CN101742052A - Method and device for suppressing pseudo color based on edge detection - Google Patents

Abstract

The invention relates to a method and a device for suppressing a pseudo color based on edge detection, which belong to the field of image processing. By analyzing the reason for generating the pseudo color and the characteristics of the pseudo color on a color gamut space, the method and the device eliminate the pseudo color by performing Gaussian smoothing on the method in a range of the color gamut space; and in order to prevent a fuzzy image edge and save computation amount, only the edge which is on the image and generates the pseudo color frequently is processed. The method comprises the following steps: firstly, detecting the edge of the image; secondly, performing morphological processing on the detected edge; and finally, performing Gaussian filter on a chromatic value to suppress the pseudo color. The technology for suppressing the pseudo color provided by the invention is simple, effective, and easy to realize without adding very high hardware cost on the basis of a single pixel point.

Description

A kind of device for suppressing pseudo color and method based on rim detection

Technical field

The invention belongs to image processing field, be specifically related to a kind of device for suppressing pseudo color and method based on rim detection.

Background technology

Digital camera in the prior art at the bigger edge of colour contrast, just can produce dazzling pseudo-coloured silk owing to adopted the color interpolation technology of Mosaic shade filter formula CCD.The pseudo-color picture quality that directly influences digital camera.The pseudo-color phenomenon that occurs on the digital camera, root cause are the color interpolation of Mosaic shade filter formula CCD.Present CCD great majority all are Mosaic shade formulas, CCD itself is perceived color not, see through RGB (perhaps CYGM) filter of each pixel front of CCD, as shown in Figure 1, what existing C CD adopted is Bayer masterplate matrix, and each pixel is only surveyed R, G, B is a kind of density of primary colors wherein, carries out the colorize interpolation processing by the camera in house software again, utilizes peripheral pixel information " conjecture " interpolation to go out other colors.Produce pseudo-color key point just in this color interpolation process, this interpolation process also can not reflect real COLOR COMPOSITION THROUGH DISTRIBUTION (being exactly the diffraction part that part of edge produces) fully with regard to pseudo-coloured silk, algorithm can only go out real panchromatic distribution by the pixel " supposition " of periphery, this has also caused blur margin clear, a series of problems such as color interference, particularly on the bigger edge of contrast ratio, it is high more to produce dazzling pseudo-color probability.Therefore suppressing pseudo-coloured silk is an important process that improves the picture quality of digital camera.

Summary of the invention

At existing problem in the prior art, the object of the present invention is to provide a kind of pseudo-color inhibition technology based on rim detection, by analyzing pseudo-color reason that produces and pseudo-color characteristics on color gamut space, employing is carried out Gauss and is smoothly eliminated pseudo-color on the method for color gamut space scope, in order to prevent the blurred picture edge and save operand that only the normal edge that produces of pseudo-experience in buying the lottery tickets handles on image.

In order to realize the foregoing invention purpose, the invention provides a kind of device for suppressing pseudo color based on the edge, this device comprises colourity and brightness separative element, is used to realize separating of colourity and brightness; A brightness processed unit is used for the luminance signal after separating is carried out processing; The pseudo-color unit that suppresses is used for that the chromatic value after separating is carried out the puppet coloured silk by gaussian filtering and suppresses.

The described pseudo-color unit that suppresses comprises an edge detection unit, is used to ask for edge of image;

An expansion process unit is used for carrying out expansion process to detecting resulting edge;

And Gauss's smoothing processing unit, be used for chromatic value is carried out filtering and suppresses pseudo-color.

