CN100559885C - Reconstruction method of picture-element color information - Google Patents
Reconstruction method of picture-element color information Download PDFInfo
- Publication number
- CN100559885C CN100559885C CNB2005101355806A CN200510135580A CN100559885C CN 100559885 C CN100559885 C CN 100559885C CN B2005101355806 A CNB2005101355806 A CN B2005101355806A CN 200510135580 A CN200510135580 A CN 200510135580A CN 100559885 C CN100559885 C CN 100559885C
- Authority
- CN
- China
- Prior art keywords
- pixel
- green
- red
- blue
- rebuild
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Abstract
The invention provides a kind of reconstruction method of picture-element color information, it is earlier according to the ratio that assigns weight of the locations of pixels distance around each red/blue pixel, and calculate chrominance signal measures of dispersion on level and the vertical direction, follow inverse proportion value of assigning weight according to the signal difference amount, to wait that along level and vertical direction two green pixel values of rebuilding pixel institute adjacency on average obtain two mean values respectively, this two mean value respectively be multiply by the weighted value addition then of its affiliated direction, the weighting sum total of gained promptly becomes the green element information in each red/blue pixel again; Supply red elemental and the blue element that other pixel is lost succeeded by bilinear interpolation method, and again with value selector correcting chromatic aberration in.Therefore the color information that the present invention rebuild has the high advantage of accuracy, can reach the uniform effect of high-resolution and color.
Description
Technical field
The present invention is relevant a kind of method of rebuilding pixel color, particularly about a kind of reconstruction method of picture-element color information that adds weight and reach chromatic aberration correction, so that color interpolation value accurately to be provided.
Background technology
Along with prevailing of audio-visual multimedia and releasing one after another of digital image apparatus, the user is healed to the requirement of colors of image and is become strict.Wherein, image sensor arrays be in the digital photography in order to the device of recording image signal, the least unit of forming picture on image sensor arrays is a photoconductive body, is commonly referred to pixel (pixel).Color because photoconductive body itself only responds to light intensity to light does not divide the ability of releasing, therefore must before each photoconductive body of image sensor arrays, add the filter or the filter of red (R), green (G), blue (B) three primary colors (primary col ó r) respectively, only to allow ruddiness, green glow and blue light respectively by obtaining colored image.
Yet, in resulting image pixel array behind image sensing array and the filter, each image pixel only has the single color element, therefore must handle image pixel briliancy number with image processing software again, makes each image pixel all contain red, blue, green three color elements.The known pixels colour reconstruction method is that utilization interpolation method (interpolation) is rebuild color element, way commonly used be utilize bilinearity method (bilinear) promptly 4 methods of average be the data that each former color pixel calculates all the other two colors, as shown in Figure 1, the periphery of each red pixel 10 has four blue pixel 12 and four green pixels 14 respectively, the pixel interpolation system utilizes the luminosity mean value of four the most contiguous blue pixel 12 and the luminosity mean value of four green pixels 14, for the red pixel 10 at center calculates blueness and the green data of being lost, so that red pixel obtains a complete rgb light degree data.In like manner, blue also identical with the interpolation principle of green pixel 12,14.Yet, this conventional practice does not consider that image edge (edge) and signal tendency change, when some have the image of careful colored lines or accurate color change when shooting, if by 4 average pixels wherein is to be positioned at image edge (edge) or signal surging part, then the ensemble average value will be influenced by this point and produce bigger error, feasible resolution variation herein, and huge dentation pigment figure occurs or produce false colour (aliasing) problem, cause the relatively poor and easy distortion of overall image image quality.
In order to improve this problem, known calculus system such as U.S. Pat 4,774,565, it is to utilize bilinearity method (bilinear) to seek out G look (channel) frequently earlier, calculates R and B look aberration territory (color difference domain) frequently with median filter (median filter) again; Yet because this patent does not only add the step of any judgement directivity with the bilinearity method when asking for G look frequency, therefore not only the G look that comes out of interpolation has the big shortcoming of error frequently, related R and the B look frequently obtained based on the G look of making frequently also has the big problem of error, and still has the shortcoming of differentiating rate variance.
Another known technology such as U.S. Pat 5,373,322, its interpolation is the level of center pixel and the signal difference amount of vertical direction judged earlier, again with the less direction calculating interpolation value of measures of dispersion, owing to do not add the suitable reconstruction in R and B aberration territory,, cause interpolation value not necessarily correctly to produce serious false colour problem so still exist the possibility of erroneous judgement direction.
Therefore, The present invention be directed to above-mentioned problem, propose a kind of reconstruction method of picture-element color information.
Summary of the invention
Main purpose of the present invention, be that a kind of reconstruction method of picture-element color information is being provided, it is an assign weight ratio and the interpolation value of these directions is weighted sum total of signal difference amount according to several specific directions, to obtain accurate green interpolation value, and cooperate median filter to set up accurate redness and blue information, make the color information of reconstruction have the high effect of accuracy, effectively to overcome the big shortcoming of known interpolation color error.
Another object of the present invention is that a kind of reconstruction method of picture-element color information is being provided, and to reach the uniform effect of resolution height and color, easily produces the problem of false colour and color distortion with the known interpolation of certain improvement system.
For reaching above-mentioned purpose, a kind of reconstruction method of picture-element color information of the present invention, in order to the color element of being lost in the red pixel in the image pixel array, blue pixel and the green pixel is supplied, make each this pixel respectively contain redness, blueness and green element, it is characterized in that this reconstruction method of picture-element color information comprises the following steps:
Choose horizontal direction for waiting each this red pixel and the blue pixel of rebuilding green element, and according to this red pixel and the blue pixel locations of pixels distance on every side ratio that assigns weight, to calculate the chrominance signal measures of dispersion on each this direction according to above-mentioned weight proportion, the ratio that red pixel and blue locations of pixels distance assign weight, obtained by level, vertical direction signal difference amount, the scope of this weighted value is 1/2 to 1/4;
The weighted value that distributes above-mentioned direction according to the inverse proportion of above-mentioned chrominance signal measures of dispersion, and follow respectively above-mentioned direction will this red pixel and two green pixel values of blue pixel institute adjacency in addition average and obtain several mean value, and then this weighted value addition then of above-mentioned mean value respectively be multiply by direction under it, the weighting of gained sum total promptly becomes this and waits to rebuild green element information in red pixel and the blue pixel, this pixel to be rebuild for each, position surrounding pixel far away is to distribute less weight proportion, and the nearer surrounding pixel in position is to distribute bigger weight proportion; And
This waits to rebuild red elemental and the blue element that pixel is lost to supply each with bilinear interpolation method, and then makes each this pixel respectively contain green, redness, blue element.
Wherein, calculating respectively in the step of chrominance signal measures of dispersion according to above-mentioned direction, is to wait to rebuild the green pixel of pixel periphery and the value of red pixel and blue pixel is calculated this chrominance signal measures of dispersion according to this.
Wherein, supply each this wait to rebuild the red elemental and blue element that pixel loses after, also comprise a step, be that this pixel to be rebuild is a benchmark with each, utilizing a median filter to choose this waits to rebuild a plurality of these pixels that the pixel surrounding matrix arranges and calculates value of chromatism individually, ask for the red of each this pixel in this matrix, green difference and indigo plant, green difference, and then by above-mentioned red, green difference and above-mentioned indigo plant, respectively choose at least one median in the green difference and with this pixel in the green element addition, further obtain the information of accurate red elemental and blue element.
Wherein, the selected matrix of this median filter be selected from 3*3,4*4 and 5*5 one of them.
Wherein, this image pixel array is green, a blue pixel column and red, a green pixel leu preface arrangement formation array.
A kind of reconstruction method of picture-element color information of the present invention, be applied to the colour reconstruction of an image pixel array, this image pixel array includes first pixel, second pixel and the 3rd pixel, supplied by the color element that this method for reconstructing is lost the pixel in this image pixel array, make each this pixel respectively contain first element, second element and element, it is characterized in that this reconstruction method of picture-element color information comprises the following steps:
Choose horizontal direction for waiting each this second pixel and the 3rd pixel of rebuilding first element, and according to this red pixel and the blue pixel locations of pixels distance on every side ratio that assigns weight, to calculate the chrominance signal measures of dispersion on each this direction according to above-mentioned weight proportion, the ratio that red pixel and blue locations of pixels distance assign weight, obtained by level, vertical direction signal difference amount, the scope of this weighted value is 1/2 to 1/4;
The weighted value that distributes above-mentioned direction according to the inverse proportion of above-mentioned chrominance signal measures of dispersion, and follow respectively above-mentioned direction will this red pixel and two first pixel values of blue pixel institute adjacency in addition average and obtain several mean value, and then this weighted value addition then of above-mentioned mean value respectively be multiply by direction under it, the weighting of gained sum total promptly becomes this and waits to rebuild first element information in second pixel and the 3rd pixel, this pixel to be rebuild for each, position surrounding pixel far away is to distribute less weight proportion, and the nearer surrounding pixel in position is to distribute bigger weight proportion; And
This waits to rebuild second element and the element that pixel is lost to supply each with bilinear interpolation method, and then makes each this pixel respectively contain first, second and third element.
Wherein, this first, second and third pixel is to be respectively green pixel, red pixel and blue pixel, and this image pixel array is green, a blue pixel column and red, a green pixel leu preface arrangement formation array.
Wherein, in step according to the chrominance signal measures of dispersion of above-mentioned direction calculating on it, be according to this wait to rebuild first pixel of pixel periphery and second pixel and
The value of the 3rd pixel is calculated.
Wherein, supply each this wait to rebuild second element and element that pixel loses after, also comprise a step, be that this pixel to be rebuild is a benchmark with each, utilizing a median filter to choose this waits to rebuild a plurality of these pixels that the pixel surrounding matrix arranges and calculates value of chromatism individually, and then by choosing at least one median in the above-mentioned value of chromatism with this pixel in the first element addition, further obtain the information of accurate second element and element.
Wherein, the selected matrix of this median filter be selected from 3*3,4*4 and 5*5 one of them.
Description of drawings
Under by the specific embodiment conjunction with figs. illustrate in detail as after, when the effect that is easier to understand purpose of the present invention, technology contents, characteristics and is reached, wherein:
Fig. 1 be known in the image pixel array schematic diagram of interpolation color.
Fig. 2 is the schematic diagram of the image pixel array of institute of the present invention interpolation.
Fig. 3 rebuilds the flow chart of picture-element color information for the present invention.
Fig. 4 rebuilds the schematic diagram of color information for the present invention in the image pixel array.
Specific implementation method
The present invention is before rebuilding a pixel with interpolation method, earlier assign weight ratio and calculate the chrominance signal measures of dispersion of level and vertical direction according to the position distance of surrounding pixel, with inverse proportion value of assigning weight according to the chrominance signal measures of dispersion, and then the green element that level and vertical direction are obtained respectively be multiply by the weighted value of direction under it, the weighting sum total of gained promptly becomes the green element interpolation value in red/blue pixel.
See also shown in Figure 2, the image pixel array 20 that after light is by a colored filter, is produced, it is made up of plural pixel 22, comprise red pixel 24, blue pixel 26 and green pixel 28, be listed as and red, green pixel (R, a G) leu preface arrangement formation array with green, a blue pixel (G, B), and then constitute the array that has four green pixels 28 on every side of each red pixel 24 or blue pixel 26.Because each former color pixel 22 only has the single color element, therefore other two color element that each pixel 22 must be lost is supplied, and makes each pixel 22 respectively contain red, blue, green element.
