CN104182947A - Low-illumination image enhancement method and system - Google Patents

Abstract

The invention discloses a low-illumination image enhancement method and system, and belongs to the technical field of image processing. The method comprises the steps of converting an input original image into an HIS space from an RGB space, keeping the saturability and the hue component of the image unchanged, enhancing the brightness component, converting the image with the brightness enhanced into the RGB space from the HIS space and outputting the image. According to the low-illumination image enhancement method and system, the image brightness is improved, extremely-dark areas (the areas close to black) are prevented from being changed into bright-colored areas after being enhanced, and the image adapts to the human visual characteristics better. Meanwhile, the complexity is low, the calculation amount is small, high-definition images can be processed in real time, and the performance requirement for hardware can be lowered.

Description

A kind of enhancement method of low-illumination image and system

Technical field

The present invention relates to technical field of image processing, relate in particular to a kind of enhancement method of low-illumination image and system.

Background technology

In the application of image acquisition or video monitoring, often have the situation of night vision or the low-light (level) such as backlight, the figure kine bias collecting is dark, can identification low, need to carry out just specifically applying after image enhancement processing.Figure image intensifying is exactly the brightness range of expanded view picture and the overall brightness that improves image, improves the subjective quality of image, make cannot identification image detail can be by human eye or machine recognition.

Common image enhancement processing method comprises luminance transformation method, homomorphic filtering method, gradient field Enhancement Method and Retinex Enhancement Method etc.The luminance transformation method that the histogram equalization method of take is representative can make image brightness distribution more even, and strengthens picture contrast, but easily produces enhancing phenomenon.Homomorphic filtering method based on illumination-reflection model is divided into high and low frequency two parts by image, reaches and strengthens the picture contrast object of compressed image dynamic range simultaneously, but there will be enhancing phenomenon, to specular, shadow enhancement weak effect.Gradient field Enhancement Method is processed the gradient of original image, by reducing image gradient value compressed image dynamic range, increases partial gradient value and strengthens image border.Shortcoming is to make image sharpening to a certain extent, and in gradient field, rebuilds the numerical algorithm that image need to be certain, is not suitable for real-time use.Retinex Enhancement Method is used the luminance component of Gaussian smoothing Function Estimation original image, application illumination compensation method is approached reflected image, can when keeping brightness of image, strengthen the information of image dark place, but exist the problem that operand is large, be difficult to process in real time high-definition image, lack universality, and directly image is strengthened and processes easily generation color loss under RGB color space, do not meet people's visual signature.

Summary of the invention

In view of this, the technical problem to be solved in the present invention is to provide a kind of strong robustness, is easy to realize, be more suitable for enhancement method of low-illumination image and the system of people's visual signature, large to solve current enhancement method of low-illumination image operand, lack universality, be difficult to process in real time the technical matters of high-definition image.

It is as follows that the present invention solves the problems of the technologies described above adopted technical scheme:

According to an aspect of the present invention, a kind of enhancement method of low-illumination image providing comprises:

The original image of input is transformed into HIS space by rgb space;

Keep saturation degree and the tone component of image constant, luminance component is strengthened;

Image after brightness is strengthened is transformed into rgb space output by HIS space.

Preferably, the original image of input being transformed into HIS space by rgb space further comprises as follows and changing:

$I=\frac{R+G+B}{3}$

$H=\arccos \left\{\frac{\frac{1}{2}[(R-G)+(R-B)]}{{[{(R-G)}^2+(R-B)(G-B)]}^{\frac{1}{2}}}\right\}$

$S=1-\frac{\min (R,G,B)+\frac{1}{A\·\mathrm{mean}(R,G,B)}}{\mathrm{mean}(R,G,B)+\frac{1}{A\·\mathrm{mean}(R,G,B)}}$

Wherein, I represents brightness, and S represents saturation degree, and H represents tone; R, G, B represent respectively three color channels of red, green, blue; Min () represents minimum value, and mean () represents average; The span of A is [200,255].

Preferably, keep saturation degree and the tone component of image constant, luminance component strengthened further and comprised:

Luminance component is carried out to auto adapted filtering, obtain illumination pattern;

Calculate the reflectogram of original image;

Adjust the contrast of illumination pattern;

According to the luminance graph after synthetic enhancing of illumination pattern after reflectogram and contrast adjustment.

Preferably, luminance component is carried out to auto adapted filtering, obtains illumination pattern and further comprise:

The template window of default M different N*N; Wherein, M gets the integer that is greater than 1, and N gets the odd number comprising more than 3; Value in template window forms by 0 and 1, and the polygonal shape that the point that is 1 by value forms is different, and all polygons combine and will cover whole template window;

For each pixel, calculate the brightness average Mean of each template window _mand corresponding variance δ _m, wherein, m is 1 to M integer;

Obtain the minimum variance δ of each pixel _mthe brightness average Mean of corresponding templates window _mas the value after the luminance filtering of this pixel, obtain the illumination value of this pixel.

Preferably, the contrast of adjustment illumination pattern further comprises:

In rejecting illumination pattern, illumination value two ends proportion is less than the pixel of default ratio value;

Calculate the average illumination value of remaining pixel;

The illumination value of remaining each pixel of pixel is deducted respectively to average illumination value, obtain illumination differential chart;

The illumination difference of each pixel in illumination differential chart is multiplied by respectively to the default adjustment factor;

Illumination difference after the adjustment of each pixel in illumination differential chart is added respectively to average illumination value, obtain the illumination pattern after contrast is adjusted.

According to another aspect of the present invention, a kind of low-light (level) image intensifier device providing comprises:

The first modular converter, for being transformed into HIS space by the original image of input by rgb space;

Luminance enhancement module, for keeping saturation degree and the tone component of image constant, strengthens luminance component;

The second modular converter, is transformed into rgb space output for the image after brightness is strengthened by HIS space.

Preferably, the first modular converter is specifically for changing as follows:

$I=\frac{R+G+B}{3}$

$H=\arccos \left\{\frac{\frac{1}{2}[(R-G)+(R-B)]}{{[{(R-G)}^2+(R-B)(G-B)]}^{\frac{1}{2}}}\right\}$

$S=1-\frac{\min (R,G,B)+\frac{1}{A\·\mathrm{mean}(R,G,B)}}{\mathrm{mean}(R,G,B)+\frac{1}{A\·\mathrm{mean}(R,G,B)}}$

Wherein, I represents brightness, and S represents saturation degree, and H represents tone; R, G, B represent respectively three color channels of red, green, blue; Min () represents minimum value, and mean () represents average; The span of A is [200,255].

Preferably, luminance enhancement module comprises:

Auto adapted filtering unit, for luminance component is carried out to auto adapted filtering, obtains illumination pattern;

Reflectogram computing unit, for calculating the reflectogram of original image;

Contrast adjustment unit, for adjusting the contrast of illumination pattern;

Image synthesis unit, for synthesizing the luminance graph after strengthening according to the illumination pattern after reflectogram and contrast adjustment.

Preferably, auto adapted filtering unit comprises:

Template-setup subelement, for the template window of default M different N*N; Wherein, M gets the integer that is greater than 1, and N gets the odd number comprising more than 3; Value in template window forms by 0 and 1, and the polygonal shape that the point that is 1 by value forms is different, and all polygons combine and will cover whole template window;

Computation subunit, for for each pixel, calculates the brightness average Mean of each template window _mand corresponding variance δ _m, wherein, m is 1 to M integer;

Determine subelement, for obtaining the minimum variance δ of each pixel _mthe brightness average Mean of corresponding templates window _mas the value after the luminance filtering of this pixel, obtain the illumination value of this pixel.

Preferably, contrast adjustment unit comprises:

Filter subelement, for rejecting illumination pattern illumination value two ends proportion, be less than the pixel of default ratio value;

Illumination mean value computation subelement, for calculating the average illumination value of remaining pixel;

Differential chart computation subunit, for the illumination value of remaining each pixel of pixel is deducted respectively to average illumination value, obtains illumination differential chart;

Adjust subelement, for the illumination difference of each pixel of illumination differential chart is multiplied by respectively to the default adjustment factor;

Syndrome unit, for the illumination difference after the adjustment of each pixel of illumination differential chart is added respectively to average illumination value, obtains the illumination pattern after contrast is adjusted.

The enhancement method of low-illumination image of the embodiment of the present invention and system, compared with prior art, by the brightness of image being strengthened in HIS space, can avoid utmost point dark areas (connecing pullous region) after enhancing, to become bright-coloured colour, more meet people's visual signature; Meanwhile, enhancing algorithm complex is low, operand is little, can not only process in real time high-definition image, and can reduce the performance requirement to hardware.In addition, by improving calculating the formula of saturation degree when the color space conversion, after the pixel of any brightness is strengthened, there will not be image to become flower phenomenon, visual effect is fine, stable performance.In addition, by adaptive filter method, in Automatic-searching neighborhood of pixel points to be filtered, pixel value changes mild region, then calculates the filtered value of pending point according to the pixel value in mild region, can keep well image edge information, avoids image blurring.

Accompanying drawing explanation

The process flow diagram of a kind of enhancement method of low-illumination image that Fig. 1 provides for the embodiment of the present invention.

The process flow diagram of a kind of luminance component Enhancement Method that Fig. 2 provides for the preferred embodiment of the present invention.

The process flow diagram of a kind of brightness adaptive filter method that Fig. 3 provides for the preferred embodiment of the present invention.

The schematic diagram of a kind of template window that Fig. 4 provides for the preferred embodiment of the present invention.

The process flow diagram of a kind of contrast method of adjustment that Fig. 5 provides for the preferred embodiment of the present invention.

The modular structure figure of a kind of low-light (level) image intensifier device that Fig. 6 provides for the embodiment of the present invention.

A kind of luminance component that Fig. 7 provides for the preferred embodiment of the present invention strengthens the structural representation of module.

The structural representation of a kind of brightness auto adapted filtering module that Fig. 8 provides for the preferred embodiment of the present invention.

The structural representation of a kind of contrast adjusting module that Fig. 9 provides for the preferred embodiment of the present invention.

Embodiment

In order to make technical matters to be solved by this invention, technical scheme and beneficial effect clearer, clear, below in conjunction with drawings and Examples, the present invention is further elaborated.Should be appreciated that specific embodiment described herein, only in order to explain the present invention, is not intended to limit the present invention.

Embodiment mono-

As shown in Figure 1, a kind of enhancement method of low-illumination image that the embodiment of the present invention provides, comprises the following steps:

S10, the original image of input is transformed into HIS space by rgb space.

Preferably, can adopt following formula to change:

$I=\frac{R+G+B}{3}---\left(1\right)$

$H=\arccos \left\{\frac{\frac{1}{2}[(R-G)+(R-B)]}{{[{(R-G)}^2+(R-B)(G-B)]}^{\frac{1}{2}}}\right\}---\left(2\right)$

$S0=1-\frac{3\·\min (R,G,B)}{R+G+B}---\left(3\right)$

$S=1-\frac{\min (R,G,B)+\frac{1}{A\·\mathrm{mean}(R,G,B)}}{\mathrm{mean}(R,G,B)+\frac{1}{A\·\mathrm{mean}(R,G,B)}}---\left(4\right)$

Wherein, I represents brightness, and S represents saturation degree, and H represents tone; R, G, B represent respectively three color channels of red, green, blue; Min () represents minimum value, and mean () represents average; The span of A is [200,255].The value of A is to make when certain pixel is very dark, and the S value of this pixel approaches 0, and when the color of certain pixel is very bright-coloured, S value is tending towards 1, and therefore, the span of A is [200,255].

Specifically, saturation degree S reflects colour purity, span from 0 to 1, and saturation degree is larger, and it is more bright-coloured that color seems.But when the value of three passages of pixel other two when larger, as (0,0.003 relatively of very little and values of having a passage all, 0), human eye looks like black, and saturation degree should be close to 0, but the saturation degree of calculating according to traditional calculations saturation degree formula (3) is 1, therefore formula (3) is inapplicable to some pixel, especially when brightness value increases, visually, this point has become green by black, therefore, may become colored after figure image intensifying, visual effect is bad.By formula (3) is improved to after above-mentioned formula (4), when in formula (4), the value of A is 255, the pixel saturation degree that is (0,0.003,0) according to formula (4) calculating pixel point triple channel value is 0.00025.Visible, the saturation degree of utilizing formula (4) to calculate more approaches human eye vision, after the pixel of any brightness is strengthened, there will not be image to become flower phenomenon, and visual effect is fine, stable performance.

It should be noted that, when the original image of input picture is not R, G, B spatial image, need first convert thereof into R, G, B space.

Saturation degree and the tone component of S20, maintenance image are constant, and luminance component is strengthened.

Refer to Fig. 2, as a kind of preferred embodiment, this step S20 is further comprising the steps:

S201, luminance component is carried out to auto adapted filtering, obtain illumination pattern.

The reflectogram of S202, calculating original image.