Described edge detection unit adopts the sobel operator to calculate the horizontal edge value EH and the vertical edge value EV of each pixel, and calculates edge strength EI;

An one threshold value Th1 and a mask are set, if edge strength EI＞Th1, then the mask value of this pixel correspondence is set to 1, otherwise is set to 0;

The expansion process that described expansion process unit is carried out be dilation (I, B)={ wherein, I is for treating the expansion plans picture for a+b|a ∈ I, b ∈ B}, and B is an expansion mask operator;

The Gaussian filter that described Gauss's smoothing processing unit adopts is:

${\stackrel{\‾}{Y}}_{m,n}=\frac{\underset{i=-2}{\overset{i=2}{\mathrm{\Σ}}}\underset{j=-2}{\overset{j=2}{\mathrm{\Σ}}}{\mathrm{\α}}_{i,j}{\mathrm{\β}}_{i,j}{Y}_{m+i,n+j}}{\underset{i=-2}{\overset{i=2}{\mathrm{\Σ}}}\underset{j=-2}{\overset{j=2}{\mathrm{\Σ}}}{\mathrm{\α}}_{i,j}{\mathrm{\β}}_{i,j}}$

Y wherein _{M, n}Be filtered chromatic value; Y _{M+i, n+j}Be the chromatic value before the filtering; α _{I, j}Coefficient for Gaussian filter; β _{I, j}Be the mask of expansion back edge on image, if mask value is 1, then this pixel participates in filtering.

The present invention also provides a kind of pseudo-color suppression method based on the edge, may further comprise the steps:

At first, the edge by sobel operator detected image;

Secondly, detected edge being carried out morphology handles;

Then, chromatic value is carried out gaussian filtering and suppress pseudo-color.

Further, the sobel operator that adopts during the edge of detected image is horizontal edge value EH and the vertical edge value EV that calculates each pixel, and calculates edge strength EI; An one threshold value Th1 and a mask are set, if edge strength EI＞Th1, then the mask value of this pixel correspondence is set to 1, otherwise the mask value is set to 0.

Detected edge is carried out morphology when handling, expansion process be dilation (I, B)={ wherein, I is for treating the expansion plans picture for a+b|a ∈ I, b ∈ B}, and B is an expansion mask operator.

Chromatic value is carried out gaussian filtering and suppresses pseudo-color Gaussian filter and be:

${\stackrel{\‾}{Y}}_{m,n}=\frac{\underset{i=-2}{\overset{i=2}{\mathrm{\Σ}}}\underset{j=-2}{\overset{j=2}{\mathrm{\Σ}}}{\mathrm{\α}}_{i,j}{\mathrm{\β}}_{i,j}{Y}_{m+i,n+j}}{\underset{i=-2}{\overset{i=2}{\mathrm{\Σ}}}\underset{j=-2}{\overset{j=2}{\mathrm{\Σ}}}{\mathrm{\α}}_{i,j}{\mathrm{\β}}_{i,j}},$

Wherein, Y _{M, n}Be filtered chromatic value; Y _{M+i, n+j}Be the chromatic value before the filtering; α _{I, j}Coefficient for Gaussian filter; β _{I, j}Be the mask of expansion back edge on image, if mask value is 1, then this pixel participates in filtering.

Effect of the present invention is: the pseudo-color technology of inhibition provided by the invention is simply effective, is highly susceptible to realizing, based on single pixel, can not increase very high hardware costs.

Description of drawings

Fig. 1 is the color interpolation schematic diagram of Mosaic shade filter formula CCD;

Fig. 2 is the gradient calculation schematic diagram of Sobel operator;

Fig. 3 is a morphology expansion schematic diagram;

Fig. 4 is the block diagram of device for suppressing pseudo color;

Fig. 5 be adopt that the present invention proposes based on the pseudo-color suppression method at edge or the sun pcolor picture of device;

Fig. 6 be adopt that the present invention proposes based on the pseudo-color suppression method at edge or the image of device.

Embodiment

Below in conjunction with specification drawings and specific embodiments the present invention is described in further detail.

Pseudo-color in order to eliminate, the present invention's employing is carried out Gauss and is smoothly eliminated pseudo-color on the method for color gamut space scope.

