CN115937021A - Polarization defogging method based on frequency domain feature separation and iterative optimization of atmospheric light - Google Patents
Polarization defogging method based on frequency domain feature separation and iterative optimization of atmospheric light Download PDFInfo
- Publication number
- CN115937021A CN115937021A CN202211393131.1A CN202211393131A CN115937021A CN 115937021 A CN115937021 A CN 115937021A CN 202211393131 A CN202211393131 A CN 202211393131A CN 115937021 A CN115937021 A CN 115937021A
- Authority
- CN
- China
- Prior art keywords
- image
- scattered light
- polarization
- atmospheric
- frequency image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
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
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Abstract
The invention discloses a polarization defogging method based on frequency domain feature separation and iterative optimization of atmospheric light, and relates to the technical field of image defogging. Acquiring a non-polarized image and a plurality of polarized images with different polarization angles in the same target scene; acquiring a low-frequency image of the polarization image through multi-scale transformation; obtaining an atmospheric scattered light image after median filtering of a low-frequency image; constructing Stokes vectors by using the atmospheric scattered light images to obtain the total light intensity of the atmospheric scattered light; constructing a physical model of foggy day imaging, and acquiring a reconstructed defogged image; performing NSP decomposition on the reconstructed defogged image to obtain a high-frequency image, and performing normalization processing; and constructing a constraint function between the total light intensity of the atmospheric scattered light and the normalized high-frequency image, and obtaining an optimized defogged image by iteratively optimizing the total light intensity of the atmospheric scattered light. The invention utilizes the high-frequency image iteration of the defogged image to optimize the atmospheric light, thereby improving the polarization defogging effect.
Description
Technical Field
The invention relates to the technical field of image defogging, in particular to a polarization defogging method based on frequency domain feature separation and iterative optimization of atmospheric light.
Background
As the adverse meteorological environment presents the characteristics of wide influence range, long duration, high concentration and the like, the work of the optical detection system is seriously influenced, the contrast of a target scene is reduced in haze weather, and the information extraction and processing of the acquired target scene image are difficult to perform by a common optical imaging means. The defogging technology for enhancing the contrast of target imaging in the haze weather, improving the image quality and improving the visibility has very important and long-term application value in various fields such as outdoor video monitoring, daily photo processing, aerial photography and underwater image processing, and the safety auxiliary driving systems of the existing automobiles and ships.
Through long-term development, three major image defogging methods based on image enhancement, physical model and deep learning appear in sequence. The image enhancement-based method has limited effect because the essential reason of image degradation is not considered, but the emerging deep learning-based methods are limited by a training set, and the robustness of the application in a real scene is still to be improved. Compared with the common imaging technology, the polarization imaging has objective advantages because additional spectral information can be obtained.
However, the existing polarization defogging method neglects the polarization characteristic of the target, so parameters such as a polarization degree correction factor and the like need to be manually regulated and controlled to obtain the optimal defogging effect under different haze concentrations, and only depending on the spatial domain information of a polarization image, it is difficult to accurately separate the atmospheric scattered light and the target reflected light.
Therefore, how to improve the stability and effectiveness of polarization defogging and improve the quality of defogged images is a problem to be solved by those skilled in the art.
Disclosure of Invention
In view of the above, the invention provides a polarization defogging method based on frequency domain feature separation and iterative optimization of atmospheric light, so as to enhance the contrast of target imaging in haze weather, improve visibility, and improve the quality of defogged images.
In order to achieve the purpose, the invention adopts the following technical scheme:
a polarization defogging method for separating and iteratively optimizing atmospheric light based on frequency domain features comprises the following steps:
acquiring a non-polarized image and a plurality of polarized images with different polarization angles in the same target scene;
acquiring a low-frequency image corresponding to each polarization image through multi-scale transformation;
performing median filtering on each low-frequency image to obtain an atmospheric scattered light image corresponding to each low-frequency image;
constructing Stokes vectors of atmospheric scattered light by using the atmospheric scattered light images corresponding to each low-frequency image, and calculating the total light intensity of the atmospheric scattered light according to the Stokes vectors of the atmospheric scattered light;
constructing a physical model of fog-day imaging according to the non-polarized image and the total light intensity of the atmospheric scattered light, and obtaining a reconstructed defogged image according to the physical model of the fog-day imaging;
performing NSP decomposition on the reconstructed defogged image to obtain a corresponding high-frequency image, and performing normalization processing on the high-frequency image to obtain a normalized high-frequency image;
and constructing a constraint function between the total light intensity of the atmospheric scattered light and the normalized high-frequency image, and obtaining an optimized defogged image by iteratively optimizing the total light intensity of the atmospheric scattered light.
Preferably, the plurality of polarization images with different polarization angles include polarization images I with three polarization angles of 0 °, 60 ° and 120 ° in the same target scene 0 、I 60 、I 120 。
Preferably, the obtaining of the low-frequency image corresponding to each polarization image through multi-scale transformation specifically includes:
and decomposing each polarized image by adopting NSP conversion, taking four decomposed layers, and acquiring a low-frequency image corresponding to each decomposed polarized image.
Preferably, the Stokes vector of the atmospheric scattered light is constructed by using the atmospheric scattered light image corresponding to each low-frequency image, and the specific expression includes:
in the formula, S A A Stokes vector representing atmospheric scattered light; s AI Represents the total light intensity of atmospheric scattered light; s AQ 、S AU Respectively representing the intensity difference of two linear polarized lights of atmospheric scattered light; a. The 0 、A 60 、A 120 And the atmosphere scattered light image is obtained by median filtering the low-frequency images respectively representing the polarization images of 0 degrees, 60 degrees and 120 degrees.
Preferably, a physical model of the foggy day imaging is constructed according to the total light intensity of the unpolarized image and the atmospheric scattered light, and a reconstructed defogged image is obtained according to the physical model of the foggy day imaging, which specifically comprises:
wherein J represents the reconstructed defogged image, I represents the non-polarized image under the target scene, and S AI Represents the total intensity of atmospheric scattered light, A ∞ Representing the intensity of atmospheric scattered light at infinity.
Preferentially, the normalization formula for performing normalization processing on the high-frequency image specifically includes:
h represents a normalized high-frequency image; norm () represents a normalization function; h' (i) represents one pixel point i in the high-frequency image; min (H') represents the minimum value of pixel points in the high-frequency image; max (H') represents the maximum value of the pixel points in the high frequency image.
Preferably, a constraint function between the total intensity of the atmospheric scattered light and the normalized high-frequency image is constructed:
in the formula, H represents a normalized high-frequency image, S AI Represents the total light intensity of atmospheric scattered light, omega represents the edge area of the image of the total light intensity of atmospheric scattered light, x 1 And y 1 Respectively representing two pixels in the border region, W (S) AI H) is a weighting function;
using the high-frequency image H after the ith normalization i Redefining the total intensity (S) of the atmospheric scattered light at the (i + 1) th time AI ) i+1 ：
In the formula, l represents Lagrange multiplier, and the total light intensity (S) of atmospheric scattered light is solved by using a gradient descent method AI ) i+1 The convergence condition is | (S) AI ) i+1 -(S AI ) i And | is less than epsilon, and epsilon represents a convergence threshold.
Preferably, the weight function expression is:
in the formula, sigma represents the variance of pixel points in the edge region of the normalized high-frequency image H, x 2 And y 2 And representing two pixel points in the edge region of the normalized high-frequency image H.
According to the technical scheme, the invention discloses a polarization defogging method for optimizing atmospheric scattering light based on frequency domain feature separation and iteration, which fully utilizes frequency domain information to optimize the process of polarization defogging parameter estimation, and has the following beneficial effects compared with the prior art, wherein the process does not need to ignore target polarization characteristics but avoids specific discussion on the target polarization characteristics:
(1) The invention utilizes the high-frequency image of the defogged image to iteratively optimize the total light intensity of the atmospheric scattered light, improves the polarization defogging effect, improves the image quality and is convenient for various complex computer vision systems to process the image in the haze weather.
