CN104299192B - A kind of single image to the fog method based on atmospheric light scattering physical model - Google Patents
A kind of single image to the fog method based on atmospheric light scattering physical model Download PDFInfo
- Publication number
- CN104299192B CN104299192B CN201410508456.9A CN201410508456A CN104299192B CN 104299192 B CN104299192 B CN 104299192B CN 201410508456 A CN201410508456 A CN 201410508456A CN 104299192 B CN104299192 B CN 104299192B
- Authority
- CN
- China
- Prior art keywords
- value
- image
- dark channel
- pixel
- atmospheric light
- 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
Abstract
The invention discloses a kind of single image to the fog method based on atmospheric light scattering physical model, is related to image processing field.Its main implementation steps is:1) input has the visible images under mist scene, obtains the variogram of original fog image;2) to there is mist image to carry out mini-value filtering twice, obtain dark channel diagram;3) according to dark channel prior knowledge, using original fog image and dark channel diagram, using variogram as criterion, solve atmospheric light value;4) transmittance figure is solved using dark channel diagram;5) mean filter is carried out on the basis of transmittance figure, obtain optimization transmittance figure;6) atmospheric light scattering physical model formed according to mist image, using the transmittance figure after solving the atmospheric light value that obtains and optimization, it is possible to obtain final fog free images.The effectiveness that atmospheric light value is chosen is this invention ensures that, fog effect is improve.
Description
Technical field
The present invention relates to image procossing, more particularly to using the single image to the fog method of atmospheric light scattering physical model.
Background technology
Aggravation recently as environmental pollution is so that SO2, nitrogen oxides and pellet in air constantly increase
Plus, the above two are gaseous contaminant, and wherein pellet is the key factor for causing haze sky, when they run into the greasy weather
When, sky can become dusky.When we are taken pictures under the conditions of haze weather, the photo taken degrades seriously, its
Main cause is that object is reflected into the light in the visual field a large amount of tiny particles contents in air were subject to before into camera gun
Scattering, refraction and reflect, become disorderly and unsystematic.Therefore, the image contrast degree of acquisition is reduced, and low color definition, picture are lost
A large amount of details are lost, the object being particularly in Vistavision from photo obtains information and greatly reduces.
In actual applications as military technology, traffic, criminal investigation, meteorology and astronomy field are frequently necessary to from outdoor collection
Clearly characteristics of image is extracted in video sequence to be used to recognize and match.Image mist elimination becomes a kind of very urgent and practical
Research topic.Image after mist elimination visually has more pleasing effect, obtains more information, and can be widely used in which
Behind its field, such as mist elimination, image is used as the effective data input of computer vision field.How figure is recovered from the image with mist
The color of picture, contrast obtain clearly image and have important Research Significance and realistic meaning.
In recent years, the method for image mist elimination achieves certain effect, wherein the image mist elimination based on physical model is studied,
It is to realize mist elimination based on information such as scene depths or using the method for multiple image mostly.Recently, take for single image mist elimination
Obtained marked improvement.But degrade because single width has mist image to be affected by mist, it is less that the information that can be utilized in scene structure becomes,
Therefore single image mist elimination is more challenging.In He Kaiming classics mist elimination algorithms, cause process time with the method for soft pick figure
Greatly increase, and dark channel prior knowledge has limitation, fog effect can be caused undesirable.
The content of the invention
It is an object of the invention to provide image clearly, contrast and high definition, details are enriched after a kind of quick, mist elimination
Single image to the fog method based on atmospheric light scattering physical model.
