CN110517327A - A kind of underwater picture Enhancement Method based on color correction and contrast stretching - Google Patents
A kind of underwater picture Enhancement Method based on color correction and contrast stretching Download PDFInfo
- Publication number
- CN110517327A CN110517327A CN201910818691.9A CN201910818691A CN110517327A CN 110517327 A CN110517327 A CN 110517327A CN 201910818691 A CN201910818691 A CN 201910818691A CN 110517327 A CN110517327 A CN 110517327A
- Authority
- CN
- China
- Prior art keywords
- image
- color
- channel
- color correction
- contrast
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 45
- 238000012937 correction Methods 0.000 title claims abstract description 32
- 230000002708 enhancing effect Effects 0.000 claims abstract description 22
- 230000008569 process Effects 0.000 claims abstract description 10
- 238000005070 sampling Methods 0.000 claims description 9
- 238000010586 diagram Methods 0.000 claims description 6
- 238000010008 shearing Methods 0.000 claims description 6
- 241001062009 Indigofera Species 0.000 claims description 4
- 238000010276 construction Methods 0.000 claims description 3
- 230000001419 dependent effect Effects 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 239000003643 water by type Substances 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims 1
- 230000003014 reinforcing effect Effects 0.000 abstract description 4
- 230000000007 visual effect Effects 0.000 abstract description 3
- 239000003086 colorant Substances 0.000 abstract description 2
- 230000000694 effects Effects 0.000 abstract description 2
- 238000012545 processing Methods 0.000 abstract description 2
- 238000010521 absorption reaction Methods 0.000 abstract 1
- 101100136092 Drosophila melanogaster peng gene Proteins 0.000 description 5
- 230000008901 benefit Effects 0.000 description 3
- 238000003384 imaging method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000007500 overflow downdraw method Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000015556 catabolic process Effects 0.000 description 1
- 230000003412 degenerative effect Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 125000001475 halogen functional group Chemical group 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
-
- 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
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20016—Hierarchical, coarse-to-fine, multiscale or multiresolution image processing; Pyramid transform
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
Abstract
The present invention provides a kind of underwater picture Enhancement Method based on color correction and contrast stretching, belong to technical field of image processing, to solve absorption and scattering process due to underwater medium, leading to underwater picture, there are cross-colors, the problems such as contrast is low, and details obscures.The method of the present invention, it include: to carry out color compensation to original underwater picture first to solve the problems, such as image color cast, colour correction is carried out again solves the problems, such as color distortion, in order to highlight image detail information, R, G, B triple channel minutia figure is carried out using difference of Gaussian pyramid to reconstruct, finally in order to which the contrast and brightness that enhance image obtain final clearly underwater enhancing image using CLAHE algorithm to image progress stretched operation.Effect is to improve image overall contrast, visual good, has preferable reinforcing effect, improves global contrast, realizes details enhancing and color fidelity.
Description
Technical field
The present invention relates to technical field of image processing, specifically, more particularly to a kind of based on color correction and contrast
The underwater picture Enhancement Method of stretching.
Background technique
In recent years, preferably developed based on underwater image restoration and the sharpening technology of enhancing.Underwater picture
Restored method is the physical model based on the imaging of a underwater optics, such method passes through the degradation mechanism of analysis underwater picture,
Model parameter and inverting degenerative process is imaged to obtain the underwater picture of a clear and natural in estimation.Underwater picture Enhancement Method is simultaneously
Underwater Imaging mechanism is not considered, such method is based on existing image enhancement technique, comes by adjusting the pixel value of degraded image
Improve contrast and color, and then obtains the underwater picture with abundant information and more preferable vision.
Underwater picture Enhancement Method based on physical model specifically includes that the method for specialized hardware, the method for priori knowledge.
Method based on specialized hardware under water image enhancing and restore aspect have certain validity, but they there is also one
A little problems.For example, the underwater picture that Underwater Imaging system needs to go capture fuzzy using optical lasers sensor.But these
It is costly and complicated that hardware acquires equipment price.Method Restoration images based on priori knowledge, since the acquisition of priori knowledge is difficult,
And model is complicated, therefore puts into practical application also with certain limitation.
Underwater picture Enhancement Method based on physical model specifically includes that histogram, color constancy method, fusion method.Histogram
The contrast that only can individually improve image is stretched in Tula, and the detailed information and color information to image can not preferably enhance, face
The constant method enhancing underwater picture of color also can not be preferably while keeping color to reply, the detailed information and contrast of image
It is improved, fusion method is a variety of enhancings of fusion or reduction technique, while solving the problems, such as multiple, and this method obtained weight in recent years
Depending on.