Particularly, at log-domain, the brightness value I of each pixel in original image is deducted to the brightness value L of corresponding pixel points in illumination pattern, obtain the brightness value R of corresponding pixel points in the reflectogram on log-domain, that is:

logR＝logI-logL

The contrast of S203, adjustment illumination pattern.

Luminance graph I after S204, the synthetic enhancing of illumination pattern after adjusting according to reflectogram and contrast _out.

Particularly, at log-domain, by the brightness value R of each pixel in reflectogram add corresponding pixel points in the illumination pattern after contrast reduces brightness value L ', the brightness value I after obtaining brightness on log-domain and strengthening _out, that is:

logI _out＝logL'+logR

S30, the image after brightness is strengthened are transformed into rgb space output by HIS space.

Particularly, H, S component remain unchanged, the I after strengthening with brightness _out, HIS space is transformed into rgb space, the RGB coloured image after being enhanced.For instance, conversion formula is as follows:

When 120 ° of 0≤H <,

$\left\{\begin{array}{c}{R}_{\mathrm{out}}=I_{\mathrm{out}}[1+\frac{S\&CenterDot;\cos H}{\cos (\mathrm{\&pi;}/3-H)}]\\ G_{\mathrm{out}}=3I_{\mathrm{out}}[1-\frac{R_{\mathrm{out}}+B_{\mathrm{out}}}{3I_{\mathrm{out}}}]\\ B_{\mathrm{out}}=I_{\mathrm{out}}(1-S)\end{array}\right.---\left(7\right)$

When 240 ° of 120 °≤H <,

When 240 °≤H≤360 °,

By above-mentioned three color channel R that calculate after enhancing _out, G _out, B _outand output image.

The enhancement method of low-illumination image of the embodiment of the present invention, by the brightness of image being strengthened in HIS space, can avoid utmost point dark areas (connecing pullous region) after enhancing, to become bright-coloured colour, more meets people's visual signature; Meanwhile, enhancing algorithm complex is low, operand is little, can not only process in real time high-definition image, and can reduce the performance requirement to hardware.In addition, by improving calculating the formula of saturation degree when the color space conversion, after the pixel of any brightness is strengthened, there will not be image to become flower phenomenon, visual effect is fine, stable performance.

Embodiment bis-

As shown in Figure 3, a kind of brightness adaptive filter method that the preferred embodiment of the present invention provides, the method is to take template window computing as basis, comprises the following steps:

The template window of S2011, default M different N*N.

Wherein, M gets the integer that is greater than 1, and N gets the odd number comprising more than 3; Value in template window forms by 0 and 1, and the polygonal shape that the point that is 1 by value forms is different, and all polygons combine and will cover whole template window;

Particularly, set in advance the template window of M different N*N, value in this template window forms by 0,1, M gets the integer that is greater than 1, N gets the odd number comprising more than 3, the polygonal shape that the point that is 1 by value in the template window of this M different N*N forms is different, and the selected principle of polygon is that all polygons will cover whole template window.Take get M as 9, N is 5, the template window of also getting 9 5*5 is example, but is not limited to this template window, makes 9 kinds of different windows, the number of window also can be set according to user's actual needs.Please refer to Fig. 4, the totally 9 kinds of different template window as shown in Fig. 4 (a)～Fig. 4 (i), 4 pentagons that the point that it is 1 that these 9 kinds of difform template window comprise by value forms, 4 hexagons, 1 square that the length of side is 3.

S2012, for each pixel, calculate the brightness average Mean of each template window _mand corresponding variance δ _m.

Wherein, m is 1 to M integer; The window of M N*N is slided in entire image, can calculate according to formula (5), (6) the brightness average Mean of the template window of this pixel to be filtered _mand corresponding variance δ _m, wherein, the point that value is 1 in these a plurality of template window represents with k, k=1, and 2,3 ..., P, the number that P is 1 for each template window intermediate value, the pixel value of some k correspondence position in luminance picture I that this value is 1 represents with I (i, j):

${\mathrm{Mean}}_m=\frac{\underset{k=1}{\overset{k=P}{\mathrm{\&Sigma;}}}I(i,j)}{P}---\left(5\right)$

The computing formula of variance is:

${\mathrm{\&delta;}}_m=\underset{k=1}{\overset{k=P}{\mathrm{\&Sigma;}}}(I^2(i,j)-{\mathrm{Mean}}_m^2)---\left(6\right)$

Thereby, for each pixel to be filtered, can obtain M variance yields σ of this pixel _m.

