CN105096271B - Based on traffic image detection method under the haze weather for improving gradient similarity core - Google Patents
Based on traffic image detection method under the haze weather for improving gradient similarity core Download PDFInfo
- Publication number
- CN105096271B CN105096271B CN201510505156.XA CN201510505156A CN105096271B CN 105096271 B CN105096271 B CN 105096271B CN 201510505156 A CN201510505156 A CN 201510505156A CN 105096271 B CN105096271 B CN 105096271B
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- mfrac
- msup
- 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.)
- Active
Links
Landscapes
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of based on traffic image detection method under the haze weather for improving gradient similarity core, traffic image under haze weather is obtained first；Then the traffic image of RGB color step 1 obtained is transformed into Lab color spaces；Then the traffic image of the Lab color spaces obtained to step 2 is filtered processing；The traffic image of Lab color spaces after finally step 3 is handled is transformed into traffic image under RGB color, the haze weather after output processing.The present invention had both met the filter effect of haze weather situation hypograph, and effectively maintained the marginal information of image, particularly important to follow-up traffic image processing and information extraction.
Description
Technical field
The invention belongs to technical field of image processing, and in particular to a kind of based on the haze weather for improving gradient similarity core
Lower traffic image detection method.
Background technology
Digital Image Processing refers to the mistake that picture signal is converted into data signal and handled using computer it
Journey.The purpose of the image procossing of early stage is to improve the quality of image, to obtain the bulk information in image.Image procossing it is conventional
Method has image enhaucament, recovery, coding, compression etc..
In recent years, continuing to develop with image processing field, obtain information using image turns into a kind of important means.
In terms of traffic, image processing techniques is widely used, such as road environment monitoring, vehicle peccancy identification.Either in road row
Car safety still plays more and more important effect in the intelligent transportation system developed.But in natural environment
Under, the traffic image of acquisition can have much noise due to factors such as weather, it is difficult to obtain accurate according to this low-quality image
Information.
At present, digital picture smoothing processing has many methods, spatial domain be smoothly more commonly use in Digital Image Processing and
Effective method.By the development of decades, airspace filter occurs in that the algorithm of many comparative maturities, such as mean filter, intermediate value
Filtering, Wiener filtering.But in itself, spatial domain filter algorithms belong to LPF, because the image after LPF is lost
Some useful high-frequency informations have been lost, have caused image detail to be lost, although filtered image eliminates noise, but causes edge
It is fuzzy, it is very unfavorable to follow-up image procossing, so, the research of various denoising methods really in denoising and retains high
The balance carried out between frequency information.
For above mentioned problem, 1998, C.Tomasi and R.Manduchi proposed a kind of non-iterative simple strategy, claimed
For bilateral filtering.It not only takes the strategy of traditional filtering method, it is contemplated that spatial positional information, is also added into codomain similarity
Influence, therefore, pixel brightness value is more or less the same in the region that image change is gentle, neighborhood, and bilateral filtering is converted into Gauss
Low pass filter；In the region that image change is violent, wave filter utilizes the brightness of the close pixel of brightness value near marginal point
Value is average to substitute former brightness value.Therefore, bilateral filtering can not only eliminate picture noise but also remain image edge information.It is double
Although side filtering has above-mentioned advantage, there is inefficiency, the not good shortcoming of filter effect in it.
The content of the invention
Detected it is an object of the invention to provide a kind of based on traffic image under the haze weather for improving gradient similarity core
Method, to overcome the defect that above-mentioned prior art is present, the present invention had both met the filter effect of haze weather situation hypograph,
The marginal information of image is effectively maintained again, it is particularly important to follow-up traffic image processing and information extraction.
To reach above-mentioned purpose, the present invention is adopted the following technical scheme that：
Based on traffic image detection method under the haze weather for improving gradient similarity core, comprise the following steps：
Step 1：Obtain traffic image under haze weather；
Step 2：The traffic image for the RGB color that step 1 is obtained is transformed into Lab color spaces；
Step 3：The traffic image of the Lab color spaces obtained to step 2 is filtered processing；
Step 4：The traffic image of Lab color spaces after step 3 is handled is transformed into RGB color, output processing
Traffic image under haze weather afterwards.
Further, traffic image under the haze weather of RGB color is transformed into Lab color spaces in step 2, handed over
Logical image is changed into tri- components of L, a, b by R, G, B color component, and wherein L represents the brightness of image, and a is represented from red to green
The scope of color, b represents the scope from yellow to blueness.
