CN104794697B - A kind of image defogging method based on dark primary priori - Google Patents
A kind of image defogging method based on dark primary priori Download PDFInfo
- Publication number
- CN104794697B CN104794697B CN201510224034.3A CN201510224034A CN104794697B CN 104794697 B CN104794697 B CN 104794697B CN 201510224034 A CN201510224034 A CN 201510224034A CN 104794697 B CN104794697 B CN 104794697B
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- image
- sky
- munder
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 21
- 238000001914 filtration Methods 0.000 claims abstract description 51
- 238000002834 transmittance Methods 0.000 claims abstract description 30
- 239000003086 colorant Substances 0.000 claims description 4
- 230000000903 blocking effect Effects 0.000 abstract description 2
- 230000011218 segmentation Effects 0.000 abstract description 2
- 238000005260 corrosion Methods 0.000 abstract 1
- 230000007797 corrosion Effects 0.000 abstract 1
- 230000000694 effects Effects 0.000 description 25
- 125000001475 halogen functional group Chemical group 0.000 description 11
- 238000003709 image segmentation Methods 0.000 description 6
- 239000002245 particle Substances 0.000 description 5
- 206010027146 Melanoderma Diseases 0.000 description 4
- 238000006731 degradation reaction Methods 0.000 description 3
- 230000002146 bilateral effect Effects 0.000 description 2
- 230000015556 catabolic process Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000014759 maintenance of location Effects 0.000 description 2
- 238000007670 refining Methods 0.000 description 2
- 206010039203 Road traffic accident Diseases 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 230000003628 erosive effect Effects 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 230000003313 weakening effect Effects 0.000 description 1
Landscapes
- Image Processing (AREA)
Abstract
The invention discloses a kind of image defogging method based on dark primary priori.Including following steps：Step 1：Segmentation is carried out to image I (x) based on Classic Clustering Algorithms and obtains candidate sky areas It(x)；Step 2：To candidate sky areas It(x) corrosion treatment obtains sky brightness；Step 3：According to the sky brightness of acquisition, the mini-value filtering being improved to image, rough estimate transmittance figure picture is obtained；Step 4：Rough estimate transmittance figure picture, the transmittance figure picture optimized are optimized by Steerable filter；Step 5：Restored image is obtained based on atmospherical scattering model.The present invention has been done to atmosphere light more reasonably to be handled, and is weakened image blocking effect, is improved the overall brightness of image, make image more natural.
Description
Technical Field
The invention belongs to the field of image processing, and particularly relates to an image defogging method based on dark channel prior, which can perform defogging processing on an image.
Background
Recently, haze weather has become a frequent weather phenomenon. Liquid drops and solid particles in haze are harmful to human health, atmospheric visibility is reduced due to the scattering effect of a large number of suspended particles, image quality is reduced, related information in an obtained image is influenced, and unnecessary traffic accidents are caused due to reduction of outdoor definition, so that an image defogging technology becomes one of the subjects which need to be researched in the field of image processing and computer vision.
Image defogging technologies mainly include two types, image restoration based and image enhancement based. The algorithm based on image restoration starts from a physical model of image degradation, and obtains relevant parameters of each degradation link by analyzing and solving the inverse process of the image degradation process, thereby restoring a clear image which is as vivid as possible. Recently, defogging algorithms based on monochromatic atmospheric scattering models have made great progress on the premise of some a priori knowledge and assumptions.
At present, the common image defogging algorithm is to estimate the atmospheric transmittance and then restore the scene color according to a monochromatic atmospheric scattering model.
Tan assumes that the ambient light of a local area is a constant and the contrast is remarkably enhanced, under the framework of a Markov random field model, a cost function related to the edge intensity is constructed, and the optimal illumination is estimated by using a graph segmentation theory, but the excessive saturation distortion of the image color caused by the optimal illumination cannot be avoided, and the serious Halo effect is generated in the boundary area with sudden depth of field.
Fattal assumes that the transmittance of a local region of an image is a constant vector, and the chromaticity of the surface of an object has statistical irrelevance with medium propagation, the algorithm needs a large amount of physical color information, but the image under the dense fog condition loses a large amount of color information, and the transmittance estimation deviation of the image is large at this time, and the algorithm fails.
Hokeming et al propose an algorithm for dark primaries priors. Assuming that the scene albedo in a local region of at least one color channel tends to zero, then a medium propagation function is roughly estimated by minimum value filtering, and then the function is refined by a soft matting algorithm, thereby obtaining a defogging image. However, the refinement algorithm is essentially a solution problem of a large-scale sparse linear equation set and has high time and space complexity; in addition, the image matting introduces an alpha channel to soften the edges of the transition areas of the foreground and the background, and the medium propagation function is an exponential attenuation factor of scene radiation, so that the soft matting algorithm is not reasonable for refining the medium propagation function; meanwhile, single minimum filtering can generate Halo effect and blocking effect.
Then, an algorithm which respectively adopts bilateral filtering and median filtering to replace minimum filtering is adopted so as to improve defogging performance. Gibson et al propose that median filtering replaces minimum filtering, Halo effect can be weakened, soft matting or bilateral filtering is not needed for thinning a transmittance diagram, time complexity of operation is remarkably reduced, and the defogged image quality is poor and black spot effect is easy to occur. The Zhang Xiao just proposes double-region filtering, defines a dark region, improves a median filtering algorithm, and weakens the black spot effect, but the algorithm has limited effect on keeping image details and reduces the fineness of the image details.
The above documents all take atmospheric light as a constant, and in images with varying depth of field, this assumption is obviously unreasonable; in addition, algorithms based on dark channel prior have dark defogged images and Halo effect, or complicated operation is adopted for eliminating the Halo effect, so that the time complexity of the algorithms is increased.
Disclosure of Invention
The invention aims to provide an image defogging method which is good in defogging effect, low in processing time complexity and based on dark channel prior.
The invention is realized by the following technical scheme:
an image defogging method based on dark channel prior comprises the following steps,
the method comprises the following steps: image I (x) is segmented based on classical clustering algorithm to obtain candidate sky area It(x)；
Step two: for candidate sky region It(x) Corroding to obtain sky brightness;
step three: according to the obtained sky brightness, carrying out improved minimum value filtering on the image to obtain a rough estimation transmissivity image;
step four: optimizing and roughly estimating a transmissivity image through guide filtering to obtain an optimized transmissivity image;
step five: and obtaining a restored image based on the atmospheric scattering model.
The invention relates to an image defogging method based on dark channel prior, which can also comprise the following steps:
1. the sky brightness is:
wherein A (x) is sky brightness, Ic(x) Is the R, G, B three primary color channel of the image I (x), Amax＝max(max(Itmin(x) ) is the sky brightness maximum, A)min＝max(min(max(Itmin(x))),0.7*Amax) Is the minimum value of sky brightness, Itmin(x) As a result of the filtering of the minimum value,
wherein x is a spatial coordinate, It(x) Is a candidate sky area and is a candidate sky area,is It(x) The R, G, B channels, Ω (x) are square areas centered around the pixel point x.
2. The minimum filtering for the improvement of the image takes the function:
wherein, Idarkm(x) Is the median value of the dark primary color, Iord(x) For coarse estimation of the transmittance image, Ic(y) is the R, G, B channel of image I (x), parameter k ∈ [0.8, 1).
3. Optimized transmittance image qiComprises the following steps:
wherein, I is a guide, i.e. image I (x) is abbreviated as I, p is an input image, omega is a square area with a pixel point k as the center, w is the pixel number of the area, IiAnd piThe values of I and P at I within Ω, μ and σ, respectively2Represents the mean and variance of I in the region, as a constraint of akThe coefficient of (a) is determined,representing the mean value for a local region of the input image.
4. The restored image is:
wherein t (x) e-βd(x)D is the depth of field of the scene, β is the atmospheric scattering coefficient,is an estimate of t (x),sky brightness a (x) is abbreviated as a.
Has the advantages that:
the invention provides an improved algorithm based on a dark channel prior defogging physical model, which is used for more reasonably processing atmospheric light, weakening the image block effect, improving the overall brightness of an image and enabling the image to be more natural. The method estimates the sky brightness based on image segmentation, so that the sky brightness estimated by directly using the brightest pixel value is prevented from being easily influenced by high brightness noise or white objects; the transmittance graph is obtained by adopting improved minimum filtering, the Halo effect and the block effect generated by independently using the minimum filtering are inhibited, and the defect of low image edge retention of the median filtering in the double-region filtering is overcome; and finally, the effective guide filtering is used for thinning the transmissivity graph, so that the time complexity and the space complexity of the algorithm are greatly reduced.
Drawings
FIG. 1 is a flow chart of the algorithm principle of the present patent.
Fig. 2a is the original image 1, fig. 2b is a candidate sky region of the original image 1, fig. 2c is the original image 2, fig. 2d is a candidate sky region of the original image 2, fig. 2e is the original image 3, and fig. 2f is a candidate sky region of the original image 3.
Fig. 3a is an original image, fig. 3b is a transmittance graph of median filtering, fig. 3c is a transmittance graph of He, fig. 3d is a transmittance graph of two-region filtering, and fig. 3e is a transmittance graph of the present invention.
Fig. 4a shows an original image 4, fig. 4b shows a restored image of the original image 4 obtained by the method of the present invention, fig. 4c shows an original image 5, and fig. 4d shows a restored image of the original image 5 obtained by the method of the present invention.
Detailed Description
The present invention will be described in further detail below with reference to the accompanying drawings.
The invention provides an effective defogging algorithm aiming at a haze image, which comprises the following steps:
(1) and estimating the sky brightness based on image segmentation of classical clustering.
(2) And acquiring a transmittance image of the haze image by using the improved minimum value filtering.
(3) The transmittance map is refined using guided filtering.
(4) And obtaining a restored image.
The image segmentation method of the classical clustering comprises the following steps:
(1) converting the haze image from an RGB space to an LAB space;
(2) performing K-means classical clustering image segmentation based on A, B components;
(3) in order to prevent the sky area from being estimated wrongly, the pixel points with the lower positions are set to be zero.
In combination with the sequential statistical filtering, the dark regions are corrected. Sequential statistical filtering is adopted in a dark area, so that the Halo effect caused by minimum filtering is weakened, and the contrast of an image is enhanced; in the non-dark area, minimum value filtering is adopted to better keep the edge of the image, and the defect of median filtering is overcome.
The specific steps in the present invention will be described in detail with reference to fig. 1.
(1) Estimating sky brightness
The sky area has the following characteristics: 1) the brightness is high; 2) the position is deviated upwards; 3) the colors are similar.
Firstly, converting a known image I (x) from an RGB space to an LAB space, and carrying out image segmentation by using a classical clustering method, wherein the image segmentation is marked as It(x) See fig. 2 a-2 f. Minimum value filtering, i.e. grey scale erosion operation, of candidate sky regions
Wherein x is a spatial coordinate, It(x) Is a candidate sky area and is a candidate sky area,is It(x) The channel of R, G, B three primary colors, omega (x) is a square area with a pixel point x as the center, Itmin(x) Is the result of the minimum filtering.
Taking the maximum value of the candidate sky area as the maximum value of brightness
Amax＝max(max(Itmin(x))) (2)
To prevent misestimation of sky regions, 0.7 × a is takenmaxObtaining the minimum value of sky brightness as the bottom limit of the minimum value of sky brightness
Amin＝max(min(max(Itmin(x))),0.7*Amax) (3)
The brightness of the image is compared with the brightness of the image by analyzing a large amount of image dataThere is some approximately linear relationship, so the sky brightness is defined as follows
In the formula (II)In time, the sky area or the area with high brightness near the sky area is mostly used as a fixed value to prevent the over enhancement of the image; a (x) is sky brightness; i isc(x) R, G and B three primary color channels of I (x).
For convenience of writing, A (x) is hereinafter referred to as A.
(2) Estimating a transmittance map
Gibson et al revises the dark channel prior function using median filtering as follows
Wherein, Idarkm(x) Referred to as dark primary median value, IcAnd (y) is the R, G and B three primary color channels of I (x). The transmittance graphs obtained by this method are shown in fig. 3a to 3 b.
Zhang Xiao just proposes a dark region, and defines the median function as follows
The above formula shows that when a pixel point is located in a dark area, the median value of the dark primary colors directly takes the minimum value of the three primary color channel of the pixel point; otherwise, taking the value between the median and the minimum of the three primary color channels. The transmittance profile obtained by this method is shown in fig. 3 c.
The transmittance plot for He is shown in fig. 3 d.
The following conclusions can be drawn from the figure: the He minimum value filtering result has the best effect on detail maintenance, but has the largest operation amount and the Halo effect; the median filtering of Gibson has the most serious damage to details, has a black spot effect, has the lowest time complexity, has no obvious Halo effect, and is a good rapid defogging algorithm; the overall effect of the dual region filtering is centered. Similar conclusions hold after a large number of image analyses.
And comprehensively considering the action and effect of the minimum value filtering and the median filtering. This document, in combination with sequential statistical filtering, corrects for dark regions. Sequential statistical filtering is adopted in a dark area, so that the Halo effect can be weakened, and the contrast of an image can be enhanced; in non-dark areas, the edge of the image can be better preserved by adopting minimum value filtering. The algorithm function is defined as follows
The above formula indicates that when a pixel is located in a dark region, minimum filtering is applied to the pixel; otherwise, adopting sequential statistical filtering, and taking the value between the median and the minimum of the three primary color channels in the field as a result; wherein k belongs to [0.8,1 ]), a more satisfactory result can be obtained, and the value is taken as 0.95; ord is the order statistical filter function.
(3) Refining transmittance images
Since the transmittance is only a function of the depth of field d (x), a region smoothing operation is performed on the coarse estimate of transmittance, preserving the edge details of the abrupt change in depth of field, which can be seen as a filtering problem. He adopts soft matting to refine the transmittance map, and although the effect is good, the algorithm has very high time and space complexity. Here, with the guide filtering, not only a fine transmittance map can be obtained, but also the time complexity of the algorithm can be greatly reduced.
In the formula, qiFor the filtered image, I is the guide, i.e. the known image, p is the input image, Ω is the square area with pixel point k as the center, w is the number of pixels in the area, IiAnd piThe values of I and P at I within Ω, μ and σ, respectively2Represents the mean and variance of I in the region, as a constraint of akThe coefficient of (a);representing the mean value for a local region of the input image.
The transmittance map after filtering is shown in fig. 3 e. It can be seen that the transmittance graph herein not only has better detail retention, but also suppresses Halo effect, and has no black spot effect caused by median filtering, and obtains better visual effect.
(4) Restoring an image
The monochromatic atmosphere scattering model under the fog and haze weather conditions, namely the gray value I (x) of the image shot by the narrow-band camera can be expressed as
I(x)＝J(x)t(x)+A(1-t(x)) (12)
In the formula, I is a known image, J is the intensity of scene light in the absence of fog, A is sky brightness, t is medium transmittance, and an atmospheric scattering model consists of two terms. The first term represents an attenuation model, also known as direct propagation or direct attenuation. Due to the scattering effect of the atmospheric particles, a part of the reflected light on the surface of the object is lost due to scattering, and the part which is not scattered directly reaches the imaging sensor, and the light intensity of the part exponentially decays along with the increase of the propagation distance. The second term represents the ambient light model because scattering of natural light by atmospheric particles causes the atmosphere to behave as a light source. The intensity of the ambient light gradually increases with increasing propagation distance. Narasiman et al indicate that the assumptions of the model are that single scattering, a homogeneous atmospheric medium, and that the scattering coefficient of atmospheric particles for visible light is independent of wavelength, and therefore, the model is not suitable for attenuation compensation for imaging scenes many kilometers away.
For the sake of simplicity (12), t (x) is defined as
t(x)＝e-βd(x)(13)
In the formula, d is the depth of field of the scene, and beta is the atmospheric scattering coefficient.
He proposes a dark channel prior based on statistics of a large number of outdoor haze-free images. In some non-sky regions of the image, at least one of the three primary color channels R, G, B has a low intensity value and can be expressed as
In the formula, Jc(y) is the R, G, B channel of J (x), and Ω (x) is a square area with pixel point x as the center. The above formula is the dark channel prior formula.
For non-sky regions, the following holds
Jdark(x)——→0 (15)
Is represented by the formula (12)
In the formula AcThe sky brightness obtained for this patent algorithm.
Minimum value operation of equation (16)
Wherein,is an estimate of t (x), whereSubstituting equation (8) with equation (15) while substituting equation (8) into the transmission image obtained by the minimum value filtering resulting in the improvement of the present invention
To make the defogged image more realistic, the formula (18) is defined as
Take ω to 0.95.
Substituting formula (19) into formula (12), and setting minimum value t to ensure that denominator is not zero00.1 to obtain
This results in a restored image j (x), see fig. 4a to 4 d.
Claims (4)
1. An image defogging method based on dark channel prior is characterized by comprising the following steps: comprises the following steps of (a) carrying out,
the method comprises the following steps: image I (x) is segmented based on classical clustering algorithm to obtain candidate sky area It(x)；
Step two: for candidate sky region It(x) Corroding to obtain sky brightness;
step three: according to the obtained sky brightness, carrying out improved minimum value filtering on the image to obtain a rough estimation transmissivity image;
step four: optimizing and roughly estimating a transmissivity image through guide filtering to obtain an optimized transmissivity image;
step five: obtaining a restored image based on the atmospheric scattering model;
the sky brightness is:
wherein A (x) is sky brightness, Ic(x) Is the R, G, B three primary color channel of the image I (x), Amax＝max(max(Itmin(x) ) is the sky brightness maximum, A)min＝max(min(max(Itmin(x))),0.7*Amax) Is the minimum value of sky brightness, Itmin(x) As a result of the filtering of the minimum value,
<mrow> <msub> <mi>I</mi> <mrow> <mi>t</mi> <mi>min</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>x</mi> <mo>)</mo> </mrow> <mo>=</mo> <munder> <mi>min</mi> <mrow> <mi>y</mi> <mo>&Element;</mo> <mi>&Omega;</mi> <mrow> <mo>(</mo> <mi>x</mi> <mo>)</mo> </mrow> </mrow> </munder> <mrow> <mo>(</mo> <munder> <mi>min</mi> <mrow> <mi>c</mi> <mo>&Element;</mo> <mo>{</mo> <mi>r</mi> <mo>,</mo> <mi>g</mi> <mo>,</mo> <mi>b</mi> <mo>}</mo> </mrow> </munder> <msup> <msub> <mi>I</mi> <mi>t</mi> </msub> <mi>c</mi> </msup> <mo>(</mo> <mi>y</mi> <mo>)</mo> <mo>)</mo> </mrow> </mrow>
wherein x is a spatial coordinate, It(x) As a candidate sky region, It c(y) is It(x) The channel of R, G, B three primary colors, omega (x) is the center of the pixel point xThe square area of the heart.
2. The image defogging method based on the dark channel prior as recited in claim 1, wherein: the minimum value filtering for improving the image adopts the following functions:
<mrow> <msub> <mi>I</mi> <mrow> <mi>d</mi> <mi>a</mi> <mi>r</mi> <mi>k</mi> <mi>m</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>x</mi> <mo>)</mo> </mrow> <mo>=</mo> <munder> <mrow> <mi>o</mi> <mi>r</mi> <mi>d</mi> </mrow> <mrow> <mi>y</mi> <mo>&Element;</mo> <mi>&Omega;</mi> <mrow> <mo>(</mo> <mi>x</mi> <mo>)</mo> </mrow> </mrow> </munder> <mrow> <mo>(</mo> <munder> <mi>min</mi> <mrow> <mi>c</mi> <mo>&Element;</mo> <mo>{</mo> <mi>r</mi> <mo>,</mo> <mi>g</mi> <mo>,</mo> <mi>b</mi> <mo>}</mo> </mrow> </munder> <msup> <mi>I</mi> <mi>c</mi> </msup> <mo>(</mo> <mi>y</mi> <mo>)</mo> <mo>)</mo> </mrow> </mrow>
wherein, Idarkm(x) Is the median value of the dark primary color, Iord(x) For coarse estimation of the transmittance image, Ic(y) is the R, G, B channel of image I (x), parameter k ∈ [0.8, 1).
