CN1610414A - Method for removing colour edge - Google Patents

Method for removing colour edge Download PDF

Info

Publication number
CN1610414A
CN1610414A CN 200310101623 CN200310101623A CN1610414A CN 1610414 A CN1610414 A CN 1610414A CN 200310101623 CN200310101623 CN 200310101623 CN 200310101623 A CN200310101623 A CN 200310101623A CN 1610414 A CN1610414 A CN 1610414A
Authority
CN
China
Prior art keywords
color
pixel
digital image
chroma
brightness
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.)
Pending
Application number
CN 200310101623
Other languages
Chinese (zh)
Inventor
丁鸣举
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Realtek Semiconductor Corp
Original Assignee
Realtek Semiconductor Corp
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Realtek Semiconductor Corp filed Critical Realtek Semiconductor Corp
Priority to CN 200310101623 priority Critical patent/CN1610414A/en
Publication of CN1610414A publication Critical patent/CN1610414A/en
Pending legal-status Critical Current

Links

Images

Landscapes

  • Image Processing (AREA)
  • Facsimile Image Signal Circuits (AREA)

Abstract

The method of eliminating color edge in digital color image includes finding out the boundary between two color blocks in digital color image and finding out one proper parameter for regulating the color expectant of the boundary. The method can avoid the appearance of unnecessary color in the boundary, or color edge.

Description

A kind of method of eliminating colored limit
Technical field
The present invention is the method for relevant a kind of color digital image processing, particularly relevant a kind of method of eliminating the colored limit of chromatic image.
Background technology
The employed inductive component of digital camera (DSC) can be Charged Coupled Device (ChargeCoupled Device is called for short CCD) or CMOS (Complementary Metal Oxide Semiconductor) (CMOS) inductive component at present.Induction installation with one 640 * 480 pixel is an example, please refer to Fig. 1, and Fig. 1 is the schematic diagram of 640 * 480 pixel induction installations.In Fig. 1, induction installation 100 is made of 640 * 480 inductive components 105, and all disposes colored filter 110 on each inductive component 105, and this colored filter only allows the light beam of single kind of color (as red R or green G or blue B) to pass through.And red R, green G and blue B) the arrangement of colored filter, be generally the arrangement mode that Bel organizes filter more.
Therefore, when light exposes to the inductive component 105 of induction installation 100 by colored filter, each inductive component 105 will only can be sensed the brightness of a kind of coloured light (R or G or B).In Fig. 1, represent the color of the light beam of inductive component 105 inductions with R or G or B.Inductive component 105 is promptly according to the corresponding charge inducing of exporting of the brightness size of the coloured light line of sensing.Afterwards, digital camera is again according to these induced currents, to form last color digital image.
Wherein, because each inductive component 105 only can be responded to wherein one and produce induced current of RGB three coloured light.Therefore, in the time of need having R, G, B numerical value simultaneously as if each pixel, the resulting brightness value of inductive component of other coloured light of induction around then need utilizing just can be tried to achieve via interpolation method.
When a certain pixel is positioned at the junction of two color lumps on the color digital image, when then utilizing above-mentioned interpolation method to try to achieve the color of this pixel, have unnecessary color and occur.In this article, this phenomenon is called as the colored limit of color digital image.
Please refer to Fig. 2, Fig. 2 represents the schematic diagram of two adjacent color lumps with R or G or B for inductive component.In Fig. 2, two adjacent color lumps 200,210 will be represented red color lump and green color lump respectively, therefore on the red color lump 200 among the R or G or B value of each inductive component representative, only there is R to have value 255, and each inductive component penetrates among the R or G or B of light representative on the green color lump 210, only has G to have value 255.
In two color lumps 200,210, the position that occurs colored limit easily then will be positioned at the B on two color lumps, 200,210 junctions 24, B 44, B 64Deng the position.Lift wherein B 44The position is an example, when calculating B with interpolation method 44During the color of position, get B usually 44The R value of location about and G value are on average with decision B 44The position the R value and the G value that should have, add B 44The B value that position itself has is with decision B 44The position the color that should have.
Therefore, B 44The R that try to achieve with interpolation method the position, G, B value are respectively:
R=(255(R 33)+0(R 35)+255(R 53)+0(R 55))/4=127.5128
G=(0(G 34)+0(G 54)+0(G 43)+255(G 45))/4=63.7564
B=0(B 44)
So last B 44The position determines its R that is had, G, B value to be respectively 128,64,0 with interpolation method, i.e. B 44The color that the position had will be partial to coffee-like.Wherein, because B 44The position belongs to and will form red block 200, therefore works as B 44When the position uses interpolation method to ask its R that has and G value, R and G value, the i.e. R that will form in the green block 200 have been quoted 35, R 55, G 45Position itself respectively the representative R value and G value, can not reflect B 44The position the R and the G value that should have.
And when representing red block 200 its boundary B 44When having coffee color on the position, just on red block 200 borders unnecessary color has appearred in the color digital image, promptly colored limit.
Therefore when red block 200 colored limit occurs with green block 210 junctions in the color digital image, the border of two blocks 200,210 thickens, in like manner, when all the chromatic colour limit occurred between each block in the color digital image, it is very poor that the sophistication of whole color digital image becomes immediately.
In view of this, the present invention proposes a kind of method of eliminating colored limit, can effectively eliminate in the color digital image between block unnecessary color on the junction, so that the quality optimization of whole color digital image.
Summary of the invention
Main purpose of the present invention provides a kind of method of eliminating colored limit, in order to eliminate the colored limit on the color digital image, its method comprises the border of finding out two color lump junctions on the color digital image, and finds out a suitable parameter to adjust the color degree of saturation on this border.
In preferred embodiment of the present invention, the method of finding out chromatic image dichromatism block boundary comprises: according to certain pixel on the color digital image and arround it between pixel the residual quantity of brightness to find out this border, adjust the color degree of saturation on this border, the colored limit of promptly eliminating two color lump intersections at boundary.
Description of drawings
Fig. 1 is the schematic diagram of 640 * 480 pixel induction installations;
To be inductive component represent the schematic diagram of two adjacent color lumps with R or G or B to Fig. 2;
Fig. 3 is that inductive component representative R, G in the preferred embodiment of the present invention, B value color conversion are the schematic diagram of color digital image pixel brightness and chroma;
Fig. 4 is Y in the color digital image AbPixel use different variable elements (Gain) adjust itself and arround unit brightness measures of dispersion E (Y between pixel Ab) graph of a relation.
The figure number explanation:
100 induction installations; 105 inductive components; 110 colored filters; 200,210 color lumps;
The R of 310 every bit pixels or G or B value; The R of 320 every bit pixels, G, B value;
Brightness (Y), the chroma (C of 330 every bit pixel representatives b, C r).
Embodiment
For making the public have further cognitive and understanding, cooperate now feature of the present invention, purpose and function
Accompanying drawing is described in detail as follows:
The practice of the present invention is, finds out the border of color lump and color lump on the color digital image earlier, and then takes out the color of representative color block boundary pixel, so that edge pixel can not occur on the chromatic colour limit.
Be familiar with color digital image skill person as can be known, form R, G, the B value of each pixel color of color digital image, also can see through brightness Y and the chroma (C that the color conversion mode is converted to each pixel color b, C r).As chroma C bAnd C rBe worth greatly more, it is bright-coloured more to can be considered this color, and it is saturated more to can be described as this color.As chroma C bAnd C rBe worth more for a short time, it is unsaturated more to can be described as this color.Therefore, if will find out the border of two adjacent color lumps on the color digital image, only need in fact to see through pixel on the color digital image with arround the luminance difference of pixel, utilize high pass filter or once (slope) differential or repeatedly the mode of differential find out its border.In preferred embodiment of the present invention, utilize the color conversion mode earlier with R, the G of each pixel, brightness (Y) and the chroma (C that the B value is converted to pixel b, C r).Then, in a subdifferential mode, just try to achieve each pixel and arround the unit residual quantity of brightness between pixel, find out the border between color lump in the color digital image.
Please refer to Fig. 3, Fig. 3 is in the preferred embodiment of the present invention, is color digital image pixel brightness (Y) and chroma (C with R, G, the B value color conversion of inductive component representative b, C r) schematic diagram.In Fig. 3, after the inductive component induction, different R or G or B value 310 will be represented respectively.Afterwards, by R or the G or the B value 310 of each inductive component representative, try to achieve last R, G, the B value 320 of taking resulting each pixel of color digital image of this chromatic image via interpolation method again.See through the mode of color conversion at last, R, G, the B value 320 of this each pixel of color digital image are changed the brightness Y and the chroma (C of each pixel representative of color digital image for this reason b, C r) 330.
After finding out the brightness value Y that this each pixel of color digital image had, can by each pixel and arround brightness poor between pixel, find out the border between color lump and color lump on this color digital image.Be Y with brightness among Fig. 3 again 11Pixel be example, Y 11Pixel and arround the residual quantity E (Y of unit of brightness between pixel 11) be equivalent to Y 11The adjacent in the horizontal direction and brightness of pixel is Y 11, Y 12The increment ratio of pixel brightness on its position, and Y 11Pixel is at the adjacent Y of vertical direction 01, Y 21The increment ratio of pixel brightness on its position, promptly
E(Y 11)=(|Y 12-Y 10|+|Y 12-Y 10|)÷2
Microcosmic, E (Y 11) generation of numerical value will represent in this color digital image and produce bright the variation between pixel and adjacent image point; Macroscopic view, E (Y 11) numerical value represents the border between color lump and color lump in this color digital image.In addition, E (Y 11) numerical value big or little also will represent the tangible degree of dichromatism block boundary on this color digital image.In other words, E (Y 11) the big more representative of numerical value two adjacent color lump borders are obvious more, E (Y 11) the more little representative two adjacent color lump heterochromias of numerical value are more little.
Therefore, E (Y 11) when numerical value is big more, represent the color digital image border of two adjacent color lumps to occur, more represent two adjacent color lumps more tangible colored limit to occur, vice versa.
So, according to above-mentioned discovery, E (Y 11) the big young pathbreaker of value determines C b, C rThe degree of revising is in preferred embodiment of the present invention, for example to be cosine function (Cos) nLevel and smooth decreasing function as revising C b, C rVariable element (Gain), and this variable element is according to E (Y 11) be worth change and correction C b, C rC B11, C R11, will become respectively after the correction:
C b11′=Cos((E(Y 11)÷2Q)π) n×C b11
And
C r11′=Cos((E(Y 11)÷2Q)π) n×C r11
In addition, according to E (Y 11) when numerical value is big more, except that representing the color digital image to occur the border of two adjacent color lumps, more represent two adjacent color lumps more tangible colored limit to occur, therefore as E (Y 11) C when numerical value is big more B11, C R11The downward revision amount is big more, and promptly Gain (obtaining increment) is more little.
And the size of n value also determines C simultaneously in the variable element (Gain) B11, C R11The speed of revising please refer to Fig. 4, and Fig. 4 is Y in the color digital image AbPixel use different variable elements (Gain) adjust itself and arround unit brightness measures of dispersion E (Y between pixel Ab) graph of a relation.In Fig. 4, suppose Y on the color digital image earlier AbPixel and arround unit brightness measures of dispersion E (Y between pixel Ab) maximum Q be 255, and Gain is Cos ((E (Y 11) ÷ 2Q) π) n
Because as E (Y Ab) C when numerical value is big more Bab, C RabThe downward revision amount is big more, and promptly Gain is more little.Therefore, as E (Y Ab)=255 o'clock, Gain=0, otherwise as E (Y Ab)=0 o'clock, Gain=1.Wherein, if the n value among the Gain is 1, then the curve distribution of Gain is smoothly, and the n value among the Gain is 4, and then the curve distribution of Gain and n value are to compare in 1 o'clock, and be more precipitous.In other words, if the n value is bigger, E (Y then Ab) curve can be with E (Y Ab) make to change faster i.e. E (Y Ab) the colored limit of representative can comparatively fast eliminate.In preferred embodiment of the present invention, the n value gets 4 for more suitable.
So, on the color digital image border of two color lump junctions (edge), can be by pixel Y on the color digital image AbWith arround the unit brightness measures of dispersion E (Y of pixel Ab) and find out, and the colored limit degree that two color lump junctions are occurred on the color digital image, more can be according to E (Y Ab) big or small to eliminate pixel Y AbColor (C Bab, C Rab), so that dichromatism block boundary pixel Y on the color digital image AbLook is adopted and is detached, and can not produce unnecessary color, and avoids occurring colored limit.
Be noted that the above embodiments are to be the value that benchmark is adjusted first degree of saturation and second degree of saturation with the cosine function, but the present invention is not as limit.Every decreasing function all can be used on the present invention after changing through suitable function.
Comprehensively above-mentioned, the present invention proposes a kind of method of eliminating colored limit, mode by a high pass filter or a subdifferential, find out the border of color digital image two color lumps, and see through a function that successively decreases as parameter adjusting the color degree of saturation of the edge pixel of two color lumps on the color digital image, and avoid the color digital image colored limit to occur and make the optimization of color digital image quality.
The above only is preferred embodiment of the present invention, when can not with limit the scope of the invention.Promptly the equalization of doing according to claim of the present invention generally changes and modifies, and will not lose main idea of the present invention place, does not also break away from the spirit and scope of the present invention, and the former capital should be considered as further enforcement situation of the present invention.