Particularly, the present invention at first provides a kind of pseudo-color suppression method based on the edge, be used on the colourity territory, suppressing pseudo-color, at first, for the soft edge that prevents that the overall situation from smoothly bringing, and the saving operand, ask for the normal edge that produces of pseudo-experience in buying the lottery tickets in the image and carry out morphology and handle, it is level and smooth to carry out Gauss then.Particularly, comprise the Image Edge-Detection process, i.e. the gradient of each point in the computed image at first, two-dimentional gradient calculation is got both direction gradient square root sum square, adopts the sobel operator here:

${\mathrm{<figure-callout\; id="1"\; label="S"\; filenames="H2009102420177E00023.png"\; state="\{\{state\}\}">S</figure-callout>}}_1=\left[\begin{array}{ccc}-1& 0& 1\\ -2& 0& 2\\ -1& 0& 1\end{array}\right]$ ${\mathrm{<figure-callout\; id="2"\; label="S"\; filenames="H2009102420177E00012.png"\; state="\{\{state\}\}">S</figure-callout>}}_2=\left[\begin{array}{ccc}-1& -2& -1\\ 0& 0& 0\\ 1& 2& 1\end{array}\right]---\left(1\right)$

When using the Sobel operator, can be detecting template S ₁And S ₂Regard one " framework " as, be enclosed within on each pixel to be detected.As shown in Figure 2, a zone in the rectangle presentation image, wherein each little lattice is represented 1 pixel, and for example the gray value of pixel 21 is a among Fig. 2 ₅, the gray value of its surrounding pixel as shown in FIG., 21 horizontal edge value EH (21) are as follows for calculating pixel:

EH(21)＝(-1)×a ₁+0×a ₂+1×a ₃+(-2)×a ₄+0×a ₅+2×a ₆+(-1)×a ₇+0×a ₈+1×a ₉

Can adopt operator S equally ₂The vertical edge value EV (21) of calculating pixel 21.

For this reason, can calculate edge strength EI: $\mathrm{EI}=\sqrt{{\mathrm{<figure-callout\; id="2"\; label="EH"\; filenames="H2009102420177E00012.png"\; state="\{\{state\}\}">EH</figure-callout>}}^2+{\mathrm{<figure-callout\; id="2"\; label="EV"\; filenames="H2009102420177E00012.png"\; state="\{\{state\}\}">EV</figure-callout>}}^2},$ One threshold value Th1 is set, if edge strength EI＞Th1, promptly think the image border, can generate an image mask in view of the above, it is consistent with original image with size to bury size, for the pixel among the former figure, if its edge strength EI＞Th1, then the mask value of this pixel correspondence is set to 1, otherwise is set to 0.

Carry out the morphology expansion process then, because pseudo-coloured silk often results from around the edge, the present invention adopts mask shown in Figure 3 to expand, this expansion process can be described as dilation (I, B)={ a+b|a ∈ I, b ∈ B}, wherein, I is for treating the expansion plans picture, and B is an expansion mask operator.

Then carry out Gauss's smoothing processing, in an embodiment of the invention, pseudo-color the inhibition is that the edge that detects and the pixel on the back mask that expands thereof are carried out gaussian filtering on Cb and the Cr passage.

The best of Gaussian function is approached by the decision of the coefficient of binomial expansion, and design Gaussian filter approach is direct calculation template weights from discrete Gaussian Profile.For convenience of calculation, wish that generally filter weights is an integer.A value is got at an angle point place in template, and to select a K to make this angle point place value be 1.Can make the filter integer by this coefficient, unaffected because the template weights sum after the integer is not equal to 1 for the even gray areas that guarantees image, can carry out the weights standardization to the filtering template.The binomial expansion coefficient of the sampled value of Gaussian filter or Gaussian filter can form discrete gaussian filter.In one embodiment, select the 5X5 Gaussian kernel that is shown below for use.

$\mathrm{\&alpha;Kernel}=\left[\begin{array}{ccccc}1& 2& 4& 2& 1\\ 2& 4& 8& 4& 2\\ 4& 8& 16& 8& 4\\ 2& 4& 8& 4& 2\\ 1& 2& 4& 2& 1\end{array}\right]---\left(2\right)$

Gaussian filter is designed to:

${\stackrel{\&OverBar;}{Y}}_{m,n}=\frac{\underset{i=-2}{\overset{i=2}{\mathrm{\&Sigma;}}}\underset{j=-2}{\overset{j=2}{\mathrm{\&Sigma;}}}{\mathrm{\&alpha;}}_{i,j}{\mathrm{\&beta;}}_{i,j}{Y}_{m+i,n+j}}{\underset{i=-2}{\overset{i=2}{\mathrm{\&Sigma;}}}\underset{j=-2}{\overset{j=2}{\mathrm{\&Sigma;}}}{\mathrm{\&alpha;}}_{i,j}{\mathrm{\&beta;}}_{i,j}}---\left(3\right)$

Wherein, Y _{M, n}Be filtered chromatic value; Y _{M+i, n+j}Be the chromatic value before the filtering; α _{I, j}Coefficient for Gaussian filter; β _{I, j}Be the mask of expansion back edge on image, if mask value is 1, then this pixel participates in filtering.Described color space is not limited to Yuv and YcbCr, all can adopt said method to realize as long as have the color space of colourity.

Because pseudo-color the inhibition occurs on the colourity territory, if image is the RGB image, need at first with image picture point R, G, B component value transform out chromatic value (u, v), conversion formula is as follows:

$\left[\begin{array}{c}Y\\ u\\ v\end{array}\right]=\left[\begin{array}{ccc}0.299& 0.587& 0.114\\ -0.17& -0.33& 0.5\\ 0.5& -0.42& -0.08\end{array}\right]\left[\begin{array}{c}R\\ G\\ B\end{array}\right]$

The present invention also provides a kind of device for suppressing pseudo color based on the edge simultaneously, as shown in Figure 4, comprise colourity and brightness separative element (not shown), be used to realize separating of colourity and brightness, a brightness processed unit (Y channel Processing) is used for the luminance signal after separating is carried out processing; And a pseudo-coloured silk inhibition unit (False Color Suppression), be used for carrying out the color inhibition of puppet behind the colourity letter after separating.

In the present embodiment, the described pseudo-color unit that suppresses comprises an edge detection unit, is used for asking for the normal edge that produces of the pseudo-experience in buying the lottery tickets of image and carries out morphology and handle, and it is level and smooth to carry out Gauss then.

The Image Edge-Detection process adopts the sobel operator:

${\mathrm{<figure-callout\; id="1"\; label="S"\; filenames="H2009102420177E00023.png"\; state="\{\{state\}\}">S</figure-callout>}}_1=\left[\begin{array}{ccc}-1& 0& 1\\ -2& 0& 2\\ -1& 0& 1\end{array}\right]$ ${\mathrm{<figure-callout\; id="2"\; label="S"\; filenames="H2009102420177E00012.png"\; state="\{\{state\}\}">S</figure-callout>}}_2=\left[\begin{array}{ccc}-1& -2& -1\\ 0& 0& 0\\ 1& 2& 1\end{array}\right]$

Calculate the horizontal edge value EH and the vertical edge value EV of each pixel, and calculate edge strength EI: $\mathrm{EI}=\sqrt{{\mathrm{EH}}^2+{\mathrm{EV}}^2},$ One threshold value Th1 is set, if edge strength EI＞TH1, promptly think the image border, can generate an image mask in view of the above, it is consistent with original image with size to bury size, for the pixel among the former figure, if its edge strength EI＞Th1, then the mask value of this pixel correspondence is set to 1, otherwise is set to 0.

Also comprise an expansion process unit, and execution expansion process: dilation (I, B)={ wherein, I is for treating the expansion plans picture for a+b|a ∈ I, b ∈ B}, and B is an expansion mask operator.