(2) The method utilizes the low-frequency image of the polarization image to obtain the total light intensity of the atmospheric scattered light, and effectively avoids the generation of a halo effect compared with the conventional method for directly obtaining the total light intensity of the atmospheric scattered light by using spatial filtering;
(3) The invention realizes polarization defogging by directly constructing the Stokes parameters of atmospheric scattered light, thereby avoiding discussing complicated target polarization characteristics.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
FIG. 1 is a schematic flow chart of a method provided by an embodiment of the present invention;
fig. 2 is a schematic diagram of an NSP decomposition principle provided by an embodiment of the present invention;
fig. 3 is a flow chart of a method provided in the embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The embodiment of the invention discloses a polarization defogging method based on frequency domain feature separation and iterative optimization of atmospheric light, which comprises the following steps as shown in figures 1-3:
step 1: the polarization imaging technology is utilized to image a target scene to acquire three polarization images in the same scene, and the used polarization camera (such as a split-focus plane type polarization camera) is required to acquire three polarization states of 0 degree, 60 degrees and 120 degreesFog polarization image, marked as I 0 、I 60 、I 120 And can simultaneously acquire the non-polarization image I.
Step 2: obtaining a low-frequency image of a polarization image through multi-scale transformation, wherein the required multi-scale transformation method is non-subsampled pyramid transformation (NSP), taking j =4 decomposition layers, and transforming the I 0 、I 60 、I 120 The low frequency part of the acquired polarization image is noted
And step 3: then, the low-frequency image is filtered in the spatial domain to obtain three atmospheric scattered light images in the polarization state
The atmospheric scattered light image obtained by median filtering is recorded as A 0 ,A 60 ,A 120 。
Here we obey the objective constraints of atmospheric scattered light, including:
(1) is not negative. Namely, each pixel of the atmosphere scattered light image is greater than or equal to 0.
(2) The total light intensity of the atmospheric scattered light must not exceed the total light intensity received by the camera. I.e. the atmospheric scattered light image must not exceed the pixel values in the corresponding polarized image at every point.
(3) Local consistency. The atmospheric scattered light images are approximately equal where the depth of field is close.
And 4, step 4: constructing Stokes vectors of atmospheric scattered light by using three atmospheric scattered light images, and estimating an atmospheric light value A at infinity according to a definitional expression of atmospheric scattered light polarization components ∞ . The expression of the constructed Stokes vector is:
wherein S is A A Stokes vector representing atmospheric scattered light; s. the AI Represents the total light intensity of atmospheric scattered light; s. the AQ 、S AU Two linear polarizations each representing atmospheric scattered lightLight intensity difference; a. The 0 、A 60 、A 120 And the atmosphere scattered light image is obtained by median filtering the low-frequency images respectively representing the polarization images of 0 degrees, 60 degrees and 120 degrees.
The median filtering method is a nonlinear smoothing technology, and sets the intensity value of each pixel point of the image as the median of all the intensity values of the pixel points in a certain neighborhood window of the point. Specifically, a moving region (n × n) of odd-numbered points is used, and the value of the center point of the region is replaced by the median value of the points in the window. However, the specific area size varies according to the size of the polarization image to be captured, and therefore, the size of the median filtering window is selected to be 199 × 199 in this example.
