CN112365425A - Low-illumination image enhancement method and system - Google Patents
Low-illumination image enhancement method and system Download PDFInfo
- Publication number
- CN112365425A CN112365425A CN202011327998.8A CN202011327998A CN112365425A CN 112365425 A CN112365425 A CN 112365425A CN 202011327998 A CN202011327998 A CN 202011327998A CN 112365425 A CN112365425 A CN 112365425A
- Authority
- CN
- China
- Prior art keywords
- image
- illumination
- component
- low
- calculating
- 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
Links
- 238000005286 illumination Methods 0.000 title claims abstract description 147
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000004364 calculation method Methods 0.000 claims abstract description 34
- 238000012937 correction Methods 0.000 claims abstract description 26
- 238000007781 pre-processing Methods 0.000 claims abstract description 21
- 238000001914 filtration Methods 0.000 claims abstract description 11
- 238000006243 chemical reaction Methods 0.000 claims description 10
- 230000002708 enhancing effect Effects 0.000 claims description 4
- 238000012545 processing Methods 0.000 abstract description 14
- 230000000007 visual effect Effects 0.000 abstract description 6
- 238000009499 grossing Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 230000002411 adverse Effects 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Images
Classifications
-
- G06T5/77—
-
- G06T5/92—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
Abstract
The invention discloses a low-illumination image enhancement method and a system, which are characterized in that firstly, color constancy preprocessing is carried out on an acquired low-illumination image so as to eliminate the color cast phenomenon caused by an artificial light source, then the low-illumination image is converted into a YCbCr color space from an RGB color space, illumination component estimation is carried out by adopting iterative multi-scale guide filtering, then, a reflection component is obtained by calculation according to a Retinex theory and global contrast correction is carried out, and finally, a processing result is converted into the RGB color space so as to obtain a final low-illumination image enhancement result. The low-illumination image enhancement method can effectively enhance the detail information of the image, improve the definition of the image and improve the visual effect of the image, and has the characteristics of rapidness, accuracy and automatic processing.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a low-illumination image enhancement method and system.
Background
Images shot under low illumination conditions such as night and dusk have poor visual quality due to low brightness and contrast, artificial light sources and the like, and great difficulty is brought to image processing and analysis in the fields of transportation, video monitoring, military reconnaissance and the like.
In the field of computer vision and image processing, in order to enhance low-illumination images and improve the visual effect of the images, currently, the mainly adopted enhancement method includes: the method comprises a low-illumination image enhancement method based on gray level transformation, a low-illumination image enhancement method based on an atmospheric scattering model, a low-illumination image enhancement method based on deep learning and a low-illumination image enhancement method based on Retinex theory.
The low-illumination image enhancement method based on gray scale transformation optimizes the gray scale distribution of the low-illumination image by changing the mapping mode or the distribution mode of the image gray scale. Although the brightness and contrast of the image can be improved, the problems of detail information loss, local over-enhancement and the like are brought.
The method for enhancing the low-illumination image based on the atmospheric scattering model comprises the steps of firstly carrying out gray level inversion on the low-illumination image, and then carrying out enhancement processing by adopting an image defogging algorithm. However, this process lacks theoretical basis and is susceptible to noise interference or adverse effects of artificial light sources, resulting in distortion of the processing results.
The low-illumination image enhancement method based on deep learning utilizes a nonlinear mapping relation between a low-illumination image and a normal-illumination image and adopts a deep neural network model to realize end-to-end mapping from the low-illumination image to the normal-illumination image. Although a high visual quality of the image can be obtained, the processing effect of this method depends to a large extent on the quality of the training data set.
The low-illumination image enhancement method based on Retinex theory takes a low-illumination image as a product of an illumination component and a reflection component, and obtains an enhanced image by estimating and eliminating the adverse effect of the illumination component. The method has the characteristics of compressing the dynamic range, meeting the requirement of color constancy and the like, so the method is widely applied to the enhancement of low-illumination images. However, the existing low-illumination image enhancement method based on the Retinex theory has defects in illumination component estimation and elimination of artificial light source interference, so that the problems of detail information loss and color distortion of an enhanced image often occur.