Summary of the invention
According to technical problem set forth above, provide a kind of based on the enhancing of the underwater picture of color correction and contrast stretching
Algorithm.The problems such as there are colour casts for the main underwater picture of the present invention, and cross-color, contrast is low, loss in detail, visual difference, point
Stage handles the above problem, is reconstructed by confluent colours compensation, colour correction, details, the operations such as contrast stretching can
Underwater picture is effectively enhanced.
The technological means that the present invention uses is as follows:
A kind of underwater picture Enhancement Method based on color correction and contrast stretching, comprising the following steps:
S1, original underwater picture is obtained, and color compensating is carried out to it；
S2, color school is carried out to the image in step S1 after color compensating based on the Retinex algorithm of Auto Laves
Just, R, G, channel B image are obtained；
S3, R, G obtained in above-mentioned steps S2, channel B image are subjected to minutia with difference of Gaussian pyramid respectively
Figure reconstruct, and image corresponding R, G and channel B that obtained each channel minutia figure is fused to color correction obtain thin
Section enhancing figure；
S4, figure progress stretched operation is enhanced to details obtained in above-mentioned steps S3 using CLAHE algorithm, obtained final clear
Clear underwater enhancing image.
Further, the color compensation method in the step S1, specific formula are as follows:
Wherein, Ir(x)、Ig(x) red, the green channel of image are respectively indicated,Respectively indicate Ir(x),Ig(x) flat
Mean value, α indicate that normal parameter, the value of α are set as 1.5.
Further, in the step S2 based on the Retinex algorithm of Auto Laves in step S1 pass through color compensating
Image afterwards carries out color correction, the specific steps of which are as follows:
S21, luminance component is obtained according to MSRCR algorithm, formula is as follows:
Wherein, c ∈ { R, G, B } respectively corresponds tri- Color Channels of R, G, B；MSRc(x, y) indicates that c-th of channel is corresponding
Export image；N indicates the corresponding Gaussian function number of various criterion difference；wnIndicate the corresponding weight of each output image；σ is high
The standard deviation of this function, σ ∈ { σ1,σ2,σ3,σ4,σ5,σ6And 0≤σ1< σ2< 50,50≤σ3< σ4< 100,100≤σ5< σ6；
S22, the Retinex bearing calibration based on Auto Laves carry out colour correction to luminance component, utilize Auto Laves
Count the grey level histogram of R, G and channel B；
S23, the high light value and shading value for cutting ratio-dependent R, G and channel B, and as clipping boundary；
S24, identical linear stretch is carried out to the middle section in each channel, formula is as follows:
Wherein, MSRCRc(x, y) indicates the gray value after color correction；Min=I_sort (m*n*percent) and Max
=I_sort (m*n*(1-percent)) respectively indicate the lower limit value and upper limit value of clipping boundary；M and n respectively indicates the row of image
And column；I_sort indicates the gray value after sequence, and I_Sort=Sort (MSRc)；Percent is the ratio cut,
Percent is set as 0.5%.
Further, the process of details enhancing figure is obtained in the step S3 specifically:
S31, the Gaussian kernel w for defining k different scale and different windows size firstk, utilize wkWith original image I1It is rolled up
Product operation obtains the image of k width same sizeIt willThe 0th layer as k-th of gaussian pyramid；
S32, l (l >=1) tomographic image for constructing k-th of gaussian pyramidBy l-1 tomographic imageAnd Gaussian kernel
wkConvolution is carried out, then convolution results are carried out with two extractions in row, column direction, constructionFormula are as follows:
Wherein, N is the maximum number of plies that gaussian pyramid decomposes；RlAnd ClIt is the line number and column of gaussian pyramid l tomographic image
Number, gaussian pyramid l tomographic imageSize compared toReduce 4 times；As k=2, two gaussian pyramids are constructedWithAnd w1To be with scale and radius be all 3 Gaussian kernel, w2To be with scale and radius be all 5 Gaussian kernel；
S33, difference of Gaussian pyramid is obtained by each layer adjacent image difference of gaussian pyramid, construct DlFormula are as follows:
2 times of up-samplings of layer progress from l layers to the 1st, are added to upper one layer for the image of this layer of up-sampling, repeat this operation
Until being added to the 0th layer, it is finally completed the reconstruct of minutia figure, formula are as follows:
Wherein, IdetailFor the minutia image of reconstruct, U (Dl(i, j)) it indicates to carry out l tomographic image 2 times of up-samplings
Operation.
S33, the minutia image I by reconstructdetailWith image I1It is merged to obtain final image I2, formula
Are as follows:
I2(i, j)=Idetail(i,j)+I1(i,j)。
Further, enhance figure to details obtained in above-mentioned steps S3 using CLAHE algorithm in the step S4 to carry out
The process of stretched operation, specifically includes:
S41, the non-overlap sub-block that k size is M × N is divided the original image into；
S42, the histogram for calculating k sub-block, each Gray Histogram grade are r, number of greyscale levels NumGray, then k it is sub
The corresponding histogram functions of block are Tm,n(r), 0≤r≤NumGray-1；
S43, shearing limiting value is determined；
The grey level histogram of S44, each sub-block of shearing, the pixel number shear off are re-assigned to each ash of each histogram
It spends in grade；
S45, the grey level histogram after each subregion contrast-limited is equalized；
S46, each subregion central point is obtained, regard these points as sample point；
S47, bilinearity difference operation is carried out to each pixel, eliminates boundary artifacts, obtains new gray value.
Further, the color compensating in the step S1 further includes when image is obtained in muddy waters, to indigo plant
The step of chrominance channel compensates carries out the formula of color compensating to blue channel are as follows:
Wherein, Ib(x),Ig(x) indigo plant of distribution table diagram picture, green channel,Respectively indicate Ib(x),Ig(x) flat
Mean value, α indicate that normal parameter, the value of α are set as 1.
Compared with the prior art, the invention has the following advantages that
1, the present invention solves the distortion of image color by color compensating and color correction first, then by using height
This difference pyramid realizes the reconstruct of details, finally schemes to carry out stretched operation to enhancing using CLAHE algorithm, final to obtain
Clearly, color balance, contrast is high and details is not lost underwater enhancing image.Solve the enhancing of traditional underwater picture and
Restoration algorithm color recovery is unbalanced, can not highlight image detailed information, that fringe region generates halo artifact, contrast is low
The problems such as.
2, the present invention has carried out color correction again after color compensating, it is therefore prevented that after color compensating, respective channel occurs
Overenhanced phenomenon can restore the realistic colour of underwater picture by color compensating and color correction.
The present invention can be widely popularized in fields such as image procossings based on the above reasons.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to do simply to introduce, it should be apparent that, the accompanying drawings in the following description is this hair
Bright some embodiments for those of ordinary skill in the art without any creative labor, can be with
It obtains other drawings based on these drawings.
Fig. 1 is flow diagram of the invention.
Fig. 2 is the reinforcing effect comparison diagram that the present invention is directed to different scenes underwater picture with other algorithms.
Specific embodiment
In order to enable those skilled in the art to better understand the solution of the present invention, below in conjunction in the embodiment of the present invention
Attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is only
The embodiment of a part of the invention, instead of all the embodiments.Based on the embodiments of the present invention, ordinary skill people
The model that the present invention protects all should belong in member's every other embodiment obtained without making creative work
It encloses.
It should be noted that description and claims of this specification and term " first " in above-mentioned attached drawing, "
Two " etc. be to be used to distinguish similar objects, without being used to describe a particular order or precedence order.It should be understood that using in this way
Data be interchangeable under appropriate circumstances, so as to the embodiment of the present invention described herein can in addition to illustrating herein or
Sequence other than those of description is implemented.In addition, term " includes " and " having " and their any deformation, it is intended that cover
Cover it is non-exclusive include, for example, the process, method, system, product or equipment for containing a series of steps or units are not necessarily limited to
Step or unit those of is clearly listed, but may include be not clearly listed or for these process, methods, product
Or other step or units that equipment is intrinsic.
Embodiment
In order to verify the validity of underwater picture of the present invention enhancing, the underwater pictures of different scenes is chosen as test set,
Simultaneously with Ancuti algorithm, Drews algorithm, Peng algorithm, Fu algorithm, Peng algorithm, the experimental result of Berman algorithm is from calmly
Property and quantitative two aspect compare and analyze.Specific steps and principle are as follows:
As shown in Figure 1, the present invention provides a kind of underwater picture Enhancement Method based on color correction and contrast stretching,
The following steps are included:
S1, original underwater picture is obtained, to solve underwater picture colour cast, color benefit is carried out to the original underwater picture of acquisition
It repays；Specific formula are as follows:
Wherein, Ir(x)、Ig(x) red, the green channel of image are respectively indicated,Respectively indicate Ir(x),Ig(x) flat
Mean value, α indicate normal parameter, it is contemplated that red channel attenuation ratio is more serious.Finally, the value of α is set as through a large number of experiments
1.5。
When image is obtained in muddy waters, image is often partially green, and blue channel is it is possible that apparent at this time
Attenuation carries out the formula of color compensating to blue channel therefore, it is necessary to compensate to blue channel are as follows:
Wherein, Ib(x),Ig(x) indigo plant of distribution table diagram picture, green channel,Respectively indicate Ib(x),Ig(x) flat
Mean value, α indicate that normal parameter, the value of α are set as 1.
S2, to prevent the underwater picture after color compensating from cross-color phenomenon occur, Retinex based on Auto Laves is calculated
Method carries out color correction to the image in step S1 after color compensating, obtains R, G, channel B image；
Face is carried out to the image in step S1 after color compensating based on the Retinex algorithm of Auto Laves in step S2
Color correction, the specific steps of which are as follows:
S21, luminance component is obtained according to MSRCR algorithm, formula is as follows:
Wherein, c ∈ { R, G, B } respectively corresponds tri- Color Channels of R, G, B；MSRc(x, y) indicates that c-th of channel is corresponding
Export image；N indicates the corresponding Gaussian function number of various criterion difference；wnIndicate the corresponding weight of each output image；σ is high
The standard deviation of this function, σ ∈ { σ1,σ2,σ3,σ4,σ5,σ6And 0≤σ1< σ2< 50,50≤σ3< σ4< 100,100≤σ5< σ6；
S22, the Retinex bearing calibration based on Auto Laves carry out colour correction to luminance component, utilize Auto Laves
Count the grey level histogram of R, G and channel B；
S23, the high light value and shading value for cutting ratio-dependent R, G and channel B, and as clipping boundary；
S24, identical linear stretch is carried out to the middle section in each channel, formula is as follows:
Wherein, MSRCRc(x, y) indicates the gray value after color correction；Min=I_sort (m*n*percent) and Max
=I_sort (m*n* (1-percent)) respectively indicates the lower limit value and upper limit value of clipping boundary；M and n respectively indicate image
Row and column；I_sort indicates the gray value after sequence, and I_Sort=Sort (MSRc)；Percent is the ratio cut,
Percent is set as 0.5%.
S3, R, G obtained in above-mentioned steps S2, channel B image are subjected to minutia with difference of Gaussian pyramid respectively
Figure reconstruct, and image corresponding R, G and channel B that obtained each channel minutia figure is fused to color correction obtain thin
Section enhancing figure；
The process of details enhancing figure is obtained in step S3 specifically:
The decomposable process of Guassian pyramid transformation:
S31, the Gaussian kernel w for defining k different scale and different windows size firstk, utilize wkWith original image I1It is rolled up
Product operation obtains the image of k width same sizeIt willThe 0th layer as k-th of gaussian pyramid；
S32, l (l >=1) tomographic image for constructing k-th of gaussian pyramidBy l-1 tomographic imageAnd Gaussian kernel
wkConvolution is carried out, then convolution results are carried out with two extractions in row, column direction, constructionFormula are as follows:
Wherein, N is the maximum number of plies that gaussian pyramid decomposes；RlAnd ClIt is the line number and column of gaussian pyramid l tomographic image
Number, gaussian pyramid l tomographic imageSize compared toReduce 4 times；As k=2, two gaussian pyramids are constructedWithAnd w1To be with scale and radius be all 3 Gaussian kernel, w2To be with scale and radius be all 5 Gaussian kernel；
S33, difference of Gaussian pyramid is obtained by each layer adjacent image difference of gaussian pyramid, construct DlFormula are as follows:
2 times of up-samplings of layer progress from l layers to the 1st, are added to upper one layer for the image of this layer of up-sampling, repeat this operation
Until being added to the 0th layer, it is finally completed the reconstruct of minutia figure, formula are as follows:
Wherein, IdetailFor the minutia image of reconstruct, U (Dl(i, j)) it indicates to carry out l tomographic image 2 times of up-samplings
Operation.
S33, the minutia image I by reconstructdetailWith image I1It is merged to obtain final image I2, formula
Are as follows:
I2(i, j)=Idetail(i,j)+I1(i,j)。
S4, contrast and shade of color in order to further increase image are obtained using CLAHE algorithm in above-mentioned steps S3
The details enhancing figure arrived carries out stretched operation, obtains final clearly underwater enhancing image.
The mistake that figure carries out stretched operation is enhanced to details obtained in above-mentioned steps S3 using CLAHE algorithm in step S4
Journey specifically includes:
S41, the non-overlap sub-block that k size is M × N is divided the original image into；
S42, the histogram for calculating k sub-block, each Gray Histogram grade are r, number of greyscale levels NumGray, then k it is sub
The corresponding histogram functions of block are Tm,n(r), 0≤r≤NumGray-1；S43, shearing limiting value is determined；
The grey level histogram of S44, each sub-block of shearing, the pixel number shear off are re-assigned to each ash of each histogram
It spends in grade；
S45, the grey level histogram after each subregion contrast-limited is equalized；
S46, each subregion central point is obtained, regard these points as sample point；
S47, bilinearity difference operation is carried out to each pixel, eliminates boundary artifacts, obtains new gray value.
As shown in Fig. 2, the reinforcing effect comparison diagram of different scenes underwater picture is directed to other algorithms for the present invention, from reality
Testing effect picture can be seen that compared to other algorithms, the present invention effective noise reduction in terms of local detail, and it is thin to enhance part
Section.In terms of color, for the image that the present invention is handled without colour cast, color restores normal.In terms of contrast, the present invention is improved
Image overall contrast is visual good.Therefore the present invention has preferable reinforcing effect, improves global contrast, realizes
Details enhancing and color fidelity.
The present embodiment carries out pair the experimental result of algorithms of different from tetra- kinds of objective indicators of IE, PCQI, UIQM and UCIQE
Than；From the data of table 1 it is found that the mean value of IE, PCQI, UIQM and UCIQE that the present invention obtains are above other algorithms, this shows
The present invention has biggish promotion to the color of original image, contrast, detail textures, clarity, and is better than other enhancing algorithms.
IE, PCQI, UIQM and UCIQE index (corresponding image of 1. present invention of table and other 6 kinds of algorithms (sequence is identical)
As shown in Figure 2)
The present embodiment is compared from experimental result of the runing time to algorithms of different；From the data of table 2 it is found that running
In terms of time, the method for the present invention is better than Drews algorithm, Li algorithm, Peng algorithm and Berman algorithm, micro- to be inferior to Fu algorithm.With
Increase image resolution ratio, this advantage be more obvious.
Table 2.Peng, Drews, Peng, Berman, Li, Fu, Ancuti algorithm and Operational Timelines of the invention
| Image Resolution | Peng 2017 | Drews | Peng 2018 | Berman | Li | Fu | Ancuti | Ours |
| 400×300 | 28.35 | 14.78 | 2.25 | 35.65 | 0.95 | 0.32 | 0.49 | 0.43 |
| 550×412 | 56.92 | 30.32 | 3.9 | 90.1 | 1.69 | 0.49 | 0.89 | 0.70 |
| 853×600 | 129.01 | 70.03 | 8.95 | 121.03 | 4.01 | 1.01 | 1.91 | 1.60 |
| 907×757 | 158.98 | 84.52 | 9.23 | 23.74 | 4.75 | 1.21 | 2.31 | 1.92 |
| 1024×738 | 183.85 | 96.98 | 13.4 | 249.92 | 5.51 | 1.28 | 2.67 | 2.35 |
| 1920×1080 | 489.52 | 259.44 | 24.62 | 251.78 | 13.01 | 2.91 | 6.24 | 4.92 |
| 2208×1474 | 789.98 | 414.76 | 51.82 | 105.19 | 21.96 | 4.59 | 9.55 | 8.78 |
Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although
Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: it still may be used
To modify to technical solution documented by previous embodiment, or some or all of the technical features are equal
Replacement；And these are modified or replaceed, the model for technical solution of the embodiment of the present invention that it does not separate the essence of the corresponding technical solution
It encloses.
Claims (6)
1. a kind of underwater picture Enhancement Method based on color correction and contrast stretching, which comprises the following steps:
S1, original underwater picture is obtained, and color compensating is carried out to it；
S2, color correction is carried out to the image in step S1 after color compensating based on the Retinex algorithm of Auto Laves, obtained
To R, G, channel B image；
S3, R, G obtained in above-mentioned steps S2, channel B image are subjected to minutia figure weight with difference of Gaussian pyramid respectively
Structure, and image corresponding R, G and channel B that obtained each channel minutia figure is fused to color correction obtain details increasing
Qiang Tu；
S4, figure progress stretched operation is enhanced to details obtained in above-mentioned steps S3 using CLAHE algorithm, obtained finally clearly
Underwater enhancing image.
2. the underwater picture Enhancement Method according to claim 1 based on color correction and contrast stretching, feature exist
In, color compensation method in the step S1, specific formula are as follows:
Wherein, Ir(x)、Ig(x) red, the green channel of image are respectively indicated,Respectively indicate Ir(x),Ig(x) average value,
α indicates that normal parameter, the value of α are set as 1.5.
3. the underwater picture Enhancement Method based on color correction and contrast stretching according to claim 1,
It is characterized in that, based on the Retinex algorithm of Auto Laves to the image in step S1 after color compensating in the step S2
Color correction is carried out, the specific steps of which are as follows:
S21, luminance component is obtained according to MSRCR algorithm, formula is as follows:
Wherein, c ∈ { R, G, B } respectively corresponds tri- Color Channels of R, G, B；MSRc(x, y) indicates the corresponding output in c-th of channel
Image；N indicates the corresponding Gaussian function number of various criterion difference；wnIndicate the corresponding weight of each output image；σ is Gaussian function
Several standard deviations, σ ∈ { σ1,σ2,σ3,σ4,σ5,σ6And 0≤σ1< σ2< 50,50≤σ3< σ4< 100,100≤σ5< σ6；
S22, the Retinex bearing calibration based on Auto Laves carry out colour correction to luminance component, are counted using Auto Laves
The grey level histogram of R, G and channel B out；
S23, the high light value and shading value for cutting ratio-dependent R, G and channel B, and as clipping boundary；
S24, identical linear stretch is carried out to the middle section in each channel, formula is as follows:
Wherein, MSRCRc(x, y) indicates the gray value after color correction；Min=I_sort (m*n*percent) and Max=I_
sort(m*n*(1-percent)) respectively indicate the lower limit value and upper limit value of clipping boundary；M and n respectively indicate image row and
Column；I_sort indicates the gray value after sequence, and I_Sort=Sort (MSRc)；Percent is the ratio cut, percent
It is set as 0.5%.
4. the underwater picture Enhancement Method according to claim 1 based on color correction and contrast stretching, feature exist
In obtaining the process of details enhancing figure in the step S3 specifically:
S31, the Gaussian kernel w for defining k different scale and different windows size firstk, utilize wkWith original image I1Carry out convolution fortune
Calculation obtains the image of k width same sizeIt willThe 0th layer as k-th of gaussian pyramid；
S32, l (l >=1) tomographic image for constructing k-th of gaussian pyramidBy l-1 tomographic imageWith Gaussian kernel wkInto
Row convolution, then convolution results are carried out with two extractions in row, column direction, constructionFormula are as follows:
Wherein, N is the maximum number of plies that gaussian pyramid decomposes；RlAnd ClIt is the line number and columns of gaussian pyramid l tomographic image,
Gaussian pyramid l tomographic imageSize compared toReduce 4 times；As k=2, two gaussian pyramids are constructed
WithAnd w1To be with scale and radius be all 3 Gaussian kernel, w2To be with scale and radius be all 5 Gaussian kernel；
S33, difference of Gaussian pyramid is obtained by each layer adjacent image difference of gaussian pyramid, construct DlFormula are as follows:
Carry out 2 times of up-samplings from l layers to the 1st layer, the image of this layer of up-sampling be added to upper one layer, repeat this operation until
It is added to the 0th layer, is finally completed the reconstruct of minutia figure, formula are as follows:
Wherein, IdetailFor the minutia image of reconstruct, U (Dl(i, j)) it indicates to carry out l tomographic image 2 times of up-sampling operations.
S33, the minutia image I by reconstructdetailWith image I1It is merged to obtain final image I2, formula are as follows:
I2(i, j)=Idetail(i,j)+I1(i,j)。
5. the underwater picture Enhancement Method according to claim 1 based on color correction and contrast stretching, feature exist
In, the process that figure carries out stretched operation is enhanced to details obtained in above-mentioned steps S3 using CLAHE algorithm in the step S4,
It specifically includes:
S41, the non-overlap sub-block that k size is M × N is divided the original image into；
S42, the histogram for calculating k sub-block, each Gray Histogram grade are r, number of greyscale levels NumGray, then k sub-block pair
The histogram functions answered are Tm,n(r), 0≤r≤NumGray-1；
S43, shearing limiting value is determined；
The grey level histogram of S44, each sub-block of shearing, the pixel number shear off are re-assigned to each gray level of each histogram
In；
S45, the grey level histogram after each subregion contrast-limited is equalized；
S46, each subregion central point is obtained, regard these points as sample point；
S47, bilinearity difference operation is carried out to each pixel, eliminates boundary artifacts, obtains new gray value.
6. the underwater picture Enhancement Method according to claim 1 or 2 based on color correction and contrast stretching, feature
It is, the color compensating in the step S1 further includes mending when image is obtained in muddy waters to blue channel
The step of repaying carries out the formula of color compensating to blue channel are as follows:
Wherein, Ib(x),Ig(x) indigo plant of distribution table diagram picture, green channel,Respectively indicate Ib(x),Ig(x) average value,
α indicates that normal parameter, the value of α are set as 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910818691.9A CN110517327B (en) | 2019-08-30 | 2019-08-30 | Underwater image enhancement method based on color correction and contrast stretching |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910818691.9A CN110517327B (en) | 2019-08-30 | 2019-08-30 | Underwater image enhancement method based on color correction and contrast stretching |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110517327A true CN110517327A (en) | 2019-11-29 |
| CN110517327B CN110517327B (en) | 2022-10-04 |
Family
ID=68629844
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910818691.9A Active CN110517327B (en) | 2019-08-30 | 2019-08-30 | Underwater image enhancement method based on color correction and contrast stretching |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110517327B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111028181A (en) * | 2019-12-25 | 2020-04-17 | 歌尔股份有限公司 | Image enhancement processing method, device, equipment and storage medium |
| CN111161170A (en) * | 2019-12-18 | 2020-05-15 | 江苏科技大学 | Underwater image comprehensive enhancement method for target recognition |
| CN111325690A (en) * | 2020-02-20 | 2020-06-23 | 大连海事大学 | Self-adaptive underwater image enhancement method based on differential evolution algorithm |
| CN112200019A (en) * | 2020-09-22 | 2021-01-08 | 江苏大学 | Rapid building night scene lighting light fault detection method |
| CN112561804A (en) * | 2020-10-09 | 2021-03-26 | 天津大学 | Low-illumination underwater image enhancement method based on multi-scale detail enhancement |
| CN113538304A (en) * | 2020-12-14 | 2021-10-22 | 腾讯科技（深圳）有限公司 | Training method and device of image enhancement model, and image enhancement method and device |
| CN114820385A (en) * | 2022-05-20 | 2022-07-29 | 河南科技学院 | Local self-adaptive underwater image color correction method |
| CN116778261A (en) * | 2023-08-21 | 2023-09-19 | 山东恒信科技发展有限公司 | Raw oil grade classification method based on image processing |
| CN116823674A (en) * | 2023-08-24 | 2023-09-29 | 湖南省水务规划设计院有限公司 | Cross-modal fusion underwater image enhancement method |
| CN117788303B (en) * | 2023-11-30 | 2024-05-31 | 江苏海洋大学 | Underwater image enhancement method based on self-adaptive histogram and G-MSRCR |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070279500A1 (en) * | 2006-06-05 | 2007-12-06 | Stmicroelectronics S.R.L. | Method for correcting a digital image |
| CN107507138A (en) * | 2017-07-27 | 2017-12-22 | 北京大学深圳研究生院 | A kind of underwater picture Enhancement Method based on Retinex model |
| CN107886486A (en) * | 2017-12-01 | 2018-04-06 | 天津大学 | Based on dark channel prior and variation Retinex underwater picture Enhancement Methods |
| CN110175964A (en) * | 2019-05-30 | 2019-08-27 | 大连海事大学 | A kind of Retinex image enchancing method based on laplacian pyramid |
-
2019
- 2019-08-30 CN CN201910818691.9A patent/CN110517327B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070279500A1 (en) * | 2006-06-05 | 2007-12-06 | Stmicroelectronics S.R.L. | Method for correcting a digital image |
| CN107507138A (en) * | 2017-07-27 | 2017-12-22 | 北京大学深圳研究生院 | A kind of underwater picture Enhancement Method based on Retinex model |
| CN107886486A (en) * | 2017-12-01 | 2018-04-06 | 天津大学 | Based on dark channel prior and variation Retinex underwater picture Enhancement Methods |
| CN110175964A (en) * | 2019-05-30 | 2019-08-27 | 大连海事大学 | A kind of Retinex image enchancing method based on laplacian pyramid |
Non-Patent Citations (1)
| Title |
|---|
| 李昌利等: "基于多通道均衡化的水下彩色图像增强算法", 《华中科技大学学报(自然科学版)》 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111161170A (en) * | 2019-12-18 | 2020-05-15 | 江苏科技大学 | Underwater image comprehensive enhancement method for target recognition |
| CN111028181B (en) * | 2019-12-25 | 2023-07-14 | 歌尔股份有限公司 | Image enhancement processing method, device, equipment and storage medium |
| CN111028181A (en) * | 2019-12-25 | 2020-04-17 | 歌尔股份有限公司 | Image enhancement processing method, device, equipment and storage medium |
| CN111325690A (en) * | 2020-02-20 | 2020-06-23 | 大连海事大学 | Self-adaptive underwater image enhancement method based on differential evolution algorithm |
| CN111325690B (en) * | 2020-02-20 | 2023-08-22 | 大连海事大学 | Self-adaptive underwater image enhancement method based on differential evolution algorithm |
| CN112200019A (en) * | 2020-09-22 | 2021-01-08 | 江苏大学 | Rapid building night scene lighting light fault detection method |
| CN112200019B (en) * | 2020-09-22 | 2024-02-09 | 上海罗曼照明科技股份有限公司 | Rapid building night scene lighting lamp fault detection method |
| CN112561804A (en) * | 2020-10-09 | 2021-03-26 | 天津大学 | Low-illumination underwater image enhancement method based on multi-scale detail enhancement |
| CN113538304B (en) * | 2020-12-14 | 2023-08-18 | 腾讯科技（深圳）有限公司 | Training method and device for image enhancement model, and image enhancement method and device |
| CN113538304A (en) * | 2020-12-14 | 2021-10-22 | 腾讯科技（深圳）有限公司 | Training method and device of image enhancement model, and image enhancement method and device |
| CN114820385B (en) * | 2022-05-20 | 2023-06-23 | 河南科技学院 | Locally adaptive underwater image color correction method |
| CN114820385A (en) * | 2022-05-20 | 2022-07-29 | 河南科技学院 | Local self-adaptive underwater image color correction method |
| CN116778261A (en) * | 2023-08-21 | 2023-09-19 | 山东恒信科技发展有限公司 | Raw oil grade classification method based on image processing |
| CN116778261B (en) * | 2023-08-21 | 2023-11-14 | 山东恒信科技发展有限公司 | Raw oil grade classification method based on image processing |
| CN116823674A (en) * | 2023-08-24 | 2023-09-29 | 湖南省水务规划设计院有限公司 | Cross-modal fusion underwater image enhancement method |
| CN116823674B (en) * | 2023-08-24 | 2024-03-12 | 湖南省水务规划设计院有限公司 | Cross-modal fusion underwater image enhancement method |
| CN117788303B (en) * | 2023-11-30 | 2024-05-31 | 江苏海洋大学 | Underwater image enhancement method based on self-adaptive histogram and G-MSRCR |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110517327B (en) | 2022-10-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110517327A (en) | A kind of underwater picture Enhancement Method based on color correction and contrast stretching | |
| CN105913396B (en) | A kind of image border holding mixing denoising method of noise estimation | |
| CN107492070A (en) | A kind of single image super-resolution computational methods of binary channels convolutional neural networks | |
| CN111105376B (en) | Single-exposure high-dynamic-range image generation method based on double-branch neural network | |
| CN104252700A (en) | Histogram equalization method for infrared image | |
| CN104574328A (en) | Color image enhancement method based on histogram segmentation | |
| CN107292830A (en) | Low-light (level) image enhaucament and evaluation method | |
| CN111179196B (en) | Multi-resolution depth network image highlight removing method based on divide-and-conquer | |
| CN113284061B (en) | Underwater image enhancement method based on gradient network | |
| CN110223251A (en) | Suitable for manually with the convolutional neural networks underwater image restoration method of lamp | |
| CN111598814B (en) | Single image defogging method based on extreme scattering channel | |
| CN111462002B (en) | Underwater image enhancement and restoration method based on convolutional neural network | |
| CN107146202B (en) | Image blind deblurring method based on L0 regularization and fuzzy kernel post-processing | |
| CN105243647A (en) | Linear spatial filtering-based image enhancement method | |
| Zhu et al. | Low-light image enhancement network with decomposition and adaptive information fusion | |
| CN114155173A (en) | Image defogging method and device and nonvolatile storage medium | |
| CN112132757B (en) | General image restoration method based on neural network | |
| CN109767407A (en) | A kind of quadratic estimate method of atmospheric transmissivity image during defogging | |
| CN116703789A (en) | Image enhancement method and system | |
| CN117333359A (en) | Mountain-water painting image super-resolution reconstruction method based on separable convolution network | |
| Park et al. | False contour reduction using neural networks and adaptive bi-directional smoothing | |
| CN104616310B (en) | The appraisal procedure and device of a kind of picture quality | |
| CN116433525A (en) | Underwater image defogging method based on edge detection function variation model | |
| CN113724139B (en) | Unsupervised infrared single-image super-resolution method for generating countermeasure network based on double discriminators | |
| CN115760640A (en) | Coal mine low-illumination image enhancement method based on noise-containing Retinex model |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |