CN101272503A - Gridding noise elimination method and device for remaining image definition - Google Patents
Gridding noise elimination method and device for remaining image definition Download PDFInfo
- Publication number
- CN101272503A CN101272503A CNA2008101057092A CN200810105709A CN101272503A CN 101272503 A CN101272503 A CN 101272503A CN A2008101057092 A CNA2008101057092 A CN A2008101057092A CN 200810105709 A CN200810105709 A CN 200810105709A CN 101272503 A CN101272503 A CN 101272503A
- Authority
- CN
- China
- Prior art keywords
- pixel
- value
- smooth region
- unit area
- similar
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 55
- 230000008030 elimination Effects 0.000 title claims abstract description 14
- 238000003379 elimination reaction Methods 0.000 title claims abstract description 14
- 238000004364 calculation method Methods 0.000 claims description 11
- 238000005516 engineering process Methods 0.000 description 6
- 230000003287 optical effect Effects 0.000 description 6
- 238000007792 addition Methods 0.000 description 5
- 238000001914 filtration Methods 0.000 description 5
- 238000004519 manufacturing process Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 3
- 238000003672 processing method Methods 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- 239000003086 colorant Substances 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012797 qualification Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Images
Landscapes
- Facsimile Image Signal Circuits (AREA)
Abstract
The invention discloses a grid noise elimination method and a device for maintaining the image definition. The method includes: when the noise elimination is implemented on each G pixel point in the area needing the noise elimination, whether the unit area is a smooth area or not is judged firstly according to all the values of non-similar pixel points in a unit area which centers on the G pixel point. If yes, the value of the G pixel point and the values of all or part of the non-similar pixel points in the unit area are weighted in average, and the result is used as the value of the G pixel point after being renewed. The device includes a judging unit, a processing unit for smooth area and a processing unit for non-smooth area. By adopting the method and the device of the invention adopted, grid noises can be eliminated, meanwhile, the image definition is maintained, more particularly, the image definition of a marginal part is maintained.
Description
Technical field
The present invention relates to digital image processing method, relate in particular to a kind of gridding noise elimination method and device thereof that keeps image definition.
Background technology
The optical pickocff of digital image pick-ups such as camera, digital camera and Digital Video adopts CCD (Charge-coupled device usually, charge coupled device) or CMOS (ComplementaryMetal Oxide Semiconductor, complementary matal-oxide semiconductor) technology, a two-dimensional matrix of forming by the photo-sensitive cell of both direction dense arrangement (CCD or CMOS) anyhow, and CCD or cmos sensor can only be responded to light luminance, can not respond to color information.Therefore (Color Filtered Array CFA) guarantees that each sensor pixel only can receive a kind of light of color: a kind of in normally red (R), green (G), blue (B) three kinds of colors must to use the color filter array.The color filter array can use different patterns, and the most frequently used is the color filter array of Bayer (Bel) pattern.The Bayer pattern is used alternatingly one group of red and green filter and one group of green and blue filter, and wherein green pixel point adds up to redness and blue pixel point sum.The line of pixels column format of original (raw) image (hereinafter to be referred as the Bayer image) of Bayer pattern as shown in Figure 1.Image for the Bayer image transitions is become can normally show need carry out cfa interpolation to it, promptly to each pixel, obtains other two kinds of color values of this pixel by the pixel value around it.For example, [m, n] locates in the position, has only the G value, utilizes the information of point on every side, can obtain the R value and the B value at this some place by interpolation.Through behind the cfa interpolation, can obtain R, G, B value on each pixel.Wherein, the G pixel can be divided into Gr pixel of going together with the R pixel and the Gb pixel of going together with the B pixel.
Because the non-fringe region of general nature scene image all is smooth, that is to say that color is slow gradual change, therefore, the G value that Bel's image is adjacent (diagonal angle is adjacent, i.e. Gr and Gb) should be very approaching.But, in the production process of digital image pick-up, because the production technology assembling alignment precision relatively poor or camera lens and optical pickocff of optical pickocff is not enough, the differ greatly phenomenon of (being G value imbalance) of adjacent G (Gr and the Gb) value in diagonal angle that makes the Bayer image that generates at optical pickocff can appear at non-fringe region.And will there be latticed noise in this Bel's image through the image that generates after the cfa interpolation, have a strong impact on picture quality.As shown in Figure 2.For above-mentioned reasons, can cause Gr pixel and Gb pixel than big-difference, but a G pixel adjacent pixels point is Gr or Gb, at smooth region, these Gr or Gb value still are slow gradual change.Claim also that below all Gr pixels or Gb pixel are similar G pixel.
The general dual mode that adopts is eliminated latticed noise in the prior art:
First kind of mode improved production technology exactly, with the quality of raising optical pickocff and the alignment precision of camera lens and optical pickocff, yet production technology improved often to be needed to drop into a large amount of funds and time, even and production technology carried out improving be difficult to also guarantee that each digital image pick-up of producing can both satisfy quality requirement.
Second method is exactly with the mode of image processing Bel's image to be carried out grid noise to eliminate, promptly the G value pixel that each original G value pixel can be adjacent is weighted on average, and with the G value of this weighted average as this pixel, eliminated the unbalanced of G value like this, also just eliminated grid noise, but it is fuzzy that said method may cause original image to produce, especially the edge in original image.
Summary of the invention
The technical problem to be solved in the present invention provides a kind of gridding noise elimination method and device thereof that keeps image definition, eliminates grid noise, and keeps the definition of image.
In order to address the above problem, the invention provides a kind of gridding noise elimination method that keeps image definition, comprise: when each the G pixel in the zone that will carry out noise removing in the image is carried out the noise removing processing, elder generation is according to the value of all non-similar pixels in the unit area that is the center with this G pixel, judge whether this unit area is smooth region, if, the value of all or part of non-similar G pixel in the value of this G pixel and this unit area is weighted on average, and the result is as the value of this G pixel after upgrading.
Further, after judging that described unit area is smooth region, 4 adjacent non-similar G pixel point values of value and diagonal angle thereof of described G pixel are weighted on average, the value after the result upgrades as this G pixel, the weights that add temporary the adjacent G pixel in 4 diagonal angles are not zero entirely.
Further, adopt a kind of in the following method to judge whether described unit area is smooth region:
With the non-similar G pixel of G pixel described in the described unit area vertically, on level and two each directions to the angular direction, which side respectively to be divided into two parts according to its described relatively G pixel in, on each direction, calculate the not absolute value of the difference of homonymy G pixel point value sum respectively, the perhaps absolute value sum of difference between a plurality of corresponding G pixel point values of homonymy not, obtain the value DV of four direction, DH, DL, DR, if these four values are all less than preset threshold value on the direction separately, think that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
Upper in all non-similar G pixels of described G pixel one side and all non-similar G pixels of opposite side to each to the angular direction, calculate the not absolute value of the difference of homonymy G pixel point value sum, the perhaps absolute value sum of difference between a plurality of corresponding G pixel point values of homonymy not, if two to the result of calculation that obtains on the angular direction all less than or smaller or equal to preset threshold value, perhaps two to the result of calculation sum that obtains on the angular direction less than or smaller or equal to preset threshold value, judge that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
Calculate the difference of any two the non-similar G pixels in the described unit area, as the maximum in the absolute value of all differences less than or during smaller or equal to preset threshold value, think that then the unit area is a smooth region, otherwise be non-smooth region; Perhaps
Calculate the maximum and the minimum value of all non-similar G pixels of G pixel described in the described unit area.If the difference of maximum and minimum value less than or during smaller or equal to preset threshold value, think that described unit area is a smooth region, otherwise be non-smooth region.