Disclosure of Invention
Based on the technical problems in the background art, the invention provides a low-illumination image enhancement method and system.
The invention provides a low-illumination image enhancement method, which comprises the following steps of;
s1, carrying out color constancy preprocessing on the low-illumination image I to obtain an image I after color correctionnew;
S2, image InewConverting the RGB color space into the YCbCr color space to obtain a brightness component Y and color components Cb and Cr;
s3, conducting iterative multi-scale guiding filtering on the brightness component Y to obtain an illumination component L;
s4, calculating a reflection component R based on Retinex theory according to the illumination component L;
s5, carrying out global contrast correction on the reflection component R to obtain an enhanced reflection component RenhI.e. enhancement result Y of luminance component Yenh;
S6, enhancing result Y of brightness component YenhAnd converting color components Cb and Cr into RGB color space to obtain low-illumination image enhancement result Ifinal。
Preferably, step S1, specifically includes;
s11, acquiring a gray pixel set P in the low-illumination image Ih;
S12, using gray pixel set PhCalculating an illumination adjustment quantity E;
s13, adjusting the low-illumination image I according to the illumination adjustment quantity E to obtain an image I after color correctionnew;
Preferably, step S11 specifically includes:
calculating a local neighborhood omega centered on an arbitrary pixel p in the image IpCorresponding to R, G, B three color channels I respectivelyR、IG、IBGradient deviation Delta I ofR(p)、△IG(p)、△IB(p), and the mean gradient deviation Δ μ (p) and the mean gray value I μ (p);
calculating the gray coefficient GI (P) of any pixel P on the image I, sorting the pixels from small to large according to the gray coefficient GI of each pixel in the image I, and taking the pixel with the lowest value of the first h% as a gray pixel set P in the image Ih;
Preferably, step S12 specifically includes:
from a set of gray pixels P in the image IhR, G, B light adjustment quantity E of three color channels is calculatedR、EG、EB;
Preferably, step S13 specifically includes:
illumination adjustment quantity E according to R, G, B three color channelsR、EG、EBFor three color channels I of the image IR、IG、IBAdjusting to obtain color corrected image Inew。
Preferably, step S3, specifically includes;
s31, calculating an initial illumination component L(0)Average gradient Gr of(0);
S32, calculating the illumination component L of the nth iteration step(n);
S33, calculating the illumination component L of the nth iteration step(n)Average gradient Gr of(n);
S34, according to Gr(n)And Gr(n-1)Calculating the average gradient difference delta Gr;
and S35, judging the termination iteration condition according to the average gradient difference delta Gr to obtain the illumination component L.
The invention also provides a low-illumination image enhancement system, which comprises;
a preprocessing module for performing color constancy preprocessing on the low-illumination image I to obtain an image I after color correctionnew;
A first conversion module, in particular for converting the image InewConverting the RGB color space into the YCbCr color space to obtain a brightness component Y and color components Cb and Cr;
the illumination component calculation module is specifically used for performing iterative multi-scale guided filtering on the brightness component Y to obtain an illumination component L;
the reflection component calculation module is specifically used for calculating and obtaining a reflection component R based on a Retinex theory according to the illumination component L;
a global correction module, specifically configured to perform global contrast correction on the reflection component R to obtain an enhanced reflection component RenhI.e. enhancement of the luminance component Yenh;
A second conversion module, in particular for converting the enhancement result Y of the luminance component YenhAnd converting color components Cb and Cr into RGB color space to obtain low-illumination image enhancement result Ifinal。
Preferably, the preprocessing module is specifically configured to: obtaining a set of gray pixels P in a low-light image Ih(ii) a Using a set of gray pixels PhCalculating an illumination adjustment quantity E; adjusting the low-illumination image I according to the illumination adjustment quantity E to obtain an image I after color correctionnew;
Preferably, the preprocessing module is specifically configured to: computing a local neighborhood ω centered on an arbitrary pixel p in an image IpCorresponding to R, G, B three color channels I respectivelyR、IG、IBGradient deviation Delta I ofR(p)、△IG(p)、 △IB(p), and the mean gradient deviation Δ μ (p) and the mean gray value I μ (p); calculating the gray coefficient GI (P) of any pixel P on the image I, sorting the pixels from small to large according to the gray coefficient GI of each pixel in the image I, and taking the pixels with the lowest value of the first h percent as a gray pixel set P in the image Ih;
Preferably, the preprocessing module is specifically configured to: from a set of gray pixels P in the image IhR, G, B light adjustment quantity E of three color channels is calculatedR、EG、EB;
Preferably, the preprocessing module is specifically configured to: illumination adjustment quantity E according to R, G, B three color channelsR、EG、EBFor three color channels I of the image IR、IG、IBAdjusting to obtain color corrected image Inew。
Preferably, the illumination component calculation module is specifically configured to:
calculating an initial illumination component L(0)Average gradient Gr of(0)(ii) a Calculating illumination component L of nth step iteration(n)(ii) a Calculating illumination component L of nth step iteration(n)Average gradient Gr of(n)(ii) a According to Gr(n)And Gr(n-1)Calculating an average gradient difference delta Gr; and judging the termination iteration condition according to the average gradient difference delta Gr to obtain an illumination component L.
In the invention, firstly, color constancy preprocessing is carried out on the acquired low-illumination image so as to eliminate the color cast phenomenon caused by an artificial light source, then the low-illumination image is converted into a YCbCr color space from an RGB color space, illumination component estimation is carried out by adopting iterative multi-scale guide filtering, then a reflection component is obtained by calculation according to a Retinex theory and global contrast correction is carried out, and finally, the processing result is converted into the RGB color space so as to obtain a final low-illumination image enhancement result. The low-illumination image enhancement method can effectively enhance the detail information of the image, improve the definition of the image and improve the visual effect of the image, and has the characteristics of rapidness, accuracy and automatic processing.
Drawings
Fig. 1 is a schematic flow chart of a low-illumination image enhancement method according to the present invention.
Detailed Description
As shown in fig. 1, fig. 1 is a schematic flow chart of a low-illumination image enhancement method according to the present invention.
Referring to fig. 1, a low-illumination image enhancement method provided by an embodiment of the present invention includes:
step S1, color constancy preprocessing is carried out on the low-illumination image I to obtain an image I after color correctionnew。
The method specifically comprises the following steps:
s11, acquiring a gray pixel set P in the low-illumination image Ih。
First, a local neighborhood ω centered on an arbitrary pixel p in an image I is calculatedpCorresponding to R, G, B three color channels I respectivelyR、IG、IBGradient deviation Delta I ofR(p)、△IG(p)、△IB(p) and the mean gradient deviation Δ μ (p) and mean gray value I μ (p), which are calculated as follows:
next, calculating a gray scale coefficient gi (p) of an arbitrary pixel p on the image I, wherein the calculation formula is as follows:
wherein epsilon is a decimal number greater than 0, and max { a, b } represents taking the maximum of a and b.
And finally, sorting according to the gray coefficient GI of each pixel in the image I from small to large, wherein the pixel with the lowest value of the first h percent is taken as a gray pixel set P in the image Ih。
S12, using gray pixel set PhThe illumination adjustment amount E is calculated.
According toSet of gray pixels P in image IhR, G, B light adjustment quantity E of three color channels is calculatedR、EG、EBThe calculation formula is as follows:
s13, adjusting the low-illumination image I according to the illumination adjustment quantity E to obtain an image I after color correctionnew;
Illumination adjustment quantity E according to R, G, B three color channelsR、EG、EBFor three color channels I of the image IR、IG、IBThe adjustment is carried out, and the calculation formula is as follows:
obtaining a color corrected image Inew。
Step S2, image InewThe conversion from the RGB color space to the YCbCr color space results in a luminance component Y and color components Cb and Cr.
In a specific embodiment, the image InewFrom RGB color spaceThe color space is converted into YCbCr color space, and the calculation formula is as follows:
and step S3, performing iterative multi-scale guiding filtering on the brightness component Y to obtain an illumination component L.
The method specifically comprises the following steps:
s31, calculating an initial illumination component L(0)Average gradient Gr of(0)。
Let L(0)Y is the initial luminance component; for L(0)Calculating L for the pixel at the upper arbitrary position (i, j)(0)Average gradient Gr of(0)The calculation formula is as follows:
wherein M and N each represent L(0)Corresponding width and height, ΔxAnd deltayRepresenting first order gradient operators in the x-direction and the y-direction, respectively.
S32, calculating the illumination component L of the nth iteration step(n)。
First, an initial scale factor s is set(1)Smoothing factor epsilon(1)And an iteration termination threshold ζ.
According to the scale factor s of the nth step iteration(n)And a smoothing factor ε(n)Illumination component L for step n-1 iteration(n-1)Conducting guiding filtering to obtain the illumination component L of the nth step(n)N is more than 1, and the calculation formula is as follows:
in the formula, Li (n)Represents L(n)Pixel i, ω of (1)kRepresenting a square filter window centered on pixel k, with a window size of (2 s)(n)+1)×(2s(n)+1), weight matrix Wij(L(n-1)) The calculation formula of (a) is as follows:
in the formula, mukAnd σkAre each L(n-1)At window omegakMean and standard deviation of (1);
s33, calculating the illumination component L of the nth iteration step(n)Average gradient Gr of(n)The calculation formula is as follows:
s34, according to Gr(n)And Gr(n-1)Calculating the average gradient difference delta Gr, delta Gr ═ Gr(n)-Gr(n-1)|;
And S35, judging the termination iteration condition according to the average gradient difference delta Gr to obtain the illumination component L.
If delta Gr is not more than zeta, the iteration is terminated, and L is output(n)I.e. the calculated illumination component L ═ L(n);
Otherwise let n ← n +1, adjust s(n)=s(1)×2nAnd epsilon(n)=ε(1)×2nReturning to step S32.
In step S4, a reflection component R is calculated based on Retinex theory according to the illumination component L.
The method specifically comprises the following steps:
image InewAnd the illumination component L is transformed to the logarithmic domain:
im=log2(Inew/255),lm=log2(L/255);
calculating a reflection component R according to the following calculation formula:
R=2(im-lm)。
step S5, global contrast correction is carried out on the reflection component R to obtain the enhanced reflection component RenhI.e. enhancement of the luminance component Yenh。
The method specifically comprises the following steps:
and (3) carrying out normalization processing on the reflection component R: rN=(R-Rmin)/(Rmax-Rmin),
Wherein R isminAnd RmaxRespectively the minimum and maximum of the reflected component R.
To the normalized reflection component RNAnd carrying out global contrast correction, wherein the calculation formula is as follows:
wherein t is an adjustment parameter.
Enhanced reflection component RenhI.e. enhancement result Y of luminance component Yenh=Renh。
Step S6, enhancing result Y of the brightness component YenhAnd converting color components Cb and Cr into RGB color space to obtain low-illumination image enhancement result Ifinal。
In a specific embodiment, the enhancement result Y of the luminance component YenhAnd the conversion of the color components Cb and Cr from the YCbCr color space to the RGB color space, which is calculated as follows:
According to the low-illumination image enhancement method provided by the embodiment of the invention, correspondingly, the embodiment of the invention also provides a low-illumination image enhancement system, which comprises the following steps of;
a preprocessing module for performing color constancy preprocessing on the low-illumination image I to obtain an image I after color correctionnew;
A first conversion module, in particular for converting the image InewConverting the RGB color space into the YCbCr color space to obtain a brightness component Y and color components Cb and Cr;
the illumination component calculation module is specifically used for performing iterative multi-scale guided filtering on the brightness component Y to obtain an illumination component L;
the reflection component calculation module is specifically used for calculating and obtaining a reflection component R based on a Retinex theory according to the illumination component L;
a global correction module, specifically configured to perform global contrast correction on the reflection component R to obtain an enhanced reflection component RenhI.e. enhancement of the luminance component Yenh;
A second conversion module, in particular for converting the enhancement result Y of the luminance component YenhAnd converting color components Cb and Cr into RGB color space to obtain low-illumination image enhancement result Ifinal。
In a further embodiment, the preprocessing module is specifically configured to:
obtaining a set of gray pixels P in a low-light image Ih(ii) a Using a set of gray pixels PhCalculating an illumination adjustment quantity E; adjusting the low-illumination image I according to the illumination adjustment quantity E to obtain the image with corrected colorImage Inew。
Wherein a set of gray pixels P in the acquisition of the low-illumination image IhThe process specifically comprises the following steps:
first, a local neighborhood ω centered on an arbitrary pixel p in an image I is calculatedpCorresponding to R, G, B three color channels I respectivelyR、IG、IBGradient deviation Delta I ofR(p)、△IG(p)、△IB(p) and the mean gradient deviation Δ μ (p) and mean gray value I μ (p), which are calculated as follows:
next, calculating a gray scale coefficient gi (p) of an arbitrary pixel p on the image I, wherein the calculation formula is as follows:
wherein epsilon is a decimal number greater than 0, and max { a, b } represents taking the maximum of a and b.
Finally, sorting according to the gray coefficient GI of each pixel in the image I from small to large, wherein the value of the top h% of the lowest valueThe pixels being a set P of gray pixels in the image Ih。
Wherein a set of gray pixels P is utilizedhIn the process of calculating the illumination adjustment amount E, the method specifically includes:
from a set of gray pixels P in the image IhR, G, B light adjustment quantity E of three color channels is calculatedR、EG、EBThe calculation formula is as follows:
wherein, adjust low light level image I according to illumination adjustment quantity E, specifically include:
illumination adjustment quantity E according to R, G, B three color channelsR、EG、EBFor three color channels I of the image IR、IG、IBThe adjustment is carried out, and the calculation formula is as follows:
obtaining a color corrected image Inew。
In a further embodiment, the first conversion module is specifically configured to:
image InewConverting from RGB color space to YCbCr color space to obtain brightness component Y and color components Cb and Cr, and calculating formula as follows:
in a further embodiment, the illumination component calculation module is specifically configured to:
calculating an initial illumination component L(0)Average gradient Gr of(0)(ii) a Calculating illumination component L of nth step iteration(n)(ii) a Calculating illumination component L of nth step iteration(n)Average gradient Gr of(n)(ii) a According to Gr(n)And Gr(n-1)Calculating an average gradient difference delta Gr; and judging the termination iteration condition according to the average gradient difference delta Gr to obtain an illumination component L.
Wherein the initial illumination intensity component L is calculated(0)Average gradient Gr of(0)The process specifically comprises the following steps:
let L(0)Y is the initial luminance component; for L(0)Calculating L for the pixel at the upper arbitrary position (i, j)(0)Average gradient Gr of(0)The calculation formula is as follows:
wherein M and N each represent L(0)Corresponding width and height, ΔxAnd deltayRepresenting first order gradient operators in the x-direction and the y-direction, respectively.
Wherein the illumination component L of the iteration of the nth step is calculated(n)The process specifically comprises the following steps:
first, an initial scale factor s is set(1)Smoothing factor epsilon(1)And an iteration termination threshold ζ.
According to the scale factor s of the nth step iteration(n)And a smoothing factor ε(n)Illumination component L for step n-1 iteration(n-1)Conducting guiding filtering to obtain the illumination component L of the nth step(n)N is more than 1, and the calculation formula is as follows:
in the formula, Li (n)Represents L(n)Pixel i, ω of (1)kRepresenting a square filter window centered on pixel k, with a window size of (2 s)(n)+1)×(2s(n)+1), weight matrix Wij(L(n-1)) The calculation formula of (a) is as follows:
in the formula, mukAnd σkAre each L(n-1)At window omegakMean and standard deviation of (1);
wherein the illumination component L of the iteration of the nth step is calculated(n)Average gradient Gr of(n)In the process, the calculation formula is as follows:
wherein, according to Gr(n)And Gr(n-1)Calculating the average gradient difference delta Gr, delta Gr ═ Gr(n)-Gr(n-1)|;
Wherein, the termination iteration condition is judged according to the average gradient difference Δ Gr to obtain the illumination component L, and the specific judgment process is as follows:
if delta Gr is not more than zeta, the iteration is terminated, and L is output(n)I.e. the illumination component L obtained in time is L(n);
Otherwise let n ← n +1, adjust s(n)=s(1)×2nAnd epsilon(n)=ε(1)×2nReturning to the above calculation of the illumination component L of the nth iteration step(n)The process of (1).
In a further embodiment, the reflection component calculation module is specifically configured to:
image InewAnd the illumination component L is transformed to the logarithmic domain: im is log2(Inew/255), lm=log2(L/255);
Calculating a reflection component R according to the following calculation formula: r is 2(im-lm)。
In a further embodiment, the global correction module is specifically configured to:
and (3) carrying out normalization processing on the reflection component R: rN=(R-Rmin)/(Rmax-Rmin),
Wherein R isminAnd RmaxRespectively the minimum and maximum of the reflected component R.
To the normalized reflection component RNAnd carrying out global contrast correction, wherein the calculation formula is as follows:
wherein t is an adjustment parameter.
Enhanced reflection component RenhI.e. enhancement result Y of luminance component Yenh=Renh。
In a further embodiment, the second conversion module is specifically configured to convert the enhancement result Y of the luminance component Y into an enhancement result YenhAnd converting the color components Cb and Cr into RGB color space, wherein the calculation formula is as follows:
in the invention, firstly, color constancy preprocessing is carried out on the acquired low-illumination image so as to eliminate the color cast phenomenon caused by an artificial light source, then the low-illumination image is converted into a YCbCr color space from an RGB color space, illumination component estimation is carried out by adopting iterative multi-scale guide filtering, then a reflection component is obtained by calculation according to a Retinex theory and global contrast correction is carried out, and finally, the processing result is converted into the RGB color space so as to obtain a final low-illumination image enhancement result. The low-illumination image enhancement method can effectively enhance the detail information of the image, improve the definition of the image and improve the visual effect of the image, and has the characteristics of rapidness, accuracy and automatic processing.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and equivalent substitutions or changes according to the technical solution and the inventive concept of the present invention should be covered by the scope of the present invention.
Claims (8)
1. A low-illumination image enhancement method is characterized by comprising the following steps of;
s1, carrying out color constancy preprocessing on the low-illumination image I to obtain an image I after color correctionnew;
S2, image InewConverting the RGB color space into the YCbCr color space to obtain a brightness component Y and color components Cb and Cr;
s3, conducting iterative multi-scale guiding filtering on the brightness component Y to obtain an illumination component L;
s4, calculating a reflection component R based on Retinex theory according to the illumination component L;
s5, carrying out global contrast correction on the reflection component R to obtain an enhanced reflection component RenhI.e. enhancement of the luminance component Yenh;
S6, enhancing result Y of brightness component YenhAnd converting color components Cb and Cr into RGB color space to obtain low-illumination image enhancement result Ifinal。
2. The low-illuminance image enhancement method according to claim 1, wherein step S1 specifically includes;
s11, acquiring a gray pixel set P in the low-illumination image Ih;
S12, using gray pixel set PhCalculating an illumination adjustment quantity E;
s13, adjusting the low-illumination image I according to the illumination adjustment quantity E to obtain an image I after color correctionnew。
3. The low-illuminance image enhancement method according to claim 2, wherein the step S11 specifically includes: calculating a local neighborhood omega centered on an arbitrary pixel p in the image IpCorresponding to R, G, B three color channels I respectivelyR、IG、IBGradient deviation Delta I ofR(p)、△IG(p)、△IB(p), and the mean gradient deviation Δ μ (p) and the mean gray value I μ (p); calculating the gray coefficient GI (p) of any pixel p on the image I; sorting according to the gray coefficient GI of each pixel in the image I from small to large, and taking the pixel with the lowest value of the first h% as a gray pixel set P in the image Ih;
Preferably, step S12 specifically includes: from a set of gray pixels P in the image IhR, G, B light adjustment quantity E of three color channels is calculatedR、EG、EB;
Preferably, step S13 specifically includes: illumination adjustment quantity E according to R, G, B three color channelsR、EG、EBFor three of image IColor channel IR、IG、IBAdjusting to obtain color corrected image Inew。
4. The low-illuminance image enhancement method according to claim 1, wherein step S3 specifically includes;
s31, calculating an initial illumination component L(0)Average gradient Gr of(0);
S32, calculating the illumination component L of the nth iteration step(n);
S33, calculating the illumination component L of the nth iteration step(n)Average gradient Gr of(n);
S34, according to Gr(n)And Gr(n-1)Calculating the average gradient difference delta Gr;
and S35, judging the termination iteration condition according to the average gradient difference delta Gr to obtain the illumination component L.
5. A low-illumination image enhancement system, comprising;
a preprocessing module for performing color constancy preprocessing on the low-illumination image I to obtain an image I after color correctionnew;
A first conversion module, in particular for converting the image InewConverting the RGB color space into the YCbCr color space to obtain a brightness component Y and color components Cb and Cr;
the illumination component calculation module is specifically used for performing iterative multi-scale guided filtering on the brightness component Y to obtain an illumination component L;
the reflection component calculation module is specifically used for calculating and obtaining a reflection component R based on Retinex theory according to the illumination component L;
a global correction module, specifically configured to perform global contrast correction on the reflection component R to obtain an enhanced reflection component RenhI.e. enhancement of the luminance component Yenh;
A second conversion module, in particular for converting the enhancement result Y of the luminance component YenhAnd converting color components Cb and Cr to RGB color space to obtain low-illumination image enhancementStrong results Ifinal。
6. The low-illuminance image enhancement system according to claim 5, wherein the pre-processing module is specifically configured to: obtaining a set of gray pixels P in a low-light image Ih(ii) a Using a set of gray pixels PhCalculating an illumination adjustment quantity E; adjusting the low-illumination image I according to the illumination adjustment quantity E to obtain an image I after color correctionnew。
7. The low-illuminance image enhancement system according to claim 6, wherein the pre-processing module is specifically configured to: calculating a local neighborhood omega centered on an arbitrary pixel p in the image IpCorresponding to R, G, B three color channels I respectivelyR、IG、IBGradient deviation Delta I ofR(p)、△IG(p)、△IB(p), and the mean gradient deviation Δ μ (p) and the mean gray value I μ (p); calculating the gray coefficient GI (p) of any pixel p on the image I; sorting according to the gray coefficient GI of each pixel in the image I from small to large, and taking the pixel with the lowest value of the first h% as a gray pixel set P in the image Ih;
Preferably, the preprocessing module is specifically configured to: from a set of gray pixels P in the image IhR, G, B light adjustment quantity E of three color channels is calculatedR、EG、EB;
Preferably, the preprocessing module is specifically configured to: illumination adjustment quantity E according to R, G, B three color channelsR、EG、EBFor three color channels I of the image IR、IG、IBAdjusting to obtain color corrected image Inew。
8. The low-illuminance image enhancement system according to claim 5, wherein the illuminance component calculation module is specifically configured to:
calculating an initial illumination component L(0)Average gradient Gr of(0)(ii) a Calculating illumination component L of nth step iteration(n)(ii) a Computing the nth iterationIllumination component L of(n)Average gradient Gr of(n)(ii) a According to Gr(n)And Gr(n-1)Calculating the average gradient difference delta Gr; and judging the termination iteration condition according to the average gradient difference delta Gr to obtain an illumination component L.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011327998.8A CN112365425A (en) | 2020-11-24 | 2020-11-24 | Low-illumination image enhancement method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011327998.8A CN112365425A (en) | 2020-11-24 | 2020-11-24 | Low-illumination image enhancement method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112365425A true CN112365425A (en) | 2021-02-12 |
Family
ID=74533211
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011327998.8A Pending CN112365425A (en) | 2020-11-24 | 2020-11-24 | Low-illumination image enhancement method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112365425A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113112418A (en) * | 2021-03-26 | 2021-07-13 | 浙江理工大学 | Low-illumination image iterative enhancement method |
CN113160096A (en) * | 2021-05-27 | 2021-07-23 | 山东中医药大学 | Low-light image enhancement method based on retina model |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104504722A (en) * | 2015-01-09 | 2015-04-08 | 电子科技大学 | Method for correcting image colors through gray points |
-
2020
- 2020-11-24 CN CN202011327998.8A patent/CN112365425A/en active Pending
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104504722A (en) * | 2015-01-09 | 2015-04-08 | 电子科技大学 | Method for correcting image colors through gray points |
Non-Patent Citations (2)
Title |
---|
张杰: "基于迭代多尺度引导滤波 Retinex 的低照度图像增强", 《图学学报》 * |
张杰;周浦城;薛模根;: "基于方向性全变分Retinex的低照度图像增强", 计算机辅助设计与图形学学报 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113112418A (en) * | 2021-03-26 | 2021-07-13 | 浙江理工大学 | Low-illumination image iterative enhancement method |
CN113112418B (en) * | 2021-03-26 | 2023-10-10 | 浙江理工大学 | Low-illumination image iteration enhancement method |
CN113160096A (en) * | 2021-05-27 | 2021-07-23 | 山东中医药大学 | Low-light image enhancement method based on retina model |
CN113160096B (en) * | 2021-05-27 | 2023-12-08 | 山东中医药大学 | Low-light image enhancement method based on retina model |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN1203677C (en) | Vision attention mode | |
CN104156921B (en) | Self-adaptive low-illuminance or non-uniform-brightness image enhancement method | |
CN107909562B (en) | Fast image fusion algorithm based on pixel level | |
CN107958465A (en) | A kind of single image to the fog method based on depth convolutional neural networks | |
CN110163807B (en) | Low-illumination image enhancement method based on expected bright channel | |
CN111968065B (en) | Self-adaptive enhancement method for image with uneven brightness | |
CN110944176B (en) | Image frame noise reduction method and computer storage medium | |
Kwok et al. | Gray world based color correction and intensity preservation for image enhancement | |
CN110110131B (en) | Airplane cable support identification and parameter acquisition method based on deep learning and binocular stereo vision | |
CN113793275A (en) | Swin Unet low-illumination image enhancement method | |
CN112365425A (en) | Low-illumination image enhancement method and system | |
CN111861896A (en) | UUV-oriented underwater image color compensation and recovery method | |
CN105809643A (en) | Image enhancing method based on self-adaptive block channel stretching | |
CN110852956A (en) | Method for enhancing high dynamic range image | |
CN108648160B (en) | Underwater sea cucumber image defogging enhancement method and system | |
CN113313702A (en) | Aerial image defogging method based on boundary constraint and color correction | |
CN115578297A (en) | Generalized attenuation image enhancement method for self-adaptive color compensation and detail optimization | |
Yu et al. | Image and video dehazing using view-based cluster segmentation | |
CN116188339A (en) | Retinex and image fusion-based scotopic vision image enhancement method | |
CN114037622A (en) | Underwater image enhancement method based on imaging model and reinforcement learning | |
CN115731166A (en) | High-voltage cable connector polishing defect detection method based on deep learning | |
Wei et al. | An image fusion dehazing algorithm based on dark channel prior and retinex | |
CN109903275B (en) | Fermented grain mildewing area detection method based on self-adaptive multi-scale filtering and histogram comparison | |
CN110298796A (en) | Based on the enhancement method of low-illumination image for improving Retinex and Logarithmic image processing | |
KR102277005B1 (en) | Low-Light Image Processing Method and Device Using Unsupervised Learning |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210212 |
|
RJ01 | Rejection of invention patent application after publication |