Claims (10)

1, a kind of method is in order to eliminating the colored limit on the color digital image, and this color digital image comprises a lot of pixels, and those pixels have brightness and chroma, comprising:
According to the luminance difference of first pixel, second pixel adjacent, define the border of this color digital image with this first pixel; And
Adjust the chroma of this first pixel according to this luminance difference.
2, the method for claim 1 wherein by differential or the mode of tabling look-up, is tried to achieve this luminance difference.
3, the method for claim 1, wherein this luminance difference is that brightness by this first pixel and second, third adjacent with this first pixel, the 4th and the 5th pixel is determined, wherein, this first, the brightness of this second, the 3rd, the 4th and the 5th pixel is respectively with Y 11, Y 10, Y 12, Y 01, Y 21Represent, then this luminance difference E (Y 11) be for:
E(Y 11)=(|Y 12-Y 10|+|Y 12-Y 10|)÷2
4, as 1 described method of claim the, wherein this method is according to adjusting this chroma with this luminance difference as the transfer function of independent variable.
5, as 4 described methods of claim the, wherein this transfer function is this cosine function of n power, n>1.
6, as 1 described method of claim the, wherein this chroma comprises the first chroma (C b) and the second chroma (C r).
7, a kind of method is that this color digital image comprises a lot of pixels in order to the colored limit on the elimination color digital image, and those pixels have brightness and chroma, comprising:
By these brightness of these pixels, try to achieve luminance difference; And
According to this luminance difference, adjust the chroma of these pixels.
8, as 7 described methods of claim the, wherein this method more comprises:
Change in the color conversion mode, with brightness and the chroma that obtains these pixels.
9,, wherein, try to achieve this luminance difference with the mode or the mode of a subdifferential to table look-up as 7 described methods of claim the.
10, as 7 described methods of claim the, wherein this method is to adjust this chroma with this luminance difference as the transfer function of independent variable.
CN 200310101623 2003-10-23 2003-10-23 Method for removing colour edge Pending CN1610414A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 200310101623 CN1610414A (en) 2003-10-23 2003-10-23 Method for removing colour edge

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 200310101623 CN1610414A (en) 2003-10-23 2003-10-23 Method for removing colour edge

Publications (1)

Publication Number Publication Date
CN1610414A true CN1610414A (en) 2005-04-27

Family

ID=34756201

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 200310101623 Pending CN1610414A (en) 2003-10-23 2003-10-23 Method for removing colour edge

Country Status (1)

Country Link
CN (1) CN1610414A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106652805A (en) * 2016-10-21 2017-05-10 上海天马微电子有限公司 Display panel and display device

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106652805A (en) * 2016-10-21 2017-05-10 上海天马微电子有限公司 Display panel and display device

Similar Documents

Publication Publication Date Title
CN106971695B (en) The brightness adjusting method and device of RGBW liquid crystal display device
US9036087B2 (en) Display device and color adjustment method thereof, and method for displaying video signal
CN101272503B (en) Gridding noise elimination method and device for remaining image definition
US20050179699A1 (en) Image display device employing selective or asymmetrical smoothing
EP3361376A1 (en) Image processing method and device for led display screen
CN106935615A (en) Display screen preparation, display screen body and electronic display unit
US20040258324A1 (en) Method for dynamic contrast enhancement by area gray-level detection
CN109686323A (en) A kind of display device and its driving method and electronic equipment
CN105306913B (en) image obtaining device and image compensation method thereof
CN110490838B (en) Method and device for processing boundaries of areas with different resolutions of display panel
CN1717003A (en) Has the more sub sampling of high display quality in the image sensing device
CN1245018C (en) Method for eliminating scintilltion of moving picture
CN1278299C (en) Method and apparatus for using low-brightness to treat non-natural signal of reducing glint
CN1432988A (en) Gamma correcting device and method for LCD
CN1610414A (en) Method for removing colour edge
CN111192227B (en) Fusion processing method for overlapped pictures
CN111965914A (en) Color electronic ink screen and display method thereof
CN104505053B (en) Show signal conversion method and device
CN1538748A (en) Method of interleave and line by line conversion of mixing two dimensional and three dimensional
CN101534450B (en) Image retrieving device, and unit and method thereof for establishing gain form
CN100454969C (en) Method for transient enhancing video image brightness
US20050025384A1 (en) Method for eliminating a color edge and apparatus thereof
CN1122423C (en) Automatic gain regulator for color video signal of display
CN1176552C (en) Colour difference signal compensation device
CN107516299A (en) Symmetrical multi thresholds error diffusion method and its application for the conversion of fabric color range

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C02 Deemed withdrawal of patent application after publication (patent law 2001)
WD01 Invention patent application deemed withdrawn after publication