Further, the filter transfer function that filtering process is used in step 3 for：
In formula, u ' (i, j) is filtered image slices vegetarian refreshments, and u (m, n) is the pixel in filter window, the spectral window
Mouth is derived from squares of the ω centered on by filtered pixel point of the traffic image for the Lab color spaces that step 2 is obtained for radius
Region, wd(m, n, i, j) is the spatial neighbor degree factor, wr(m, n, i, j) is the codomain similarity factor, and (i, j) is that step 2 is obtained
Lab color spaces traffic image pixel coordinate value, (m, n) is the coordinate value of the pixel in filter window, and ω is
Filter window radius, Cs,rFor normalization coefficient, and
Further, the spatial neighbor degree factor w in filter transfer functiondComputational methods be：
In formula, (m, n) is the coordinate of pixel in filter window, and (i, j) is the friendship for the Lab color spaces that step 2 is obtained
The pixel coordinate value of logical image, σdFor Gauss variance.
Further, the codomain similarity factor w in filter transfer functionrComputational methods be：
In formula, σrFor gaussian coefficient, gL、gaAnd gbRespectively L, a, b Grad.
Further, gL、gaWith and gbComputational methods be：
In formula,WithIt is the obtained L * component image of step 2 respectively to xLSeek partial derivative and to yLSeek partial derivative,WithIt is a component images respectively to xaSeek partial derivative and to yaSeek partial derivative,WithIt is b component images respectively
To xbSeek partial derivative and to ybAsk partial derivative, gL、gaWith and gbRespectively L, a, b Grad.
Further, Gauss variances sigmad=3, gaussian coefficient σr=0.1, filter window radius ω=5.
Compared with prior art, the present invention has following beneficial technique effect：
The present invention constructs gradient similarity core using the gradient of adjacent pixel brightness value, and gradient method is at traditional images edge
Context of detection is widely used, and is that, because gradient is more sensitive to edge, can more efficiently be detected specifically using gradient method
The image border in direction, and because the too strong image denoising effect that result in gradient detection edge declines, the bilateral filter of simple gradient
Wave method filter effect difference loses the essential meaning of wave filter, therefore the present invention improves gradient calculation formula, public with reference to gradient
Formula weaken in itself the image after gradient edge Detection results, filtering process not only edge keep preferably and also filter effect more preferably.
The present invention is average to be weighted to image neighborhood pixels by geometry adjacency core and improved gradient similarity core, so that real
Now filter, both met the filter effect of haze weather situation hypograph, and effectively maintain the marginal information of image, to rear
Continuous traffic image processing and information extraction are particularly important.
Brief description of the drawings
Fig. 1 is the schematic flow sheet of the present invention；
Fig. 2 is of the invention and other filtering methods are contrasted to traffic image denoising effect under haze weather, wherein, (a) is former
Begin noisy acoustic image, (b) mean filter image, (c) tradition bilateral filtering image, (d) gradient bilateral filtering image, (e) this hair
Bright bilateral filtering image；
Fig. 3 is Fig. 2 partial enlargement, wherein, (a) original noisy acoustic image, (b) mean filter image, (c) is double for tradition
Side filtering image, (d) gradient bilateral filtering image, the bilateral filtering image of (e) present invention.
Embodiment
The present invention is described in further detail below in conjunction with the accompanying drawings：
Referring to Fig. 1, based on traffic image detection method under the haze weather for improving gradient similarity core, including following step
Suddenly：
Traffic image under step 1, acquisition haze weather.
Using image capture device, the traffic image degraded under haze weather is obtained.
Step 2, the image of RGB color is transformed into Lab color spaces, image is L, a, b by R, G, B color transition
Three components, wherein L represents the brightness of image, and a represents the scope from red to green, and b represents the scope from yellow to blueness.
Step 3, by step 1 and step 2, we obtain the traffic image under pretreated haze weather, due to
The image that we collect in practice all contains noise, and this useful information obtained to us in image is extremely disadvantageous, and
And noise also has certain interference to further image procossing, therefore it is extremely to have to be filtered processing to the image of acquisition
It is necessary.
Traditional filtering has medium filtering, mean filter etc..Medium filtering and mean filter are carried out to a certain pixel
All it is that the general pixel in filter window is not added with averaging or intermediate value for differentiation during processing.From the principle of bilateral filtering
In we just can be seen that two-sided filter and may be considered what is be made up of two functions, a function be by geometric space away from
From the coefficient for determining wave filter, another is determined the coefficient of wave filter by pixel value difference.