S2013, obtain the minimum variance δ of each pixel _mthe brightness average Mean of corresponding templates window _mas the value after the luminance filtering of this pixel.

Particularly, for each pixel, by the M a calculating σ _msort, find minimum σ _mthe brightness average Mean of corresponding templates _mas filtering result.The window of N*N is slided in entire image, utilize said method just can realize the filtered value to all pixels, obtain image irradiation information.

In the embodiment of the present invention, by pixel value in Automatic-searching neighborhood of pixel points to be filtered, change mild region, then calculate the filtered value of pending point according to the pixel value in mild region, can keep well image edge information, avoid fuzzy.

Embodiment tri-

As shown in Figure 5, a kind of illumination pattern contrast method of adjustment that the preferred embodiment of the present invention provides, comprises the following steps:

In S2031, rejecting illumination pattern, illumination value two ends proportion is less than the pixel of default ratio value.

Particularly, reject the impact that the very little pixel of illumination value proportion can further be eliminated noise, default ratio value is generally got 1% pixel.

S2032, calculate the average illumination value of remaining pixel.

S2033, the illumination value of remaining each pixel of pixel is deducted respectively to average illumination value, obtain illumination differential chart.

S2034, the illumination difference of each pixel in illumination differential chart is multiplied by respectively to the default adjustment factor.

Wherein, the default adjustment factor is got the numerical value between 0 to 2.

S2035, the illumination difference after the adjustment of each pixel in illumination differential chart is added respectively to average illumination value, obtain the illumination pattern after contrast is adjusted.

In the present embodiment, the ultimate facts being formed by illumination information, object edge information and three parts of random noise based on image, by in the situation that Retain edge information is constant, when adjusting illumination contrast, can keep preferably the detailed information of the original image of input.

Embodiment tetra-

As shown in Figure 6, a kind of low-light (level) image intensifier device that the embodiment of the present invention provides, comprises with lower module:

The first modular converter 10, for being transformed into HIS space by the original image of input by rgb space.

Luminance enhancement module 20, for keeping saturation degree and the tone component of image constant, strengthens luminance component.

The second modular converter 30, is transformed into rgb space output for the image after brightness is strengthened by HIS space.

Preferably, the first modular converter 10 can adopt following formula to change:

$I=\frac{R+G+B}{3}---\left(1\right)$

$H=\arccos \left\{\frac{\frac{1}{2}[(R-G)+(R-B)]}{{[{(R-G)}^2+(R-B)(G-B)]}^{\frac{1}{2}}}\right\}---\left(2\right)$

$S0=1-\frac{3\&CenterDot;\min (R,G,B)}{R+G+B}---\left(3\right)$

$S=1-\frac{\min (R,G,B)+\frac{1}{A\&CenterDot;\mathrm{mean}(R,G,B)}}{\mathrm{mean}(R,G,B)+\frac{1}{A\&CenterDot;\mathrm{mean}(R,G,B)}}---\left(4\right)$

Wherein, wherein, I represents brightness, and S represents saturation degree, and H represents tone; R, G, B represent respectively three color channels of red, green, blue; Min () represents minimum value, and mean () represents average; The span of A is [200,255].The value of A is to make when certain pixel is very dark, and the S value of this pixel approaches 0, and when the color of certain pixel is very bright-coloured, S value is tending towards 1, and therefore, the span of A is [200,255].

Specifically, saturation degree S reflects colour purity, span from 0 to 1, and saturation degree is larger, and it is more bright-coloured that color seems.But when the value of three passages of pixel other two when larger, as (0,0.003 relatively of very little and values of having a passage all, 0), human eye looks like black, and saturation degree should be close to 0, but the saturation degree of calculating according to traditional calculations saturation degree formula (3) is 1, therefore formula (3) is inapplicable to some pixel, especially when brightness value increases, visually, this point has become green by black, therefore, may become colored after figure image intensifying, visual effect is bad.By formula (3) is improved to after above-mentioned formula (4), when in formula (4), the value of A is 255, the pixel saturation degree that is (0,0.003,0) according to formula (4) calculating pixel point triple channel value is 0.00025.Visible, the saturation degree of utilizing formula (4) to calculate more approaches human eye vision, after the pixel of any brightness is strengthened, there will not be image to become flower phenomenon, and visual effect is fine, stable performance.

Preferably, refer to Fig. 7, luminance enhancement module 20 comprises auto adapted filtering unit 201, reflectogram computing unit 202, contrast adjustment unit 203 and image synthesis unit 204, wherein:

Auto adapted filtering unit 201, for luminance component is carried out to auto adapted filtering, obtains illumination pattern;

Reflectogram computing unit 202, for calculating the reflectogram of original image; Particularly, at log-domain, the brightness value I of each pixel in original image is deducted to the brightness value L of corresponding pixel points in illumination pattern, obtain the brightness value R of corresponding pixel points in the reflectogram on log-domain, that is:

logR＝logI-logL

Contrast adjustment unit 203, for adjusting the contrast of illumination pattern;

Image synthesis unit 204, for synthesizing the luminance graph after strengthening according to the illumination pattern after reflectogram and contrast adjustment.Particularly, at log-domain, by the brightness value R of each pixel in reflectogram add corresponding pixel points in the illumination pattern after contrast reduces brightness value L ', the brightness value I after obtaining brightness on log-domain and strengthening _out, that is:

logI _out＝logL'+logR

The second 30 pairs of modular converters H, S component remain unchanged, with the I after strengthening _out, HIS space is transformed into rgb space, the RGB coloured image after being enhanced.For instance, conversion formula is as follows:

When 120 ° of 0≤H <,

$\left\{\begin{array}{c}{R}_{\mathrm{out}}=I_{\mathrm{out}}[1+\frac{S\&CenterDot;\cos H}{\cos (\mathrm{\&pi;}/3-H)}]\\ G_{\mathrm{out}}=3I_{\mathrm{out}}[1-\frac{R_{\mathrm{out}}+B_{\mathrm{out}}}{3I_{\mathrm{out}}}]\\ B_{\mathrm{out}}=I_{\mathrm{out}}(1-S)\end{array}\right.---\left(7\right)$

When 240 ° of 120 °≤H <,

When 240 °≤H≤360 °,

By above-mentioned three color channel R that calculate after enhancing _out, G _out, B _outand output image.

It should be noted that, when the original image of input picture is not R, G, B spatial image, need first convert thereof into R, G, B space.

The low-light (level) image intensifier device of the embodiment of the present invention, by the brightness of image being strengthened in HIS space, can avoid utmost point dark areas (connecing pullous region) after enhancing, to become bright-coloured colour, more meets people's visual signature; Meanwhile, enhancing algorithm complex is low, operand is little, can not only process in real time high-definition image, and can reduce the performance requirement to hardware.In addition, by improving calculating the formula of saturation degree when the color space conversion, after the pixel of any brightness is strengthened, there will not be image to become flower phenomenon, visual effect is fine, stable performance.

Embodiment five

As shown in Figure 8, a kind of auto adapted filtering unit that the embodiment of the present invention provides comprises following subelement:

Template-setup subelement 2011, for the template window of default M different N*N; Wherein, M gets the integer that is greater than 1, and N gets the odd number comprising more than 3; Value in template window forms by 0 and 1, and the polygonal shape that the point that is 1 by value forms is different, and all polygons combine and will cover whole template window;

Particularly, set in advance the template window of M different N*N, value in this template window forms by 0,1, M gets the integer that is greater than 1, N gets the odd number comprising more than 3, the polygonal shape that the point that is 1 by value in the template window of this M different N*N forms is different, and the selected principle of polygon is the window that all polygons will cover whole template window.Take get M as 9, N is 5, the template window of also getting 9 5*5 is example, but is not limited to this template window, makes 9 kinds of different windows, the number of window also can be set according to user's actual needs.Please refer to Fig. 4, the totally 9 kinds of different template window as shown in Fig. 4 (a)～Fig. 4 (i), 4 pentagons that the point that it is 1 that these 9 kinds of difform template window comprise by value forms, 4 hexagons, 1 square that the length of side is 3.

Computation subunit 2012, for for each pixel, calculates the brightness average Mean of each template window _mand corresponding variance δ _m, wherein, m is 1 to M integer.

Particularly, the window of M N*N is slided in entire image, can calculate according to formula (5), (6) the brightness average Mean of the template window of this pixel to be filtered _mand corresponding variance δ _m, wherein, the point that value is 1 in these a plurality of template window represents with k, k=1, and 2,3 ..., P, the number that P is 1 for each template window intermediate value, the pixel value of some k correspondence position in luminance picture I that this value is 1 represents with I (i, j):

${\mathrm{Mean}}_m=\frac{\underset{k=1}{\overset{k=P}{\mathrm{\&Sigma;}}}I(i,j)}{P}---\left(5\right)$

The computing formula of variance is:

${\mathrm{\&delta;}}_m=\underset{k=1}{\overset{k=P}{\mathrm{\&Sigma;}}}(I^2(i,j)-{\mathrm{Mean}}_m^2)---\left(6\right)$

Thereby, for each pixel to be filtered, can obtain M variance yields σ of this pixel _m.

Determine subelement 2013, for obtaining the minimum variance δ of each pixel _mthe brightness average Mean of corresponding templates window _mas the value after the luminance filtering of this pixel, obtain the illumination value of this pixel.

Particularly, for each pixel, by the M a calculating σ _msort, find minimum σ _mthe brightness average Mean of corresponding templates _mas filtering result.The window of N*N is slided in entire image, utilize said method just can realize the filtering to all pixels, obtain image irradiation information.

In the embodiment of the present invention, by pixel value in Automatic-searching neighborhood of pixel points to be filtered, change mild region, then calculate the filtered value of pending point according to the pixel value in mild region, can keep well image edge information, avoid fuzzy.

Embodiment six

As shown in Figure 7, the contrast adjustment unit that the preferred embodiment of the present invention provides, comprises following subelement:

Filter subelement 2031, for rejecting illumination pattern illumination value two ends proportion, be less than the pixel of default ratio value;

Illumination mean value computation subelement 2032, for calculating the average illumination value of remaining pixel;

Differential chart computation subunit 2033, for the illumination value of remaining each pixel of pixel is deducted respectively to average illumination value, obtains illumination differential chart;

Adjust subelement 2034, for the illumination difference of each pixel of illumination differential chart is multiplied by respectively to the default adjustment factor;

Syndrome unit 2035, for the illumination difference after the adjustment of each pixel of illumination differential chart is added respectively to average illumination value, obtains the illumination pattern after contrast is adjusted.

In the present embodiment, the ultimate facts being formed by illumination information, object edge information and three parts of random noise based on image, by in the situation that Retain edge information is constant, when adjusting illumination contrast, can keep preferably the detailed information of the original image of input.

One of ordinary skill in the art will appreciate that all or part of step realizing in above-described embodiment method is can control relevant hardware by program to complete, described program can be in being stored in a computer read/write memory medium, described storage medium, as ROM/RAM, disk, CD etc.

With reference to the accompanying drawings of the preferred embodiments of the present invention, not thereby limit to interest field of the present invention above.Those skilled in the art do not depart from the scope and spirit of the present invention, and can have multiple flexible program to realize the present invention, such as the feature as an embodiment can be used for another embodiment, obtain another embodiment.Allly using any modification of doing within technical conceive of the present invention, be equal to and replace and improve, all should be within interest field of the present invention.

Claims (10)

1. an enhancement method of low-illumination image, is characterized in that, the method comprises:

The original image of input is transformed into HIS space by rgb space;

Keep saturation degree and the tone component of image constant, luminance component is strengthened;

Image after brightness is strengthened is transformed into rgb space output by HIS space.

2. enhancement method of low-illumination image according to claim 1, is characterized in that, describedly the original image of input is transformed into HIS space by rgb space further comprises as follows and changing:

$I=\frac{R+G+B}{3}$

$H=\arccos \left\{\frac{\frac{1}{2}[(R-G)+(R-B)]}{{[{(R-G)}^2+(R-B)(G-B)]}^{\frac{1}{2}}}\right\}$

$S=1-\frac{\min (R,G,B)+\frac{1}{A\&CenterDot;\mathrm{mean}(R,G,B)}}{\mathrm{mean}(R,G,B)+\frac{1}{A\&CenterDot;\mathrm{mean}(R,G,B)}}$

Wherein, I represents brightness, and S represents saturation degree, and H represents tone; R, G, B represent respectively three color channels of red, green, blue; Min () represents minimum value, and mean () represents average; The span of A is [200,255].

3. enhancement method of low-illumination image according to claim 1, is characterized in that, the saturation degree of described maintenance image and tone component are constant, and luminance component is strengthened further and comprised:

Luminance component is carried out to auto adapted filtering, obtain illumination pattern;

Calculate the reflectogram of described original image;

Adjust the contrast of described illumination pattern;

According to the luminance graph after synthetic enhancing of illumination pattern after described reflectogram and contrast adjustment.