3. The image defogging method based on the dark channel prior as recited in claim 1, wherein: the advantages mentionedNormalized transmittance image qiComprises the following steps:
<mrow> <msub> <mi>q</mi> <mi>i</mi> </msub> <mo>=</mo> <mfrac> <mn>1</mn> <mrow> <mo>|</mo> <mi>w</mi> <mo>|</mo> </mrow> </mfrac> <munder> <mo>&Sigma;</mo> <mrow> <mi>k</mi> <mo>:</mo> <mi>i</mi> <mo>&Element;</mo> <msub> <mi>&Omega;</mi> <mi>k</mi> </msub> </mrow> </munder> <mrow> <mo>(</mo> <msub> <mi>a</mi> <mi>k</mi> </msub> <mi>I</mi> <mo>+</mo> <msub> <mi>b</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> </mrow>1
<mrow> <msub> <mi>a</mi> <mi>k</mi> </msub> <mo>=</mo> <mfrac> <mrow> <mo>(</mo> <msub> <mi>&Sigma;</mi> <mrow> <mi>i</mi> <mo>&Element;</mo> <msub> <mi>&Omega;</mi> <mi>k</mi> </msub> </mrow> </msub> <msub> <mi>I</mi> <mi>i</mi> </msub> <msub> <mi>p</mi> <mi>i</mi> </msub> <mo>-</mo> <msub> <mi>&mu;</mi> <mi>k</mi> </msub> <msub> <mover> <mi>p</mi> <mo>&OverBar;</mo> </mover> <mi>k</mi> </msub> <mo>)</mo> <mo>/</mo> <mo>|</mo> <mi>w</mi> <mo>|</mo> </mrow> <mrow> <msubsup> <mi>&sigma;</mi> <mi>k</mi> <mn>2</mn> </msubsup> <mo>+</mo> <mi>&epsiv;</mi> </mrow> </mfrac> </mrow>
<mrow> <msub> <mi>b</mi> <mi>k</mi> </msub> <mo>=</mo> <mover> <msub> <mi>p</mi> <mi>k</mi> </msub> <mo>&OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>a</mi> <mi>k</mi> </msub> <msub> <mi>&mu;</mi> <mi>k</mi> </msub> </mrow>
wherein, I is a guide, i.e. image I (x) is abbreviated as I, p is an input image, omega is a square area with a pixel point k as the center, w is the pixel number of the area, IiAnd piThe values of I and P at I within Ω, μ and σ, respectively2Represents the mean and variance of I in the region, as a constraint of akThe coefficient of (a) is determined,representing the mean value for a local region of the input image.
4. The image defogging method based on the dark channel prior as recited in claim 1, wherein: the restored image is as follows:
<mrow> <mi>J</mi> <mrow> <mo>(</mo> <mi>x</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mrow> <mi>I</mi> <mrow> <mo>(</mo> <mi>x</mi> <mo>)</mo> </mrow> <mo>-</mo> <mi>A</mi> </mrow> <mrow> <mi>m</mi> <mi>a</mi> <mi>x</mi> <mrow> <mo>(</mo> <mover> <mi>t</mi> <mo>~</mo> </mover> <mo>(</mo> <mi>x</mi> <mo>)</mo> <mo>,</mo> <msub> <mi>t</mi> <mn>0</mn> </msub> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>+</mo> <mi>A</mi> </mrow>
wherein t (x) e-βd(x)D is the depth of field of the scene, β is the atmospheric scattering coefficient,the estimated value of t (x), the sky brightness A (x) is abbreviated as A; minimum value t0＝0.1。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510224034.3A CN104794697B (en) | 2015-05-05 | 2015-05-05 | A kind of image defogging method based on dark primary priori |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510224034.3A CN104794697B (en) | 2015-05-05 | 2015-05-05 | A kind of image defogging method based on dark primary priori |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104794697A CN104794697A (en) | 2015-07-22 |
| CN104794697B true CN104794697B (en) | 2017-11-21 |
Family
ID=53559477
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510224034.3A Expired - Fee Related CN104794697B (en) | 2015-05-05 | 2015-05-05 | A kind of image defogging method based on dark primary priori |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104794697B (en) |
Families Citing this family (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105354806B (en) * | 2015-11-20 | 2018-05-11 | 上海熙菱信息技术有限公司 | Rapid defogging method and system based on dark |
| CN105574827B (en) * | 2015-12-17 | 2018-04-13 | 中国科学院深圳先进技术研究院 | A kind of method, apparatus of image defogging |
| CN105701785B (en) * | 2016-02-17 | 2018-08-21 | 天津大学 | The image haze minimizing technology of Weighted T V transmissivities optimization is divided based on sky areas |
| CN105894466B (en) * | 2016-03-29 | 2019-01-11 | 青岛海信移动通信技术股份有限公司 | A kind of processing method of image data, device and terminal device |
| CN106251301B (en) * | 2016-07-26 | 2019-10-15 | 北京工业大学 | A kind of single image to the fog method based on dark primary priori |
| CN107038691A (en) * | 2017-04-12 | 2017-08-11 | 北京爱特拉斯信息科技有限公司 | The dark primary remote sensing image defogging method aided in based on cloud detection |
| CN107451966B (en) * | 2017-07-25 | 2021-05-11 | 四川大学 | Real-time video defogging method implemented by guiding filtering through gray level image |
| CN107578386A (en) * | 2017-09-01 | 2018-01-12 | 荆门程远电子科技有限公司 | A kind of optimization defogging processing method of unmanned plane shooting image |
| CN107527329A (en) * | 2017-09-19 | 2017-12-29 | 重庆邮电大学 | A kind of image defogging method based on dark channel prior and Markov random field |
| CN108022225A (en) * | 2017-12-26 | 2018-05-11 | 哈尔滨理工大学 | Based on the improved dark channel prior image defogging algorithm of quick Steerable filter |
| CN108280809B (en) * | 2017-12-26 | 2021-07-30 | 浙江工商大学 | Foggy image sky region estimation method based on atmospheric scattering physical model |
| TWI674804B (en) | 2018-03-15 | 2019-10-11 | 國立交通大學 | Video dehazing device and method |
| CN110532918B (en) * | 2019-08-21 | 2022-02-25 | 南京大学 | Off-shore wind farm space-time attribute determination method based on time series remote sensing images |
| CN112529802B (en) * | 2020-12-09 | 2021-07-20 | 电子科技大学 | Atmospheric scattering degraded image recovery method based on scattering coefficient ratio estimation |
| CN115439494B (en) * | 2022-11-08 | 2023-01-31 | 山东大拇指喷雾设备有限公司 | Spray image processing method for quality inspection of sprayer |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104346783A (en) * | 2014-10-29 | 2015-02-11 | 中国科学院深圳先进技术研究院 | Processing method and processing device for defogging image |
-
2015
- 2015-05-05 CN CN201510224034.3A patent/CN104794697B/en not_active Expired - Fee Related
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104346783A (en) * | 2014-10-29 | 2015-02-11 | 中国科学院深圳先进技术研究院 | Processing method and processing device for defogging image |
Non-Patent Citations (3)
| Title |
|---|
| 《Guided Image Filtering》;Kaiming He et al;《ECCV 2010：Computer Vision-ECCV 2010》;20101231;第35卷(第6期);第1-14页 * |
| 《基于暗通道先验的快速图像去雾算法》;肖钟捷等;《吉林师范大学学报（自然科学版）》;20140830(第3期);第106-110页 * |
| 《面向内河雾天图像的大气亮度值估算方法研究》;黄明晶等;《交通信息与安全》;20130620;第31卷(第3期);第33-38页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104794697A (en) | 2015-07-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104794697B (en) | A kind of image defogging method based on dark primary priori | |
| CN110148095B (en) | Underwater image enhancement method and enhancement device | |
| Yu et al. | Fast single image fog removal using edge-preserving smoothing | |
| CN111292258B (en) | Image defogging method based on dark channel prior and bright channel prior | |
| CN102831591B (en) | Gaussian filter-based real-time defogging method for single image | |
| CN109118446B (en) | Underwater image restoration and denoising method | |
| CN104318524A (en) | Method, device and system for image enhancement based on YCbCr color space | |
| CN111161167B (en) | Single image defogging method based on middle channel compensation and self-adaptive atmospheric light estimation | |
| CN111861896A (en) | UUV-oriented underwater image color compensation and recovery method | |
| CN111462022B (en) | Underwater image sharpness enhancement method | |
| Kumari et al. | Single image fog removal using gamma transformation and median filtering | |
| CN107093173A (en) | A kind of method of estimation of image haze concentration | |
| CN105023246B (en) | A kind of image enchancing method based on contrast and structural similarity | |
| CN111311502A (en) | Method for processing foggy day image by using bidirectional weighted fusion | |
| Wei et al. | An image fusion dehazing algorithm based on dark channel prior and retinex | |
| CN111598814B (en) | Single image defogging method based on extreme scattering channel | |
| CN114693548B (en) | Dark channel defogging method based on bright area detection | |
| CN109345479B (en) | Real-time preprocessing method and storage medium for video monitoring data | |
| CN111489302B (en) | Maritime image enhancement method in fog environment | |
| CN109544470A (en) | A kind of convolutional neural networks single image to the fog method of boundary constraint | |
| CN113920136B (en) | Improved dark channel priori defogging algorithm | |
| Negru et al. | Exponential image enhancement in daytime fog conditions | |
| Lee et al. | A 4K-capable hardware accelerator of haze removal algorithm using haze-relevant features | |
| CN115170404A (en) | Night image defogging method and device based on brightness compensation and layer separation | |
| CN112686825A (en) | Ground dense fog scene gray image defogging method based on FPGA |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20171121 |