CN110458769B - Color polarization image restoration method based on elimination of inter-channel crosstalk - Google Patents
Color polarization image restoration method based on elimination of inter-channel crosstalk Download PDFInfo
- Publication number
- CN110458769B CN110458769B CN201910660103.3A CN201910660103A CN110458769B CN 110458769 B CN110458769 B CN 110458769B CN 201910660103 A CN201910660103 A CN 201910660103A CN 110458769 B CN110458769 B CN 110458769B
- Authority
- CN
- China
- Prior art keywords
- crosstalk
- image
- elimination
- matrix
- color
- 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.)
- Active
Links
- 230000010287 polarization Effects 0.000 title claims abstract description 37
- 230000008030 elimination Effects 0.000 title claims abstract description 25
- 238000003379 elimination reaction Methods 0.000 title claims abstract description 25
- 238000000034 method Methods 0.000 title claims abstract description 19
- 239000011159 matrix material Substances 0.000 claims abstract description 22
- 238000010586 diagram Methods 0.000 claims abstract description 9
- 238000002834 transmittance Methods 0.000 claims description 11
- 238000001514 detection method Methods 0.000 claims description 8
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 7
- 230000005540 biological transmission Effects 0.000 claims description 6
- 238000005286 illumination Methods 0.000 claims description 4
- 230000003287 optical effect Effects 0.000 abstract description 4
- 230000000694 effects Effects 0.000 description 10
- 238000003384 imaging method Methods 0.000 description 10
- 238000005516 engineering process Methods 0.000 description 4
- 239000002245 particle Substances 0.000 description 3
- 230000002238 attenuated effect Effects 0.000 description 2
- 239000003086 colorant Substances 0.000 description 2
- 238000011084 recovery Methods 0.000 description 2
- 238000010521 absorption reaction Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Images
Classifications
-
- G06T5/73—
-
- 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
Abstract
The invention discloses a color polarization image restoration method based on elimination of crosstalk between channels, which comprises the following steps of 1, photographing a detected target object 3; step 2, finding the brightest and darkest images in the obtained color polarization images, and respectively corresponding to the maximum and minimum conditions of the back scattering light; step 3, correcting the transmissivity deviation caused by the crosstalk of each channel by utilizing a crosstalk elimination matrix of the transmissivity among the RGB channels, and step 4, optimizing a matrix T opt Substituting into the crosstalk elimination matrix formula and the crosstalk elimination light intensity diagram formula, and calculating step by step to obtain the final restored image after crosstalk elimination restoration. Compared with the prior art, the invention combines the channel crosstalk of the color image with the polarization optical model in the traditional scattering medium for the first time, is applied to the field of defogging of the color polarization image in the environment of the scattering medium, and organically integrates the polarization image restoration and the color correction.
Description
Technical Field
The invention belongs to the technical field of polarization imaging restoration, and particularly relates to a color polarization image restoration method based on elimination of inter-channel crosstalk in a scattering environment.
Background
The polarization information is one of the basic physical information of the light wave, and can provide the information of the object which can not be provided by other light wave information. The polarization imaging technology is a novel optical detection technology developed based on the idea, and especially under the scattering environment, the target detection and identification based on the polarization imaging technology has unique advantages and special applications incomparable with other imaging modes. In the traditional polarization recovery method, a color image is recovered by generally adopting a channel division processing mode, however, in the imaging process of a color camera, the recovery effect is usually large in color deviation due to crosstalk between RGB channels, and the phenomenon is particularly obvious when the water body presents heavy colors.
Disclosure of Invention
The invention provides a color polarization image restoration method based on the elimination of the crosstalk between channels, which eliminates the deviation caused by the crosstalk between RGB channels in the traditional color polarization image restoration technology, achieves the aims of improving the restoration quality and restoring the balance between color image channels in the scattering medium environment, and has the advantages of simple steps, wide application range and obvious restoration effect.
The color polarization image restoration method based on elimination of the crosstalk between the RGB channels corrects the transmissivity deviation caused by the crosstalk of each channel by utilizing the crosstalk elimination matrix of the transmissivity between the RGB channels, thereby restoring a color image under a scattering environment, improving the imaging definition and simultaneously restoring real colors.
The invention discloses a color polarization image restoration method based on elimination of crosstalk between channels, which comprises the following steps:
determining the transmission of RGB channels
The formula is as follows:
wherein (x, y) represents the coordinates of a pixel in the diagram,
and &>
Representing the backscattered light at infinity in the parallel and vertical directions, I || (x,y)、I ⊥ (x, y) respectively representing the light intensity received by the detector in the parallel direction and the light intensity received by the detector in the vertical direction, and i represents the RGB channel number;
will be provided with
Multiplying a 3 × 3 matrix T to eliminate the inter-channel crosstalk, and obtaining a 3 × 3 crosstalk elimination matrix which makes the image restoration effect optimal, where the expression is as follows:
thereby obtaining RGB channel transmittance value after eliminating crosstalk
And calculating a light intensity graph after crosstalk elimination according to the formula:
and (3) optimizing the result by using an EME function as the criterion of the image, wherein the expression of the EME function is as follows:
wherein k is 1 ×k 2 The representation is divided into blocks, each block is represented byThe position on the picture is determined by the position,
and &>
Representing the maximum and minimum intensity values of the pixels in a block ω, which is (k, l) in the image, q is set to a small constant value to prevent the divisor from being 0;
the final matrix result optimized by SCE is:
Compared with the prior art, the invention has the following beneficial effects:
the channel crosstalk of the color image is combined with a polarization optical model in the traditional scattering medium for the first time, the method is applied to the field of defogging of the color polarization image in the environment of the scattering medium, and the polarization image restoration and the color correction are organically integrated;
the light intensity modulation effect of the scattering medium on the object reflected light is reduced, the color distortion of the RGB channels of the color image is recovered, and the more accurate reflected light information of the RGB channels of the object is obtained, so that the image definition and the color accuracy of the defogged image are obviously improved, and the more clear and more real restored image can be obtained in the scattering medium.
Drawings
Fig. 1 is a schematic diagram of an underwater imaging system according to an embodiment of the present invention.
FIG. 2 is a graph of RGB channel quantum efficiency distribution for a color camera;
FIG. 3 is a graph of the quadrature intensity obtained with a PSA system;
FIG. 4 is a defogged image restored by the color polarization image restoration method based on the elimination of the crosstalk between channels of the invention: the method comprises the steps of (a) obtaining an original light intensity graph, and (b) defogging and restoring the graph;
fig. 5 is a schematic diagram of the effects of the embodiment, (a) (c) other common scattering medium scenes, and (b) (d) restored images after defogging.
Reference numerals:
1. an illumination system 2, a water tank filled with scattering medium (turbid water), 3, a detected target object 4, a polarization analyzer (PSA), 5 and a color light intensity detection device (CCD).
Detailed Description
The following detailed description of embodiments of the invention will be made with reference to the accompanying drawings.
Fig. 1 is a schematic diagram of an underwater imaging system according to an embodiment of the present invention; the detected target object 3 is placed in a water tank 2 filled with a scattering medium (turbid water), an illumination system 1 (including an active type or a passive type) 1 illuminates the direction of the detected target object 3, and a color light intensity detection device (CCD) 5 shoots the detected target object 3 through a polarization analyzer 4 (PSA) so as to obtain a color polarization image of the detected target object. The method of the invention is used for imaging the measured object in the scattering medium, the polarization analyzer (PSA) is used for extracting the polarization information of each channel, and the polarization information is used for calculating the transmission rate t of the RGB channels of the color image, thereby calculating the original reflection light intensity of the object. Compared with the previous research, the invention particularly considers the crosstalk elimination of each channel, and a 3 x 3 matrix is used for compensating and correcting specific t values of each channel. Particularly, under the scattering medium environment with high color deviation, the method can further improve the defogging restoration quality of the color polarization image and realize the improvement of the detection effect.
The theoretical basis of the invention comprises the following specific contents:
when detecting an object in a scattering medium, the light received by the detector can be divided into two parts: part of the light is object reflected light D (x, y), which is attenuated by absorption and scattering by scattering particles when transmitted in a scattering medium:
D(x,y)＝L(x,y)t(x,y) (1)
t(x,y)＝e -β(x,y)z(x,y) (2)
where (x, y) represents the coordinates of the pixel in the diagram, L (x, y) represents the object reflected light that has not been attenuated by the scattering particles, t (x, y) represents the transmittance of the medium, β (x, y) represents the attenuation coefficient, and z (x, y) represents the transmission distance of the light in the scattering medium.
The other part is the light scattered by the particles into the detector, called background light or backscattered light, which is expressed by:
B(x,y)＝A ∞ [1-t(x,y)] (3)
wherein, A ∞ Representing the backscattering value in the scattering medium extending to infinity. The total light intensity I (x, y) received by the detector is obtained, and the expression is as follows:
in most cases, the two images corresponding to the maximum and minimum intensities of polarized light generated by scattering are approximately orthogonal, whereas in the conventional polarization method, the object reflected light D (x, y) can be considered to be equal in the orthogonal directions, and the backscattered light B is respectively in the orthogonal directions || (x, y) and B ⊥ (x, y). The light intensity received by the detector is I in the parallel direction || (x, y) in the vertical direction is I ⊥ (x, y), the expression is as follows:
according to the formula, the transmissivity obtained by calculating the crosstalk of the RGB channels without eliminating can be obtained by using two light intensity graphs in the orthogonal directions
Comprises the following steps:
wherein, the first and the second end of the pipe are connected with each other,
and/or>
Representing backscattered light at infinity in the parallel and vertical directions, respectively.
The RGB channel quantum efficiency profile of the color camera shown in fig. 2 indicates that: in a color camera, where there is an overlapping range of the perceived wavelengths of the RGB channels, meaning that there is an unavoidable crosstalk between the channels, the detected RGB channel intensities are determined by equation (7):
I i (x,y)＝∫Q i (λ)L(x,y,λ)t(x,y,λ)+Q i (λ)A ∞ (λ)[1-t(x,y,λ)]dλ,i∈[R,G,B]
(7)
where Q (λ) represents the quantum efficiency of the camera for the wavelength λ, and t (x, y, λ) represents the transmittance of the wavelength λ. In this case, if the transmittance calculated by each of the three channels is continuously used to directly calculate the light transmitted directly from the object, a relatively large color deviation occurs, and a good defogging effect cannot be obtained. Analyzing the crosstalk between channels of the polarized image, and searching to obtain a 3 multiplied by 3 crosstalk elimination matrix which enables the image restoration effect to be optimal; therefore, it will be directly estimated
A 3 x 3 matrix T is multiplied to achieve inter-channel crosstalk cancellation.
Wherein the content of the first and second substances,
the transmittance value after eliminating the crosstalk is represented, and then a light intensity graph after eliminating the crosstalk is calculated according to the transmittance value, wherein the expression is as follows:
in formula (8), T is a 3 × 3 matrix that can be searched to obtain an optimized result, and an EME function (measure effect) is used as a criterion of an image to perform result optimization, where the EME function expression is:
where the image is divided into k 1 ×k 2 Small squares, (k, l) indicate the position of each square on the picture,
and &>
Representing the maximum and minimum intensity values of a pixel within a block omega, which is (k, l) in the image, q is set to a small constant value to prevent the divisor from being 0. The final matrix result optimized by SCE is:
will optimize the matrix T opt And substituting the formula (8) and the formula (9) into the image to obtain a final restored image after crosstalk elimination restoration through gradual calculation. The matrix is adopted to eliminate crosstalk in the imaging process among RGB channels, simultaneously realize the processes of image quality restoration and color balance of each channel, realize image restoration in the final scattering medium environment by combining a traditional polarized optical model in the scattering medium, obtain more accurate object reflected light information and solve the problem of the crosstalk among the channels neglected in the defogging process of a color image. Particularly, under the common scattering medium environment with large color deviation (such as a deep blue water body under a deep sea environment), the invention can further improve the defogging restoration quality of the color polarization image of the object and realize the improvement of the detection effect.
Claims (1)
1. A color polarization image restoration method based on elimination of crosstalk between channels is characterized by comprising the following steps:
step 1, placing a detected target object in a water tank filled with scattering media, illuminating the detected target object by an illumination system, and photographing the detected target object by a light intensity detection transmission polarization analyzer to obtain a color polarization image of the detected target object;
step 2, finding the brightest and darkest images in the obtained color polarization images, and respectively corresponding to the maximum and minimum conditions of the back scattering light;
step 3, correcting the transmittance deviation caused by crosstalk of each channel by using a crosstalk elimination matrix of the transmittance among the RGB channels, and specifically comprising the following steps:
determining the transmission of RGB channels
The formula is as follows:
wherein (x, y) represents the coordinates of a pixel in the diagram,
and/or>
Representing the backscattered light at infinity in the parallel and vertical directions, I || (x,y)、I ⊥ (x, y) respectively representing the light intensity received by the detector in the parallel direction and the light intensity received by the detector in the vertical direction, and i represents the RGB channel number;
will be provided with
Multiplying a 3 x 3 matrix T to eliminate the crosstalk between channels to obtain a 3 x 3 crosstalk eliminating moment for optimizing the image restoration effectArray, the expression is as follows:
thereby obtaining RGB channel transmittance value after eliminating crosstalk
Accordingly, a light intensity graph after crosstalk elimination is calculated, and the expression is as follows:
and (3) optimizing the result by using an EME function as the criterion of the image, wherein the expression of the EME function is as follows:
where the image is divided into k 1 ×k 2 Small squares, (k, l) denote the position of each square on the picture,
and &>
Representing the maximum and minimum intensity values of the pixels in a block omega, positioned (k, l) in the image, setting q to a small constant value to prevent the divisor from being 0;
the final matrix result optimized by SCE is:
step 4, optimizing the matrix T opt Substituting into the crosstalk eliminating matrix formula and the crosstalk eliminating light intensity diagram formula, and calculating step by step to obtain the final crosstalk eliminating recoveryAnd finally, restoring the image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910660103.3A CN110458769B (en) | 2019-07-22 | 2019-07-22 | Color polarization image restoration method based on elimination of inter-channel crosstalk |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910660103.3A CN110458769B (en) | 2019-07-22 | 2019-07-22 | Color polarization image restoration method based on elimination of inter-channel crosstalk |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110458769A CN110458769A (en) | 2019-11-15 |
| CN110458769B true CN110458769B (en) | 2023-04-07 |
Family
ID=68481601
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910660103.3A Active CN110458769B (en) | 2019-07-22 | 2019-07-22 | Color polarization image restoration method based on elimination of inter-channel crosstalk |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110458769B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111429540B (en) * | 2020-04-22 | 2020-12-08 | 清华大学 | Temperature field and deformation field synchronous measurement device and method |
| CN112419179A (en) * | 2020-11-18 | 2021-02-26 | 北京字跳网络技术有限公司 | Method, device, equipment and computer readable medium for repairing image |
| CN114324185A (en) * | 2022-01-04 | 2022-04-12 | 浙江大学 | Underwater polarization detection device based on Stokes vector |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1630834A (en) * | 2002-06-24 | 2005-06-22 | 西尔弗布鲁克研究有限公司 | Photographic prints having magnetically recordable media |
| CN105486408A (en) * | 2015-11-26 | 2016-04-13 | 天津大学 | Three-wavelength Stokes vector polarization measurement system and method bases on color CCD |
| US9423602B1 (en) * | 2009-12-31 | 2016-08-23 | Gene Dolgoff | Practical stereoscopic 3-D television display system |
| CN107945141A (en) * | 2017-12-28 | 2018-04-20 | 天津大学 | A kind of polarization image restored method in scattering environments based on circularly polarized light |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI489416B (en) * | 2013-03-06 | 2015-06-21 | Novatek Microelectronics Corp | Image recovery method |
-
2019
- 2019-07-22 CN CN201910660103.3A patent/CN110458769B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1630834A (en) * | 2002-06-24 | 2005-06-22 | 西尔弗布鲁克研究有限公司 | Photographic prints having magnetically recordable media |
| US9423602B1 (en) * | 2009-12-31 | 2016-08-23 | Gene Dolgoff | Practical stereoscopic 3-D television display system |
| CN105486408A (en) * | 2015-11-26 | 2016-04-13 | 天津大学 | Three-wavelength Stokes vector polarization measurement system and method bases on color CCD |
| CN107945141A (en) * | 2017-12-28 | 2018-04-20 | 天津大学 | A kind of polarization image restored method in scattering environments based on circularly polarized light |
Non-Patent Citations (3)
| Title |
|---|
| Lei Kong等.R-OFDM for RGBA-LED-Based Visible Light Communication With Illumination Constraints.Journal of lightwave technology.2016,第第34卷卷(第第34卷期),5412-5422. * |
| 张晶等.一种减少空间调制快拍成像测偏仪伪信息的方法.物理学报.2018,第67卷(第4期),49-58. * |
| 王铁城 ; 姚晓天 ; 万木森 ; 刘铁根.偏振相关损耗对偏振复用系统信道正交性的影响.中国激光.2019,第36卷(第4期),879-883. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110458769A (en) | 2019-11-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110458769B (en) | Color polarization image restoration method based on elimination of inter-channel crosstalk | |
| Wang et al. | Underwater image restoration via maximum attenuation identification | |
| EP2278788B1 (en) | Method and apparatus for correcting lens shading | |
| EP2461576B1 (en) | Image processing apparatus and image processing program | |
| Namer et al. | Advanced visibility improvement based on polarization filtered images | |
| TWI516132B (en) | An image processing apparatus, an image processing method, and a program | |
| EP2464125B1 (en) | Image processing device | |
| EP3031202B1 (en) | Dynamic color shading correction | |
| CN110322410B (en) | Underwater image defogging and color cast correction method based on bright channel transmissivity compensation | |
| US20120257821A1 (en) | Image processing apparatus and image processing method, and program | |
| US8804013B2 (en) | Method of calculating lens shading compensation factor and method and apparatus for compensating for lens shading by using the method | |
| CN103209330B (en) | For the treatment of the method and apparatus of aberration and purple fringe | |
| CN102844788A (en) | Image processing apparatus and image pickup apparatus using the same | |
| CN106485681A (en) | Color image restoration method under water based on color correction and red channel prior | |
| CN102598682B (en) | Three-dimensional Imaging Device | |
| CN109949247A (en) | A kind of gradient field adaptive gain underwater picture Enhancement Method based on YIQ space optics imaging model | |
| CN111738941B (en) | Underwater image optimization method integrating light field and polarization information | |
| TW201308251A (en) | Underwater image enhancement system | |
| CN112488948B (en) | Underwater image restoration method based on black pixel point estimation back scattering | |
| CN107197152A (en) | Focusing method, device, computer-readable recording medium and mobile terminal | |
| Liu et al. | Polarimetric underwater image recovery for color image with crosstalk compensation | |
| TW201919385A (en) | Defective pixel compensation method and device | |
| CN103500457B (en) | A kind of method of video image color cast detection | |
| CN101510306B (en) | Estimation method for video image illumination distribution | |
| CN102946501A (en) | Color distortion correction method and device in imaging system or image output system |
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 |