4. enhancement method of low-illumination image according to claim 3, is characterized in that, described luminance component is carried out to auto adapted filtering, obtains illumination pattern and further comprises:

The template window of default M different N*N; Wherein, M gets the integer that is greater than 1, and N gets the odd number comprising more than 3; Value in described template window forms by 0 and 1, and the polygonal shape that the point that is 1 by value forms is different, and all polygons combine and will cover whole template window;

For each pixel, calculate the brightness average Mean of each template window _mand corresponding variance δ _m, wherein, m is 1 to M integer;

Obtain the minimum variance δ of each pixel _mthe brightness average Mean of corresponding templates window _mas the value after the luminance filtering of this pixel, obtain the illumination value of this pixel.

5. enhancement method of low-illumination image according to claim 3, is characterized in that, the contrast of the described illumination pattern of described adjustment further comprises:

Reject illumination value two ends proportion in described illumination pattern and be less than the pixel of default ratio value;

Calculate the average illumination value of remaining pixel;

The illumination value of remaining each pixel of pixel is deducted respectively to described average illumination value, obtain illumination differential chart;

The illumination difference of each pixel in described illumination differential chart is multiplied by respectively to the default adjustment factor;

Illumination difference after the adjustment of each pixel in described illumination differential chart is added respectively to average illumination value, obtain the illumination pattern after contrast is adjusted.

6. a low-light (level) image intensifier device, is characterized in that, this device comprises:

The first modular converter, for being transformed into HIS space by the original image of input by rgb space;

Luminance enhancement module, for keeping saturation degree and the tone component of image constant, strengthens luminance component;

The second modular converter, is transformed into rgb space output for the image after brightness is strengthened by HIS space.

7. low-light (level) image intensifier device according to claim 6, is characterized in that, described the first modular converter is specifically for changing as follows:

$I=\frac{R+G+B}{3}$

$H=\arccos \left\{\frac{\frac{1}{2}[(R-G)+(R-B)]}{{[{(R-G)}^2+(R-B)(G-B)]}^{\frac{1}{2}}}\right\}$

$S=1-\frac{\min (R,G,B)+\frac{1}{A\&CenterDot;\mathrm{mean}(R,G,B)}}{\mathrm{mean}(R,G,B)+\frac{1}{A\&CenterDot;\mathrm{mean}(R,G,B)}}$

Wherein, I represents brightness, and S represents saturation degree, and H represents tone; R, G, B represent respectively three color channels of red, green, blue; Min () represents minimum value, and mean () represents average; The span of A is [200,255].

8. low-light (level) image intensifier device according to claim 6, is characterized in that, described luminance enhancement module comprises:

Auto adapted filtering unit, for luminance component is carried out to auto adapted filtering, obtains illumination pattern;

Reflectogram computing unit, for calculating the reflectogram of described original image;

Contrast adjustment unit, for adjusting the contrast of described illumination pattern;

Image synthesis unit, for synthesizing the luminance graph after strengthening according to the illumination pattern after described reflectogram and contrast adjustment.

9. low-light (level) image intensifier device according to claim 8, is characterized in that, described auto adapted filtering unit comprises:

Template-setup subelement, for the template window of default M different N*N; Wherein, M gets the integer that is greater than 1, and N gets the odd number comprising more than 3; Value in described template window forms by 0 and 1, and the polygonal shape that the point that is 1 by value forms is different, and all polygons combine and will cover whole template window;

Computation subunit, for for each pixel, calculates the brightness average Mean of each template window _mand corresponding variance δ _m, wherein, m is 1 to M integer;

Determine subelement, for obtaining the minimum variance δ of each pixel _mthe brightness average Mean of corresponding templates window _mas the value after the luminance filtering of this pixel, obtain the illumination value of this pixel.

10. low-light (level) image intensifier device according to claim 8, is characterized in that, described contrast adjustment unit comprises:

Filter subelement, for rejecting described illumination pattern illumination value two ends proportion, be less than the pixel of default ratio value;

Illumination mean value computation subelement, for calculating the average illumination value of remaining pixel;

Differential chart computation subunit, for the illumination value of remaining each pixel of pixel being deducted respectively to described average illumination value, obtains illumination differential chart;

Adjust subelement, for the illumination difference of described each pixel of illumination differential chart is multiplied by respectively to the default adjustment factor;

Syndrome unit, for the illumination difference after the adjustment of described each pixel of illumination differential chart is added respectively to average illumination value, obtains the illumination pattern after contrast is adjusted.