Because in Red, the brightness value of green primary color Gao Erke provide luminosity data more accurately, therefore interpolation method of the present invention is to rebuild green look (Gchannel) frequently with suitable weight calculation earlier, again with the red and blue aberration territory (colordifference domain) of accurate green look data calculating frequently, to obtain accurate green, redness and blue look data frequently.
Reconstruction method of picture-element color information of the present invention is to be software, normally realizes with an image processor.The flow chart of this reconstruction method of picture-element color information sees also shown in Figure 3, it is to use interpolation method and add suitable weighting at interpolation process to judge, please cooperate Fig. 2 simultaneously, method of the present invention comprises the following steps: at first, flow process as step S10 starting pixel color information re-establishing, reception is by the signal of the image pixel array 20 that image sensor arrays transmitted, then shown in step S12, scan-image pel array 20 in regular turn from left to right, so that carry out the reconstruction of green element earlier, wherein whenever scan a blue pixel 26 or a red pixel 24, be to choose level and vertical direction for this pixel, and according to the ratio that assigns weight of the locations of pixels distance around it, position surrounding pixel far away is to distribute less weight proportion, and the nearer surrounding pixel in position then distributes bigger weight proportion; Next promptly as step S14, calculate chrominance signal measures of dispersion Δ H, Δ V on level and the vertical direction according to these weight proportions, this calculates the step of chrominance signal measures of dispersion, is calculated according to this value of waiting to rebuild green pixel 28, red pixel 24 and the blue pixel 26 of pixel periphery to obtain the chrominance signal measures of dispersion.
After calculating the chrominance signal measures of dispersion of level and vertical direction, soon shown in step S16, be weighted sum total according to the inverse proportion distribution level of these chrominance signal measuress of dispersion and the weighted value of vertical direction, be that the big person of measures of dispersion distributes less weighted value, and then follow respectively level and vertical direction will wait to rebuild pixel red/two green pixel values of 24/26 adjacency of blue pixel are in addition average and obtain two mean values, and with the weighted value addition then that this two mean value respectively multiply by its affiliated direction, the weighting of gained sum total promptly becomes the green element information in red/blue pixel 24/26.So repeating step S12 to S16 scans and rebuilds the green element of being lost in each red, blue pixel 24,26 in regular turn, and the green element in all red, blue pixels 24,26 is rebuild and finished.The direction of above-mentioned scanning also can be from right to left or is waited scanning direction from top to bottom.
Then continue and carry out step S18, supply each in regular turn with bilinearity (bilinear) interpolation method and wait to rebuild red elemental and the blue element that pixel is lost, this bilinear interpolation method promptly is that 4 or 2 ask for linear interpolation and treat interpolation value around utilizing, and then makes each pixel 22 in the image pixel array 20 respectively contain green, redness, blue element.
After completing steps S18, each pixel 22 promptly has green, blue and red three kinds of color informations, for making blueness and the red elemental value of utilizing the bilinearity method to try to achieve more accurate, therefore the present invention is after completing steps S18, also comprise the step that an aberration is proofreaied and correct, shown in step S20, be to be benchmark with a pixel to be rebuild, utilize a median filter (median filter) to choose a plurality of pixels that this pixel surrounding matrix is arranged, and ask for the red of each pixel 22 in this matrix, green difference (color difference) and indigo plant, green difference, and then it is red by these, green difference and these indigo plant, respectively choose a median in the green difference, and this median and this are waited to rebuild green element addition in the pixel, further to obtain the information of accurate red elemental and blue element.Finish in regular turn that each is red, after the chromatic aberration correction of blue interpolation value, can make each pixel 22 respectively have green accurately, red, blue color information, and then finish whole colour reconstruction flow process as step S22.Wherein, the selected matrix of median filter normally is 3*3,4*4 or 5*5.
So far, the spirit of reconstruction method of picture-element color information of the present invention has illustrated and has finished, following spy verifies explanation interpolation method of the present invention in detail with a concrete example, obtains enough knowledge so that those who familiarize themselves with the technology can consider the description of this example in light of actual conditions and enforcement according to this.
Utilizing reconstruction method of picture-element color information of the present invention is as shown in Figure 4 as the practical application of colour reconstruction in the image processing system, image is by colour filter array (color filter array, CFA) and form as shown in the figure image pixel array after the image sensor arrays, this image carry out subsequent treatment before, the colour component of being lost must be rebuild, method for reconstructing comprises following four key steps.
Step 1: by the total occlusal reconstruction of weighting green (G) color value.
In Fig. 4, G
44Originally be non-existent, R
44The G of position
44Reconstruction need with reference to R
44G all around
34, G
54, G
43, G
45Four green shading values, and through certain weighted, the calculating formula of this weighted principle is as follows:
ΔH＝w
1×abs(G
43-G
45)+w
2×[abs(R
44-R
42)+abs(R
44-R
46)] (1)
ΔV＝w
1×abs(G
34-G
54)+w
2×[abs(R
44-R
24)+abs(R
44-R
64)] (2)
Wherein, Δ H represents R
44Pixel is in the signal difference amount of horizontal direction all around, and promptly the signal tendency of horizontal direction changes, and Δ V represents R
44Pixel is in the signal difference amount of vertical direction all around; w
1And w
2Each represents weighted value, is respectively w
1=1/2 and w
2=1/4.
As shown in the figure, for R
44Horizontal direction because G
43, G
45Nearer, so weighted value is bigger, be 1/2; And R
42And R
46Far away, so weighted value is less, be 1/4, see through above-mentioned according to the assign weight mode of ratio of position distance, be to make the level obtained, vertical direction signal difference amount Δ H, Δ V, can effectively solve and suitably judge or handle and ask for the shortcoming of Δ H, Δ V known the adding near real situation.
After calculating level and vertical direction change amount signal Δ H, Δ V, then carry out the weighted calculation of interpolation value again, its calculating formula is as follows:
Wherein,
In calculating formula (3), k
1And k
2Value is respectively the weighted value of level and vertical component, it is along with level and vertical direction signal difference amount Δ H, Δ V and change, make the G value come the weighted value of distribution level and vertical component and be weighted sum total according to the inverse proportion of these signal difference amount Δs H, Δ V, have less weighted value with the G component that makes the measures of dispersion general orientation thus, and the G component of the little direction of measures of dispersion has bigger weighted value.Therefore formula (3) is will wait to rebuild pixel G in order to follow level and vertical direction
44Two green pixel value G of institute's adjacency
43, G
45And G
34, G
54Respectively in addition average, and obtain two mean values, i.e. horizontal component and vertical component, and this two mean value respectively be multiply by the weighted value k of direction under it
1And k
2, addition weighted value then, the weighting sum total G of gained
44Promptly become and wait to rebuild red pixel R
44In green element information, and then with B
44The G that is lost of position
44Supplied.
The G value that R in so in regular turn each being listed as and B position are lost is obtained with aforementioned weighting interpolating method, makes R pixel and B pixel in the image pixel array respectively contain the green element G that loses to some extent.So far, finished the green color reconstruction of all pixels in the image pixel array.
Step 2: with bilinearity method decision red (G), blue (B) color value.
Finishing after the G look rebuilds frequently, then, utilize the bilinearity method to obtain red (G) of loss, blue (B) color value, please join again shown in Figure 4, R
33Originally be non-existent, B
33The R of position
33Reconstruction be with reference to B
334 the most contiguous R values, calculating formula is as follows:
Because B
334 the most contiguous R positions are respectively R
22, R
24, R
42, R
44So, with these 4 average in addition, i.e. interpolation obtains R
33Value.
Wherein, be positioned at the interpolation of the pixel of image edge, then average with the R value of its level or vertical two sides, for example:
The R value that G in the image pixel array and B pixel are lacked is calculated one by one so in regular turn, makes each G position and B position respectively have red elemental R.
In like manner, the reconstruction of blue element is also same, so do not give unnecessary details in this.So far, finished the preliminary reconstruction of image pixel array medium blue (B) color and red (R) color.
Step 3: ask for colour difference information by median filter.
After the interpolation of finishing the green of each pixel, blueness and red elemental is calculated, for making blueness and the red elemental value of utilizing the bilinearity method to try to achieve more accurate, therefore the present invention and be benchmark with each pixel to be rebuild, utilize a median filter (median filter) by choosing 5*5 block (pixel) in the image pixel array, individually each pixel is set up earlier aberration territory (color differencedomain), by choosing the aberration intermediate value in these aberration territories, its calculation procedure is as follows again:
(Ri-Gi)
median＝median?filter(R-G) (4)
(Bi-Gi)
median＝median?filter(B-G) (5)
Above-mentioned program (4), (5) are in order to cooperate the intermediate value of median filter to select based on accurate G value, to select best redness difference territory intermediate value and blue difference territory intermediate value.
Step 4: calculate accurate R and B value.
Utilizing after median filter selects best redness difference territory intermediate value and blue difference territory intermediate value, then soon utilize following formula (6), (7) to calculate accurate R and B value.
R
i＝(R
i-G
i)
median+G
i (6)
B
i＝(B
i-G
i)
median+G
i (7)
So each R pixel and each B pixel of institute's interpolation in the image pixel array are carried out chromatic aberration correction one by one, make each pixel respectively contain accurate RGB three colouring informations, and then finish the program of whole image reconstruction.
Therefore, the present invention utilizes before interpolation calculating, earlier according to the position distance ratio that assigns weight, to calculate out the signal difference amount for the treatment of interpolation G value position peripheral horizontal and vertical direction, and then by signal difference amount distribution horizontal weights value and vertical weight value, provide horizontal G value that interpolation obtains and vertical G value respectively to multiply by the weighted value of direction under it, ask for the green element interpolation value to utilize the weighting sum total for red/blue pixel, then carry out the red elemental that other pixel loses and/or the interpolation of blue element again and calculate, and set up more accurate redness and blue information with a median filter.So, the present invention can make the calculation of signal numerical value of whole interpolation process near truth, so that interpolation color-values accurately to be provided, reaching the uniform effect of resolution height and color, and then can positively improve problems such as false colour that known interpolation system produces greatly because of error, distortion.
The above is by embodiment characteristics of the present invention to be described, its purpose is had the knack of this operator and can Liao be separated content of the present invention and implement according to this making, and non-limiting claim of the present invention, so, all other do not break away from equivalence modification or the modification that disclosed spirit is finished, and must be included in the claim of the following stated.
Claims (10)
1. reconstruction method of picture-element color information, in order to the color element of being lost in the red pixel in the image pixel array, blue pixel and the green pixel is supplied, make each this pixel respectively contain redness, blueness and green element, it is characterized in that this reconstruction method of picture-element color information comprises the following steps:
Choose horizontal direction and vertical direction for waiting each this red pixel and the blue pixel of rebuilding green element, and according to this red pixel and the blue pixel locations of pixels distance on every side ratio that assigns weight, to calculate the chrominance signal measures of dispersion on each this direction according to above-mentioned weight proportion, the ratio that red pixel and blue locations of pixels distance assign weight, obtain by level, vertical direction signal difference amount, this its weighted value of nearer measures of dispersion is 1/2, and its weighted value of measures of dispersion far away is 1/4;
The weighted value that distributes above-mentioned direction according to the inverse proportion of above-mentioned chrominance signal measures of dispersion, and follow respectively above-mentioned direction will this red pixel and two green pixel values of blue pixel institute adjacency in addition average and obtain several mean value, and then this weighted value addition then of above-mentioned mean value respectively be multiply by direction under it, the weighting of gained sum total promptly becomes this and waits to rebuild green element information in red pixel and the blue pixel, this pixel to be rebuild for each, position surrounding pixel far away is to distribute less weight proportion, and the nearer surrounding pixel in position is to distribute bigger weight proportion; And
This waits to rebuild red elemental and the blue element that pixel is lost to supply each with bilinear interpolation method, and then makes each this pixel respectively contain green, redness, blue element.
2. reconstruction method of picture-element color information as claimed in claim 1, it is characterized in that, wherein, calculating respectively in the step of chrominance signal measures of dispersion according to above-mentioned direction, is to wait to rebuild the green pixel of pixel periphery and the value of red pixel and blue pixel is calculated this chrominance signal measures of dispersion according to this.
3. reconstruction method of picture-element color information as claimed in claim 1, it is characterized in that, wherein, supply each this wait to rebuild the red elemental and blue element that pixel loses after, also comprise a step, be that this pixel to be rebuild is a benchmark with each, utilizing a median filter to choose this waits to rebuild a plurality of these pixels that the pixel surrounding matrix arranges and calculates value of chromatism individually, ask for the red of each this pixel in this matrix, green difference and indigo plant, green difference, and then by above-mentioned red, green difference and above-mentioned indigo plant, respectively choose at least one median in the green difference and with this pixel in the green element addition, further obtain the information of accurate red elemental and blue element.
4. reconstruction method of picture-element color information as claimed in claim 3 is characterized in that, wherein, the selected matrix of this median filter be selected from 3*3,4*4 and 5*5 one of them.
5. reconstruction method of picture-element color information as claimed in claim 1 is characterized in that, wherein, this image pixel array is green, a blue pixel column and red, a green pixel leu preface arrangement formation array.
6. reconstruction method of picture-element color information, be applied to the colour reconstruction of an image pixel array, this image pixel array includes first pixel, second pixel and the 3rd pixel, supplied by the color element that this method for reconstructing is lost the pixel in this image pixel array, make each this pixel respectively contain first element, second element and element, it is characterized in that this reconstruction method of picture-element color information comprises the following steps:
Choose horizontal direction and vertical direction for waiting each this second pixel and the 3rd pixel of rebuilding first element, and according to the ratio that assigns weight of the locations of pixels distance around this second pixel and the 3rd pixel, to calculate the chrominance signal measures of dispersion on each this direction according to above-mentioned weight proportion, the ratio that second pixel and the 3rd locations of pixels distance assign weight, obtain by level, vertical direction signal difference amount, this its weighted value of nearer measures of dispersion is 1/2, and its weighted value of measures of dispersion far away is 1/4;
The weighted value that distributes above-mentioned direction according to the inverse proportion of above-mentioned chrominance signal measures of dispersion, and it is in addition average and obtain several mean value with two first pixel values of this second pixel and the 3rd pixel institute adjacency to follow above-mentioned direction respectively, and then this weighted value addition then of above-mentioned mean value respectively be multiply by direction under it, the weighting of gained sum total promptly becomes this and waits to rebuild first element information in second pixel and the 3rd pixel, this pixel to be rebuild for each, position surrounding pixel far away is to distribute less weight proportion, and the nearer surrounding pixel in position is to distribute bigger weight proportion; And
This waits to rebuild second element and the element that pixel is lost to supply each with bilinear interpolation method, and then makes each this pixel respectively contain first, second and third element.
7. reconstruction method of picture-element color information as claimed in claim 6, it is characterized in that, wherein, this first, second and third pixel is to be respectively green pixel, red pixel and blue pixel, and this image pixel array is green, a blue pixel column and red, a green pixel leu preface arrangement formation array.
8. reconstruction method of picture-element color information as claimed in claim 6, it is characterized in that, wherein, in step, be to wait to rebuild first pixel of pixel periphery and the value of second pixel and the 3rd pixel is calculated according to this according to the chrominance signal measures of dispersion of above-mentioned direction calculating on it.
9. reconstruction method of picture-element color information as claimed in claim 6, it is characterized in that, wherein, supply each this wait to rebuild second element and element that pixel loses after, also comprise a step, be that this pixel to be rebuild is a benchmark with each, utilizing a median filter to choose this waits to rebuild a plurality of these pixels that the pixel surrounding matrix arranges and calculates value of chromatism individually, and then by choosing at least one median in the above-mentioned value of chromatism with this pixel in the first element addition, further obtain the information of accurate second element and element.
10. reconstruction method of picture-element color information as claimed in claim 9 is characterized in that, wherein, the selected matrix of this median filter be selected from 3*3,4*4 and 5*5 one of them.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005101355806A CN100559885C (en) | 2005-12-29 | 2005-12-29 | Reconstruction method of picture-element color information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005101355806A CN100559885C (en) | 2005-12-29 | 2005-12-29 | Reconstruction method of picture-element color information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1992909A CN1992909A (en) | 2007-07-04 |
| CN100559885C true CN100559885C (en) | 2009-11-11 |
Family
ID=38214783
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2005101355806A Expired - Fee Related CN100559885C (en) | 2005-12-29 | 2005-12-29 | Reconstruction method of picture-element color information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100559885C (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI664605B (en) * | 2018-08-16 | 2019-07-01 | 瑞昱半導體股份有限公司 | Color reconstruction device and method |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101815220B (en) * | 2009-02-20 | 2011-08-10 | 华晶科技股份有限公司 | Method for correcting image color distortion |
| CN102098520A (en) * | 2009-12-15 | 2011-06-15 | 佳能企业股份有限公司 | Color interpolating system and method |
| CN105430231B (en) * | 2015-11-24 | 2018-09-04 | 努比亚技术有限公司 | Image processing apparatus and method |
| CN113454687A (en) * | 2020-09-28 | 2021-09-28 | 深圳市大疆创新科技有限公司 | Image processing method, apparatus and system, computer readable storage medium |
-
2005
- 2005-12-29 CN CNB2005101355806A patent/CN100559885C/en not_active Expired - Fee Related
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI664605B (en) * | 2018-08-16 | 2019-07-01 | 瑞昱半導體股份有限公司 | Color reconstruction device and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1992909A (en) | 2007-07-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102006484B (en) | Image processing device, image processing method, and program | |
| CN101455075B (en) | System and method for a high performance color filter mosaic array | |
| CN102870417B (en) | Color imaging device | |
| CN1312939C (en) | Image processing method, image processing program, and image processor | |
| US9767742B2 (en) | Image display device | |
| CN1610412B (en) | Image processing apparatus and image processing method | |
| CN100544450C (en) | Image processing method | |
| TWI303723B (en) | Method for rendering data of a color space onto the display of another color space | |
| CN102036091B (en) | Method of and apparatus for processing image data for display by a multiple-view display device | |
| CN102959958A (en) | Color imaging device | |
| CN103733220B (en) | A kind of processing method based on bayer format-pattern and device | |
| EP0804037A2 (en) | Process and system for generating a full color image or multispectral image from the image data of a CCD image sensor with a mosaic color filter. | |
| CN100559885C (en) | Reconstruction method of picture-element color information | |
| EP2523160A1 (en) | Image processing device, image processing method, and program | |
| CN102292975A (en) | Solid state imaging element, camera system and method for driving solid state imaging element | |
| KR20040066051A (en) | Weighted gradient based and color corrected interpolation | |
| CN102893609B (en) | Image processing apparatus and control method for image processing apparatus | |
| CN102663741A (en) | Method for carrying out visual stereo perception enhancement on color digit image and system thereof | |
| CN110246080A (en) | Demosaicing methods and its system | |
| KR101999546B1 (en) | Method of correcting colors, machine-implemented method for a multi-primary color matrix display apparatus, and imgae data signal processing apparatus | |
| US7102789B2 (en) | Method for rendering an image comprising multi-level pixels | |
| CN1326088C (en) | Interpolation method for complementing pixel color | |
| TWI308461B (en) | ||
| CN101411205A (en) | Method and arrangement for generating a color video signal | |
| CN101489142B (en) | Colored interpolation method for digital image |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20091111 Termination date: 20151229 |
|
| EXPY | Termination of patent right or utility model |