We set windows radius size to be 5, calculate spatial neighbor degree factor wd, formula is as follows
In formula, (m, n) is the coordinate of pixel in filter window, and (i, j) is input by the coordinate of filtered pixel point.Formula
Substantially Gaussian function calculation formula, σdFor Gauss variance.
It is empty as the pixel in window is increased to input by the distance of filtered pixel point from above Gaussian function
Between adjacency factor wdReduce, i.e. neighbor pixel is larger on filtering influence.
Step 4, by the filtering in step 3 pixel of geometric space different distance is take into account to pixel to be processed
The difference of point influence.In order to keep the boundary information of image, we are contemplated that the shadow of pixel value difference between adjacent pixels point
Ring, because gradient is very sensitive to edge, we calculate pixel similarity by modified hydrothermal process from a kind of improved Grad
Coefficient.Traditional gradient calculation formula is as follows：
Wherein GxAnd GyIt is the partial derivative to X and Y respectively, the influence of the derivative of its own is not considered, therefore, we
It is improved that can to obtain new gradient formula as follows to traditional gradient formula：
In formula,It is that image seeks partial derivative to x,It is that image seeks y partial derivative, g is the Grad of output.
We have obtained new gradient formula in step 5, step 4, and using new gradient calculation formula, we can obtain
Lab space Grad gL、gaAnd gb, the gray scale difference value when Grad is larger between adjacent pixel is larger, then can be determined that this pixel
Point is edge pixel point, now, and formula needs to weaken to the filter effect of edge pixel, then we introduce codomain similarity system
Number, when Grad is larger, codomain coefficient of similarity reduces, and function filter effect suitably reduces to keep marginal information.
Codomain similarity factor w is calculated according to the improved Grad of Lab spacerIt is as follows：
In formula, σrIt is used to adjust edge holding effect, g for the gaussian coefficient of codomain similarity factor formulaL、gaAnd gbPoint
Wei not be using the gradient calculation value for improving L, a, b that gradient formula calculating is obtained.
Step 6, general space adjacency factor wdWith improved codomain similarity factor wrFilter transfer function is obtained, is entered
Row filtering process.Determine whether marginal point by calculating the Grad of adjacent pixel, and then adjust codomain similarity factor wr
Size, function pair edge and non-edge is had different filter effects.Weaken filter effect in boundary, reach holding
The purpose at edge.
Filter transfer function：
In formula, u (m, n) is the pixel in filter window, wdFor the spatial neighbor degree factor, wrFor the codomain similarity factor.
(i, j) is that, by the pixel coordinate value of filtering image, (m, n) is the coordinate value of the pixel in filter window, and ω is spectral window
Port radius.
Normalization coefficient：
Horizontal traversing graph is filtered processing to image as pixel.To formula basic parameter assignment, σ in this exampled=3,
σr=0.1, filter window ω=5.
Step 7, the image of Lab color spaces is transformed into RGB color, traffic under the haze weather after output processing
Image.
Treatment effect can more intuitively be seen by Fig. 3 partial enlarged drawing, Fig. 3 (a) is grandfather tape noise image, it is seen that
With much noise it is difficult to carry out follow-up defogging processing in image；Fig. 3 (b) uses radius for the mean filter of 5 filter window,
As can be seen that mean filter soft edge, loss in detail, filter effect is poor；Traditional bilateral filtering result such as Fig. 3 (c), together
Sample use radius for 5 filter window, although maintain a part of edge details of image but filter effect is poor；Fig. 3 (d) is
Gradient bilateral filtering result, filter effect has strengthened；Using filtering method effect such as Fig. 3 (e) of the present invention, filtering effect
Fruit has greatly improved relative to traditional bilateral filtering, and edge keeps preferable.
In summary, filtering method of the invention is directed to traffic image denoising under haze weather, and denoising effect is good, and edge is protected
Hold preferably, to the effective ideal of traffic picture processing under haze weather, further processing to image and accurately obtain image
Information has great significance.
