 图像 之 借 借 借 借 借 借 借 借 借 借 借 借 借 借In the peripheral correlation, the prime color calculation unit interpolates the interpolation: r number, the G-determined number acquisition unit -25-201234866, the addition unit, which is generated by the pixel interpolation unit. Image: plus the second G interpolated pixel of the capture k of the low-frequency component by the aforementioned high-pass chopper; wherein the image signal causes the second G interpolation 昼$ mt } = The pixel value is determined to be the pixel value obtained by the specific value and the other average value, and the correlation coefficient calculation unit is adjacent to each of the correlations calculated by the second calculation. Draw the book on f~ then the mean value of the horizontal and vertical values of the other flattened pixels . Pre-target pixel neighbor 2. The pixel interpolation circuit according to the first item of the patent application, which further includes: a first reference image generating unit, a peripheral region of 1 乂*, and a neighboring prime value is used: And the reference image generation unit; and the reference image generation unit, the peripheral region of the pelt, the calculation of the pixel value of the interpolated object, and the second reference image is generated in the straight direction. The phase correlation calculation unit calculates the *quasi-image generation based on the first reference image generated by the front portion, and based on the second reference, the correlation in the j + direction, the reference image, Calculate the phase in the second 罝 direction generated by the vertical generating portion. 3. As claimed in the patent scope 拴. The ith complement circuit, the k-division 1 reference 产 image production unit, obtains the 昼 昼 昼 】 周边 周边 周边 周边 周边 周边 周边 周边 周边 周边 周边 周边 周边 周边 周边 周边 -26 -26 -26 -26 -26 -26 -26 -26 -26 -26 In the horizontal direction and the difference value of the second value, the further average value is the second element of the respective pixel values adjacent to the second element and the third pixel is based on the aforementioned level #八生前第一第一基The difference between the obtained difference values and the radiance: 第 The second reference image generation unit includes a vertical difference calculation unit that obtains the vertical direction of the first pixel by drawing the pixel value and the other region of the interpolation target. And the "differential value of the + sentence value" is further the average value is in the picture; the average value of the pixel values of the fourth pixel and the fifth element adjacent to the first pixel: The second reference image is generated by each of the difference values obtained by the vertical difference calculation unit. The pixel interpolation circuit of the second aspect of the present invention, wherein the first soil image generation unit includes: a horizontal difference The computing unit' obtains the pixel value of the R pixel: the G picture adjacent to the R pixel in the horizontal direction The difference value of the mean value; the tf 2 horizontal difference operation unit' obtains a difference value between the pixel value of the B pixel and the mean value of the pixel value of the G pixel adjacent to the B pixel in the horizontal direction; The difference between the average value of the pixel value of the R element adjacent to the 0 pixel in the horizontal direction and the difference between the value of the pixel element and the fourth level difference operation unit is obtained in The difference between the average value of the elemental value of the halogen adjacent to the G element in the horizontal direction and the elemental value of the G element; -27- 201234866 based on the difference value of the above-mentioned 4th to 4th water, the left运 运 产生 产生 产生 产生 产生 产生 产生 产生 产生 产生 产生 产生 产生 产生 产生 产生 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第The value of the difference between the average value of the G-sound of the pixel of the state; the second vertical difference operation of the constant pixel value, which obtains the mean value of the pixel value of β and the adjacent two in the vertical direction. The difference value; the vertical difference calculation unit of the pixel value of the G-pixel of the Beijing, which is loaded with rain The book of the R-character of G-pixels is used to estimate the difference between the difference value of the neighboring value and the value of the pixel value in the vertical direction and the 垂直=4 vertical difference operation unit of the G-pixel, which is connected to the G pixel. 3 book 伢隹! The difference value of the adjacent prime value in the mutual direction "the average value of the prime value of the prime and the basis of the G pixel are obtained by the difference value of the first $, and the vertical difference operation unit of the flute is generated. The obtained reference image calculation unit includes: a parametric interpolation circuit of the item, wherein the correlation first calculation unit, 1 is followed by the value I- to: 'first correlation value': the absolute value of the first correlation score i, the difference I obtained by the horizontal difference calculation unit, the sum of the two values multiplied by the value of the specific coefficient; and the second calculation unit The total value of the total value is the sum of the second correlation value, the absolute value of the second correlation score, and the sum of the values of the difference and the specific coefficient obtained by the straight difference calculation unit; 28-201234866 Calculate the correlation between the first correlation value calculated by the second calculation unit and the correlation between the borrowing and the The correlation value of the direction is calculated, and the horizontal direction is calculated. 