CN111861896A - UUV-oriented underwater image color compensation and recovery method - Google Patents
UUV-oriented underwater image color compensation and recovery method Download PDFInfo
- Publication number
- CN111861896A CN111861896A CN201910358840.8A CN201910358840A CN111861896A CN 111861896 A CN111861896 A CN 111861896A CN 201910358840 A CN201910358840 A CN 201910358840A CN 111861896 A CN111861896 A CN 111861896A
- Authority
- CN
- China
- Prior art keywords
- image
- underwater
- color
- pixel
- channel
- 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
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000011084 recovery Methods 0.000 title claims abstract description 29
- 230000000694 effects Effects 0.000 claims abstract description 28
- 238000012545 processing Methods 0.000 claims abstract description 26
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 24
- 238000012634 optical imaging Methods 0.000 claims abstract description 13
- 238000002834 transmittance Methods 0.000 claims abstract description 10
- 230000004927 fusion Effects 0.000 claims abstract description 9
- 238000004422 calculation algorithm Methods 0.000 claims description 22
- 230000002238 attenuated effect Effects 0.000 claims description 11
- 238000003708 edge detection Methods 0.000 claims description 6
- 230000005540 biological transmission Effects 0.000 claims description 5
- 238000001914 filtration Methods 0.000 claims description 5
- 230000001965 increasing effect Effects 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 claims description 4
- 230000015572 biosynthetic process Effects 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 4
- 239000003086 colorant Substances 0.000 claims description 4
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 claims description 3
- 238000010521 absorption reaction Methods 0.000 claims description 3
- 150000001875 compounds Chemical class 0.000 claims description 3
- 238000002474 experimental method Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 230000008033 biological extinction Effects 0.000 claims description 2
- 230000001186 cumulative effect Effects 0.000 claims description 2
- 230000001419 dependent effect Effects 0.000 claims description 2
- 238000005315 distribution function Methods 0.000 claims description 2
- 238000005562 fading Methods 0.000 claims description 2
- 238000007499 fusion processing Methods 0.000 claims description 2
- 230000008569 process Effects 0.000 claims description 2
- 239000000049 pigment Substances 0.000 claims 1
- 238000012958 reprocessing Methods 0.000 claims 1
- 229920006395 saturated elastomer Polymers 0.000 claims 1
- 230000000007 visual effect Effects 0.000 abstract description 3
- 238000000889 atomisation Methods 0.000 abstract description 2
- 238000005516 engineering process Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 8
- 238000012360 testing method Methods 0.000 description 6
- 238000012937 correction Methods 0.000 description 4
- 230000002146 bilateral effect Effects 0.000 description 2
- 125000001475 halogen functional group Chemical group 0.000 description 2
- 230000035772 mutation Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000007500 overflow downdraw method Methods 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 238000011158 quantitative evaluation Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/90—Dynamic range modification of images or parts thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration using histogram techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- 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
- G06T2207/10024—Color 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
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
- Facsimile Image Signal Circuits (AREA)
Abstract
The invention belongs to the field of underwater optical imaging and image enhancement, and particularly relates to a UUV-oriented underwater image color compensation and recovery method. Firstly, acquiring a bright channel underwater image through an underwater optical imaging model, carrying out color compensation on the bright channel underwater image, and further carrying out image visual effect enhancement by utilizing histogram equalization; secondly, on the basis of an underwater optical imaging model, considering various water area types, analyzing the red, green and blue channel attenuation ratio, estimating background light and underwater transmittance, and selecting an optimal underwater recovery image by utilizing a white balance hypothesis; and finally, performing weighted fusion on the color compensation image and the color recovery image of the original image based on pixel level fusion and weighted processing, solving the problems of serious image atomization, low contrast and the like, and realizing the effect of improving the image quality.
Description
Technical Field
The invention belongs to the field of underwater optical imaging and image enhancement, and particularly relates to a UUV-oriented underwater image color compensation and recovery method.
Background
An Underwater Unmanned Vehicle (UUV) is a novel Underwater weapon following a torpedo, a torpedo and the like, is an Underwater Unmanned system integrating functions of environment sensing, dynamic planning, behavior control, energy adaptation and the like, and has important significance in the aspects of military surveying, Underwater reconnaissance, attack defense and the like. At present, the underwater vision technology applied to the UUV can be mainly divided into acoustic vision and optical vision, wherein the optical vision image enhancement technology becomes a research hotspot for enhancing the underwater image of the UUV.