Claims (3)
1. based on traffic image detection method under the haze weather for improving gradient similarity core, it is characterised in that including following step
Suddenly：
Step 1：Obtain traffic image under haze weather；
Step 2：The traffic image for the RGB color that step 1 is obtained is transformed into Lab color spaces；
Step 3：The traffic image of the Lab color spaces obtained to step 2 is filtered processing, the filter that wherein filtering process is used
Ripple transmission function is：
<mrow>
<msup>
<mi>u</mi>
<mo>&prime;</mo>
</msup>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<msub>
<mi>C</mi>
<mrow>
<mi>s</mi>
<mo>,</mo>
<mi>r</mi>
</mrow>
</msub>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>m</mi>
<mo>=</mo>
<mi>i</mi>
<mo>-</mo>
<mi>&omega;</mi>
</mrow>
<mrow>
<mi>i</mi>
<mo>+</mo>
<mi>&omega;</mi>
</mrow>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>m</mi>
<mo>=</mo>
<mi>j</mi>
<mo>-</mo>
<mi>&omega;</mi>
</mrow>
<mrow>
<mi>j</mi>
<mo>+</mo>
<mi>&omega;</mi>
</mrow>
</munderover>
<msub>
<mi>w</mi>
<mi>d</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>,</mo>
<mi>n</mi>
<mo>,</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
<msub>
<mi>w</mi>
<mi>r</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>,</mo>
<mi>n</mi>
<mo>,</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
<mi>u</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>,</mo>
<mi>n</mi>
<mo>)</mo>
</mrow>
</mrow>
In formula, u ' (i, j) is filtered image slices vegetarian refreshments, and u (m, n) is the pixel in filter window, and the filter window takes
The ω centered on by filtered pixel point of the traffic image of the Lab color spaces obtained from step 2 for radius square area,
wd(m, n, i, j) is the spatial neighbor degree factor, wr(m, n, i, j) is the codomain similarity factor, and (i, j) is the Lab that step 2 is obtained
The pixel coordinate value of the traffic image of color space, (m, n) is the coordinate value of the pixel in filter window, and ω is spectral window
Port radius, Cs,rFor normalization coefficient, and
Spatial neighbor degree factor w in filter transfer functiondComputational methods be：
<mrow>
<msub>
<mi>w</mi>
<mi>d</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>,</mo>
<mi>m</mi>
<mo>,</mo>
<mi>n</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<mfrac>
<mrow>
<mo>|</mo>
<mi>i</mi>
<mo>-</mo>
<mi>m</mi>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>+</mo>
<mo>|</mo>
<mi>j</mi>
<mo>-</mo>
<mi>n</mi>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
</mrow>
<mrow>
<mn>2</mn>
<msubsup>
<mi>&sigma;</mi>
<mi>d</mi>
<mn>2</mn>
</msubsup>
</mrow>
</mfrac>
</mrow>
</msup>
</mrow>
In formula, (m, n) is the coordinate value of the interior pixel of filter window, and (i, j) is the friendship for the Lab color spaces that step 2 is obtained
The pixel coordinate value of logical image, σdFor Gauss variance；
Codomain similarity factor w in filter transfer functionrComputational methods be：
<mrow>
<msub>
<mi>w</mi>
<mi>r</mi>
</msub>
<mo>=</mo>
<mi>exp</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<mfrac>
<mrow>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>g</mi>
<mi>L</mi>
</msub>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>+</mo>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>g</mi>
<mi>a</mi>
</msub>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>+</mo>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>g</mi>
<mi>b</mi>
</msub>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
<mrow>
<mn>2</mn>
<msubsup>
<mi>&sigma;</mi>
<mi>r</mi>
<mn>2</mn>
</msubsup>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
</mrow>
In formula, σrFor gaussian coefficient, gL、gaAnd gbRespectively L, a, b Grad；
gL、gaAnd gbComputational methods be：