To solve above-mentioned technical problem, the technical solution used in the present invention is:It is a kind of to be based on atmospheric light scattering physical model
Single image to the fog method, it is characterised in that comprise the steps:
Step 1:Input has the otherwise visible light color RGB image under mist scene, calculates the gray variance figure of original fog image, side
Method is as follows:
The mist image that has of definition input is three-channel rgb format image I (x), in RGB Cube space, variance S
It is defined asWherein m is single pixel point triple channel
The average gray value of { r, g, b }, in the range from (0,255), S is the variance of single pixel point, and k is proportionality coefficient;By to I
X in (), each pixel seeks S, variogram S (x) of width original image I (x) is obtained;Proportionality coefficient λ is introduced, is redefinedWherein
Step 2:To there is mist image to carry out mini-value filtering twice, dark channel diagram is obtained, method is as follows:
For any one input picture I (x), its dark channel diagram IdarkX () is defined by formula asWherein, Ω (x) represents one piece of square region centered on pixel x;IcRepresent defeated
Enter numerical value of image I (x) in c (c ∈ { r, g, b }) Color Channel;Dark channel diagram seems that input picture takes minimum through twice
Value computing is obtained,It is to take minima therein for each pixel in three passages { r, g, b },Be one most
Little value filter;
Step 3:It is according to dark channel prior knowledge, using original fog image and dark channel diagram, accurate as differentiating using variogram
Then, atmospheric light value is solved, method is as follows:
Threshold value selection is carried out using variogram S proposed by the invention, introduce Δ as the threshold value chosen, in the present invention
Threshold value Δ=36, if S≤Δ, abandon this data;If S >=Δ, this numerical value is effective, is shone as atmosphere light is weighed
The foundation of value;
It is, by the order sequence successively decreased, to determine intensity in dark channel diagram by the brightness value of pixel to solve atmospheric light value
Value size is front 0.1% position of the pixel in dark channel image, is judged whether effectively, by corresponding to effective data
The average intensity value of the pixel in original fog image region is used as atmospheric light value;
Step 4:Transmittance figure is solved using dark channel diagram, method is as follows:
According toW (0 < w≤1), obtains the transmittance figure of input picture, its
In, Ω (x) represents one piece of square region centered on pixel x;IcRepresenting input images I (x) are in c (c ∈ { r, g, b }) color
Numerical value in passage, dark channel diagram seem that input picture is obtained through taking minimum operation twice,It is for each pixel
Minima therein is taken in three passages { r, g, b },It is a minimum filtering device, AcFor the c (c of atmospheric light value
∈ { r, g, b }) passage component, w is constant coefficient, intermediate value of the present invention be 0.98;
Step 5:Mean filter is carried out on the basis of transmittance figure, optimization transmittance figure is obtained, method is as follows:
The transmittance figure obtained in step 4 is carried out into mean filter operation, filtering size P (x) is 60 × 60;
Step 6:According to the atmospheric light scattering physical model that mist image is formed, using solve the atmospheric light value that obtains and
Transmittance figure after optimization, it is possible to obtain final fog free images, method are as follows:
According to formula atmospheric light scattering physical model formulaFor the transmittance figure after optimization
T (x) sets a lower limit t0, intermediate value of the present invention is 0.1;
It is described according to dark channel prior knowledge, it is using original fog image and dark channel diagram, accurate as differentiating using variogram
Then, solve atmospheric light value, it is characterised in that including dark channel prior knowledge, variogram as criterion, solve it is big
Gas illumination value;
Described dark channel prior knowledge, by being analyzed and summarizing its statistical property to a large amount of outdoor fogless images
The rule for drawing:In the regional area of most non-skies, pixel as always existing, they have at least one
Color Channel has intensity very low and is close to 0 value, and the minima of the light intensity in the region is the number of a very little, for any one
Individual input picture I (x), its dark channel diagram IdarkX () is defined by formula asWherein, Ω
X () represents one piece of square region centered on pixel x;IcRepresenting input images I (x) are in c (c ∈ { r, g, b }) Color Channel
In numerical value, if I is the outdoor fogless image of a width, except sky areas, the dark channel value very little of input picture, connect substantially
0 is bordering on, i.e.,:Idark→ 0, as dark channel prior knowledge;If occurring the higher pixel of a large amount of brightness in dark channel diagram,
These brightness should be that, from sky or fog, the brightness of the denseer dark channel image of fog will be higher, by dark channel diagram
