CN102800060A - Quick self-adaption optimizing method for digital image at low illumination level - Google Patents

Abstract

The invention discloses a quick self-adaption optimizing method for a digital image at a low illumination level. The method comprises the following steps of: selecting the digital image acquired at the low illumination level as a source image, acquiring chromatic values of red, green and blue components of each pixel point of the source image to make chroma spectrum of the red, green and blue components, and calculating the mean brightness of the source image; generating a standard image of the source image; and quickly optimizing the standard image to acquire an optimal image. The quick self-adaption optimizing method has the beneficial effects that the optimization of the digital image at the low illumination level can be realized in a self-adaption mode according to the mean brightness of the original image without multiple times of scanning, so that the processing time is greatly saved, and the real-time processing on quality optimization of the digital image is realized; and the method is effectively applied to the field of video processing.

Description

The quick self-adapted optimization method of digital picture under the low-light (level)

Technical field

The invention belongs to technical field of image processing, is to utilize the Zadeh-X transform method to obtain the method for the best in qualityization digital picture adaptively fast.

Background technology

People generally hope to obtain image the best in quality when using the photography and vedio recording equipment.But seldom method relates to the best in qualityization of coloured image.A kind of method (patent No.: ZL 200910190834.2) of obtaining color image with best quality has been proposed before the inventor; The value of transformation parameter Delta that can be through scanning Zadeh-X conversion obtains the corresponding best quality image of original color image to obtain the maximal value of color image quality evaluation functions NCAF.But this method need change the value of the transformation parameter Delta of Zadeh-X conversion one by one to be realized, length consuming time is unfavorable for the real-time processing of coloured image, can not be used for video field.

Summary of the invention

The purpose of this invention is to provide and a kind ofly can realize low-light (level) digital picture method for optimizing apace, adaptively.

For reaching the object of the invention, the present invention proposes the quick self-adapted optimization method of digital picture under a kind of low-light (level), and its key is to comprise the following steps:

Step 1: the digital picture of selecting to obtain under the low-light (level) is source images S, obtain three kinds of components of red, green, blue of each pixel of source images S chromatic value R (x, y), G (x; Y), B (x; Y), make the colourity spectrum of three kinds of components of red, green, blue, and calculate the brightness value L (x of each pixel; Y), obtain the mean flow rate AL of source images S;

Colourity is spectrum can (patent No.: carry out according to Chinese patent " high-resolution detection method that is used for the gradation of image/chrominance information of bottom layer image mining " by method 200610054324.9); Use this method to obtain the colourity spectrum; Can obtain having the high-resolution colourity spectrum of a gamut of chromaticities, more help confirming of boundary value in the step 2.

The mean flow rate AL of source images S obtains through computes:

$\mathrm{AL}=\frac{1}{M\×N}\underset{y=0}{\overset{N-1}{\mathrm{\Σ}}}\underset{x=0}{\overset{M-1}{\mathrm{\Σ}}}L(x,y)$

Wherein, (x is that (x, brightness value y), M, N are the pixel count of source images S on X, Y direction to source images S pixel y) to L.

Pixel (x, brightness value L y) (x, y) can pass through this pixel (x, y) the chromatic value R of three kinds of components of red, green, blue (x, y), G (x, y), B (x y) obtains according to computes:

$L(x,y)=\frac{1}{\sqrt{3}}(\sqrt{R^2(x.y)+G^2(x,y)+B^2(x,y)}$

In the formula;

is field of definition constraint constant; Make L (x; Y) in [0,255], change.

What the present invention handled is the digital picture of obtaining under the low-light (level), so AL≤127.5.

Step 2: generate the standardized images B of source images S, standardized images B can obtain through following method:

(1) left side dividing value Leftr, Leftg, Leftb and the right dividing value Rightr, Rightg, the Rightb of three kinds of component colourities spectrums of difference search source image S red, green, blue; The maximal value of getting among Leftr, Leftg, the Leftb is set at Left, gets that minimum value is set at Right among Rightr, Rightg, the Rightb;

(2) confirm the transformation parameter Theta and the Delta of Zadeh-X conversion:

Theta＝Left；

Delta＝Right-Left；

(3) the red, green, blue three-component to source images S carries out the Zadeh-X conversion respectively, and the digital picture after the generating transformation is standardized images B;

Said Zadeh-X transform method is following:

$T(k,x,y)=k\frac{O(x,y)-\mathrm{Theta}}{\mathrm{Delta}}$

Constraint condition is:

$T(k,x,y)=\left\{\begin{array}{cc}k,& T(k,x,y)>k\\ 0,& T(k,x,y)<0\end{array}\right.$

In the following formula, O (x, y) expression source images S pixel (x, the chromatic value of three kinds of components of red, green, blue y); T (k, x, y) standardized images B pixel (x, the chromatic value of three kinds of components of red, green, blue y) after the expression conversion; K=255.

Standardized images B is for source images S, and the brightness and contrast is improved, and from the visual effect of human eye, quality is improved.

Step 3: standardized images B is carried out quick optimization, obtain optimization image Z;

Fast optimization realizes through the Zadeh-X conversion, and in the Zadeh-X conversion, two transformation parameter Delta and Theta value respectively are:

Delta＝5×AL ^0.811

Theta＝0

Be that the Zadeh-X conversion can be exchanged into: $T(k,x,y)=k\frac{O(x,y)-0}{5\&times;{\mathrm{AL}}^{0.811}}$

Constraint condition is: $T(k,x,y)=\left\{\begin{array}{cc}k,& T(k,x,y)>k\\ 0,& T(k,x,y)<0\end{array}\right.$

In the following formula, O (x, y) expression standardized images B pixel (x, the chromatic value of three kinds of components of red, green, blue y); T (k, x, y) optimization image Z pixel (x, the chromatic value of three kinds of components of red, green, blue y) after the expression conversion; K=255; AL is the mean flow rate of source images S.

The Delta that the present invention chooses in the quick optimization procedures of step 3 is after the characteristics of digital picture under the research low-light (level); The process lot of test is summed up and is obtained; In practical application, do not need manual scanning, just the optimization of the adaptive realization picture quality of ability.The optimized adaptive transformation of picture quality key also of the present invention, adaptive basis is the mean flow rate of original image, and that mean flow rate is an image is special.

The three-component chromatic value of the red, green, blue of gray level image all is identical, and the colourity spectrum of three kinds of components that obtain also is identical, and we can think that it is a kind of special coloured image.The brightness of gray level image is gray scale.In standardisation process, can adopt above-mentioned standardization transform method that the gray-scale value of this gray level image is carried out conversion, promptly; At first obtain the gray scale spectrum, give Theta with the left side dividing value Left assignment of gray scale spectrum then, gray scale is composed the poor of the right dividing value Right and left side dividing value Left; Be Right-Left; Assignment is given Delta, then through the Zadeh-X conversion, just can obtain the standardized images of gray level image.In optimization procedures, directly the gray scale of standardization gray level image is carried out the Zadeh-X conversion, just can realize the best in qualityization of gray level image.

Remarkable result of the present invention is: utilize the Zadeh-X transform method, through two main steps, i.e. standardization and quick optimization just can obtain the optimization image of low-light (level) digital picture.The present invention need be through repeatedly scanning; Just can realize the optimization of low-light (level) digital picture adaptively based on the mean flow rate of original image; Practiced thrift the processing time greatly, realized real-time processing, effectively be applied to field of video processing the best in qualityization of digital picture.

Description of drawings

Fig. 1 process flow diagram of the present invention;

Fig. 2 (a) is the source images among the embodiment 1, and Fig. 2 (b) is the optimization image of Fig. 2 (a);

Fig. 3 (a) is the source images among the embodiment 2, and Fig. 3 (b) is the optimization image of Fig. 3 (a);

Fig. 4 (a) is the source images among the embodiment 3, and Fig. 4 (b) is the optimization image of Fig. 4 (a).

Embodiment

Below in conjunction with accompanying drawing and specific embodiment the present invention is done further explain.

Embodiment 1:

Flow process as shown in Figure 1: the quick self-adapted optimization method of digital picture under a kind of low-light (level) comprises the following steps:

Step 1: selecting the digital picture Fig. 2 (a) that obtains under the low-light (level) is source images S; Obtain three kinds of components of red, green, blue of each pixel of source images S chromatic value R (x, y), G (x, y), B (x; Y); According to Chinese patent " high-resolution detection method that is used for the gradation of image/chrominance information of bottom layer image mining " (patent No.: method 200610054324.9), make the colourity spectrum of three kinds of components of red, green, blue, and calculate the brightness value L (x of each pixel; Y), obtain the mean flow rate AL=6.7573 of source images S;

Step 2: generate the standardized images B of source images S, standardized images B can obtain through following method:

(1) left side dividing value Leftr, Leftg, Leftb and the right dividing value Rightr, Rightg, the Rightb of three kinds of component colourities spectrums of difference search source image S red, green, blue; The maximal value of getting among Leftr, Leftg, the Leftb is set at Left, gets that minimum value is set at Right among Rightr, Rightg, the Rightb;

(2) confirm the transformation parameter Theta and the Delta of Zadeh-X conversion:

Theta＝Left；

Delta＝Right-Left；

(3) the red, green, blue three-component to source images S carries out the Zadeh-X conversion respectively, and the digital picture after the generating transformation is standardized images B;

Said Zadeh-X transform method is following:

$T(k,x,y)=k\frac{O(x,y)-\mathrm{Theta}}{\mathrm{Delta}}$

Constraint condition is:

$T(k,x,y)=\left\{\begin{array}{cc}k,& T(k,x,y)>k\\ 0,& T(k,x,y)<0\end{array}\right.$

In the following formula, O (x, y) expression source images S pixel (x, the chromatic value of three kinds of components of red, green, blue y); T (k, x, y) standardized images B pixel (x, the chromatic value of three kinds of components of red, green, blue y) after the expression conversion; K=255.

Step 3: standardized images B is carried out quick optimization, obtain the optimization image Z shown in Fig. 2 (b);

Fast optimization realizes through the Zadeh-X conversion, and in the Zadeh-X conversion, two transformation parameter Delta and Theta value respectively are:

Delta＝5×AL ^0.811

Theta＝0

Be that the Zadeh-X conversion can be exchanged into: $T(k,x,y)=k\frac{O(x,y)-0}{5\&times;{\mathrm{AL}}^{0.811}}$

Constraint condition is: $T(k,x,y)=\left\{\begin{array}{cc}k,& T(k,x,y)>k\\ 0,& T(k,x,y)<0\end{array}\right.$

In the following formula, O (x, y) expression standardized images B pixel (x, the chromatic value of three kinds of components of red, green, blue y); T (k, x, y) optimization image Z pixel (x, the chromatic value of three kinds of components of red, green, blue y) after the expression conversion; K=255; AL is the mean flow rate of source images S.

The mean flow rate of the optimization image shown in Fig. 2 (b) is 109.6411, compares the mean flow rate 6.7573 of Fig. 2 (a), and the mean flow rate of Fig. 2 (b) is improved, and from the visual effect of human eye, picture quality is better.

Embodiment 2:

Present embodiment and embodiment 1 are roughly the same, and its difference is: the source images S of present embodiment is shown in Fig. 3 (a), and the mean flow rate that calculates is 47.2394; The optimization image that obtains at last is shown in Fig. 3 (b), and mean flow rate is 106.4164, and the optimization image is than source images, and mean flow rate is improved, and from people's visual effect, quality is also better.

Embodiment 3:

The source images S that present embodiment adopted is a width of cloth gray level image shown in Fig. 4 (a), and its quick self-adapted optimized process is as shown in Figure 1, comprises the following steps:

Step 1: selecting the digital picture Fig. 4 (a) that obtains under the low-light (level) is source images S; Because the chromatic value R of three kinds of components of the red, green, blue of gray level image (x, y), G (x, y), B (x; Y) be identical; According to Chinese patent " high-resolution detection method that is used for the gradation of image/chrominance information of bottom layer image mining " (patent No.: method 200610054324.9), make the gray scale spectrum of this image, and calculate the brightness value L (x of each pixel; Y), obtain the mean flow rate AL=8.8283 of source images S;

Step 2: generate the standardized images B of source images S, standardized images B can obtain through following method:

(1) left side dividing value of search source image S gray scale spectrum is set at Left respectively, and the right dividing value of gray scale spectrum is set at Right;

(2) confirm the transformation parameter Theta and the Delta of Zadeh-X conversion:

Theta＝Left；

Delta＝Right-Left；

(3) the Zadeh-X conversion is carried out in the brightness (being gray scale) of each pixel of source images S, the digital picture after the generating transformation is standardized images B;

Said Zadeh-X transform method is following:

$T(k,x,y)=k\frac{O(x,y)-\mathrm{Theta}}{\mathrm{Delta}}$

Constraint condition is:

$T(k,x,y)=\left\{\begin{array}{cc}k,& T(k,x,y)>k\\ 0,& T(k,x,y)<0\end{array}\right.$

In the following formula, O (x, y) expression source images S pixel (x, brightness value y); T (k, x, y) standardized images B pixel (x, brightness value y) after the expression conversion; K=255.

Step 3: standardized images B is carried out quick optimization, obtain the optimization image Z shown in Fig. 4 (b);

Fast optimization realizes through the Zadeh-X conversion, and in the Zadeh-X conversion, two transformation parameter Delta and Theta value respectively are:

Delta＝5×AL ^0.811

Theta＝0

Be that the Zadeh-X conversion can be exchanged into: $T(k,x,y)=k\frac{O(x,y)-0}{5\&times;{\mathrm{AL}}^{0.811}}$

Constraint condition is: $T(k,x,y)=\left\{\begin{array}{cc}k,& T(k,x,y)>k\\ 0,& T(k,x,y)<0\end{array}\right.$

In the following formula, O (x, y) expression standardized images B pixel (x, brightness value y); T (k, x, y) optimization image Z pixel (x, brightness value y) after the expression conversion; K=255; AL is the mean flow rate of source images S.

The mean flow rate of the optimization image shown in Fig. 4 (b) is 100.0388, compares the mean flow rate 8.8283 of Fig. 4 (a), and the mean flow rate of Fig. 4 (b) is improved, and from the visual effect of human eye, picture quality is better.

Claims (1)

1. the quick self-adapted optimization method of digital picture under the low-light (level) is characterized in that comprising the following steps:

Step 1: the digital picture of selecting to obtain under the low-light (level) is source images S, obtain three kinds of components of red, green, blue of each pixel of source images S chromatic value R (x, y), G (x; Y), B (x; Y), make the colourity spectrum of three kinds of components of red, green, blue, and calculate the brightness value L (x of each pixel; Y), obtain the mean flow rate AL of source images S;

Step 2: generate the standardized images B of source images S, standardized images B can obtain through following method:

(1) left side dividing value Leftr, Leftg, Leftb and the right dividing value Rightr, Rightg, the Rightb of three kinds of component colourities spectrums of difference search source image S red, green, blue; The maximal value of getting among Leftr, Leftg, the Leftb is set at Left, gets that minimum value is set at Right among Rightr, Rightg, the Rightb;

(2) confirm the transformation parameter Theta and the Delta of Zadeh-X conversion:

Theta＝Left；

Delta＝Right-Left；

(3) the red, green, blue three-component to source images S carries out the Zadeh-X conversion respectively, and the digital picture after the generating transformation is standardized images B;

Said Zadeh-X transform method is following:

$T(k,x,y)=k\frac{O(x,y)-\mathrm{Theta}}{\mathrm{Delta}}$

Constraint condition is:

$T(k,x,y)=\left\{\begin{array}{cc}k,& T(k,x,y)>k\\ 0,& T(k,x,y)<0\end{array}\right.$

In the following formula, O (x, y) expression source images S pixel (x, the chromatic value of three kinds of components of red, green, blue y); T (k, x, y) standardized images B pixel (x, the chromatic value of three kinds of components of red, green, blue y) after the expression conversion; K=255;

Step 3: standardized images B is carried out quick optimization, obtain optimization image Z;

Fast optimization realizes through the Zadeh-X conversion, and in the Zadeh-X conversion, two transformation parameter Delta and Theta value respectively are:

Delta＝5×AL ^0.811

Theta＝0

Be that the Zadeh-X conversion can be exchanged into: $T(k,x,y)=k\frac{O(x,y)-0}{5\&times;{\mathrm{AL}}^{0811}}$

Constraint condition is: $T(k,x,y)=\left\{\begin{array}{cc}k,& T(k,x,y)>k\\ 0,& T(k,x,y)<0\end{array}\right.$

In the following formula, O (x, y) expression standardized images B pixel (x, the chromatic value of three kinds of components of red, green, blue y); T (k, x, y) optimization image Z pixel (x, the chromatic value of three kinds of components of red, green, blue y) after the expression conversion; K=255; AL is the mean flow rate of source images S.

Images