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>g</mi>
<mi>L</mi>
</msub>
<mo>=</mo>
<msqrt>
<mfrac>
<mrow>
<msup>
<mrow>
<mo>(</mo>
<mfrac>
<mrow>
<mo>&part;</mo>
<msub>
<mi>u</mi>
<mi>L</mi>
</msub>
</mrow>
<mrow>
<mo>&part;</mo>
<msub>
<mi>x</mi>
<mi>L</mi>
</msub>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>+</mo>
<msup>
<mrow>
<mo>(</mo>
<mfrac>
<mrow>
<mo>&part;</mo>
<msub>
<mi>u</mi>
<mi>L</mi>
</msub>
</mrow>
<mrow>
<mo>&part;</mo>
<msub>
<mi>y</mi>
<mi>L</mi>
</msub>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
<mrow>
<mfrac>
<mrow>
<mo>&part;</mo>
<msub>
<mi>u</mi>
<mi>L</mi>
</msub>
</mrow>
<mrow>
<mo>&part;</mo>
<msub>
<mi>x</mi>
<mi>L</mi>
</msub>
</mrow>
</mfrac>
<mo>&times;</mo>
<mfrac>
<mrow>
<mo>&part;</mo>
<msub>
<mi>u</mi>
<mi>L</mi>
</msub>
</mrow>
<mrow>
<mo>&part;</mo>
<msub>
<mi>y</mi>
<mi>L</mi>
</msub>
</mrow>
</mfrac>
</mrow>
</mfrac>
</msqrt>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>g</mi>
<mi>a</mi>
</msub>
<mo>=</mo>
<msqrt>
<mfrac>
<mrow>
<msup>
<mrow>
<mo>(</mo>
<mfrac>
<mrow>
<mo>&part;</mo>
<msub>
<mi>u</mi>
<mi>a</mi>
</msub>
</mrow>
<mrow>
<mo>&part;</mo>
<msub>
<mi>x</mi>
<mi>a</mi>
</msub>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>+</mo>
<msup>
<mrow>
<mo>(</mo>
<mfrac>
<mrow>
<mo>&part;</mo>
<msub>
<mi>u</mi>
<mi>a</mi>
</msub>
</mrow>
<mrow>
<mo>&part;</mo>
<msub>
<mi>y</mi>
<mi>a</mi>
</msub>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
<mrow>
<mfrac>
<mrow>
<mo>&part;</mo>
<msub>
<mi>u</mi>
<mi>a</mi>
</msub>
</mrow>
<mrow>
<mo>&part;</mo>
<msub>
<mi>x</mi>
<mi>a</mi>
</msub>
</mrow>
</mfrac>
<mo>&times;</mo>
<mfrac>
<mrow>
<mo>&part;</mo>
<msub>
<mi>u</mi>
<mi>a</mi>
</msub>
</mrow>
<mrow>
<mo>&part;</mo>
<msub>
<mi>y</mi>
<mi>a</mi>
</msub>
</mrow>
</mfrac>
</mrow>
</mfrac>
</msqrt>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>g</mi>
<mi>b</mi>
</msub>
<mo>=</mo>
<msqrt>
<mfrac>
<mrow>
<msup>
<mrow>
<mo>(</mo>
<mfrac>
<mrow>
<mo>&part;</mo>
<msub>
<mi>u</mi>
<mi>b</mi>
</msub>
</mrow>
<mrow>
<mo>&part;</mo>
<msub>
<mi>x</mi>
<mi>b</mi>
</msub>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>+</mo>
<msup>
<mrow>
<mo>(</mo>
<mfrac>
<mrow>
<mo>&part;</mo>
<msub>
<mi>u</mi>
<mi>b</mi>
</msub>
</mrow>
<mrow>
<mo>&part;</mo>
<msub>
<mi>y</mi>
<mi>b</mi>
</msub>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
<mrow>
<mfrac>
<mrow>
<mo>&part;</mo>
<msub>
<mi>u</mi>
<mi>b</mi>
</msub>
</mrow>
<mrow>
<mo>&part;</mo>
<msub>
<mi>x</mi>
<mi>b</mi>
</msub>
</mrow>
</mfrac>
<mo>&times;</mo>
<mfrac>
<mrow>
<mo>&part;</mo>
<msub>
<mi>u</mi>
<mi>b</mi>
</msub>
</mrow>
<mrow>
<mo>&part;</mo>
<msub>
<mi>y</mi>
<mi>b</mi>
</msub>
</mrow>
</mfrac>
</mrow>
</mfrac>
</msqrt>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
1
In formula,WithIt is the obtained L * component image of step 2 respectively to xLSeek partial derivative and to yLSeek partial derivative,WithIt is a component images respectively to xaSeek partial derivative and to yaSeek partial derivative,WithIt is b component images respectively to xbAsk inclined
Derivative and to ybAsk partial derivative, gL、gaAnd gbRespectively L, a, b Grad；
Step 4：The traffic image of Lab color spaces after step 3 is handled is transformed into RGB color, after output processing
Traffic image under haze weather.
2. traffic image detection method under the haze weather according to claim 1 based on improvement gradient similarity core, its
Be characterised by, traffic image under the haze weather of RGB color be transformed into Lab color spaces in step 2, traffic image by
R, G, B color component are changed into tri- components of L, a, b, and wherein L represents the brightness of image, and a represents the scope from red to green,
B represents the scope from yellow to blueness.