Further, described unit area is to be the zones that comprise 3 row * 3 row at center with described G pixel, adopts a kind of in the following method to judge whether this unit area is smooth region:
Press the adjacent G pixel in diagonal angle of described G pixel vertical respectively, horizontal direction is divided into two groups respectively, by two the angular direction is divided into one group respectively, every group comprises the adjacent G pixel in two diagonal angles on this direction, on each direction, calculate the absolute value of every group of G pixel value difference value respectively, the absolute value addition that when organize arranged more each group is obtained, obtain the value DV of four direction respectively, DH, DL, DR, if these four values are all less than preset threshold value on the direction separately, think that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
Calculate each absolute difference of described G pixel respectively to two adjacent G pixel point values on the angular direction, if these two absolute difference all less than or smaller or equal to preset threshold value, think that then this G pixel and its all neighbor pixels are in same smooth region, otherwise be unsmooth zone; Perhaps
Calculate each absolute difference of described G pixel respectively to two adjacent G pixel point values on the angular direction, if these two absolute difference and less than or smaller or equal to preset threshold value, think that then the unit area that described G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
Calculate the difference between any two pixels in four G pixels adjacent with described G pixel diagonal angle, if the maximum in the absolute value of all differences less than or during smaller or equal to preset threshold value, think that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
In four the G pixels adjacent, find out maximum and minimum value with described G pixel diagonal angle, if the difference of maximum and minimum value less than or during smaller or equal to preset threshold value, think that the unit area that described G pixel is the center is a smooth region, otherwise be unsmooth zone.
Further, if described unit area is not a smooth region, then keep the value of this G pixel constant.
Further, if described unit area is not a smooth region, judge the non-similar G pixel of part that is in same smooth region with described G pixel in the non-similar G pixel of G pixel described in the described unit area again, with the value of described G pixel and the value of this part non-similar G pixel be weighted average after, as the value of the described G pixel after upgrading.
Further, described unit area is to be the zone that comprises 3 row * 3 row at center with described G pixel, if described unit area is not a smooth region, judge the part G pixel that is in same smooth region in the adjacent G pixel in described diagonal angle with described G pixel again, with the value of the adjacent G pixel of value of described G pixel with this part be weighted average after, as the value of the described G pixel after upgrading.
Further, judge that unit area that described G pixel is the center is non-smooth region after, a kind of in such a way upgrades described G pixel point value:
Select its corresponding direction according to the little value in vertical direction DV value and the horizontal direction DH value, obtain the adjacent G pixel in 2 groups of diagonal angles on this direction, calculate the difference between the similar G pixel of two nearest on described G pixel and this direction vertical direction described G pixels more respectively, determine the similar G pixel that its absolute difference is little, to be the one group of diagonal angle adjacent G pixel point value of the similar G pixel of benchmark homonymy and the value weighted average of this G pixel with described G pixel, as the value of the described G pixel after upgrading with this; Perhaps
Calculate the DV of vertical direction, the DH of horizontal direction, two DL to the angular direction, minimum value among the DR, when if this minimum value is DV or DH, select its corresponding direction, obtain the adjacent G pixel in 2 groups of diagonal angles on this direction, and then calculate G (i respectively, j) with this direction vertical direction on difference between the similar G pixel of nearest two and described G pixel, determine the similar G pixel that its absolute difference is less, to be one group of diagonal angle adjacent G pixel point value of the similar G pixel of benchmark homonymy and this G (i with described G pixel with this, j) value weighted average is as the value of the described G pixel after upgrading; When if this minimum value is DL or DR, with the value of adjacent G pixel in two diagonal angles on its respective direction and described G pixel be weighted average after, as the value of the described G pixel after upgrading; Perhaps
Calculate described G pixel difference between any two pixels in four adjacent G pixel point values on to the angular direction, determine the minimization of the sum of absolute value of all differences, then with the value weighted average of the value of described G pixel and two diagonal angle adjacent G pixels corresponding, as the value of the described G pixel after upgrading with this minimum value.
The present invention also provides a kind of grid noise elimination device that keeps image definition, comprises judging unit and smooth region processing unit, wherein:
Described judging unit, be used for to carry out each G pixel in the zone of noise removing and carry out noise removing when handling image, elder generation is according to the value of all non-similar pixels in the unit area that is the center with this G pixel, judge whether this unit area is smooth region, if send triggering signal to described smooth region processing unit;
Described smooth region processing unit is used for after receiving the triggering signal of described judging unit, and the value of all or part of non-similar G pixel in the value of this G pixel and this unit area is weighted on average, and the result is as the value of this G pixel after upgrading.
Further, also comprise non-smooth region processing unit;
Described judging unit also is used for when judging described unit area and be non-smooth region, sends triggering signal to described non-smooth region processing unit;
After described non-smooth region processing unit is used to receive the triggering signal of described judging unit, keep the value of described G pixel constant, perhaps judge the non-similar G pixel of part that is in same smooth region with described G pixel in the non-similar G pixel of G pixel described in the described unit area again, with the value of described G pixel and the value of this part non-similar G pixel be weighted average after, as the value of the described G pixel after upgrading.
Further, also comprise non-smooth region processing unit;
Described smooth region processing unit is that 4 adjacent non-similar G pixel point values of the value of described G pixel and diagonal angle thereof are weighted on average, the value after the result upgrades as this G pixel, and the weights that add temporary the adjacent G pixel in 4 diagonal angles are not zero entirely.
Further, described judging unit adopts a kind of in the following method to judge whether described unit area is smooth region:
With the non-similar G pixel of G pixel described in the described unit area vertically, on level and two each directions to the angular direction, which side respectively to be divided into two parts according to its described relatively G pixel in, on each direction, calculate the not absolute value of the difference of homonymy G pixel point value sum respectively, the perhaps absolute value sum of difference between a plurality of corresponding G pixel point values of homonymy not, obtain the value DV of four direction, DH, DL, DR, if these four values are all less than preset threshold value on the direction separately, think that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
Upper in all non-similar G pixels of described G pixel one side and all non-similar G pixels of opposite side to each to the angular direction, calculate the not absolute value of the difference of homonymy G pixel point value sum, the perhaps absolute value sum of difference between a plurality of corresponding G pixel point values of homonymy not, if two to the result of calculation that obtains on the angular direction all less than or smaller or equal to preset threshold value, perhaps two to the result of calculation sum that obtains on the angular direction less than or smaller or equal to preset threshold value, judge that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
Calculate the difference of any two the non-similar G pixels in the described unit area, as the maximum in the absolute value of all differences less than or during smaller or equal to preset threshold value, think that then the unit area is a smooth region, otherwise be non-smooth region; Perhaps
Calculate the maximum and the minimum value of all non-similar G pixels of G pixel described in the described unit area.If the difference of maximum and minimum value less than or during smaller or equal to preset threshold value, think that described unit area is a smooth region, otherwise be non-smooth region.