And Gauss's smoothing processing unit, Gaussian filter is designed to:

${\stackrel{\&OverBar;}{Y}}_{m,n}=\frac{\underset{i=-2}{\overset{i=2}{\mathrm{\&Sigma;}}}\underset{j=-2}{\overset{j=2}{\mathrm{\&Sigma;}}}{\mathrm{\&alpha;}}_{i,j}{\mathrm{\&beta;}}_{i,j}{Y}_{m+i,n+j}}{\underset{i=-2}{\overset{i=2}{\mathrm{\&Sigma;}}}\underset{j=-2}{\overset{j=2}{\mathrm{\&Sigma;}}}{\mathrm{\&alpha;}}_{i,j}{\mathrm{\&beta;}}_{i,j}},$

Wherein, Y _{M, n}Be filtered chromatic value; Y _{M+i, n+j}Be the chromatic value before the filtering; α _{I, j}Coefficient for Gaussian filter; β _{I, j}Be the mask of expansion back edge on image, if mask value is 1, then this pixel participates in filtering.

In the optional execution mode, should can also comprise a color space converting unit based on the device for suppressing pseudo color at edge, be arranged on before the YC separative element,, then be converted into the color space that has chromatic value if input image colors does not comprise chromatic value.

Fig. 5 for do not adopt that the present invention proposes based on the pseudo-color suppression method at edge or the sun pcolor picture of device, and Fig. 6 be adopt that the present invention proposes based on the pseudo-color suppression method at edge or the image of device, can see that the edge at Fig. 5 has many pseudo-color, and Fig. 6 is for adopting the effect of 5*5 Gaussian kernel after pseudo-color the inhibition, as can be seen, the present invention can improve the effect of visualization of image, reduce pseudo-color influence to image, can suppress pseudo-color effectively, can not bring bluring on image detail and the edge simultaneously.

Obviously, those skilled in the art can carry out various changes and modification to the present invention and not break away from the spirit and scope of the present invention.Like this, if of the present invention these are revised and modification belongs within the scope of claim of the present invention and equivalent technology thereof, then the present invention also is intended to comprise these changes and modification interior.

Claims (10)

1. device for suppressing pseudo color based on rim detection, it is characterized in that: this device comprises colourity and brightness separative element, is used to realize separating of colourity and brightness; A brightness processed unit is used for the luminance signal after separating is carried out processing; The pseudo-color unit that suppresses is used for that the chromatic value after separating is carried out the puppet coloured silk by gaussian filtering and suppresses.

2. the device for suppressing pseudo color based on rim detection as claimed in claim 1 is characterized in that: the described pseudo-color unit that suppresses comprises an edge detection unit, is used to ask for edge of image;

With an expansion process unit, be used for carrying out expansion process to detecting resulting edge;

And Gauss's smoothing processing unit, be used for chromatic value is carried out filtering and suppresses pseudo-color.

3. the device for suppressing pseudo color based on rim detection as claimed in claim 2 is characterized in that: the sobel operator that described edge detection unit adopts is

$S_1=\left[\begin{array}{ccc}-1& 0& 1\\ -2& 0& 2\\ -1& 0& 1\end{array}\right]$ $S_2=\left[\begin{array}{ccc}-1& -2& -1\\ 0& 0& 0\\ 1& 2& 1\end{array}\right],$

Utilize above-mentioned operator to calculate the horizontal edge value EH and the vertical edge value EV of each pixel, and calculate edge strength EI: $\mathrm{EI}=\sqrt{{\mathrm{EH}}^2+{\mathrm{EV}}^2};$

An one threshold value Th1 and a mask are set, if edge strength EI＞Th1, then the mask value of this pixel correspondence is set to 1, otherwise the mask value is set to 0;

The expansion process that described expansion process unit is carried out be dilation (I, B)={ wherein, I is for treating the expansion plans picture for a+b|a ∈ I, b ∈ B}, and B is an expansion mask operator.

4. the device for suppressing pseudo color based on rim detection as claimed in claim 2 is characterized in that: the Gaussian filter that described Gauss's smoothing processing unit adopts is:

$\stackrel{\&OverBar;}{Y}m,n=\frac{\underset{i=-2}{\overset{i=2}{\mathrm{\&Sigma;}}}\underset{j=-2}{\overset{j=2}{\mathrm{\&Sigma;}}}{\mathrm{\&alpha;}}_{i,j}{\mathrm{\&beta;}}_{i,j}{Y}_{m+i,n+j}}{\underset{i=-2}{\overset{i=2}{\mathrm{\&Sigma;}}}\underset{j=-2}{\overset{j=2}{\mathrm{\&Sigma;}}}{\mathrm{\&alpha;}}_{i,j}{\mathrm{\&beta;}}_{i,j}}$

Wherein, Y _{M, n}Be filtered chromatic value; Y _{M+i, n+j}Be the chromatic value before the filtering; α _{I, j}Coefficient for Gaussian filter; β _{I, j}Be the mask of expansion back edge on image, if mask value is 1, then this pixel participates in filtering.

5. the device for suppressing pseudo color based on rim detection as claimed in claim 1, it is characterized in that: this device also comprises a color space converting unit, be arranged on before the YC separative element, if input image colors does not comprise chromatic value, then be converted into the color space that has chromatic value.

6. pseudo-color suppression method based on rim detection is characterized in that may further comprise the steps:

At first, the edge by sobel operator detected image;

Secondly, detected edge being carried out morphology handles;

Then, chromatic value is carried out gaussian filtering and suppress pseudo-color.

7. the pseudo-color suppression method based on rim detection as claimed in claim 6 is characterized in that: during the edge of detected image, the sobel operator of employing is

$S_1=\left[\begin{array}{ccc}-1& 0& 1\\ -2& 0& 2\\ -1& 0& 1\end{array}\right]$ $S_2=\left[\begin{array}{ccc}-1& -2& -1\\ 0& 0& 0\\ 1& 2& 1\end{array}\right],$

Utilize above-mentioned operator to calculate the horizontal edge value EH and the vertical edge value EV of each pixel, and calculate edge strength EI: $\mathrm{EI}=\sqrt{{\mathrm{EH}}^2+{\mathrm{EV}}^2};$

An one threshold value Th1 and a mask are set, if edge strength EI＞Th1, then the mask value of this pixel correspondence is set to 1, otherwise the mask value is set to 0.

8. the pseudo-color suppression method based on rim detection as claimed in claim 6 is characterized in that: when detected edge was carried out the morphology processing, expansion process was dilation (I, B)=a+b|a ∈ I, b ∈ B}, wherein, I is for treating the expansion plans picture, and B is an expansion mask operator.

9. the pseudo-color suppression method based on rim detection as claimed in claim 6 is characterized in that: chromatic value is carried out gaussian filtering and suppresses pseudo-color Gaussian filter and be:

$\stackrel{\&OverBar;}{Y}m,n=\frac{\underset{i=-2}{\overset{i=2}{\mathrm{\&Sigma;}}}\underset{j=-2}{\overset{j=2}{\mathrm{\&Sigma;}}}{\mathrm{\&alpha;}}_{i,j}{\mathrm{\&beta;}}_{i,j}{Y}_{m+i,n+j}}{\underset{i=-2}{\overset{i=2}{\mathrm{\&Sigma;}}}\underset{j=-2}{\overset{j=2}{\mathrm{\&Sigma;}}}{\mathrm{\&alpha;}}_{i,j}{\mathrm{\&beta;}}_{i,j}},$

Wherein, Y _{M, n}Be filtered chromatic value; Y _{M+i, n+j}Be the chromatic value before the filtering; α _{I, j}Coefficient for Gaussian filter; β _{I, j}Be the mask of expansion back edge on image, if mask value is 1, then this pixel participates in filtering.

10. the pseudo-color suppression method based on rim detection as claimed in claim 6 is characterized in that: if input image colors does not comprise chromatic value, then be converted into the color space that has chromatic value.

Images