3. traffic image detection method under the haze weather according to claim 1 based on improvement gradient similarity core, its
It is characterised by, Gauss variances sigmad=3, gaussian coefficient σr=0.1, filter window radius ω=5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510505156.XA CN105096271B (en) | 2015-08-17 | 2015-08-17 | Based on traffic image detection method under the haze weather for improving gradient similarity core |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510505156.XA CN105096271B (en) | 2015-08-17 | 2015-08-17 | Based on traffic image detection method under the haze weather for improving gradient similarity core |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105096271A CN105096271A (en) | 2015-11-25 |
| CN105096271B true CN105096271B (en) | 2017-08-25 |
Family
ID=54576622
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510505156.XA Active CN105096271B (en) | 2015-08-17 | 2015-08-17 | Based on traffic image detection method under the haze weather for improving gradient similarity core |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105096271B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2568438A2 (en) * | 2011-09-08 | 2013-03-13 | Fujitsu Limited | Image defogging method and system |
| CN103996178A (en) * | 2014-05-30 | 2014-08-20 | 天津大学 | Sand and dust weather color image enhancing method |
-
2015
- 2015-08-17 CN CN201510505156.XA patent/CN105096271B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2568438A2 (en) * | 2011-09-08 | 2013-03-13 | Fujitsu Limited | Image defogging method and system |
| CN103996178A (en) * | 2014-05-30 | 2014-08-20 | 天津大学 | Sand and dust weather color image enhancing method |
Non-Patent Citations (2)
| Title |
|---|
| Single image haze removal using dark channel prior;Kaiming He et al;《2009 IEEE Conference on Computer Vision and Pattern Recognition》;20090620;1956-1963 * |
| 梯度双边滤波的图像去噪;蒋辉;《计算机工程与应用http://www.cnki.net/kcms/doi/10.3778/j.issn.1002-8331.1402-0317.html》;20140618;摘要，第1-3节 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105096271A (en) | 2015-11-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103942758B (en) | Dark channel prior image dehazing method based on multiscale fusion | |
| CN113436168B (en) | Crack defect detection method for industrial welding line X-ray diagram | |
| CN100474337C (en) | Noise-possessing movement fuzzy image restoration method based on radial basis nerve network | |
| CN102982542B (en) | Fundus image vascular segmentation method based on phase congruency | |
| CN104899876B (en) | A kind of eye fundus image blood vessel segmentation method based on adaptive Gauss difference | |
| CN104253930B (en) | A kind of real-time video defogging method | |
| CN102693426B (en) | Method for detecting image salient regions | |
| CN108564597B (en) | Video foreground object extraction method fusing Gaussian mixture model and H-S optical flow method | |
| CN102855617B (en) | Method and system for processing adaptive images | |
| CN102479383B (en) | Method and device for removing salt and pepper noise | |
| CN101783012A (en) | Automatic image defogging method based on dark primary colour | |
| CN104794696B (en) | A kind of image goes motion blur method and device | |
| CN104200437A (en) | Image defogging method | |
| CN111047556B (en) | Strip steel surface defect detection method and device | |
| CN106096491B (en) | Automatic identification method for microaneurysms in fundus color photographic image | |
| CN106446905A (en) | Surface crack texture extraction method based on fusion of seepage algorithm and adaptive Canny algorithm | |
| CN115797872A (en) | Machine vision-based packaging defect identification method, system, equipment and medium | |
| CN105225243B (en) | One kind can antimierophonic method for detecting image edge | |
| CN106530254A (en) | Algorithm for inhibiting mixed noise of images based on wavelet threshold function and improved median filtering fusion | |
| CN104835127B (en) | A kind of self-adaptive smooth filtering method | |
| CN105719275A (en) | Parallel combination image defect segmentation method | |
| CN110351453A (en) | A kind of computer video data processing method | |
| CN102670251B (en) | Numerical sorting weighted average-based ultrasonic image spatial filter method | |
| CN105787912B (en) | Classification-based step type edge sub-pixel positioning method | |
| CN107369144A (en) | Based on the multiple dimensioned Retinex images defogging method for improving homomorphic filtering |
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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20210128 Address after: Room 508, block a, Rongcheng cloud Valley, 57 Keji 3rd road, Zhangba Street office, high tech Zone, Xi'an City, Shaanxi Province, 710075 Patentee after: Xi'an Huizhi Information Technology Co.,Ltd. Address before: 710064 No. 33, South Second Ring Road, Shaanxi, Xi'an Patentee before: CHANG'AN University |
|
| TR01 | Transfer of patent right |