Transmittance figure can be estimated, the dense thin of mist is estimated with this;
Described variogram is used as criterion, it is characterised in that the present invention utilizes proposed variogram to carry out threshold value
Select, threshold value of the introducing Δ as selection, in the present invention threshold value Δ=36, if S≤Δ, then it is assumed that most bright point comes from
On sky areas or white object, this data is abandoned;If S >=Δ, this numerical value is effective, used as measurement atmospheric light value
Foundation;
Described solves atmospheric light value, it is characterised in that definition atmospheric light value is A, will be each in dark channel diagram
The brightness value of pixel determines pixel that intensity level size is front 0.1% in dark channel image by the order sequence successively decreased
Position, then the average intensity value of the pixel of front 0.1% maximum in the original fog image region corresponding to these positions
As atmospheric light value, by the use of variogram as criterion, if the position of selected point come from sky areas or other
On white object, i.e. S≤Δ then abandons this data, if S >=Δ, this numerical value be it is effective, can be used as weighing atmosphere light
According to the foundation of value, this operation is independently carried out to triple channel { r, g, b }, obtains A respectivelyr, Ag, Ab;
Described utilization dark channel diagram solves transmittance figure, it is characterised in that utilize atmospheric scattering physical model, according toW (0 < w≤1), obtain input picture transmittance figure, wherein, Ω (x) represent with
One piece of square region centered on pixel x;IcNumerical value of representing input images I (x) in c (c ∈ { r, g, b }) Color Channel,
Dark channel diagram seems that input picture is obtained through taking minimum operation twice,It is in three passages for each pixel
Minima therein is taken in { r, g, b },It is a minimum filtering device, AcC (c ∈ { r, g, b }) for atmospheric light value
The component of passage, w are constant coefficient;
Mean filter is carried out on the basis of described transmittance figure, optimization transmittance figure is obtained, it is characterised in that due to t
X () is in an image block center not always constant, therefore estimate that the transmittance figure for obtaining occurs in that many blocking effects, causes
Go fog effect not ideal, carried out mean filter operation on the basis of rough estimate transmittance figure, filtering size P (x) by
Image size is determined;
Described atmospheric light scattering physical model, it is characterised in that the optical model quilt of the degraded image obtained in the greasy weather
It is described as:I (x)=t (x) J (x)+(1-t (x)) A, wherein, I (x) represents the image that collects, and A is atmospheric light value, J (x)
It is the radiant illumination of scene, that is, clearly fog free images to be obtained, t (x) are transmittance figure, for describing light by medium
The part not being scattered during being transmitted to imaging device, the target of mist elimination are exactly have mist image I from what known observation was obtained
Clearly mist elimination image J (x) is obtained in (x).
The present invention compared with prior art, with following obvious advantage and beneficial effect:
It is that mist elimination is realized based on physical model in the prior art mostly, instant invention overcomes dark elder generation during this
Test the defect that knowledge is not applied on white object, it is ensured that the effectiveness that atmospheric light value is chosen, improve fog effect, be
The useful supplement of prior art.
Description of the drawings
The implementation process diagram of Fig. 1 embodiments of the invention;
The visible images for having mist scene of Fig. 2 inputs;
Fig. 3 RGB color space figures;
The variogram of Fig. 4 input pictures;
Fig. 5 input pictures take the dark channel diagram of minimum Value Operations for the first time;
Fig. 6 input pictures take the dark channel diagram of mini-value filtering operation for the second time;
The rough transmittance figure that Fig. 7 is obtained according to dark channel prior;
The optimization transmittance figure that Fig. 8 is obtained according to smooth median filter;
The final mist elimination image that Fig. 9 is obtained according to atmospheric light scattering physical model.
Specific embodiment
The present invention will be further described with reference to the accompanying drawings and detailed description.
Implementation steps of the embodiment of the present invention based on the single image to the fog method of physical model are as follows:
Step 1:Input has the visible images under mist scene, obtains the variogram of original fog image;
The mist image that has of definition input is three-channel rgb format image I (x) (as shown in Figure 3), according toR, g, b are image three channel intensity level, and wherein λ is
Constant coefficient, λ=17 in the present embodiment, the variogram for obtaining are as shown in Figure 4.
Step 2:To there is mist image to carry out mini-value filtering twice, dark channel diagram is obtained;
With evidenceWherein, I (x) has mist input picture, I for a widthdarkX () is which
Dark channel diagram, Ω (x) represent one piece of square region centered on pixel x;IcRepresenting input images I (x) c (c ∈ r, g,
B }) numerical value in Color Channel.Dark channel diagram seem input picture through taking what minimum operation was obtained twice,It is pin
Minima therein is taken in three passages { r, g, b } to each pixel,It is a minimum filtering device, the filter taken
Ripple size is Ω (x)=15 × 15.The image after minimum Value Operations is taken for the first time as shown in figure 5, taking minimum Value Operations for the second time
Image afterwards is as shown in Figure 6.
Step 3:It is according to dark channel prior knowledge, using original fog image and dark channel diagram, accurate as differentiating using variogram
Then, solve atmospheric light value;
By the brightness value of each pixel by the order sequence successively decreased in dark channel diagram, determine that intensity level size is front
Position of 0.1% pixel in dark channel image, then front 0.1% in the original fog image region corresponding to these positions
The average intensity value of the pixel of maximum is used as atmospheric light value.By the use of variogram as criterion, if selected point
Position comes from sky areas or other white objects, i.e. S≤Δ, then abandon this data, if S >=Δ, we recognize
It is effective for this numerical value, triple channel { r, g, b } can be independently carried out as the foundation for weighing atmospheric light value, this operation,
A is obtained respectivelyr, Ag, Ab。
Step 4:Transmittance figure is solved using dark channel diagram;
Using atmospherical scattering model, according toW (0 < w≤1), obtains input figure
The transmittance figure of picture.Wherein,For dark channel diagram Idark, one piece centered on pixel x of Ω (x) representatives
Square region;IcNumerical value of representing input images I (x) in c (c ∈ { r, g, b }) Color Channel.Dark channel diagram seems input figure
As obtaining through taking minimum operation twice,Be take in three passages { r, g, b } for each pixel it is therein
Minima,It is a minimum filtering device, the filtering size taken is Ω (x)=15 × 15.AcFor the c of atmospheric light value
The component of (c ∈ { r, g, b }) passage, w is constant coefficient, takes 0.95, and the rough transmittance figure for obtaining is as shown in Figure 7.
Step 5:Mean filter is carried out on the basis of transmittance figure, optimization transmittance figure is obtained;
Mean filter operation is carried out on the basis of rough estimate transmittance figure, filtering size P (x) is determined by image size
Fixed, generally filtering size P (x)=30 × 30, the transmittance figure after optimization are as shown in Figure 8.
Step 6:According to the atmospheric light scattering physical model that mist image is formed, using solve the atmospheric light value that obtains and
Transmittance figure after optimization, it is possible to obtain final fog free images.
According toWhen the absorbance of sky areas levels off to 0, and t (x) is when be close to 0,
Directly restore the original image for obtaining to tend to comprising noise.Therefore, under one will being set for transmittance figure t (x) after optimization
Limit t0, the t in the implementation case0Take 0.1.Wherein, I (x) represents the image that collects of acquisition system, and A is atmospheric light value, J
X () is the clearly fog free images for obtaining, t (x) is transmittance figure.Final mist elimination figure is as shown in Figure 9.
Finally it should be noted that:Above example only not limits technical side described in the invention to illustrate the present invention
Case;Therefore, although this specification with reference to above-mentioned example to present invention has been detailed description, this area it is general
Lead to it will be appreciated by the skilled person that still can modify to the present invention or equivalent;And all without departing from invention spirit
With technical scheme and its improvement of scope, which all should be covered in the middle of scope of the presently claimed invention.
Claims (1)
1. a kind of single image to the fog method based on atmospheric light scattering physical model, it is characterised in that comprise the steps:
Step 1:Input has the otherwise visible light color RGB image under mist scene, calculates the gray variance figure of original fog image, and method is such as
Under:
The mist image that has of definition input is three-channel rgb format image I (x), in RGB Cube space, the definition of variance S
ForWherein m is single pixel point triple channel { r, g, b }
Average gray value, in the range from (0,255), S is the variance of single pixel point, and k is proportionality coefficient;By to I (x)
In each pixel seek S, be obtained width original image I (x) variogram S (x);Proportionality coefficient λ is introduced, is redefinedWherein
Step 2:To there is mist image to carry out mini-value filtering twice, dark channel diagram is obtained, method is as follows:
For any one input picture I (x), its dark channel diagram IdarkX () is defined by formula asWherein, Ω (x) represents one piece of square region centered on pixel x;IcRepresent defeated
Enter numerical value of image I (x) in c (c ∈ { r, g, b }) Color Channel;Dark channel diagram seems that input picture takes minimum through twice
Value computing is obtained,It is to take minima therein for each pixel in three passages { r, g, b },Be one most
Little value filter;
Step 3:According to dark channel prior knowledge, using original fog image and dark channel diagram, using variogram as criterion, ask
Atmospheric light value is solved, method is as follows:
Threshold value selection is carried out using variogram S, Δ is introduced as the threshold value chosen, threshold value Δ=36, if S≤Δ, then it is assumed that
Most bright point comes from sky areas or white object, abandons this data;If S>Δ, then this numerical value be it is effective, as
Weigh the foundation of atmospheric light value;
It is, by the order sequence successively decreased, to determine that intensity level is big in dark channel diagram by the brightness value of pixel to solve atmospheric light value
It is little for front 0.1% position of the pixel in dark channel image, judge whether effectively, will be original corresponding to effective data
The average intensity value of the pixel in mist image-region is used as atmospheric light value;
Step 4:Transmittance figure is solved using dark channel diagram, method is as follows:
According toW (0 < w≤1), obtains the transmittance figure of input picture, wherein, Ω
X () represents one piece of square region centered on pixel x;IcRepresenting input images I (x) are in c (c ∈ { r, g, b }) Color Channel
In numerical value, dark channel diagram seems that input picture is obtained through taking minimum operation twice,It is three for each pixel
Minima therein is taken in individual passage { r, g, b },It is a minimum filtering device, AcC for atmospheric light value (c ∈ r,
G, b }) component of passage, w is constant coefficient, and its value is 0.98;
Step 5:Mean filter is carried out on the basis of transmittance figure, optimization transmittance figure is obtained, method is as follows:
The transmittance figure obtained in step 4 is carried out into mean filter operation, filtering size P (x) is 60 × 60;
Step 6:According to the atmospheric light scattering physical model that mist image is formed, using solving the atmospheric light value and optimization that obtain
Transmittance figure afterwards, it is possible to obtain final fog free images, method are as follows:
According to formula atmospheric light scattering physical model formulaFor transmittance figure t (x) after optimization
One lower limit t of setting0, its value is 0.1;
It is described according to dark channel prior knowledge, using original fog image and dark channel diagram, using variogram as criterion, ask
Solve atmospheric light value, it is characterised in that including dark channel prior knowledge, variogram as criterion, solve atmosphere light photograph
Value;
Described dark channel prior knowledge, is drawn by being analyzed and summarizing its statistical property to a large amount of outdoor fogless images
A rule:In the regional area of most non-skies, pixel as always existing, they have at least one color
Passage has intensity very low and is close to 0 value, and the minima of the light intensity in the region is the number of a very little, for any one is defeated
Enter image I (x), its dark channel diagram IdarkX () is defined by formula asWherein, Ω (x) generations
One piece square region of the table centered on pixel x;IcNumber of representing input images I (x) in c (c ∈ { r, g, b }) Color Channel
Value, if I is the outdoor fogless image of a width, except sky areas, the dark channel value very little of input picture, is essentially close to 0,
I.e.:Idark→ 0, as dark channel prior knowledge;If occurring the higher pixel of a large amount of brightness in dark channel diagram, these are bright
Degree should be that, from sky or fog, the brightness of the denseer dark channel image of fog will be higher, can be estimated by dark channel diagram
Meter transmittance figure, estimates the dense thin of mist with this;
Described solves atmospheric light value, and definition atmospheric light value is A, in dark channel diagram is pressed the brightness value of each pixel
The order sequence successively decreased, determines the position of pixel that intensity level size is front 0.1% in dark channel image, then these positions
The average intensity value of the pixel of front 0.1% maximum in corresponding original fog image region is used as atmospheric light value, profit
With variogram as criterion, if the position of selected point comes from sky areas or other white objects, i.e. S≤
Δ, then abandon this data, if S>Δ, then this numerical value be it is effective, can as weigh atmospheric light value foundation, this operation
Triple channel { r, g, b } is independently carried out, A is obtained respectivelyr, Ag, Ab;
Described utilization dark channel diagram solves transmittance figure, it is characterised in that utilize atmospheric scattering physical model, according toW (0 < w≤1), obtain input picture transmittance figure, wherein, Ω (x) represent with
One piece of square region centered on pixel x;IcNumerical value of representing input images I (x) in c (c ∈ { r, g, b }) Color Channel,
Dark channel diagram seems that input picture is obtained through taking minimum operation twice,Be for each pixel three passages r,
G, b } in take minima therein,It is a minimum filtering device, AcC (c ∈ { r, g, b }) passage for atmospheric light value
Component, w is constant coefficient;
Described atmospheric light scattering physical model, the optical model of the degraded image obtained in the greasy weather are described as:I (x)=max
T (x) J (x)+(1-max t (x)) A, wherein, I (x) represents the image for collecting, and A is atmospheric light value, and J (x) is the spoke of scene
Illumination is penetrated, that is, clearly fog free images to be obtained, t (x) are transmittance figure, are transmitted to imaging by medium for describing light
The part not being scattered in device procedures, the target of mist elimination are exactly to observe having in mist image I (x) of obtaining and obtain clear from known
Clear mist elimination image J (x).
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410508456.9A CN104299192B (en) | 2014-09-28 | 2014-09-28 | A kind of single image to the fog method based on atmospheric light scattering physical model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410508456.9A CN104299192B (en) | 2014-09-28 | 2014-09-28 | A kind of single image to the fog method based on atmospheric light scattering physical model |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104299192A CN104299192A (en) | 2015-01-21 |
CN104299192B true CN104299192B (en) | 2017-03-29 |
Family
ID=52318913
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410508456.9A Active CN104299192B (en) | 2014-09-28 | 2014-09-28 | A kind of single image to the fog method based on atmospheric light scattering physical model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104299192B (en) |
Families Citing this family (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105512623B (en) * | 2015-12-02 | 2018-09-14 | 吉林大学 | Based on multisensor travelling in fog day vision enhancement and visibility early warning system and method |
CN105488769B (en) * | 2015-12-08 | 2019-02-19 | 中国航空工业集团公司西安航空计算技术研究所 | A kind of real-time video defogging method |
CN105741246A (en) * | 2016-01-29 | 2016-07-06 | 南京航空航天大学 | Method for eliminating influence on optical picture by dust and mist |
CN105608683B (en) * | 2016-03-11 | 2018-02-16 | 北京理工大学 | A kind of single image to the fog method |
CN106846259A (en) * | 2016-12-15 | 2017-06-13 | 上海交通大学 | A kind of the real-time of laparoscopic surgery video frame rate self adaptation goes smog method |
CN108765302B (en) * | 2018-03-29 | 2022-07-22 | 西安电子科技大学 | Image real-time defogging method based on GPU |
CN109919859B (en) * | 2019-01-25 | 2021-09-07 | 暨南大学 | Outdoor scene image defogging enhancement method, computing device and storage medium thereof |
CN110910319B (en) * | 2019-10-30 | 2022-10-21 | 中国医学科学院北京协和医院 | Operation video real-time defogging enhancement method based on atmospheric scattering model |
CN111192210B (en) * | 2019-12-23 | 2023-05-26 | 杭州当虹科技股份有限公司 | Self-adaptive enhanced video defogging method |
CN111489302B (en) * | 2020-03-27 | 2023-12-05 | 武汉理工大学 | Maritime image enhancement method in fog environment |
CN111738938B (en) * | 2020-06-01 | 2022-09-09 | 余姚市浙江大学机器人研究中心 | Nonuniform atomization video optimization method based on prior target identification |
CN111754433B (en) * | 2020-06-22 | 2024-03-05 | 哈尔滨理工大学 | Defogging method for aerial image |
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 |
CN113298732A (en) * | 2021-06-08 | 2021-08-24 | 北京联合大学 | Image defogging method and system based on regional similarity |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102289791A (en) * | 2011-06-29 | 2011-12-21 | 清华大学 | Method for quickly demisting single image |
CN102411774A (en) * | 2011-08-08 | 2012-04-11 | 安防科技(中国)有限公司 | Processing method, device and system based on single-image defogging |
CN102930514A (en) * | 2012-09-27 | 2013-02-13 | 西安电子科技大学 | Rapid image defogging method based on atmospheric physical scattering model |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8588551B2 (en) * | 2010-03-01 | 2013-11-19 | Microsoft Corp. | Multi-image sharpening and denoising using lucky imaging |
-
2014
- 2014-09-28 CN CN201410508456.9A patent/CN104299192B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102289791A (en) * | 2011-06-29 | 2011-12-21 | 清华大学 | Method for quickly demisting single image |
CN102411774A (en) * | 2011-08-08 | 2012-04-11 | 安防科技(中国)有限公司 | Processing method, device and system based on single-image defogging |
CN102930514A (en) * | 2012-09-27 | 2013-02-13 | 西安电子科技大学 | Rapid image defogging method based on atmospheric physical scattering model |
Non-Patent Citations (2)
Title |
---|
single image haze removal using dark channel prior;Kaiming He et al.;《IEEE Transactions on Pattern Analysis and Machine Intelligence》;20111231;第33卷(第12期);2341-2353 * |
基于暗通道先验图像去雾的方法改进;崔冰琪 等;《信息通信》;20131231(第6期);60-61 * |
Also Published As
Publication number | Publication date |
---|---|
CN104299192A (en) | 2015-01-21 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104299192B (en) | A kind of single image to the fog method based on atmospheric light scattering physical model | |
Tripathi et al. | Single image fog removal using bilateral filter | |
KR101470831B1 (en) | Appatatus for image dehazing using the user controllable radical root operation | |
CN107301624B (en) | Convolutional neural network defogging method based on region division and dense fog pretreatment | |
CN104050637B (en) | Quick image defogging method based on two times of guide filtration | |
CN102930514A (en) | Rapid image defogging method based on atmospheric physical scattering model | |
CN103186887B (en) | Image demister and image haze removal method | |
CN111062293B (en) | Unmanned aerial vehicle forest flame identification method based on deep learning | |
CN103985091A (en) | Single image defogging method based on luminance dark priori method and bilateral filtering | |
CN103049890A (en) | Real-time image defogging method based on CUDA (Compute Unified Device Architecture) | |
CN102063704B (en) | Airborne vision enhancement method and device | |
CN103198459A (en) | Haze image rapid haze removal method | |
CN107451966A (en) | A kind of real-time video defogging method realized using gray-scale map guiding filtering | |
CN103578083A (en) | Single image defogging method based on joint mean shift | |
CN106023118B (en) | A kind of image defogging method and its implementation method on FPGA | |
CN107730472A (en) | A kind of image defogging optimized algorithm based on dark primary priori | |
CN112365467B (en) | Foggy image visibility estimation method based on single image depth estimation | |
CN105447825A (en) | Image defogging method and system | |
CN106657948A (en) | low illumination level Bayer image enhancing method and enhancing device | |
CN112950589A (en) | Dark channel prior defogging algorithm of multi-scale convolution neural network | |
CN102156968A (en) | Color cubic priori based single image visibility restoration method | |
Al-Zubaidy et al. | Removal of atmospheric particles in poor visibility outdoor images | |
CN105631825B (en) | Based on the image defogging method for rolling guiding | |
Putra et al. | Visibility restoration of lake crater hazy image based on dark channel prior | |
CN109166081B (en) | Method for adjusting target brightness in video visibility detection process |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |