CN109767399A - A kind of underwater picture Enhancement Method based on unsupervised colour correction - Google Patents
A kind of underwater picture Enhancement Method based on unsupervised colour correction Download PDFInfo
- Publication number
- CN109767399A CN109767399A CN201910032545.3A CN201910032545A CN109767399A CN 109767399 A CN109767399 A CN 109767399A CN 201910032545 A CN201910032545 A CN 201910032545A CN 109767399 A CN109767399 A CN 109767399A
- Authority
- CN
- China
- Prior art keywords
- image
- channel
- avg
- colour
- underwater
- 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
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/30—Assessment of water resources
Landscapes
- Image Processing (AREA)
Abstract
The invention belongs to ocean engineering fields, and in particular to a kind of underwater picture Enhancement Method based on unsupervised colour correction.Present invention improves over the cross-color problems for limiting to the underwater picture to the non-blue colour cast of effective solution processing for taking channel B as prominent channel；The present invention includes: RGB color model, contrast correction method, HIS color model, the underwater picture Enhancement Method based on unsupervised colour correction.The advantages of present invention more general underwater picture Enhancement Method, is: being improved to a kind of adaptive method, different colour casts are suffered from good treatment effect；Image is effectively balanced, colour cast is eliminated, improve illumination and increases true color.This method is easily achieved, and calculation amount is small, high reliablity, improves the feasibility of underwater optics image recognition, and to autonomous underwater vehicle from now on, job task etc. development has positive effect under water.
Description
Technical field
The invention belongs to ocean engineering fields, and in particular to a kind of underwater picture enhancing side based on unsupervised colour correction
Method.
Background technique
Since by underwater complex and severe imaging circumstances are influenced, taken underwater picture generally has low comparison
The deficiencies of degree, fuzzy and colour cast.The lower general who has surrendered of visual quality of images extracts subsequent characteristics and the performance of target identification process has
Serious adverse effect, therefore to underwater picture carry out enhancing pretreatment have very big significance.
At home, Harbin Engineering University's marine technology research center is by the processing to optics acoustic picture in 2003 Hes
The information such as orientation and the size of submarine target have been obtained in 2004 ocean engineering experiment, have achieved huge achievement, to demonstrate,prove
It is true and reliable that the Processing Algorithm of underwater picture, which is illustrated,.It Beijing Institute of Technology's photoelectronic imaging laboratory can be by studying sequence
Image can rebuild underwater 3 D image to obtain range information；2-d wavelet is used in the Central China University of Science and Technology in 2008
Successful division underwater picture is converted, has reached and has enhanced the contrast of local direction.It is 2012, old to propose one kind from equality people
Effective low contrast underwater picture enhances algorithm, successfully improves high-frequency gain and inhibits low-frequency component.
In deep-sea, sunlight is can not to be transferred to here, we can only remove shooting image by point light source herein,
Its center is most bright, and light emits from surrounding, even naked eyes it can be seen that, but remoter from light source, light is weaker.Cause
This just will appear the non-uniform situation of intensity profile when indicating luminous energy information on the image.Due to the shadow of underwater complex environment
It rings, light is constantly absorbed by water body, results in the continuous reduction of light energy, the scattering of hydrone and impurity in water leads to light
The offset of line.Exactly because collected information is also imperfect also in this way, underwater picture imaging just can be so difficult.Different waves
The scattering of the underwater medium of long-wave band and absorption characteristic are inconsistent, and water body causes underwater picture basic because reflecting the blue of sky
Blue is showed, which results in underwater picture cross-color (partially green or partially blue).
Summary of the invention
It is an object of the invention to solve the deficiencies in the prior art, provide a kind of based on unsupervised colour correction
Underwater picture Enhancement Method, the problem of distortion when can handle non-blue colour cast image efficiently against original method.
The beneficial effects of the present invention are:
A kind of underwater picture Enhancement Method based on unsupervised colour correction of the invention effectively identifies image color cast, solution
Original method of having determined handles the cross-color problem of the underwater picture of non-blue colour cast.The present invention transports rgb space in conjunction with HIS space
With improving the photograph of original image in addition the contrast correction method in all directions effectively eliminates the colour cast problem of underwater picture
Degree and true color.It is compared with other methods, this method can significantly improve the visual quality of underwater picture, consistent to enhance under water
The objective quality of image.
Detailed description of the invention
Fig. 1 is a kind of implementation steps figure that bearing calibration is supervised based on unsupervised color.
Fig. 2 (a) is that A schemes underwater typical target visual pattern.
Fig. 2 (b) is that B schemes underwater typical target visual pattern.
Fig. 3 (a) is the image of original method treated A figure.
Fig. 3 (b) is the image of original method treated B figure.
Fig. 4 (a) is the image of the present invention treated A figure.
Fig. 4 (b) is the image of the present invention treated B figure.
Specific embodiment
Present invention will be further explained below with reference to specific examples.It should be understood that these embodiments are merely to illustrate the present invention
Rather than it limits the scope of the invention.In addition, it should also be understood that, after having read the content that the present invention is told about, those skilled in the art
Member can make various changes or modifications the present invention, and such equivalent forms equally fall within the application the appended claims and limited
Range.
With the exploration of ocean, lake, river resources, underwater picture has become an important field of research.However, by
In by underwater complex and severe imaging circumstances are influenced, taken underwater picture usually will appear color distortion and comparison
Spend the problems such as low.Underwater Imaging can be expressed as the linear superposition of forward scattered elements and back scattering component, and forward scattering is led
The fuzzy of characteristics of image is caused, and back scattering masks the details of scene.Due to each color subgraph have different wavelength and
Energy rank can be absorbed in water with different rates.The light component of the longer wavelengths such as feux rouges or orange light is often in water
It is rapidly absorbed, therefore underwater picture typically exhibits out green or blue keynote.The lower general who has surrendered of visual quality of images is to subsequent spy
Sign is extracted and the performance of target identification process has serious adverse effect.Image enhancement is the basic skills for handling image, is
Solve that underwater photograph technical image is smudgy, one of the serious low effective means of quality.In order to ensure the information of image is complete, open
Issue a kind of better processing effect underwater picture Enhancement Method become very it is necessary to.
In order to enhance original underwater picture, the invention proposes a kind of, and the underwater picture based on unsupervised colour correction increases
Strong method, this method include following key step:
Step 1: calculation processing is carried out to image using RGB color model.If IR(i,j),IG(i,j),IJ(i, j) difference
For the RGB image red, green, blue component having a size of M × N, wherein i=1 ..., M；J=1 ..., N.Then each color point is calculated
Measure Rmax, Rmin, Gmax, GminAnd Bmax, BminMaximum and minimum pixel value:
It is most protruded using above-mentioned equation and colour cast channel least outstanding.
Step 2: Color Channel is determined using multiplier to match m rank and generates balance image.RGB face is calculated first
Each color component R in color modelavg, GavgAnd BavgAverage value:
Secondly colour cast channel outstanding is obtained using step 1, and increases other colors so that image is flat using high component
Weighing apparatus, works as RmaxWhen maximum, image is red colour cast；Work as GmaxWhen maximum, image is green colour cast；Work as BmaxWhen maximum, image is blue inclined
Color.Based on leading colour cast, two gain factors are calculated, set target average for highest Color Channel, and use multiplication
Device generates balance image to determine Color Channel to match m rank.The method proposed reduced using two Color Channels by
Influence the colour cast of image.
When the channel R is prominent channel:
A=Ravg/Gavg
B=Ravg/Bavg
G'=a × G
B'=b × B
Wherein G and B is the pixel value in original image, and G' and B' be it is adjusted after pixel value.
When the channel G is prominent channel:
A=Gavg/Ravg
B=Gavg/Bavg
R'=a × R
B'=b × B
Wherein R and B is the pixel value in original image, and R' and B' be it is adjusted after pixel value.
When channel B is prominent channel:
A=Bavg/Ravg
B=Bavg/Gavg
R'=a × R
G'=b × G
Wherein R and G is the pixel value in original image, and R' and G' be it is adjusted after pixel value.