Further, described unit area is to be the zones that comprise 3 row * 3 row at center with described G pixel, and described judging unit adopts a kind of in the following method to judge whether described unit area is smooth region:
Press the adjacent G pixel in diagonal angle of described G pixel vertical respectively, horizontal direction is divided into two groups respectively, by two the angular direction is divided into one group respectively, every group comprises the adjacent G pixel in two diagonal angles on this direction, on each direction, calculate the absolute value of every group of G pixel value difference value respectively, the absolute value addition that when organize arranged more each group is obtained, obtain the value DV of four direction respectively, DH, DL, DR, if these four values are all less than preset threshold value on the direction separately, think that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
Calculate each absolute difference of described G pixel respectively to two adjacent G pixel point values on the angular direction, if these two absolute difference all less than or smaller or equal to preset threshold value, think that then this G pixel and its all neighbor pixels are in same smooth region, otherwise be unsmooth zone; Perhaps
Calculate each absolute difference of described G pixel respectively to two adjacent G pixel point values on the angular direction, if these two absolute difference and less than or smaller or equal to preset threshold value, think that then the unit area that described G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
Calculate the difference between any two pixels in four G pixels adjacent with described G pixel diagonal angle, if the maximum in the absolute value of all differences less than or during smaller or equal to preset threshold value, think that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
In four the G pixels adjacent, find out maximum and minimum value with described G pixel diagonal angle, if the difference of maximum and minimum value less than or during smaller or equal to preset threshold value, think that the unit area that described G pixel is the center is a smooth region, otherwise be unsmooth zone.
Further, described unit area is to be the zone that comprises 3 row * 3 row at center with described G pixel, after described non-smooth region processing unit is received the triggering signal of described judging unit, judge the part G pixel that is in same smooth region in the adjacent G pixel in described diagonal angle with described G pixel again, with the value of the adjacent G pixel of value of described G pixel with this part be weighted average after, as the value of the described G pixel after upgrading.
Further, after described non-smooth region processing unit was received the triggering signal of described judging unit, a kind of value to described G pixel in was upgraded in the following ways:
Select its corresponding direction according to the little value in vertical direction DV value and the horizontal direction DH value, obtain the adjacent G pixel in 2 groups of diagonal angles on this direction, calculate the difference between the similar G pixel of two nearest on described G pixel and this direction vertical direction described G pixels more respectively, determine the similar G pixel that its absolute difference is little, to be the one group of diagonal angle adjacent G pixel point value of the similar G pixel of benchmark homonymy and the value weighted average of this G pixel with described G pixel, as the value of the described G pixel after upgrading with this; Perhaps
Calculate the DV of vertical direction, the DH of horizontal direction, two DL to the angular direction, minimum value among the DR, when if this minimum value is DV or DH, select its corresponding direction, obtain the adjacent G pixel in 2 groups of diagonal angles on this direction, and then calculate G (i respectively, j) with this direction vertical direction on difference between the similar G pixel of nearest two and described G pixel, determine the similar G pixel that its absolute difference is less, to be one group of diagonal angle adjacent G pixel point value of the similar G pixel of benchmark homonymy and this G (i with described G pixel with this, j) value weighted average is as the value of the described G pixel after upgrading; When if this minimum value is DL or DR, with the value of adjacent G pixel in two diagonal angles on its respective direction and described G pixel be weighted average after, as the value of the described G pixel after upgrading; Perhaps
Calculate described G pixel difference between any two pixels in four adjacent G pixel point values on to the angular direction, determine the minimization of the sum of absolute value of all differences, then with the value weighted average of the value of described G pixel and two diagonal angle adjacent G pixels corresponding, as the value of the described G pixel after upgrading with this minimum value.
Adopt method and apparatus of the present invention, can eliminate grid noise, keep the definition of image simultaneously, especially keep the definition of marginal portion in the image.
Description of drawings
Fig. 1 is the line of pixels column format schematic diagram of the original image of Bayer pattern;
Fig. 2 is the schematic diagram of normal Bayer image and latticed noise Bayer image;
Fig. 3 is a flow chart of eliminating the grid noise method among the present invention;
Fig. 4 is (i, j) the pixel position view of G in the present embodiment;
Fig. 5 is a structure chart of eliminating the grid noise device in the present embodiment;
Fig. 6 is the pixel position view the when unit area is set at 7 row * 7 row in the present embodiment.
Embodiment
In the present invention, the Bayer original image is carried out subarea processing, promptly in image, distinguish smooth region and unsmooth zone (being fringe region or non-smooth region), different processing methods is taked in above-mentioned two zones, grid noise can be eliminated, the definition of image can be kept again.
Below in conjunction with accompanying drawing, the specific embodiment of the present invention is elaborated:
Among the present invention the G pixel of selection area in the Bayer original image is handled, at first needed to set the unit area size that to consider when 1 G pixel handled.Concerning certain G pixel, the unit area is meant that with this G pixel be the zone that the center comprises 3 row * 3 row at least, except that 3 row * 3 row, also can be 5 row * 5 row, the square area of 7 row * 7 row, or 3 row * 5 row, the rectangular area of 5 row * 3 row, or the zone of other size and/or shape.The present invention does not do qualification to the size and the shape of unit area.
In the present embodiment, the unit area is the square area of 3 row * 3 row.As shown in Figure 3, present embodiment method that each G pixel in the selection area is handled may further comprise the steps:
As shown in Figure 4, (i, j) the G pixel at capable, the j row place of i in the presentation video also show (i, j) adjacent R, G, B pixel and corresponding to G (i, position j) with G to G among the figure.
Judge (i, j) pixel is whether the unit area at center is smooth region to current G, can adopt following determination methods.
Determination methods one:
(vertical, horizontal direction can be divided into 2 groups to the angular direction grouping by vertical, level and two respectively for i, neighbor pixel j), and every group comprises the adjacent G pixel in two diagonal angles on this direction, the angular direction is divided into one group of two neighbor pixel with G.On each direction, calculate after the absolute difference of respectively organizing G pixel point value respectively and addition, that is:
Current G (i, j) the adjacent G pixel in the diagonal angle of pixel difference in vertical direction is DV,
DV＝|G(i-1，j-1)-G(i+1，j-1)|+|G(i-1，j+1)-G(i+1，j+1)|
Current G (i, j) the adjacent G pixel in the diagonal angle of pixel difference in the horizontal direction is DH,
DH＝|G(i-1，j-1)-G(i-1，j+1)|+|G(i+1，j-1)-G(i+1，j+1)|
Current G (i, j) the adjacent G pixel in the diagonal angle of pixel first pair of angular direction (with the horizontal direction angle be 45 the degree) on difference be DL,
DL=|G (i-1, j+1)-G (i+1, j-1) | * 2, herein * the 2nd, in order to be in a magnitude with above-mentioned DV, DH.
Current G (i, j) the adjacent G pixel in the diagonal angle of pixel second pair of angular direction (with horizontal direction be 145 the degree) on difference be DR,
DR=|G (i-1, j-1)-G (i+1, j+1) | * 2, herein * the 2nd, in order to be in a magnitude with above-mentioned DV, DH.
If the value of DH, DV, DL, DR all less than or smaller or equal to preset threshold value TH, then think current G (i, j) pixel is that the unit area at center is a smooth region, otherwise, think that (i, j) pixel is that the unit area at center is non-smooth region to current G.Notice that the predetermined threshold value of 4 directions is set to equate, but does not need so in other embodiments here, can be provided with respectively for each direction, press above for example DL=|G (i-1, j+1)-G (i+1, j-1) |, DR=|G (i-1, j-1)-and G (i+1, j+1) | during calculating, the threshold value of this both direction can be made as TH/2, promptly the threshold value of 4 directions can correspondingly be provided with respectively, the threshold value of 4 directions all is set to TH can thinks one of them special case.
Determination methods two:
Calculate G (i respectively, j) two of pixel absolute difference to (promptly on 45 directions and 145 directions) neighbor on the angular direction, if these two absolute difference all less than or smaller or equal to preset threshold value, then think G (i, j) pixel is that the unit area at center is a smooth region; Otherwise, think current G (i, j) pixel is that the unit area at center is non-smooth region; Perhaps, if these two absolute difference and less than or smaller or equal to described preset threshold value, then think this G (i, j) pixel is that the unit area at center is a smooth region; Otherwise, think that (i, j) pixel is that the unit area at center is non-smooth region to current G.That is:
D1＝|G(i-1，j-1)-G(i+1，j+1)|
D2＝|G(i-1，j+1)-G(i+1，j-1)|
If D1＜=Th and D2＜=Th, then think current G (i, j) pixel is that the unit area at center is a smooth region; Otherwise, think current G (i, j) pixel is that the unit area at center is non-smooth region, Th is a preset threshold value; Perhaps
If D1+D2＜=Th, then think current G (i, j) pixel is that the unit area at center is a smooth region; Otherwise, think that (i, j) pixel is that the unit area at center is non-smooth region to current G.
Determination methods three:
Calculate and G (i, j) four G pixels that pixel is adjacent at the diagonal angle be G (i-1, j-1), G (i-1, j+1), G (i+1, j-1), G (i+1, j+1) difference between any two pixels, as the maximum in the absolute value of all differences less than or during smaller or equal to preset threshold value TH2, then think current G (i, j) pixel is that the unit area at center is a smooth region, otherwise, think that (i, j) pixel is that the unit area at center is non-smooth region to current G.
Determination methods four:
With G (i, j) pixel on to the angular direction adjacent four G pixels promptly G (i-1, j-1), G (i-1, j+1), G (i+1, j-1), G (i+1, j+1) } find out maximum max and minimum value min in, if the difference of maximum and minimum value less than or during smaller or equal to preset threshold value, think current G (i, j) pixel is that the unit area at center is a smooth region, otherwise, think that (i, j) pixel is that the unit area at center is non-smooth region to current G.
The judgement of above-mentioned smooth region is based in the smooth region pixel point value and changes slowly, the characteristics of Bel's image G value distribution, and Gr and the different characteristic of Gb value difference, (i, the j) difference between the neighbor point value is as thinking then that less than corresponding threshold value this zone is a smooth region for the G that calculates.This moment can not be as conventional method (i, j) value is adjacent the comparison of pixel point value with G.
Concrete, with current G (i, j) value of pixel be updated to G (i, j) '
G(i，j)′
＝[a×G(i，j)+b×G(i-1，j-1)+c×G(i-1，j+1)+d×(i+1，j-1)+e×G(i+1，j+1)]
/(a+b+c+d+e)
Following formula a, b, c, d, e are filter factor, and a is the real number greater than 0, and b, c, d, e are the real number (if revise, b, c, d, e can not be 0 entirely) more than or equal to 0.The value of a, b, c, d, e is in an application example: a=4, b=1, c=1, d=1, e=1.
If do not belong to same smooth region, illustrate that this zone is that promptly there is the edge in the zone of brightness or color change in the image, as being the intersection of two objects in the details of image or the image herein.The thinking of the G pixel that belongs to fringe region being carried out Filtering Processing is: judge G (i earlier, j) be in same smooth region (being in other words on the subject image of the same side at edge) with any part neighbor pixel, then with G (i, j) with this part neighbor point value weighted average, as the G after upgrading (i, j) value.
Specifically can adopt following several method:
Filtering method one:
This method is to select its corresponding direction according to the little value in vertical direction DV value and the horizontal direction DH value earlier, obtain 2 groups of neighbor pixels on this direction, and then calculate G (i respectively, j) difference between the G pixel similar with nearest two and described G pixel on this direction vertical direction, determine the similar G pixel that its absolute difference is little, will be with G (i, j) be 1 group of neighbor point value and this G (i of the similar G pixel of benchmark homonymy with this, j) value weighted average, as the G after upgrading (i, j) value.
Particularly,
If DV≤DH and | G (i, j)-G (i, j-2) |≤| G (i, j)-G (i, j+2) | then current G (i, j) value of pixel be updated to G (i, j) '=[(2 * G (and i, j)+G (i-1, j-1)+G (i+1, j-1)]/4;
If DV≤DH and | G (i, j)-G (i, j-2) |＞| G (i, j)-G (i, j+2) | then current G (i, j) value of pixel be updated to G (i, j) '=[(2 * G (and i, j)+G (i-1, j+1)+G (i+1, j+1)]/4;
If DV＞DH and | G (i, j)-G (i-2, j) |≤| G (i, j)-G (i+2, j) | then current G (i, j) value of pixel be updated to G (i, j) '=[(2 * G (and i, j)+G (i-1, j-1)+G (i-1, j+1)]/4;
If DV＞DH and | G (i, j)-G (i-2, j) |＞| G (i, j)-G (i+2, j) | then current G (i, j) value of pixel be updated to G (i, j) '=[(2 * G (and i, j)+G (i+1, j-1)+G (i+1, j+1)]/4;
Filtering method two:
Calculate vertical direction DV, horizontal direction DH, two to angular direction DL, minimum value among the DR, when if this minimum value is DV or DH, select its corresponding direction, obtain 2 groups of neighbor pixels on this direction, and then calculate G (i respectively, j) difference between the G pixel similar with nearest two and described G pixel on this direction vertical direction, determine the similar G pixel that its absolute difference is less, will be with G (i, j) be 1 group of neighbor point value and this G (i of the similar G pixel of benchmark homonymy with this, j) value weighted average is as the G after upgrading (i, j) value; When if this minimum value is DL or DR, then (i, j) value is weighted on average, as the G after upgrading (i, j) value with two neighbor pixels on the respective direction and G.
Particularly,
If among DV, DH, DL, the DR DV minimum and | G (i, j)-G (i, j-2) |≤| G (i, j)-G (i, j+2) | then current G (i, j) value of pixel be updated to G (i, j) '=[(2 * G (and i, j)+G (i-1, j-1)+G (i+1, j-1)]/4;
If among DV, DH, DL, the DR DV minimum and | G (i, j)-G (i, j-2) |＞| G (i, j)-G (i, j+2) | then current G (i, j) value of pixel be updated to G (i, j) '=[(2 * G (and i, j)+G (i-1, j+1)+G (i+1, j+1)]/4;
If among DV, DH, DL, the DR DH minimum and | G (i, j)-G (i-2, j) |≤| G (i, j)-G (i+2, j) | then current G (i, j) value of pixel be updated to G (i, j) '=[(2 * G (and i, j)+G (i-1, j-1)+G (i-1, j+1)]/4;
If among DV, DH, DL, the DR DH minimum and | G (i, j)-G (i-2, j) |＞| G (i, j)-G (i+2, j) | then current G (i, j) value of pixel be updated to G (i, j) '=[(2 * G (and i, j)+G (i+1, j-1)+G (i+1, j+1)]/4.
If DL minimum among DV, DH, DL, the DR then current G (i, j) value of pixel be updated to G (i, j) '=[(2 * G (and i, j)+G (i-1, j+1)+G (i+1, j-1)]/4;
If DR minimum among DV, DH, DL, the DR then current G (i, j) value of pixel be updated to G (i, j) '=[(2 * G (and i, j)+G (i-1, j-1)+G (i+1, j+1)]/4.
Filtering method three:
Calculate G (i earlier, j) pixel (comprises G (i-1 herein to the non-similar pixel point value of angular direction, j-1), G (i-1, j+1), G (i+1, j-1), G (i+1, j+1)) difference between any two pixels, determine the minimization of the sum of absolute value of all differences, with G (i, j) pixel point value and therewith after two neighbor point value weighted averages of minimum value correspondence, as current G (i, j) value of pixel after upgrading.
Said method two is identical with method three effects, and the amount of calculation of method two is less.
G pixel value on four sidelines of the selected image-region that will carry out noise processed can not deal with, and can adopt the outer G point of selection area that is adjacent to carry out same treatment in the said method yet.When above-mentioned selection area was entire image, the G pixel value on its four sidelines did not process.
Another embodiment:
The present embodiment unit area is set at 7 row * 7 row, when carrying out the judgement of smooth region, what use in a last embodiment is adjacent 4 the inhomogeneity G pixels (is Gb to Gr) in diagonal angle, expand to 16 inhomogeneity G pixels on every side in the present embodiment, for the Gr pixel is an example, its schematic diagram as shown in Figure 6 with the central point.
In this embodiment, when whether being smooth region in the judging unit zone, as press the thinking of first kind of determination methods of the foregoing description, then can be with the non-similar G pixel of G pixel described in the described unit area vertically, on level and two each directions to the angular direction, which side respectively to be divided into two parts according to its described relatively G pixel in, on each direction, calculate the not absolute value of the difference of homonymy G pixel point value sum respectively, the perhaps absolute value sum of difference between a plurality of corresponding G pixel point values of homonymy not, obtain the value DV of four direction, DH, DL, DR, if these four values are all less than preset threshold value on the direction separately, think that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone.
Particularly, can adopt following formula:
For vertical direction, the capable and i+1 line operate with i-1, formula is:
DV＝|G(i-1，j-3)-G(i+1，j-3)|+|G(i-1，j-1)-G(i+1，j-1)|+|G(i-1，j+1)-G(i+1，j+1|+|G(i-1，j+3)-G(i+1，j+3)|
Perhaps
DV=|G (i-1, j-3)+G (i-1, j-1)+G (i-1, j+1)+G (i-1, j+3)-(G (i+1, j-3)+G (i+1, j-1)+G (i+1, j+1)+G (i+1, j+3)) |, on average subtracting earlier here is influence for less noise again.
In like manner,, can obtain DH, no longer repeat here with j-1 row and the operation of j+1 row for horizontal direction.
DL operates in four Gb on a pair of angular direction, and two Gb that deduct opposite side after can two Gb additions with a side take absolute value, and also can be with two Gb of a side takes absolute value after two corresponding with opposite side respectively Gb subtract each other, again with two absolute value additions.
DR operates in another to four Gb on the angular direction, and algorithm can be identical with DL.
As thinking by a last embodiment determination methods two, in the present embodiment, can be upper to the angular direction in all non-similar G pixels of described G pixel one side and all non-similar G pixels of opposite side to each, calculate the not absolute value of the difference of homonymy G pixel point value sum, the perhaps absolute value sum of difference between a plurality of corresponding G pixel point values of homonymy not, if two to the result of calculation that obtains on the angular direction all less than or smaller or equal to preset threshold value, perhaps two to the result of calculation sum that obtains on the angular direction less than or smaller or equal to preset threshold value, judge that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone;
In like manner, thinking according to the determination methods three of a last embodiment, the difference of any two the non-similar G pixels of present embodiment in can the computing unit zone, as the maximum in the absolute value of all differences less than or during smaller or equal to preset threshold value, think that then the unit area is a smooth region, otherwise be non-smooth region.
In like manner, according to the thinking of the determination methods four of a last embodiment, present embodiment can the computing unit zone described in the maximum and the minimum value of non-similar G pixel.If the difference of maximum and minimum value less than or during smaller or equal to preset threshold value, think that the unit area is a smooth region, otherwise be non-smooth region.
All or part of non-similar pixel weighted-average method in G pixel point value and the unit area is adopted in the renewal of the G pixel of smooth region, as still adopting G pixel point value and 4 diagonal angle neighbor point value weighted-average method.
G pixel to non-smooth region can not upgrade, and can upgrade yet, and can adopt the identical method with a last embodiment during as renewal.Certainly, also can do some variations.
As shown in Figure 5, keep the grid noise elimination device of image definition, comprise judging unit, smooth region processing unit and non-smooth region processing unit;
Described judging unit, be used for to carry out each G pixel in the zone of noise removing and carry out noise removing when handling image, elder generation is according to the value of all non-similar pixels in the unit area that is the center with this G pixel, judge whether this unit area is smooth region, if, send triggering signal to described smooth region processing unit, otherwise send triggering signal to described non-smooth region.The description that sees above of concrete determination methods.
Described smooth region processing unit is used for after receiving the triggering signal of described judging unit, and the value of all or part of non-similar G pixel in the value of this G pixel and this unit area is weighted on average, and the result is as the value of this G pixel after upgrading.
Described non-smooth region processing unit after being used to receive the triggering signal of described judging unit, keeps the value of this G pixel constant.Perhaps judge the non-similar G pixel of part that is in same smooth region with described G pixel in the non-similar G pixel of G pixel described in the described unit area again, with the value of described G pixel and the value of this part non-similar G pixel be weighted average after, as the value of the described G pixel after upgrading.The description that sees above of concrete update method.
The present invention also can have other various embodiments; under the situation that does not deviate from spirit of the present invention and essence thereof; those of ordinary skill in the art are when making various corresponding changes and distortion according to the present invention, and these change and be out of shape the protection range that all should belong to the appended claim of the present invention accordingly.
Claims (15)
1, a kind of gridding noise elimination method that keeps image definition is characterized in that,
When each the G pixel in the zone that will carry out noise removing in the image is carried out the noise removing processing, elder generation is according to the value of all non-similar pixels in the unit area that is the center with this G pixel, judge whether this unit area is smooth region, if, the value of all or part of non-similar G pixel in the value of this G pixel and this unit area is weighted on average, and the result is as the value of this G pixel after upgrading.
2, the method for claim 1 is characterized in that,
After judging that described unit area is smooth region, 4 adjacent non-similar G pixel point values of value and diagonal angle thereof of described G pixel are weighted on average, value after the result upgrades as this G pixel, the weights that add temporary the adjacent G pixel in 4 diagonal angles are not zero entirely.
3, method as claimed in claim 1 or 2 is characterized in that, adopts a kind of in the following method to judge whether described unit area is smooth region:
With the non-similar G pixel of G pixel described in the described unit area vertically, on level and two each directions to the angular direction, which side respectively to be divided into two parts according to its described relatively G pixel in, on each direction, calculate the not absolute value of the difference of homonymy G pixel point value sum respectively, the perhaps absolute value sum of difference between a plurality of corresponding G pixel point values of homonymy not, obtain the value DV of four direction, DH, DL, DR, if these four values are all less than preset threshold value on the direction separately, think that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
Upper in all non-similar G pixels of described G pixel one side and all non-similar G pixels of opposite side to each to the angular direction, calculate the not absolute value of the difference of homonymy G pixel point value sum, the perhaps absolute value sum of difference between a plurality of corresponding G pixel point values of homonymy not, if two to the result of calculation that obtains on the angular direction all less than or smaller or equal to preset threshold value, perhaps two to the result of calculation sum that obtains on the angular direction less than or smaller or equal to preset threshold value, judge that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
Calculate the difference of any two the non-similar G pixels in the described unit area, as the maximum in the absolute value of all differences less than or during smaller or equal to preset threshold value, think that then the unit area is a smooth region, otherwise be non-smooth region; Perhaps
Calculate the maximum and the minimum value of all non-similar G pixels of G pixel described in the described unit area.If the difference of maximum and minimum value less than or during smaller or equal to preset threshold value, think that described unit area is a smooth region, otherwise be non-smooth region.
4, method as claimed in claim 1 or 2 is characterized in that:
Described unit area is to be the zones that comprise 3 row * 3 row at center with described G pixel, adopts a kind of in the following method to judge whether this unit area is smooth region:
Press the adjacent G pixel in diagonal angle of described G pixel vertical respectively, horizontal direction is divided into two groups respectively, by two the angular direction is divided into one group respectively, every group comprises the adjacent G pixel in two diagonal angles on this direction, on each direction, calculate the absolute value of every group of G pixel value difference value respectively, the absolute value addition that when organize arranged more each group is obtained, obtain the value DV of four direction respectively, DH, DL, DR, if these four values are all less than preset threshold value on the direction separately, think that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
Calculate each absolute difference of described G pixel respectively to two adjacent G pixel point values on the angular direction, if these two absolute difference all less than or smaller or equal to preset threshold value, think that then this G pixel and its all neighbor pixels are in same smooth region, otherwise be unsmooth zone; Perhaps
Calculate each absolute difference of described G pixel respectively to two adjacent G pixel point values on the angular direction, if these two absolute difference and less than or smaller or equal to preset threshold value, think that then the unit area that described G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
Calculate the difference between any two pixels in four G pixels adjacent with described G pixel diagonal angle, if the maximum in the absolute value of all differences less than or during smaller or equal to preset threshold value, think that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
In four the G pixels adjacent, find out maximum and minimum value with described G pixel diagonal angle, if the difference of maximum and minimum value less than or during smaller or equal to preset threshold value, think that the unit area that described G pixel is the center is a smooth region, otherwise be unsmooth zone.
5, method as claimed in claim 1 or 2 is characterized in that,
If described unit area is not a smooth region, then keep the value of this G pixel constant.
6, method as claimed in claim 1 or 2 is characterized in that,
If described unit area is not a smooth region, judge the non-similar G pixel of part that is in same smooth region with described G pixel in the non-similar G pixel of G pixel described in the described unit area again, with the value of described G pixel and the value of this part non-similar G pixel be weighted average after, as the value of the described G pixel after upgrading.
7, method as claimed in claim 6 is characterized in that,
Described unit area is to be the zone that comprises 3 row * 3 row at center with described G pixel, if described unit area is not a smooth region, judge the part G pixel that is in same smooth region in the adjacent G pixel in described diagonal angle with described G pixel again, with the value of the adjacent G pixel of value of described G pixel with this part be weighted average after, as the value of the described G pixel after upgrading.
8, method as claimed in claim 6 is characterized in that,
After judging that unit area that described G pixel is the center is non-smooth region, a kind of in such a way upgrades described G pixel point value:
Select its corresponding direction according to the little value in vertical direction DV value and the horizontal direction DH value, obtain the adjacent G pixel in 2 groups of diagonal angles on this direction, calculate the difference between the similar G pixel of two nearest on described G pixel and this direction vertical direction described G pixels more respectively, determine the similar G pixel that its absolute difference is little, to be the one group of diagonal angle adjacent G pixel point value of the similar G pixel of benchmark homonymy and the value weighted average of this G pixel with described G pixel, as the value of the described G pixel after upgrading with this; Perhaps
Calculate the DV of vertical direction, the DH of horizontal direction, two DL to the angular direction, minimum value among the DR, when if this minimum value is DV or DH, select its corresponding direction, obtain the adjacent G pixel in 2 groups of diagonal angles on this direction, and then calculate G (i respectively, j) with this direction vertical direction on difference between the similar G pixel of nearest two and described G pixel, determine the similar G pixel that its absolute difference is less, to be one group of diagonal angle adjacent G pixel point value of the similar G pixel of benchmark homonymy and this G (i with described G pixel with this, j) value weighted average is as the value of the described G pixel after upgrading; When if this minimum value is DL or DR, with the value of adjacent G pixel in two diagonal angles on its respective direction and described G pixel be weighted average after, as the value of the described G pixel after upgrading; Perhaps
Calculate described G pixel difference between any two pixels in four adjacent G pixel point values on to the angular direction, determine the minimization of the sum of absolute value of all differences, then with the value weighted average of the value of described G pixel and two diagonal angle adjacent G pixels corresponding, as the value of the described G pixel after upgrading with this minimum value.
9, a kind of grid noise elimination device that keeps image definition is characterized in that, comprises judging unit and smooth region processing unit, wherein:
Described judging unit, be used for to carry out each G pixel in the zone of noise removing and carry out noise removing when handling image, elder generation is according to the value of all non-similar pixels in the unit area that is the center with this G pixel, judge whether this unit area is smooth region, if send triggering signal to described smooth region processing unit;
Described smooth region processing unit is used for after receiving the triggering signal of described judging unit, and the value of all or part of non-similar G pixel in the value of this G pixel and this unit area is weighted on average, and the result is as the value of this G pixel after upgrading.
10, grid noise elimination device as claimed in claim 9 is characterized in that, also comprises non-smooth region processing unit;
Described judging unit also is used for when judging described unit area and be non-smooth region, sends triggering signal to described non-smooth region processing unit;
After described non-smooth region processing unit is used to receive the triggering signal of described judging unit, keep the value of described G pixel constant, perhaps judge the non-similar G pixel of part that is in same smooth region with described G pixel in the non-similar G pixel of G pixel described in the described unit area again, with the value of described G pixel and the value of this part non-similar G pixel be weighted average after, as the value of the described G pixel after upgrading.
11, as claim 9 or 10 described grid noise elimination devices, it is characterized in that, also comprise non-smooth region processing unit;
Described smooth region processing unit is that 4 adjacent non-similar G pixel point values of the value of described G pixel and diagonal angle thereof are weighted on average, the value after the result upgrades as this G pixel, and the weights that add temporary the adjacent G pixel in 4 diagonal angles are not zero entirely.
As claim 9 or 10 described methods, it is characterized in that 12, described judging unit adopts a kind of in the following method to judge whether described unit area is smooth region:
With the non-similar G pixel of G pixel described in the described unit area vertically, on level and two each directions to the angular direction, which side respectively to be divided into two parts according to its described relatively G pixel in, on each direction, calculate the not absolute value of the difference of homonymy G pixel point value sum respectively, the perhaps absolute value sum of difference between a plurality of corresponding G pixel point values of homonymy not, obtain the value DV of four direction, DH, DL, DR, if these four values are all less than preset threshold value on the direction separately, think that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
Upper in all non-similar G pixels of described G pixel one side and all non-similar G pixels of opposite side to each to the angular direction, calculate the not absolute value of the difference of homonymy G pixel point value sum, the perhaps absolute value sum of difference between a plurality of corresponding G pixel point values of homonymy not, if two to the result of calculation that obtains on the angular direction all less than or smaller or equal to preset threshold value, perhaps two to the result of calculation sum that obtains on the angular direction less than or smaller or equal to preset threshold value, judge that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
Calculate the difference of any two the non-similar G pixels in the described unit area, as the maximum in the absolute value of all differences less than or during smaller or equal to preset threshold value, think that then the unit area is a smooth region, otherwise be non-smooth region; Perhaps
Calculate the maximum and the minimum value of all non-similar G pixels of G pixel described in the described unit area.If the difference of maximum and minimum value less than or during smaller or equal to preset threshold value, think that described unit area is a smooth region, otherwise be non-smooth region.
13, as claim 9 or 10 described methods, it is characterized in that: described unit area is to be the zones that comprise 3 row * 3 row at center with described G pixel, and described judging unit adopts a kind of in the following method to judge whether described unit area is smooth region:
Press the adjacent G pixel in diagonal angle of described G pixel vertical respectively, horizontal direction is divided into two groups respectively, by two the angular direction is divided into one group respectively, every group comprises the adjacent G pixel in two diagonal angles on this direction, on each direction, calculate the absolute value of every group of G pixel value difference value respectively, the absolute value addition that when organize arranged more each group is obtained, obtain the value DV of four direction respectively, DH, DL, DR, if these four values are all less than preset threshold value on the direction separately, think that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
Calculate each absolute difference of described G pixel respectively to two adjacent G pixel point values on the angular direction, if these two absolute difference all less than or smaller or equal to preset threshold value, think that then this G pixel and its all neighbor pixels are in same smooth region, otherwise be unsmooth zone; Perhaps
Calculate each absolute difference of described G pixel respectively to two adjacent G pixel point values on the angular direction, if these two absolute difference and less than or smaller or equal to preset threshold value, think that then the unit area that described G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
Calculate the difference between any two pixels in four G pixels adjacent with described G pixel diagonal angle, if the maximum in the absolute value of all differences less than or during smaller or equal to preset threshold value, think that then the unit area that current G pixel is the center is a smooth region, otherwise be unsmooth zone; Perhaps
In four the G pixels adjacent, find out maximum and minimum value with described G pixel diagonal angle, if the difference of maximum and minimum value less than or during smaller or equal to preset threshold value, think that the unit area that described G pixel is the center is a smooth region, otherwise be unsmooth zone.
14, device as claimed in claim 10 is characterized in that,
Described unit area is to be the zone that comprises 3 row * 3 row at center with described G pixel, after described non-smooth region processing unit is received the triggering signal of described judging unit, judge the part G pixel that is in same smooth region in the adjacent G pixel in described diagonal angle with described G pixel again, with the value of the adjacent G pixel of value of described G pixel with this part be weighted average after, as the value of the described G pixel after upgrading.
15, as claim 10 or 14 described devices, it is characterized in that,
After described non-smooth region processing unit was received the triggering signal of described judging unit, a kind of value to described G pixel in was upgraded in the following ways:
Select its corresponding direction according to the little value in vertical direction DV value and the horizontal direction DH value, obtain the adjacent G pixel in 2 groups of diagonal angles on this direction, calculate the difference between the similar G pixel of two nearest on described G pixel and this direction vertical direction described G pixels more respectively, determine the similar G pixel that its absolute difference is little, to be the one group of diagonal angle adjacent G pixel point value of the similar G pixel of benchmark homonymy and the value weighted average of this G pixel with described G pixel, as the value of the described G pixel after upgrading with this; Perhaps
Calculate the DV of vertical direction, the DH of horizontal direction, two DL to the angular direction, minimum value among the DR, when if this minimum value is DV or DH, select its corresponding direction, obtain the adjacent G pixel in 2 groups of diagonal angles on this direction, and then calculate G (i respectively, j) with this direction vertical direction on difference between the similar G pixel of nearest two and described G pixel, determine the similar G pixel that its absolute difference is less, to be one group of diagonal angle adjacent G pixel point value of the similar G pixel of benchmark homonymy and this G (i with described G pixel with this, j) value weighted average is as the value of the described G pixel after upgrading; When if this minimum value is DL or DR, with the value of adjacent G pixel in two diagonal angles on its respective direction and described G pixel be weighted average after, as the value of the described G pixel after upgrading; Perhaps
Calculate described G pixel difference between any two pixels in four adjacent G pixel point values on to the angular direction, determine the minimization of the sum of absolute value of all differences, then with the value weighted average of the value of described G pixel and two diagonal angle adjacent G pixels corresponding, as the value of the described G pixel after upgrading with this minimum value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2008101057092A CN101272503B (en) | 2008-04-30 | 2008-04-30 | Gridding noise elimination method and device for remaining image definition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2008101057092A CN101272503B (en) | 2008-04-30 | 2008-04-30 | Gridding noise elimination method and device for remaining image definition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101272503A true CN101272503A (en) | 2008-09-24 |
| CN101272503B CN101272503B (en) | 2011-04-06 |
Family
ID=40006149
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2008101057092A Active CN101272503B (en) | 2008-04-30 | 2008-04-30 | Gridding noise elimination method and device for remaining image definition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101272503B (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101859444A (en) * | 2009-04-08 | 2010-10-13 | 佳能株式会社 | Image processing apparatus and image processing method |
| CN102625133A (en) * | 2011-01-31 | 2012-08-01 | 英属开曼群岛商恒景科技股份有限公司 | Detection method of image sensing of bad pixel |
| CN102663702A (en) * | 2012-04-20 | 2012-09-12 | 西安电子科技大学 | Natural image denoising method based on regional division |
| CN102957940A (en) * | 2012-10-30 | 2013-03-06 | 武汉烽火众智数字技术有限责任公司 | Improvement-based four-way operator video noise detection method |
| CN103338337A (en) * | 2013-05-30 | 2013-10-02 | 益海芯电子技术江苏有限公司 | Image noise elimination method in optical indicating device |
| CN104202583A (en) * | 2014-08-07 | 2014-12-10 | 华为技术有限公司 | Image processing device and method |
| CN104486618A (en) * | 2014-12-30 | 2015-04-01 | 浙江宇视科技有限公司 | Video image noise detection method and device |
| CN107424131A (en) * | 2017-07-14 | 2017-12-01 | 北京智慧眼科技股份有限公司 | Image based on deep learning removes grid method and device |
| CN107909555A (en) * | 2017-11-27 | 2018-04-13 | 北京大恒图像视觉有限公司 | A kind of gridding noise elimination method for keeping acutance |
| CN109561255A (en) * | 2018-12-20 | 2019-04-02 | 惠州Tcl移动通信有限公司 | Terminal photographic method, device and storage medium |
| CN109816597A (en) * | 2017-11-22 | 2019-05-28 | 展讯通信（上海）有限公司 | The method and apparatus for removing image GrGb noise |
| CN110786009A (en) * | 2018-09-19 | 2020-02-11 | 深圳市大疆创新科技有限公司 | Method, device and machine-readable storage medium for detecting Bell image |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107018343B (en) * | 2017-03-15 | 2020-07-17 | 珠海全志科技股份有限公司 | Bayer image green channel balancing method and device |
-
2008
- 2008-04-30 CN CN2008101057092A patent/CN101272503B/en active Active
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101859444B (en) * | 2009-04-08 | 2012-08-29 | 佳能株式会社 | Image processing apparatus and image processing method |
| CN101859444A (en) * | 2009-04-08 | 2010-10-13 | 佳能株式会社 | Image processing apparatus and image processing method |
| CN102625133A (en) * | 2011-01-31 | 2012-08-01 | 英属开曼群岛商恒景科技股份有限公司 | Detection method of image sensing of bad pixel |
| CN102625133B (en) * | 2011-01-31 | 2014-11-05 | 英属开曼群岛商恒景科技股份有限公司 | Detection method of image sensing of bad pixel |
| CN102663702A (en) * | 2012-04-20 | 2012-09-12 | 西安电子科技大学 | Natural image denoising method based on regional division |
| CN102663702B (en) * | 2012-04-20 | 2014-07-09 | 西安电子科技大学 | Natural image denoising method based on regional division |
| CN102957940A (en) * | 2012-10-30 | 2013-03-06 | 武汉烽火众智数字技术有限责任公司 | Improvement-based four-way operator video noise detection method |
| CN102957940B (en) * | 2012-10-30 | 2015-03-25 | 武汉烽火众智数字技术有限责任公司 | Improvement-based four-way operator video noise detection method |
| CN103338337A (en) * | 2013-05-30 | 2013-10-02 | 益海芯电子技术江苏有限公司 | Image noise elimination method in optical indicating device |
| CN104202583B (en) * | 2014-08-07 | 2017-01-11 | 华为技术有限公司 | Image processing device and method |
| CN104202583A (en) * | 2014-08-07 | 2014-12-10 | 华为技术有限公司 | Image processing device and method |
| CN104486618A (en) * | 2014-12-30 | 2015-04-01 | 浙江宇视科技有限公司 | Video image noise detection method and device |
| CN107424131A (en) * | 2017-07-14 | 2017-12-01 | 北京智慧眼科技股份有限公司 | Image based on deep learning removes grid method and device |
| CN107424131B (en) * | 2017-07-14 | 2019-11-12 | 智慧眼科技股份有限公司 | Image based on deep learning removes grid method and device |
| CN109816597A (en) * | 2017-11-22 | 2019-05-28 | 展讯通信（上海）有限公司 | The method and apparatus for removing image GrGb noise |
| CN109816597B (en) * | 2017-11-22 | 2020-11-20 | 展讯通信（上海）有限公司 | Method and device for removing GrGb noise of image |
| CN107909555A (en) * | 2017-11-27 | 2018-04-13 | 北京大恒图像视觉有限公司 | A kind of gridding noise elimination method for keeping acutance |
| CN107909555B (en) * | 2017-11-27 | 2020-06-02 | 北京大恒图像视觉有限公司 | Sharpness-keeping grid noise elimination method |
| CN110786009A (en) * | 2018-09-19 | 2020-02-11 | 深圳市大疆创新科技有限公司 | Method, device and machine-readable storage medium for detecting Bell image |
| CN110786009B (en) * | 2018-09-19 | 2021-08-06 | 深圳市大疆创新科技有限公司 | Method, device and machine-readable storage medium for detecting Bell image |
| CN109561255A (en) * | 2018-12-20 | 2019-04-02 | 惠州Tcl移动通信有限公司 | Terminal photographic method, device and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101272503B (en) | 2011-04-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101272503B (en) | Gridding noise elimination method and device for remaining image definition | |
| CN102959958B (en) | Color imaging device | |
| CN101490694B (en) | Image enhancement in the mosaic domain | |
| TWI449027B (en) | Adaptive pixel compensation method | |
| US10417955B2 (en) | Image processing method and device for LED display screen | |
| CN104350744A (en) | Camera system with multi-spectral filter array and image processing method thereof | |
| JP2013048410A5 (en) | Color imaging apparatus and image processing method | |
| WO2016031597A1 (en) | Image processing device, image processing method, and image processing system | |
| CN101335900A (en) | Image processing device, image processing method, program, and imaging device | |
| CN103905802A (en) | Method and device for mosaic removal based on P-mode color filter array | |
| CN100563302C (en) | A kind of bad point self-adapted grid noise elimination device and method | |
| CN110248095B (en) | Focusing device, focusing method and storage medium | |
| CN100508556C (en) | Grid noise detection and elimination device of the digital image and its method | |
| CN103621070B (en) | Imaging device and control method thereof | |
| CN107644437B (en) | Color cast detection system and method based on blocks | |
| WO2008007729A1 (en) | Image analyzing method, image analyzing apparatus, inspecting apparatus, image analyzing program and computer readable recording medium | |
| CN103430552B (en) | Imaging device | |
| CN103563360B (en) | Imaging device and imaging method | |
| CN109285113B (en) | Improved color image interpolation method based on gradient | |
| CN109788261A (en) | Color displacement bearing calibration and device | |
| CN103002230A (en) | Adaptive pixel compensating method | |
| US8938120B2 (en) | Image sensing device and image data processing method using the same | |
| CN1326402C (en) | Image signal processing device, circuit, method, program, and recording medium | |
| JP5400573B2 (en) | Image processing apparatus and imaging apparatus | |
| CN114511469B (en) | Intelligent image noise reduction prior detection method |
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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20210128 Address after: No. 602, 6th floor, shining building, 35 Xueyuan Road, Haidian District, Beijing 100083 Patentee after: BEIJING ZHONGXINGTIANSHI TECHNOLOGY Co.,Ltd. Address before: 100083, Haidian District, Xueyuan Road, Beijing No. 35, Nanjing Ning building, 16 floor Patentee before: Vimicro Corp. |
|
| TR01 | Transfer of patent right |