6. The check value determining unit of the third paragraph of the patent application scope will be in the direction of the interpolation circuit, in which the aforementioned correlation is strong. The average value of the pixel values of the pixels of each phase determines the adjacent pixel values of the object pixels. In the case of the second application of the second calculation unit, Calculate the first rate. The ratio of the second correlation value is 8. The pixel interpolation method includes the following steps: Only the interpolation value of the g-study of the peripheral region of the object pixel is used, and the interpolation object is interpolated. The element of G is a prime value, and the interpolated object is interpolated by the interpolation of the determined pixel of the G-pixel of the G-pixel, which is determined by the G-pixel, to generate an interpolated image. a signal in which the pixel value of the G pixel in the Bayer data is non-existent; 'for the aforementioned Bayer data', the r-pixel, the 〇 素, B 昼The pixel value of the pixel of the pixel is calculated by calculating the pixel value of the surrounding element of the pixel of the interpolation target, thereby calculating the correlation in the horizontal direction of the peripheral region of the interpolation target pixel. Correlation between sex and vertical direction, wherein the pixel value of the surrounding pheromone is a pixel value of a peripheral pixel of a color component different from the color component of the above-mentioned attention pixel; -29- 201234866 will be specified The ratio is added to the average book tone and the 4 J value and the other value are obtained. The second correlation of the pixel: the median value, where 'the average value is the average value of the pixel values adjacent to the interpolation target pixel in the direction in which the correlation between the calculated phases is strong' The other value is an average value of the pixels adjacent to the interpolation target pixel in the horizontal direction; ... obtains a G image signal, and the G image signal is only determined by the former fan. The G pixel of the interpolation pixel value is configured; the low frequency component of the G image signal obtained as described above is removed; and the interpolation image signal generated by the October description is added to the low frequency component, '! The second G interpolated pixel value of the G image signal removed. 9. The type of recorded media 'memory has a computer readable program, which causes the computer to perform the following processing: only use the pixel value of the G pixel of the peripheral region of the interpolation object and interpolate And interpolating the pixel value of the G element of the pixel of the interpolation object, and interpolating the pixel value of the determined G element into the interpolation object A to generate an interpolation image signal, wherein the interpolation image signal is generated. The complement of the object is a non-existent element of the g-pixel of the g-pixel in the Bayer data; for the Bayer data, the R-pixel, the G-pixel, and the b-crystal are used as the pixel of interest. The pixel value of the pixel is calculated by calculating the pixel value of the surrounding pixel of one of the peripheral regions of the pixel of interest, thereby calculating the correlation between the horizontal direction of the peripheral region of the interpolation target -30-201234866. In the peripheral paintings of the pixels, the painting direction and the vertical and vertical values of the vertical and vertical values are determined by a specific value, and the mean value of the phase in the picture is the straight direction; The determined determination removes the aforementioned and generated G image removed as described above Correlation in the straight direction, wherein the color of the peripheral element of the pixel is different; the prime value; the ratio of the addition - the average value and the other value are interpolated to the interpolation The second of the target pixels, t, the average value is the average of the interpolated ❹ prime pixel values of the above-mentioned calculated strong correlations, and the other average is horizontally The flat image k number of the pixel adjacent to the interpolation target pixel is composed of only the G pixel having the first two G interpolation pixel values; the obtained G image signal The low frequency component; the interpolated image signal plus the second G interpolated pixel value of the low frequency component 唬. -31·