Because light is attenuated and scattered in water, the UUV light vision system obtains uneven illumination and obtains an underwater image with the problems of blurring, low contrast, color distortion and noise. At present, methods for researching underwater image enhancement at home and abroad are mainly divided into three methods, namely a spatial domain method, a frequency domain method and a color correction method. The spatial domain method is basically based on gray mapping transformation, and directly processes pixels of an image, the underwater image enhancement technology is mature, and has the advantages of simple algorithm realization, obvious contrast enhancement of the processed underwater image and the like, but the color cast condition is not considered, and the problems of introduction of artifacts, color distortion, amplified noise and the like are not well solved; the transform domain method is to use transform technology and adjust the definition of an image in a digital filtering mode, and the method has a good effect on removing underwater image noise, but does not have a perfect effect on the comprehensive problems of low contrast, color deviation, blur and the like of the underwater image, and compared with the airspace enhancement technology, the method has less research work and slow progress on the enhancement of the underwater image based on a frequency domain; the color correction method extracts red, green and blue channels based on physical hypothesis, and adopts color compensation according to wavelength propagation characteristics and absorption characteristics to achieve the aim of recovering underwater images.
The method mainly aims at various problems of underwater image blurring, low contrast, color distortion and the like caused by underwater attenuation of light, utilizes an underwater optical imaging model Jaffe-McGlamry to separate red, green and blue channels to respectively compensate and restore colors, and utilizes a pixel-level weighting fusion method to fuse a color compensation image and a color restoration image of an original image, so that various problems of underwater image blurring, low contrast, color distortion and the like are solved.
Disclosure of Invention
In order to overcome the problems of fuzzy underwater images, low contrast and color distortion pointed out in the background art, the invention provides a UUV-oriented underwater image color compensation and recovery method, and aims to achieve the aims of improving the quality of degraded images and highlighting detailed information in the images.
A UUV-oriented underwater image color compensation and recovery method comprises the following steps:
the method comprises the following steps: extraction of bright channel for image shot by UUV
The expression of the total strong light can be obtained according to the Jaffe-McGlamry underwater optical imaging model, namely
In the formula (1), (x, y) represents the specific position of a pixel in an image, C epsilon { R, G, B } represents three channels of red, green and blue, Jc(x, y) is the original image, I c(x, y) is the attenuated image, A∞Is background light; ed(x, y) is the direct light intensity received by the camera; c (lambda) is the total attenuation coefficient, including the attenuation caused by absorption and scattering, lambda is one of the three color channels of image R, G, B; d (x) is the distance between the object and the UUV; g (x, y) is a point spread function; ef(x, y) is the forward scattered intensity; eb(x, y) represents the intensity of the backscattered light;
the light received by the UUV consists of a direct component, a forward scattering component and a backward scattering component, and the transmissivity S is setc(x) Comprises the following steps:
Sc(x)=exp[-c(λ)d(x)](2)
when the object is very small from the camera, the forward scattering light causes negligible blur, so equation (1)
Can be simplified as follows:
Ic(x,y)=Jc(x,y)Sc(x)+A∞[1-Sc(x)](3)
taking the maximum value from two sides of equation (3) and three channel neighborhood of red, green and blue to extract the bright channel, i.e.
based on the above analysis, it is proposed to compensate the color of each channel with reference to the bright channel. The bright channel acquisition method comprises the following steps:
in the formula (I), the compound is shown in the specification,for obtaining an image IcThe maximum value of the red, green and blue channels at each pixel point in (x, y),for maximum filtering, the filter template Ω (x, y) size is taken to be 15.
Step two: carrying out bright channel color compensation on the image obtained by extracting the bright channel in the step one
Gray values with the same mean value of the red channel, the green channel and the blue channel in a noise-free image can be obtained by the gray world white balance hypothesis, and the difference value of the mean value of the compensated image gray value and the mean value of the attenuated image gray value is in direct proportion to the mean value of the original image gray value through a large number of experiments and theoretical derivation. Therefore, the attenuation image needs to be multiplied by the difference value between the average value of the gray value of the compensated image and the average value of the gray value of the attenuated image, and meanwhile, the attenuation of the channels of the three colors is different underwater.
Based on the above analysis, a color recovery algorithm based on a bright channel is proposed, namely:
wherein C is in the form of { R, G, B } channel of red, green and blue, Ic(x, y) is the attenuated image, IBC(x, y) is the image after color compensation,the mean value of the gray values of the three color channels after compensation for the bright channel,the average of the grey values of the three channels of the image is attenuated. All image pixel values involved in the calculation in equation (6) are reduced to [0,1 ] by dividing by 255]In the interval, the average brightness value of the naturally exposed image is about 125 statistically, η can adjust the brightness of the compensated image, η is 3.5, so that the average brightness value of the compensated image satisfies the above rule, η is 3.5,for rounding-down operations, N<>Indicating that the pixel value is linearly stretched to 0,255]An interval.
Step three: performing histogram equalization processing on the image subjected to the two-color compensation in the step
Under the complex underwater environment, the scattering effects caused by different degrees of turbidity of water are different, so that the pictures show different degrees of blur, and further processing is required to be carried out by using a histogram equalization algorithm.
In discrete form, the probability can be replaced by a frequency, where r is usedkRepresents IBCDiscrete gray levels of (x, y), and k is the number of gray levels. And the following holds:
Wherein r is more than or equal to 0k≤1,k=0,1,2,3,…,n-1,nkFor the occurrence of r in the imagekThe number of pixels of such a gray scale, n being the total number of pixels in the image, andis the frequency of the probabilities.
The function expression of histogram equalization of the image is as follows:
the functional expression for histogram equalization needs to satisfy two conditions: (a) t is a single value monotone increasing function, (b) r is more than or equal to 0 and less than or equal to L-1, and T is more than or equal to 0 and less than or equal to (r) and less than or equal to L-1. When the two conditions are satisfied, the cumulative distribution function can convert the original histogram with uneven gray level distribution into the histogram with even gray level distribution, and the corresponding inverse transformation is as follows:
ri=T-1(Si) (9)
using transformed discrete grey levels riRecovering the image Iz=F(ri) Wherein; f is a mapping function; step four: improving an underwater image forming model, and estimating background light and underwater transmissivity in the improved model to further estimate scene recovery Jc
Step 4.1, obtaining a simplified underwater optical imaging model according to a Jaffe-McGlamry underwater optical imaging model formula (1):
Ic(x)=tc(x)Jc(x)+(1-tc(x))·Ac(10)
wherein x represents a pixel point in the image, IcRepresenting images taken by UUV, tcRepresenting the underwater transmission of the color channel, JcRepresenting a clear original image, AcIs the background light, c ∈ { R, G, B }.
Transmission t under waterc(x) Dependent on the distance z (x) of the object from the UUV and the water attenuation coefficient beta of each channel cI.e. by
tc(x)=exp(-βcz(x)) (11)
Step 4.2: estimation of background light in the underwater image formation model improved in step 4.1
Estimation of background light A using linear contrast stretching and edge detection modelscThe specific principle is as follows: first, I is obtained by linear contrast stretching and structured edge detectioncEdge detection image IeAnd to IeThe pixels of (a) are arranged from large to small; secondly, cutting the pixel according to a pixel point judgment threshold Th, if the pixel x is larger than the threshold, the pixel belongs to the maximum connected component Lc, otherwise, the pixel does not belong to Lc, and the method specifically comprises the following steps:
further, an average Av of the pixels in Lc and Sum is calculated as the estimated background light.
Step 4.3: underwater Transmission estimation in the Underwater image formation model improved for step 4.1
Combining equations (10) and (11), the increment of the blue channel can be calculated:
the increment of the red channel is calculated:
The ratio between the attenuation coefficients is defined as:
further, according to the compensation space formula (16), a structure similar to the formula (10) is formed by the underwater transmittance:
due to I in the formula (10)c-AcPossibly negative, and therefore increasing β by increasing the absolute value and preserving the signBcWhile avoiding the occurrence of I c-AcNegative value problem.
The boundary J is replaced in equation (10) to account for the different attenuation coefficients of the different color channelscNot less than 0 and obtaining lower limit t on the underwater transmissivity of the blue channelLB:
Setting pixel x of the underwater transmittance to t calculated in equation (18)LB. Since binary classification of pixels x to Lc often results in abrupt discontinuities, the underwater transmittance is estimated using the algorithm soft-matching:
in the formula (I), the compound is shown in the specification,∑x∈LcDM(I(x) Represents the average of the mahalanobis distance for the background light pixels,representing the maximum mahalanobis distance, σMRepresents the standard deviation, alpha (x) represents the extinction coefficient of a pixel belonging to an object or water area,
step 4.4: estimation scene recovery Jc
Color fading compensation is performed and a clear original image is estimated according to equations (12), (18), and (19):
step five: for J recovered in step fourcSelecting the optimal underwater recovery image based on the white balance hypothesis
If the wrong attenuation coefficient beta is usedcCan lead to color deviation and repair of underwater transmittance image errors, for which the best results are automatically selected by grey world assumption according to different water area types of JerlovAssuming that the water area type is N, i is 1,2, … N, piIndicates the ith water area type, J cAnd if three color channels of R, G and B exist, the specific gray world algorithm processing steps are as follows:
(a) calculating water area type image JcThree-channel mean value Gray of R, G, Bi:
(b) Calculation of JcGain coefficients of three channels of images R, G, B:
(c) according to the Von Kries diagonal model, for each pixel x in the image, the R, G and B components are adjusted to be Cr、CgAnd Cb:
(d) According to the formula (23), Cr、CgAnd CbThe number of pixels is M, for CrPixel C included inrj、CgPixel C included ingjAnd CbPixel C included inbjSumming and taking the average value avr:
(e) differencing each channel avr and taking the absolute value pi:
pi=|avrr-avrg-avrb| (25)
Comparing the calculated difference p of different water area typesiSelecting the image corresponding to the water area type with the minimum difference value as
Step six: pixel-based image fusion
IzFor bright channel color compensated and histogram processed images,and (3) reconstructing and recovering the image of the underwater 3D scene, wherein J is the image after fusion processing:
in the formula, α and β are respective weights and satisfy α + β ═ 1.
The method mainly aims at various problems of fuzzy underwater images, low contrast, color distortion and the like caused by the underwater attenuation of light, obtains a bright channel image by using red, green and blue channels, performs color compensation on each channel by taking the bright channel image as a reference, and obtains a relatively ideal underwater enhanced image by multi-scale fusion of images by adopting histogram enhancement and bilateral filtering denoising technologies.
Compared with a single underwater image enhancement algorithm, the method has the following beneficial effects:
1. the color information is complete, the compensation factor is increased along with the larger the depth of field of the shot by the bright channel color compensation method and is not a negative value, so that the color information cannot be lost;
2. the contrast is high, and the problems of darker images and low contrast are solved by carrying out equalization processing on the underwater images subjected to histogram color compensation;
3. the method has strong comprehensiveness, adopts multi-aspect technical fusion, and efficiently solves various problems of underwater image blur, low contrast, color distortion and the like.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a comparison of color restoration effect of color chart, original image and four methods; wherein (a) is a standard color card; (b) is an original picture; (c) is the bright channel processing result; (d) is the histogram processing result; (e) is the 3D scene restoration processing result; (f) is the algorithm processing result of the invention;
FIG. 3 is a comparison of the restoration effect of various underwater images according to the present invention; wherein, the group a of pictures are original pictures; b group of diagrams are processing effect diagrams of bright channel color compensation technology; c group of diagrams are processing effect diagrams of the histogram stretching technology; the D group of diagrams are the processing effect diagrams of the 3D scene color recovery technology, and the e group of diagrams are the effect diagrams of the technical processing proposed by the invention;
FIG. 4 is a graph of application test effects; wherein, the group (a) is SURF feature matching of the original image; (b) the set is the SURF feature matching of the present invention.
Detailed Description
The present invention will be described in further detail with reference to specific examples, but the embodiments of the present invention are not limited thereto.
The invention provides a UUV-oriented underwater image color compensation and recovery method, which comprises the steps that firstly, an underwater optical imaging model obtains a bright channel underwater image and performs color compensation on the bright channel underwater image, and further, histogram equalization is utilized to enhance the visual effect of the image; secondly, on the basis of an underwater optical imaging model, considering various water area types, analyzing the red, green and blue channel attenuation ratio, estimating background light and underwater transmittance, and selecting an optimal underwater recovery image by utilizing a white balance hypothesis; and finally, performing weighted fusion on the color compensation image and the color recovery image of the original image based on pixel level fusion and weighted processing, solving the problems of serious image atomization, low contrast and the like, and realizing the effect of improving the image quality.
The color correction is realized by utilizing the color compensation of the bright channel, the color information is not easy to lose, and on the basis, the image standard deviation mean value is kept at about 0.5 by adopting a self-adaptive contrast stretching and bilateral filtering method; better underwater image edge details can be kept while noise is reduced. Reconstructing an underwater 3D scene by using a method for estimating unknown parameters of an image, recovering the color of a distant scene, and reducing edge mutation halo; by weighting and re-fusing the two images, the contrast of the underwater image is greatly improved, the color recovery is obvious, the detail processing is more exquisite, and the quality of the underwater image is enhanced.
Example 1:
1. color recovery test
The color recovery result is shown in FIG. 2. As can be seen from FIG. 2, the contrast of the image processed by the histogram algorithm is enhanced, but the color of the image processed by the histogram algorithm has distortion of different degrees, and the image processed by the bright channel algorithm has white light; the color of the image processed by the underwater 3D scene recovery algorithm is reddish, and the color recovery effect is poor; comparing the color test card, the image color processed by the method of the invention is recovered to be correct.
2. Visual effect contrast
In order to detect the effect of the invention on processing the problems of underwater image fogging, blurring and the like, 6 images 1-6 are selected, as shown in fig. 3, wherein a group of images are original images, b group of images are bright channel color compensation technical processing effect images, c group of images are histogram stretching technical processing effect images, D group of images are 3D scene color recovery technical processing effect images, and e group of images are effect images of the technical processing proposed by the invention.
It is not easy to see that bright channel color compensation improves the brightness of the picture to a certain extent, but has poor defogging effect on the picture and cannot well restore the real color of the image; the defogging effect is obvious by adopting a histogram algorithm, so that the contrast of the picture is enhanced, but a large amount of white light is generated at the edge of the picture, and the color distortion of the whole picture is generated to different degrees; the 3D underwater scene reconstruction algorithm aims to solve the problems of image far color distortion, halo appearing in edge mutation and the like by reconstructing an underwater 3D scene to obtain two global attenuation ratio parameters to obtain a far color calibration image, the processed image color correction effect is better, but the processed image has low contrast ratio and is not obvious enough compared with a histogram in the image defogging effect, and the color of the processed image is slightly reddish; the method has a good processing effect, solves the problems of excessive white light and fuzzy edges of the pictures processed by the histogram, solves the problems of dim color tone and weak image contrast of a 3D underwater scene reconstruction algorithm, and better restores the underwater real images.
3. Image quality evaluation
Quantitative evaluation is carried out on underwater images 1-image6 in fig. 2 by peak signal to noise ratio (PSNR), and a higher PSNR represents that the algorithm introduces a small amount of noise and retains more valuable image information.
The evaluation results are shown in table 1, and data show that the image processed by the algorithm of the invention has low introduced noise, retains more valuable information, and the algorithm processing effect is superior to histogram equalization and 3D scene recovery algorithm; less information is retained compared to the bright channel compensation method.
TABLE 1 Peak SNR PSNR contrast analysis
4. Application testing
In order to verify the application effect of the image processed by the method, application contrast experiments before and after image processing are set. The method adopts an accelerated robust feature (SURF) feature matching test, compares the matching number of feature points before and after image processing under the condition of the same feature similarity threshold, applies the test effect as shown in FIG. 4, performs SURF feature matching on the original image after rotating the original image by right angle degrees and only can correctly match a small number of feature points under the same threshold condition, and increases the number of correctly matched feature points of the image processed by the method. Referring to fig. 4, the result shows that the image processed by the method of the present invention has a good application effect in the subsequent feature extraction process.
The foregoing is a more detailed description of the invention in connection with specific preferred embodiments and it is not intended that the invention be limited to these specific details. For those skilled in the art to which the invention pertains, several simple deductions or substitutions can be made without departing from the spirit of the invention, and all shall be considered as belonging to the protection scope of the invention.
Claims (2)
1. A UUV-oriented underwater image color compensation and recovery method is characterized by comprising the following steps:
the method comprises the following steps: extraction of bright channel for image shot by UUV
Obtaining an expression of total strong light according to the Jaffe-McGlamry underwater optical imaging model, namely
In the formula (1), (x, y) represents the specific position of a pixel in an image, C epsilon { R, G, B } represents three channels of red, green and blue, Jc(x, y) is the original image, Ic(x, y) is the attenuated image, A∞Is background light; ed(x, y) is the direct light intensity received by the camera; c (λ) is the overall attenuation coefficient, including the attenuation caused by absorption and scattering; lambda is one of three color channels of the images R, G and B; d (x) is the distance between the object and the UUV; g (x, y) is a point spread function; ef(x, y) is the forward scattered intensity; e b(x, y) represents the intensity of the backscattered light;
the light received by the UUV consists of a direct component, a forward scattering component and a backward scattering component, and the transmissivity S is setc(x) Comprises the following steps:
Sc(x)=exp[-c(λ)d(x)](2)
when the object is a little away from the camera, the forward scattering light causes blurring to be ignored, so equation (1) is simplified as:
Ic(x,y)=Jc(x,y)Sc(x)+A∞[1-Sc(x)](3)
taking the maximum value from two sides of equation (3) and three channel neighborhood of red, green and blue to extract the bright channel, i.e.
the bright channel acquisition method is expressed as equation (5):
for obtaining an image IcThe maximum value of the red, green and blue channels at each pixel point in (x, y),for maximum filtering, the filter template size is Ω (x, y);
step two: carrying out bright channel color compensation on the image obtained by extracting the bright channel in the step one
Obtaining gray values with the same mean value of red, green and blue channels in a noise-free image by assuming gray world white balance, and obtaining that the difference value of the mean value of the compensated image gray value and the mean value of the attenuated image gray value is in direct proportion to the mean value of the original image gray value through a large number of experiments and theories, so that the attenuated image is multiplied by the difference value of the mean value of the compensated image gray value and the mean value of the attenuated image gray value; meanwhile, under water, the attenuation of the channels of the three colors is different; if the images are processed uniformly, the values of some channels are saturated to generate a large amount of white light, so that pigments with lower gray values in the images need to be compensated;
Based on the above analysis, a color recovery algorithm based on a bright channel is proposed, namely:
in the formula (6), C is in the { R, G, B } channel of three colors of red, green and blue, Ic(x, y) is the attenuated image, IBC(x, y) is the image after color compensation,the mean value of the gray values of the three color channels after compensation for the bright channel,attenuating the average value of the gray values of three channels of the image;
all image pixel values involved in the calculation in equation (6) are reduced to [0,1 ] by dividing by 255]The interval, η, is 3.5,for rounding-down operations, N<>Indicating that the pixel value is linearly stretched to 0,255]An interval;
step three: performing histogram equalization processing on the image subjected to the two-color compensation in the step
Under the complex underwater environment, the scattering effects caused by different water turbidity degrees are different, so that the pictures show different degrees of blur and need to be further processed by a histogram equalization algorithm;
in discrete form, the probability is replaced by frequency, where r is usedkRepresents IBCDiscrete gray levels of (x, y), k is the number of gray levels, and the following equation (7) holds:
wherein r is more than or equal to 0k≤1,k=0,1,2,3,…,n-1,nkFor the occurrence of r in the imagekThe number of pixels of such a gray scale, n being the total number of pixels in the image, andis the frequency of the probabilities;
the function expression of histogram equalization of the image is as follows:
The functional expression (8) of histogram equalization needs to satisfy two conditions: (a) t is a single value monotone increasing function, (b) r is more than or equal to 0 and less than or equal to L-1, and T is more than or equal to 0 and less than or equal to (r) and less than or equal to L-1;
when the two conditions are met, the cumulative distribution function converts the original histogram with uneven gray level distribution into the histogram with even gray level distribution, and the corresponding inverse transformation is as follows:
ri=T-1(Si) (9)
using transformed discrete grey levels riRecovering the image Iz=F(ri) Wherein; f is a mapping function;
step four: reprocessing the original image shot by UUV, improving the underwater image forming model, estimating the background light and the underwater transmissivity in the improved model, and further estimating the scene recovery Jc
Step 4.1, obtaining a simplified underwater optical imaging model according to a Jaffe-McGlamry underwater optical imaging model formula (1):
Ic(x)=tc(x)Jc(x)+(1-tc(x))·Ac(10)
wherein x represents a pixel point in the image, IcRepresenting images taken by UUV, tcRepresenting the underwater transmission of the color channel, JcRepresenting a clear original image, AcIs background light, c ∈ { R, G, B };
transmission t under waterc(x) Dependent on the distance z (x) of the object from the UUV and the water attenuation coefficient beta of each channelcI.e. by
tc(x)=exp(-βcz(x)) (11)
Step 4.2 estimation of background light in the underwater image formation model improved in step 4.1
Estimation of background light A using linear contrast stretching and edge detection models cThe specific principle is as follows: first, I is obtained by linear contrast stretching and structured edge detectioncEdge detection image IeAnd to IeThe pixels of (a) are arranged from large to small; secondly, cutting the pixel according to a pixel point judgment threshold Th, if the pixel x is larger than the threshold, the pixel belongs to the maximum connected component Lc, otherwise, the pixel does not belong to Lc, and the method specifically comprises the following steps:
further calculating the average value Av of the pixels and Sum in Lc, and taking Av as the estimated background light;
step 4.3, the underwater transmittance estimation in the underwater image formation model improved in step 4.1 is combined with the formula (10) and the formula (11), and the increment of the blue channel is calculated:
the increment of the red channel is calculated:
the ratio between the attenuation coefficients is defined as:
further, according to the compensation space formula (16), a structure similar to the formula (10) is formed by the underwater transmittance:
the boundary J is replaced in equation (10) to account for the different attenuation coefficients of the different color channelscNot less than 0 and obtaining lower limit t on the underwater transmissivity of the blue channelLB:
Setting pixel x of the underwater transmittance to t calculated in equation (18)LB(ii) a Since binary classification of pixels x to Lc often results in abrupt discontinuities, the underwater transmittance is estimated using the algorithm soft-matching:
In the formula (19), the compound represented by the formula (I),∑x∈LcDM(I (x)) represents the average of the Mahalanobis distances for the background light pixels,representing the maximum mahalanobis distance, σMRepresents the standard deviation, alpha (x) represents the extinction coefficient of a pixel belonging to an object or water area,
step 4.4 estimation of scene recovery Jc
Color fading compensation is performed according to equations (12), (18) and (19), and a clear original image is estimated:
step five: for J recovered in step fourcSelecting the optimal underwater recovery image based on the white balance hypothesis
Automatically selecting the best result by the grey world assumption according to different water area types of Jerlov
Step six: based on pixel level image fusion, specifically, carrying out image fusion on the image recovered in the step three and the optimal scene image selected in the step five;
Izfor bright channel color compensated and histogram processed images,and (3) reconstructing and recovering the image of the underwater 3D scene, wherein J is the image after fusion processing:
in formula (26), α and β are respective weights and satisfy α + β ═ 1.
2. According to the rightThe UUV-oriented underwater image color compensation and recovery method of claim 1, wherein the optimal result is automatically selected through grey world assumptionThe specific process comprises the following steps:
assuming that the water area type is N, i is 1,2, … N, p iIndicates the ith water area type, JcAnd if three color channels of R, G and B exist, the specific gray world algorithm processing steps are as follows:
(a) calculating water area type image JcThree-channel mean value Gray of R, G, Bi:
(b) Calculation of JcGain coefficients of three channels of images R, G, B:
(c) according to the Von Kries diagonal model, for each pixel x in the image, the R, G and B components are adjusted to be Cr、CgAnd Cb:
(d) According to the formula (23), Cr、CgAnd CbThe number of pixels is M, for CrPixel C included inrj、CgPixel C included ingjAnd CbPixel C included inbjSumming and taking the average value avr:
(e) differencing each channel avr and taking the absolute value pi:
pi=|avrr-avrg-avrb| (25)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910358840.8A CN111861896A (en) | 2019-04-30 | 2019-04-30 | UUV-oriented underwater image color compensation and recovery method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910358840.8A CN111861896A (en) | 2019-04-30 | 2019-04-30 | UUV-oriented underwater image color compensation and recovery method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111861896A true CN111861896A (en) | 2020-10-30 |
Family
ID=72965460
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910358840.8A Pending CN111861896A (en) | 2019-04-30 | 2019-04-30 | UUV-oriented underwater image color compensation and recovery method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111861896A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112132925A (en) * | 2020-11-24 | 2020-12-25 | 上海彩虹鱼海洋科技股份有限公司 | Method and device for reconstructing underwater image color |
CN112446841A (en) * | 2020-12-14 | 2021-03-05 | 中国科学院长春光学精密机械与物理研究所 | Self-adaptive image recovery method |
CN112804510A (en) * | 2021-01-08 | 2021-05-14 | 海南省海洋与渔业科学院 | Color fidelity processing method and device for deep water image, storage medium and camera |
CN114511462A (en) * | 2022-02-11 | 2022-05-17 | 电子科技大学 | Visual image enhancement method |
CN114821021A (en) * | 2022-04-28 | 2022-07-29 | 昆明理工大学 | Underwater image enhancement method combining multichannel equalization and multi-scale fusion |
WO2023070958A1 (en) * | 2021-10-28 | 2023-05-04 | 中国科学院沈阳自动化研究所 | Image restoration method based on physical scattering model |
-
2019
- 2019-04-30 CN CN201910358840.8A patent/CN111861896A/en active Pending
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112132925A (en) * | 2020-11-24 | 2020-12-25 | 上海彩虹鱼海洋科技股份有限公司 | Method and device for reconstructing underwater image color |
CN112446841A (en) * | 2020-12-14 | 2021-03-05 | 中国科学院长春光学精密机械与物理研究所 | Self-adaptive image recovery method |
CN112804510A (en) * | 2021-01-08 | 2021-05-14 | 海南省海洋与渔业科学院 | Color fidelity processing method and device for deep water image, storage medium and camera |
WO2023070958A1 (en) * | 2021-10-28 | 2023-05-04 | 中国科学院沈阳自动化研究所 | Image restoration method based on physical scattering model |
CN114511462A (en) * | 2022-02-11 | 2022-05-17 | 电子科技大学 | Visual image enhancement method |
CN114511462B (en) * | 2022-02-11 | 2023-04-18 | 电子科技大学 | Visual image enhancement method |
CN114821021A (en) * | 2022-04-28 | 2022-07-29 | 昆明理工大学 | Underwater image enhancement method combining multichannel equalization and multi-scale fusion |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zhou et al. | Auto color correction of underwater images utilizing depth information | |
CN111861896A (en) | UUV-oriented underwater image color compensation and recovery method | |
CN107358585B (en) | Foggy day image enhancement method based on fractional order differential and dark channel prior | |
Chen et al. | Hazy image restoration by bi-histogram modification | |
CN109447917B (en) | Remote sensing image haze eliminating method based on content, characteristics and multi-scale model | |
Dharejo et al. | A color enhancement scene estimation approach for single image haze removal | |
Soni et al. | An improved image dehazing technique using CLAHE and guided filter | |
Yu et al. | Image and video dehazing using view-based cluster segmentation | |
CN109003238B (en) | Image haze removal method based on model, histogram and gray level enhancement | |
CN111598814B (en) | Single image defogging method based on extreme scattering channel | |
Tang et al. | A local flatness based variational approach to retinex | |
CN117252773A (en) | Image enhancement method and system based on self-adaptive color correction and guided filtering | |
CN110335210B (en) | Underwater image restoration method | |
Wei et al. | An image fusion dehazing algorithm based on dark channel prior and retinex | |
CN111598886A (en) | Pixel-level transmittance estimation method based on single image | |
CN112419163A (en) | Single image weak supervision defogging method based on priori knowledge and deep learning | |
Wang et al. | Nighttime image dehazing using color cast removal and dual path multi-scale fusion strategy | |
CN116757949A (en) | Atmosphere-ocean scattering environment degradation image restoration method and system | |
CN116797484A (en) | Image defogging method and system based on improved dark channel prior | |
CN116883259A (en) | Underwater image enhancement method based on denoising diffusion probability model | |
Abbaspour et al. | A new fast method for foggy image enhancement | |
Yuan et al. | Defogging Technology Based on Dual‐Channel Sensor Information Fusion of Near‐Infrared and Visible Light | |
CN113379631B (en) | Image defogging method and device | |
Zheng et al. | An illumination adaptive underwater image enhancement method | |
Wang et al. | Fast visibility restoration using a single degradation image in scattering media |
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 |