Calculating the polarization angle theta of atmospheric scattered light by using Stokes vector of atmospheric light A And degree of polarization P A ：
Atmospheric light polarization component A P Is defined as:
to obtain A ∞ Finding out the pixel value with the ratio of 0.95 or more in the matrix and the non-polarized image I, and averaging to obtain the atmospheric light intensity A at infinity ∞ 。
And 5: constructing a physical model of foggy day imaging according to the total light intensity of the non-polarized image and the atmospheric scattered light to obtain a reconstructed defogged image J:
and 6: performing NSP decomposition on the reconstructed defogged image J to obtain a high-frequency image, and recording the high-frequency image as H after normalization, wherein the normalization formula is as follows:
h represents a normalized high-frequency image; norm () represents a normalization function; h' (i) represents a pixel point i in the high-frequency image; min (H') represents the minimum value of pixel points in the high-frequency image; max (H') represents the maximum value of the pixel points in the high-frequency image, and each pixel point of each high-frequency image is converted into a value in [0,1] after normalization processing.
And 7: and constructing a constraint function between the total light intensity of the atmospheric scattered light and the normalized high-frequency image, and obtaining a more accurate reconstructed image by iteratively optimizing the total light intensity of the atmospheric scattered light until an iteration result is converged, so as to obtain a final reconstructed image. The constraint function between the normalized high-frequency image H and the total light intensity of the atmospheric scattered light is as follows:
wherein omega is a certain edge area of the total light intensity image of the atmospheric scattered light, x 1 And y 1 And two pixel points in the edge region of the total light intensity of the atmospheric scattered light are represented. W (S) AI And H) is that the weight function is a weight function, and the expression is as follows:
sigma represents the variance of pixel points in the edge region of the normalized high-frequency image H, x 2 And y 2 And expressing two pixel points in the edge region of the normalized high-frequency image H.
Using the high-frequency image H after the ith normalization i Redefining the total light intensity (S) of the (i + 1) th atmospheric scattered light AI ) i+1 ：
l represents Lagrange multiplier, and the total light intensity (S) of the atmospheric scattered light of the (i + 1) th time can be solved by using a gradient descent method AI ) i+1 The convergence condition is | (S) AI ) i+1 -(S AI ) i And | < epsilon. Here, epsilon represents the set convergence threshold.
In the invention, the edge region division can be detected by some common edge detection algorithms, but the specific region size does not meet the uniform requirement of the invention according to the size of the shot polarization image, and the size of omega in the example is 200 multiplied by 200.
Total light intensity S of atmospheric scattered light in the present invention AI The size of the represented image is completely consistent with that of the normalized high-frequency image H, the coordinate positions of the selected edge region omega relative to the two images are the same, and only the pixel amplitude value and S of the coordinate point in H are the same AI Are different.
In the embodiment of the invention, the atmosphere scattering light is also referred to as the atmosphere light for short, all processing objects in the invention are matrixes, and the matrixes are processed and calculated, such as polarized images, non-polarized images, high-frequency images, low-frequency images, total light intensity of the atmosphere scattering light and the like.
Once the total light intensity S of the atmospheric light AI Convergence, meaning S AI And the defogged image obtained by using the formula (11) does not change too much with the iteration, so that the atmospheric light intensity convergence and the defogged image convergence are equivalent, and the obtained defogged image is the final optimized defogged image.
In the embodiment of the present invention, the polarization angles of the three polarization images are 0 °, 60 °, and 120 °, and in other embodiments, the polarization angles of the three polarization images may also be other polarization angles, for example, the polarization angles of the three polarization images are respectively 0 °, 45 °, and 90 °, so that corresponding parameters in the equation for calculating the Stokes vector of the atmospheric scattering light in step 4 are also changed accordingly. Theoretically, the polarization angles of the three polarization images can be arbitrarily selected from three non-repetitive polarization angles in the range of [0,180], but the three polarization angles are recommended to be selected in the range of directly calculating the angle value of the trigonometric function, such as 0 degrees, 45 degrees, 60 degrees, 90 degrees, 120 degrees and 135 degrees, and the polarization images of three of the angles can be simultaneously acquired by adopting the split-focal-plane type polarization camera.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (8)
1. A polarization defogging method for separating and iteratively optimizing atmospheric light based on frequency domain characteristics, the method comprising the steps of:
acquiring a non-polarized image and a plurality of polarized images with different polarization angles in the same target scene;
obtaining a low-frequency image corresponding to each polarization image through multi-scale transformation;
performing median filtering on each low-frequency image to obtain an atmosphere scattered light image corresponding to each low-frequency image;
constructing a Stokes vector of atmospheric scattered light by using the atmospheric scattered light image corresponding to each low-frequency image, and calculating the total light intensity of the atmospheric scattered light according to the Stokes vector of the atmospheric scattered light;
constructing a physical model of fog day imaging according to the total light intensity of the non-polarized image and the atmospheric scattered light, and obtaining a reconstructed defogged image according to the physical model of the fog day imaging;
performing NSP decomposition on the reconstructed defogged image to obtain a corresponding high-frequency image, and performing normalization processing on the high-frequency image to obtain a normalized high-frequency image;
and constructing a constraint function between the total light intensity of the atmospheric scattered light and the normalized high-frequency image, and obtaining an optimized defogged image by iteratively optimizing the total light intensity of the atmospheric scattered light.
2. The polarization defogging method according to claim 1, wherein the plurality of polarization images with different polarization angles comprise polarization images with three polarization angles of 0 °, 60 ° and 120 ° in the same target scene.
3. The polarization defogging method according to claim 1, wherein the obtaining of the corresponding low frequency image of each polarization image through multi-scale transformation specifically comprises:
and decomposing each polarized image by adopting NSP conversion, taking four decomposed layers, and acquiring a low-frequency image corresponding to each decomposed polarized image.
4. The polarization defogging method according to claim 2, wherein a Stokes vector of the atmospheric scattered light is constructed by using the atmospheric scattered light image corresponding to each low-frequency image, and the specific expression comprises:
in the formula, S A A Stokes vector representing atmospheric scattered light; s AI Represents the total light intensity of atmospheric scattered light; s. the AQ 、S AU Respectively representing the intensity difference of two linear polarized lights of atmospheric scattered light; a. The 0 、A 60 、A 120 And the low-frequency images respectively represent the polarized images of 0 degrees, 60 degrees and 120 degrees and are subjected to median filtering to obtain atmosphere scattered light images.
5. The polarization defogging method according to claim 1, wherein a physical model of fog-day imaging is constructed according to the total light intensity of the non-polarized image and the atmosphere scattered light, and a reconstructed defogged image is obtained according to the physical model of fog-day imaging, and the method specifically comprises the following steps:
wherein J represents the reconstructed defogged image, I represents the non-polarized image under the target scene, and S AI Represents the total intensity of atmospheric scattered light, A ∞ Representing the intensity of atmospheric scattered light at infinity.
6. The polarization defogging method according to claim 1, wherein the normalization formula for normalizing the high frequency image specifically comprises:
h represents a normalized high-frequency image; norm () represents a normalization function; h' (i) represents one pixel point i in the high-frequency image; min (H') represents the minimum value of pixel points in the high-frequency image; max (H') represents the maximum value of the pixel points in the high frequency image.
7. The polarization defogging method according to claim 1, wherein a constraint function between the atmospheric scattered light and the normalized high-frequency image is constructed, and the optimized defogged image is obtained by iteratively optimizing the atmospheric scattered light, specifically comprising:
constructing a constraint function between the total light intensity of the atmospheric scattered light and the normalized high-frequency image:
in the formula, H represents a normalized high-frequency image, S AI Represents the total light intensity of atmospheric scattered light, and omega represents the total light intensity of atmospheric scattered lightImage edge region, x 1 And y 1 Respectively represent two pixel points in the edge region omega, W (S) AI H) is a weighting function;
using the high-frequency image H after the ith normalization i Redefining the total intensity (S) of the atmospheric scattered light at the (i + 1) th time AI ) i+1 ：
In the formula, l represents Lagrange multiplier, and the total light intensity (S) of atmospheric scattered light is solved by using a gradient descent method AI ) i+1 The convergence condition is | (S) AI ) i+1 -(S AI ) i And | | < ε, which represents the convergence threshold.
8. The polarized defogging method according to claim 7, wherein the weight function expression is:
in the formula, sigma represents the variance of pixel points in the edge region of the normalized high-frequency image H, x 2 And y 2 And expressing two pixel points in the edge region of the normalized high-frequency image H.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211393131.1A CN115937021A (en) | 2022-11-08 | 2022-11-08 | Polarization defogging method based on frequency domain feature separation and iterative optimization of atmospheric light |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211393131.1A CN115937021A (en) | 2022-11-08 | 2022-11-08 | Polarization defogging method based on frequency domain feature separation and iterative optimization of atmospheric light |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115937021A true CN115937021A (en) | 2023-04-07 |
Family
ID=86698473
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211393131.1A Pending CN115937021A (en) | 2022-11-08 | 2022-11-08 | Polarization defogging method based on frequency domain feature separation and iterative optimization of atmospheric light |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115937021A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116337088A (en) * | 2023-05-30 | 2023-06-27 | 中国人民解放军国防科技大学 | Foggy scene relative motion estimation method and device based on bionic polarization vision |
-
2022
- 2022-11-08 CN CN202211393131.1A patent/CN115937021A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116337088A (en) * | 2023-05-30 | 2023-06-27 | 中国人民解放军国防科技大学 | Foggy scene relative motion estimation method and device based on bionic polarization vision |
| CN116337088B (en) * | 2023-05-30 | 2023-08-11 | 中国人民解放军国防科技大学 | Foggy scene relative motion estimation method and device based on bionic polarization vision |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chen et al. | Visual depth guided color image rain streaks removal using sparse coding | |
| CN104683767B (en) | Penetrating Fog image generating method and device | |
| Li et al. | Single image dehazing using the change of detail prior | |
| TWI808406B (en) | Image dehazing method and image dehazing apparatus using the same | |
| Krishnan et al. | A survey on different edge detection techniques for image segmentation | |
| CN111539246B (en) | Cross-spectrum face recognition method and device, electronic equipment and storage medium thereof | |
| CN111861896A (en) | UUV-oriented underwater image color compensation and recovery method | |
| Bansal et al. | A review of image restoration based image defogging algorithms | |
| Farhadifard et al. | Learning-based underwater image enhancement with adaptive color mapping | |
| John et al. | Enhancement of weather degraded video sequences using wavelet fusion | |
| CN115937021A (en) | Polarization defogging method based on frequency domain feature separation and iterative optimization of atmospheric light | |
| CN113487509B (en) | Remote sensing image fog removal method based on pixel clustering and transmissivity fusion | |
| Wu et al. | Image haze removal: Status, challenges and prospects | |
| Chen et al. | Visual depth guided image rain streaks removal via sparse coding | |
| Fu et al. | An anisotropic Gaussian filtering model for image de-hazing | |
| Babu et al. | An efficient image dahazing using Googlenet based convolution neural networks | |
| CN116823694B (en) | Infrared and visible light image fusion method and system based on multi-focus information integration | |
| Ajith et al. | Dark Channel Prior based Single Image Dehazing of Daylight Captures | |
| CN110009575B (en) | Infrared image stripe noise suppression method based on sparse representation | |
| Raigonda et al. | Haze Removal Of Underwater Images Using Fusion Technique | |
| Wang et al. | Rapid nighttime haze removal with color-gray layer decomposition | |
| CN114140361A (en) | Generation type anti-network image defogging method fusing multi-stage features | |
| Liu et al. | Research on lane line segmentation algorithm based on deeplabv3 | |
| CN109767399B (en) | Underwater image enhancement method based on unsupervised color correction | |
| Chen et al. | GADO-Net: an improved AOD-Net single image dehazing algorithm |
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 |