Step 3: the bearing calibration of Comparison study degree in the image after processing.Contrast correction formula are as follows:
Wherein P0It is the pixel value by contrast correction；PiIt is considered pixel value；A is 0 lower limit value；B is 255
Upper limit value；C be in image there is currently minimum pixel value；D be in image there is currently max pixel value.Successively to upper
Side, downside, two sides use contrast correction method, and when being applied to upside, lower limit is changed to the minimum of minimum color value component
Value；When being applied to upside, the upper limit is changed to most protrude the maximum value of Color Channel；When being applied to two sides, formula is constant.
Step 4: image is transformed into HIS space from rgb space.Its conversion formula are as follows:
Step 5: in the bearing calibration of Comparison study degree in the image of HIS color space.We only carry out two sides herein
Contrast correction, the same step 3 of method.
By the operation of above five steps, enhanced underwater picture is finally obtained.Compared with other prior arts, this
Invention effectively utilizes the high colour cast in image and increases other colors to balance image；Bearing calibration is compared in color enhancement side
Face plays a crucial role because in this way can by by high colour cast color histogram be stretched to minimum side come
Reduce high colour cast.Similarly, increase low colour cast by the way that low colour cast color histogram is pulled to maximum side, to generate high quality
Image.The illumination and true color of underwater picture, therefore image are increased by the intensity and saturation parameters of HSI color model
Seem brighter richer.It considers the attribute of image and enhances image according to its characteristic rather than static criteria, therefore should
Invention preferably can carry out enhancing processing to underwater picture than existing certain methods.It is compared to original method, is solved well
Original method of having determined handles the cross-color problem of the underwater picture of non-blue colour cast.
Claims (10)
1. a kind of underwater picture Enhancement Method based on unsupervised colour correction, which is characterized in that method includes the following steps:
Step 1: acquisition underwater picture collects the underwater object figure under different background color by underwater picture acquisition system
Picture；
Step 2: calculation processing being carried out to image using RGB color model, most prominent and colour cast channel least outstanding；
Step 3: determining Color Channel using multiplier to match m rank and generate balance image；
Step 4: the bearing calibration of Comparison study degree in the image after processing；
Step 5: image is transformed into HIS space from rgb space；
Step 6: in the bearing calibration of Comparison study degree in the image of HIS color space.
2. a kind of underwater picture Enhancement Method based on unsupervised colour correction according to claim 1, which is characterized in that
Underwater picture acquisition system in the step 1 includes underwater camera, submersible lighting fixture, image pick-up card, desk-top calculating
Machine, independent 12V DC power supply and image processing platform；Underwater camera and submersible lighting fixture are placed under water, 12V DC electricity
It presses and sets on the coast while providing power supply for underwater camera and submersible lighting fixture, image pick-up card is mounted on desktop computer
PCI-E slot on.
3. a kind of underwater picture Enhancement Method based on unsupervised colour correction according to claim 2, which is characterized in that
Calculation processing is carried out to image using RGB color model described in the step 2 method particularly includes: set IR(i, j) is size
For the red component of RGB image of M × N, if IG(i, j) is the green component of RGB image having a size of M × N, if IJ(i, j) is having a size of M
The RGB image indigo plant component of × N, wherein i=1 ..., M；J=1 ..., N then calculate the maximum and minimum image of each color component
Element value, the maximum value R of red componentmaxAre as follows:
The minimum value R of red componentminAre as follows:
The maximum value G of green componentmaxAre as follows:
The minimum value G of green componentminAre as follows:
The maximum value B of blue componentmaxAre as follows:
The minimum value B of blue componentminAre as follows:
It is most protruded and colour cast channel least outstanding using formula above.
4. a kind of underwater picture Enhancement Method based on unsupervised colour correction according to claim 3, which is characterized in that
Determine that Color Channel generates balance image to match m rank described in step 3 using multiplier, comprising the following steps:
Step 3.1: calculating the average value of red component in RGB color model, blue component and green component；
Step 3.2: increasing other colors using high component with colour cast channel least outstanding according to most prominent, to protrude colour cast
To dominate colour cast, the first gain molecule and the second gain molecule are calculated；
Step 3.3: setting target average for highest Color Channel, and determine Color Channel using multiplier to match
M rank generates balance image.
5. a kind of underwater picture Enhancement Method based on unsupervised colour correction according to claim 4, which is characterized in that
The average value of red component, blue component and green component in RGB color model described in step 3.1 are as follows:
Red components R in RGB color modelavgAverage value indicate are as follows:
Blue component G in RGB color modelavgAverage value:
Green component B in RGB color modelavgAverage value:
6. a kind of underwater picture Enhancement Method based on unsupervised colour correction according to claim 4, which is characterized in that
Work as RmaxWhen maximum, image is red colour cast；Work as GmaxWhen maximum, image is green colour cast；Work as BmaxWhen maximum, image is blue colour cast, with
Prominent colour cast is leading colour cast, calculates the first gain molecule and the second gain molecule described in step 3.2；
When the channel R is prominent channel, the first gain molecule a in the channel RRWith the second gain molecule b in the channel RRAre as follows:
aR=Ravg/Gavg
bR=Ravg/Bavg
G'=aR×G
B'=bR×B
Wherein G and B is the pixel value in original image, and G' and B' be it is adjusted after pixel value；
When the channel G is prominent channel, the first gain molecule a in the channel GGWith the second gain molecule b in the channel GGAre as follows:
aG=Gavg/Ravg
bG=Gavg/Bavg
R'=aG×R
B'=bG×B
Wherein R and B is the pixel value in original image, and R' and B' be it is adjusted after pixel value；
When channel B is prominent channel, the first gain molecule a of channel BBWith the second gain molecule b of channel BBAre as follows:
aB=Bavg/Ravg
bB=Bavg/Gavg
R'=aB×R
G'=bB×G
Wherein R and G is the pixel value in original image, and R' and G' be it is adjusted after pixel value.
7. a kind of underwater picture Enhancement Method based on unsupervised colour correction according to claim 4, which is characterized in that
Contrast correction method described in step 4 the following steps are included:
Step 4.1: to the contrast correction of upside；
Step 4.2: to the contrast correction of downside；
Step 4.3: to the contrast correction of two sides.
8. a kind of underwater picture Enhancement Method based on unsupervised colour correction according to claim 7, which is characterized in that
Following formula is expressed as to the contrast correction of upside described in step 4.1:
Following formula is expressed as to the contrast correction of downside described in step 4.2:
The contrast correction of two sides is indicated described in step 4.3 are as follows:
Wherein, P0It is the pixel value by contrast correction；PiIt is considered pixel value；A is 0 lower limit value；B be 255 it is upper
Limit value；C be in image there is currently minimum pixel value；D be in image there is currently max pixel value.
9. a kind of underwater picture Enhancement Method based on unsupervised colour correction according to claim 7, which is characterized in that
Image is transformed into HIS space, conversion formula from rgb space described in step 5 are as follows:
10. a kind of underwater picture Enhancement Method based on unsupervised colour correction according to claim 9, feature exist
In,
Contrast correction method described in step 6 is identical as contrast correction method described in step 4, but only to two sides degree of comparing
Correction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910032545.3A CN109767399B (en) | 2019-01-14 | 2019-01-14 | Underwater image enhancement method based on unsupervised color correction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910032545.3A CN109767399B (en) | 2019-01-14 | 2019-01-14 | Underwater image enhancement method based on unsupervised color correction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109767399A true CN109767399A (en) | 2019-05-17 |
| CN109767399B CN109767399B (en) | 2023-04-07 |
Family
ID=66453650
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910032545.3A Active CN109767399B (en) | 2019-01-14 | 2019-01-14 | Underwater image enhancement method based on unsupervised color correction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109767399B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110189268A (en) * | 2019-05-23 | 2019-08-30 | 西安电子科技大学 | Underwater picture color correcting method based on GAN network |
| CN114418876A (en) * | 2022-01-05 | 2022-04-29 | 大连海事大学 | Underwater image enhancement method based on light scattering characteristics |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0720390A2 (en) * | 1994-12-30 | 1996-07-03 | Matsushita Electric Industrial Co., Ltd. | Non-linear color corrector and method thereof |
| CN107886486A (en) * | 2017-12-01 | 2018-04-06 | 天津大学 | Based on dark channel prior and variation Retinex underwater picture Enhancement Methods |
| CN108053374A (en) * | 2017-12-05 | 2018-05-18 | 天津大学 | A kind of underwater picture Enhancement Method of combination bilateral filtering and Retinex |
-
2019
- 2019-01-14 CN CN201910032545.3A patent/CN109767399B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0720390A2 (en) * | 1994-12-30 | 1996-07-03 | Matsushita Electric Industrial Co., Ltd. | Non-linear color corrector and method thereof |
| CN107886486A (en) * | 2017-12-01 | 2018-04-06 | 天津大学 | Based on dark channel prior and variation Retinex underwater picture Enhancement Methods |
| CN108053374A (en) * | 2017-12-05 | 2018-05-18 | 天津大学 | A kind of underwater picture Enhancement Method of combination bilateral filtering and Retinex |
Non-Patent Citations (2)
| Title |
|---|
| 李社蕾等: "基于改进的UCM水下图像增强算法研究（英文）", 《机床与液压》 * |
| 黄冬梅等: "不同颜色模型下自适应直方图拉伸的水下图像增强", 《中国图象图形学报》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110189268A (en) * | 2019-05-23 | 2019-08-30 | 西安电子科技大学 | Underwater picture color correcting method based on GAN network |
| CN114418876A (en) * | 2022-01-05 | 2022-04-29 | 大连海事大学 | Underwater image enhancement method based on light scattering characteristics |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109767399B (en) | 2023-04-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106981053A (en) | A kind of underwater picture Enhancement Method based on Weighted Fusion | |
| Sidorov | Conditional gans for multi-illuminant color constancy: Revolution or yet another approach? | |
| CN110232661A (en) | Low illumination colour-image reinforcing method based on Retinex and convolutional neural networks | |
| Wang et al. | Variational single nighttime image haze removal with a gray haze-line prior | |
| CN106780368A (en) | A kind of underwater picture Enhancement Method based on foreground model | |
| CN102063706A (en) | Rapid defogging method | |
| CN103149214B (en) | Method for detecting flaw on surface of fruit | |
| CN107316278A (en) | A kind of underwater picture clearness processing method | |
| CN106296620B (en) | A kind of color rendition method based on histogram translation | |
| CN107067386B (en) | Shallow sea underwater image enhancement method based on relative global histogram stretching | |
| CN113850747B (en) | Underwater image sharpening processing method based on light attenuation and depth estimation | |
| CN109767399A (en) | A kind of underwater picture Enhancement Method based on unsupervised colour correction | |
| Zhang et al. | Illumination decomposition for photograph with multiple light sources | |
| CN109187364A (en) | A kind of polarization imaging method under high concentration water | |
| Peng et al. | Multi-algorithm fusion of RGB and HSV color spaces for image enhancement | |
| CN103295206A (en) | low-light-level image enhancement method and device based on Retinex | |
| CN107609603A (en) | A kind of image matching method of multiple color spaces difference fusion | |
| CN111968065A (en) | Self-adaptive enhancement method for image with uneven brightness | |
| Parihar et al. | A comprehensive analysis of fusion-based image enhancement techniques | |
| CN111263044A (en) | Underwater shooting and image processing device and method | |
| Mi et al. | A generalized enhancement framework for hazy images with complex illumination | |
| Zhao et al. | Enhancing underwater imagery via latent low-rank decomposition and image fusion | |
| CN112907470A (en) | Underwater image recovery method based on Lab color gamut transformation, classification and white balance | |
| CN113160066A (en) | Low-illumination image efficient enhancement method | |
| Bi et al. | Perception and Recognition of Underwater Cable Based on Deep 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |