CN107301623A - A kind of traffic image defogging method split based on dark and image and system - Google Patents
A kind of traffic image defogging method split based on dark and image and system Download PDFInfo
- Publication number
- CN107301623A CN107301623A CN201710330823.4A CN201710330823A CN107301623A CN 107301623 A CN107301623 A CN 107301623A CN 201710330823 A CN201710330823 A CN 201710330823A CN 107301623 A CN107301623 A CN 107301623A
- Authority
- CN
- China
- Prior art keywords
- image
- mist
- light transmission
- traffic image
- sky areas
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 230000005540 biological transmission Effects 0.000 claims abstract description 164
- 239000003595 mist Substances 0.000 claims abstract description 149
- 230000003044 adaptive effect Effects 0.000 claims abstract description 81
- 238000012937 correction Methods 0.000 claims abstract description 78
- 230000008033 biological extinction Effects 0.000 claims abstract description 52
- 230000015556 catabolic process Effects 0.000 claims abstract description 30
- 238000006731 degradation reaction Methods 0.000 claims abstract description 30
- 238000010586 diagram Methods 0.000 claims abstract description 19
- 238000003709 image segmentation Methods 0.000 claims abstract description 12
- 238000005457 optimization Methods 0.000 claims description 22
- 238000004364 calculation method Methods 0.000 claims description 18
- 230000011218 segmentation Effects 0.000 claims description 17
- 230000000877 morphologic effect Effects 0.000 claims description 11
- 238000002834 transmittance Methods 0.000 claims description 5
- 238000001914 filtration Methods 0.000 claims description 4
- 238000009738 saturating Methods 0.000 claims description 4
- 230000007850 degeneration Effects 0.000 claims 1
- 238000012545 processing Methods 0.000 abstract description 5
- 239000002131 composite material Substances 0.000 abstract 1
- 230000000694 effects Effects 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 238000012544 monitoring process Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 2
- 241000626238 Cepora Species 0.000 description 1
- 241000283070 Equus zebra Species 0.000 description 1
- 206010039203 Road traffic accident Diseases 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000007935 neutral effect Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000005728 strengthening Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G06T5/73—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20172—Image enhancement details
- G06T2207/20192—Edge enhancement; Edge preservation
Abstract
The invention discloses a kind of traffic image defogging method split based on dark and image and system, method includes:To there is mist traffic image to split, the close shot region and sky areas of mist traffic image are obtained;The average intensity value of mist traffic image sky areas is taken as air backscatter extinction logarithmic ratio, and is combined with the dark channel diagram of mist traffic image and calculates scene air light transmission;Adaptive correction is carried out to the air light transmission of sky areas in scene air light transmission, and the degradation model for combining air backscatter extinction logarithmic ratio and having mist traffic image obtains the traffic image after defogging.The present invention is comprehensive to employ image segmentation and the method for dark carries out traffic image defogging, and avoids the transmissivity of sky areas after defogging relatively low by adaptive correction or occur for the situation of negative value, more efficient and versatility is more preferable;The average intensity value of mist traffic image sky areas is taken as air backscatter extinction logarithmic ratio, more rationally.It the composite can be widely applied to image processing field.
Description
Technical field
The present invention relates to image processing field, especially a kind of traffic image defogging side split based on dark and image
Method and system.
Background technology
In recent years, with the propulsion and industrialized fast development of urbanization all over China, environmental problem is increasingly serious, together
When haze weather also more and more continually occur so that the image information accuracy that monitoring system is obtained is by serious shadow
Ring.By taking freeway monitoring system as an example, due to having a dense fog, the visibility of road is substantially reduced, and driver is obtained by vision
Traffic information is often inaccurate, have impact on its correct judgement to environment, is easily caused the generation of traffic accident.Meanwhile, the greasy weather
The degraded image of acquisition also result in very big difficulty to the situation of monitoring traffic route so that traffic department can not pass through shooting
Head obtains useful transport information, and huge resistance is brought to point duty.Therefore, high-quality image defogging is realized, it is right
Being applied in image procossing and computer vision has very important realistic meaning.
Image defogging algorithm mainly has two major classes:One class is the defogging algorithm based on atmospherical scattering model, from the physics of mist
The origin cause of formation is analyzed atmospheric scattering rule, sets up corresponding degradation model, and image scene is carried out using priori
Restore;Another kind of is the defogging algorithm based on image enhaucament, by strengthening scape in the contrast of degraded image itself, prominent image
The feature of thing and valuable information, so as to improve the quality of image, and reach the purpose of sharpening.Going based on image enhaucament
Mist algorithm comparison is ripe, and is easily achieved, but it does not account for the physical of image imaging, therefore when haze concentration is larger,
It just can not recover the mist elimination image of sharpening;And the defogging algorithm based on atmospherical scattering model is built using atmospheric scattering theory
Vertical image degradation model, the reason for having taken into full account image degradation, and image degradation and the relation of atmospheric scattering, image are gone
Mist can obtain more preferable effect.
At present, in image defogging field, widely used method is that the single image defogging based on atmospherical scattering model is calculated
Method, such algorithm uses the prior information included in single image or proposes some reasonably it is assumed that realizing the defogging to image.
Defogging algorithm based on dark primary priori, is calculated as one kind typical case of the single image defogging algorithm based on atmospherical scattering model
Method, by combining degradation model and the dark principle that the greasy weather is imaged, realizes image defogging.However, being currently based on dark primary
The defogging algorithm of priori still suffers from following defect:
1) setting up on dark primary priori theoretical basis, the bright district of dark primary being not present to there are sky, white clouds etc.
For the image in domain, its easily because the transmissivity using the sky areas for causing dark primary priori theoretical to calculate it is relatively low (because
By the depth of field of sky areas it is approximately infinite for dark primary priori theoretical), even negative value is (because dark primary priori theoretical is to big
The estimate of gas backscatter extinction logarithmic ratio is small compared with actual value), it is difficult to effective defogging of image is realized, is not suitable for handling this kind of bag of traffic image
Image containing large area sky areas, versatility is poor.
2) preceding 0.1% pixel is extracted according to the size of brightness generally from the dark channel diagram of foggy image, then takes this
The maximum in original image corresponding to a little pixels so does the air backscatter extinction logarithmic ratio that each passage occurs as air backscatter extinction logarithmic ratio
All close to the situation of max pixel value 255, cause the image after defogging local color spot and color offset phenomenon occur, not enough close
Reason.
The content of the invention
In order to solve the above technical problems, it is an object of the invention to:There is provided a kind of effective, versatility good and rational, base
The traffic image defogging method split in dark and image.
Another object of the present invention is to:A kind of effective, versatility is provided good and rational, based on dark and image point
The traffic image defogging system cut.
The technical solution used in the present invention is:
A kind of traffic image defogging method split based on dark and image, is comprised the following steps:
To there is mist traffic image to split, the close shot region and sky areas of mist traffic image are obtained;
The average intensity value of mist traffic image sky areas is taken as air backscatter extinction logarithmic ratio, and is combined with mist traffic image
Dark channel diagram calculates scene air light transmission;
Adaptive correction is carried out to the air light transmission of sky areas in scene air light transmission, and combines atmosphere light
Coefficient and the degradation model for having mist traffic image obtain the traffic image after defogging.
Further, described pair has mist traffic image to split, and obtains the close shot region and day dead zone of mist traffic image
The step for domain, it includes:
It is false to there is the region in mist traffic image beyond sky areas as close shot region using sky areas as background
If t is any segmentation threshold for having mist traffic image, the probability ω in close shot region is calculated respectivelytWith the probability ω of sky areasB;
Calculate the average gray value μ in close shot regiontWith the average gray value μ of sky areasB;
Calculating view picture has the overall average gray value μ of mist traffic imagerWith inter-class variance σ2, wherein, μrAnd σ2Calculation formula
Respectively:μr=ωt×μt+ωB×μB, σ2=ωt×(μr-μt)2+ωB×(μr-μB)2;
Ask for making inter-class variance σ using the method for traversal2Corresponding t values are as sky areas and close shot area when taking maximum
The optimal segmenting threshold T in domain;
According to optimal segmenting threshold T to there is mist traffic image to split, candidate's close shot region and candidate's sky are partitioned into
Region, wherein, the gray value in candidate's close shot region is 0, and the gray value of candidate sky areas is 1;
Morphological scale-space is carried out to the image after segmentation, mist traffic image final close shot region and day dead zone is obtained
Domain.
Further, the average intensity value for taking mist traffic image sky areas is as air backscatter extinction logarithmic ratio, and is combined with
The step for dark channel diagram of mist traffic image calculates scene air light transmission, it includes:
Calculating has the air backscatter extinction logarithmic ratio of mist traffic image sky areas, the atmosphere light for having a mist traffic image sky areas
Coefficient A expression formula is:Wherein, Ig(v) gray-scale map of foggy image is represented, Ψ (v) indicates mist
The sky areas of traffic image, mean is used for the average value for asking for all pixels point;
The dark for having mist traffic image is calculated based on dark primary priori theoretical, wherein, dark primary priori theoretical will be any
Width image J dark Jdark(x) it is defined as:Wherein, c is image J
{ r, g, b } in one of Color Channel, JcJ Color Channel c components are represented, Ω (x) represents one piece of side centered on x
Shape region, y is any pixel point in region Ω (x), Jc(y) it is image JcPixel value at pixel y,For
The minimum value in r, g, tri- passages of b is sought,For minimum filtering device;
Calculating has the scene air light transmission of mist traffic image, the scene air light transmission for having a mist traffic image
T (x) calculation formula is:Wherein, Ic(y) it is to have mist communication chart
As I (y) c channel components, AcFor A c channel components, ω to characterize the constant of defogging degree, ω ∈ (0,1].
Further, the air light transmission of sky areas carries out adaptive correction in the air light transmission to scene,
And combine air backscatter extinction logarithmic ratio and the step for the degradation model that has mist traffic image obtains the traffic image after defogging, it includes:
Adaptive correction is carried out to the air light transmission of sky areas in scene air light transmission, adaptively repaiied
Scene air light transmission after just;
The scene air light transmission after adaptive correction is optimized using the method for Steerable filter, obtained after optimization
Scene air light transmission;
According to the scene air light transmission after air backscatter extinction logarithmic ratio and optimization, the degradation model for being combined with mist traffic image is obtained
Traffic image after to defogging.
Further, the air light transmission of sky areas carries out adaptive correction in the air light transmission to scene,
The step for obtaining the scene air light transmission after adaptive correction, it includes:
The air light transmission t (v) of sky areas is obtained from scene air light transmission t (x);
Adaptive correction is carried out to the air light transmission t (v) of sky areas, the sky areas after adaptive correction is obtained
Air light transmission t ' (v), the expression formula of the transmissivity t ' (v) is:T ' (v)=min (| λ × t (v) |, μ), wherein, v ∈
Ψ (x), λ are that, for keeping the successional constant of sky areas transmissivity, μ is the thresholding of transmissivity amendment;
The scene air after adaptive correction is obtained according to the sky areas air light transmission t ' (v) after adaptive correction
Light transmission t ' (x).
Further, the scene air light transmission according to after air backscatter extinction logarithmic ratio and optimization, is combined with mist traffic image
Degradation model the step for obtain the traffic image after defogging, it is specially:
According to the scene air light transmission t " (x) after air backscatter extinction logarithmic ratio A and optimization, moving back for mist traffic image I (x) is combined with
Change model and obtain the expression formula of the traffic image J (x) after the traffic image after defogging, the defogging and be:
Another technical scheme for being taken of the present invention is:
A kind of traffic image defogging system split based on dark and image, including:
Image segmentation module, for there is mist traffic image to split, obtain mist traffic image close shot region and
Sky areas;
Scene atmosphere light transmittance calculation module, for taking the average intensity value of mist traffic image sky areas as big
Gas backscatter extinction logarithmic ratio, and it is combined with the dark channel diagram calculating scene air light transmission of mist traffic image;
Adaptive correction and defogging module, enter for the air light transmission to sky areas in scene air light transmission
Row adaptive correction, and the degradation model for combining air backscatter extinction logarithmic ratio and having mist traffic image obtains the traffic image after defogging.
Further, described image segmentation module includes:
Probability calculation unit, for using sky areas as background, to there is the area in mist traffic image beyond sky areas
Domain is used as close shot region, it is assumed that t is any segmentation threshold for having mist traffic image, and the probability ω in close shot region is calculated respectivelytWith
The probability ω of sky areasB;
Average gray value computing unit, for calculating the average gray value μ t in close shot region and the average gray of sky areas
Value μB;
Overall average gray value and inter-class variance computing unit, have the overall average gray value of mist traffic image for calculating view picture
μrWith inter-class variance σ2, wherein, μrAnd σ2Calculation formula be respectively:μr=ωt×μt+ωB×μB, σ2=ωt×(μr-μt)2+
ωB×(μr-μB)2;
Optimal segmenting threshold asks for unit, asks for making inter-class variance σ for the method using traversal2Correspondence when taking maximum
T values as sky areas and close shot region optimal segmenting threshold T;
Cutting unit, for, to there is mist traffic image to split, being partitioned into candidate's close shot area according to optimal segmenting threshold T
Domain and candidate sky areas, wherein, the gray value in candidate's close shot region is 0, and the gray value of candidate sky areas is 1;
Morphological scale-space unit, for carrying out Morphological scale-space to the image after segmentation, has obtained mist traffic image final
Close shot region and sky areas.
Further, the adaptive correction includes with defogging module:
Adaptive correction unit, is carried out adaptive for the air light transmission to sky areas in scene air light transmission
It should correct, obtain the scene air light transmission after adaptive correction;
Optimize unit, the scene air light transmission after adaptive correction is carried out for the method using Steerable filter excellent
Change, the scene air light transmission after being optimized;
Defogging unit, for according to the scene air light transmission after air backscatter extinction logarithmic ratio and optimization, being combined with mist communication chart
The degradation model of picture obtains the traffic image after defogging.
Further, the adaptive correction unit includes:
First obtains subelement, the air light transmission for obtaining sky areas from scene air light transmission t (x)
t(v);
Sky areas adaptive correction subelement, adaptive correction is carried out to the air light transmission t (v) of sky areas,
The sky areas air light transmission t ' (v) after adaptive correction is obtained, the expression formula of the transmissivity t ' (v) is:T ' (v)=
Min (| λ × t (v) |, μ), wherein, v ∈ Ψ (x), λ is that, for keeping the successional constant of sky areas transmissivity, μ is transmission
The thresholding of rate amendment;
Second obtains subelement, for being obtained certainly according to the sky areas air light transmission t ' (v) after adaptive correction
Adapt to revised scene air light transmission t ' (x).
The beneficial effects of the method for the present invention is:Including to there is mist traffic image to split, having taken mist traffic image day
The average intensity value of dummy section is as air backscatter extinction logarithmic ratio, and it is saturating to be combined with the dark channel diagram calculating scene atmosphere light of mist traffic image
Penetrate rate and the air light transmission to sky areas in scene air light transmission carries out adaptive correction, and combine atmosphere light
The step of coefficient and the degradation model for having mist traffic image obtain the traffic image after defogging, it is comprehensive to employ image segmentation and dark
The method of passage carries out traffic image defogging, and is additionally arranged the air light transmission to sky areas in scene air light transmission
The step of carrying out adaptive correction, avoids the transmissivity of sky areas after defogging relatively low or for negative value by adaptive correction
Situation occurs, it is adaptable to handle this kind of image for including large area sky areas of traffic image, more efficient and versatility is more
It is good;The average intensity value of mist traffic image sky areas is taken as air backscatter extinction logarithmic ratio, it is to avoid the air backscatter extinction logarithmic ratio of each passage
All the situation close to max pixel value 255 occurs, more rationally.Adopted when further, to there is mist traffic image to split
With improved Otsu algorithm, not only have the advantages that traditional Otsu algorithm is calculated simply and real-time is good, and pass through form
Processing avoids the situation appearance that close shot region is mistakenly divided into sky areas, and precision is higher.Further, including use
The step of method of Steerable filter is optimized to the scene air light transmission after adaptive correction, employs Steerable filter
Optimization method, not only increase air light transmission optimization speed, also efficiently solve occur after image defogging halation and
The problem of color distortion.
The beneficial effect of system of the present invention is:Including image segmentation module, scene atmosphere light transmittance calculation module and
Adaptive correction and defogging module, the comprehensive method progress traffic image defogging for employing image segmentation and dark, and certainly
The air light transmission that amendment is adapted to being additionally arranged in defogging module to sky areas in scene air light transmission carries out adaptive
The operation that should be corrected, is avoided the transmissivity of sky areas after defogging relatively low by adaptive correction or gone out for the situation of negative value
It is existing, it is adaptable to handle this kind of image for including large area sky areas of traffic image, it is more efficient and versatility is more preferable;It is on the scene
The average intensity value of mist traffic image sky areas is taken in scape atmosphere light transmittance calculation module as air backscatter extinction logarithmic ratio, it is to avoid
The air backscatter extinction logarithmic ratio of each passage all occurs close to the situation of max pixel value 255, more rationally.Further, in image
Improved Otsu algorithm is employed when in segmentation module to there is mist traffic image to split, not only with traditional Otsu algorithm meter
The simple and good advantage of real-time is calculated, and close shot region is avoided by Morphological scale-space and is mistakenly divided into sky areas
Situation occur, precision is higher.Further, adaptive correction includes being used for the method using Steerable filter to certainly with defogging module
The optimization unit that revised scene air light transmission is optimized is adapted to, the optimization method of Steerable filter is employed, not only
The speed of air light transmission optimization is improved, also efficiently solves and occurs asking for halation and color distortion after image defogging
Topic.
Brief description of the drawings
Fig. 1 is a kind of overall flow figure for the traffic image defogging method split based on dark and image of the present invention;
Fig. 2 is the frame diagram of the image defogging algorithm of the embodiment of the present invention one;
Fig. 3 is that Otsu algorithm and improved Otsu algorithm is respectively adopted to there is mist traffic image sky in the embodiment of the present invention one
The effect contrast figure that region is split;
Fig. 4 has the process that mist traffic image scene transfer figure carries out adaptive correction and optimization a pair for the embodiment of the present invention
Schematic diagram;
Fig. 5 is the traffic image schematic diagram before and after the defogging of the embodiment of the present invention one.
Embodiment
A kind of reference picture 1, traffic image defogging method split based on dark and image, is comprised the following steps:
To there is mist traffic image to split, the close shot region and sky areas of mist traffic image are obtained;
The average intensity value of mist traffic image sky areas is taken as air backscatter extinction logarithmic ratio, and is combined with mist traffic image
Dark channel diagram calculates scene air light transmission;
Adaptive correction is carried out to the air light transmission of sky areas in scene air light transmission, and combines atmosphere light
Coefficient and the degradation model for having mist traffic image obtain the traffic image after defogging.
It is further used as preferred embodiment, described pair has mist traffic image to split, and has obtained mist traffic image
Close shot region and sky areas the step for, it includes:
It is false to there is the region in mist traffic image beyond sky areas as close shot region using sky areas as background
If t is any segmentation threshold for having mist traffic image, the probability ω in close shot region is calculated respectivelytWith the probability ω of sky areasB;
Calculate the average gray value μ in close shot regiontWith the average gray value μ of sky areasB;
Calculating view picture has the overall average gray value μ of mist traffic imagerWith inter-class variance σ2, wherein, μrAnd σ2Calculation formula
Respectively:μr=ωt×μt+ωB×μB, σ2=ωt×(μr-μt)2+ωB×(μr-μB)2;
Ask for making inter-class variance σ using the method for traversal2Corresponding t values are as sky areas and close shot area when taking maximum
The optimal segmenting threshold T in domain;
According to optimal segmenting threshold T to there is mist traffic image to split, candidate's close shot region and candidate's sky are partitioned into
Region, wherein, the gray value in candidate's close shot region is 0, and the gray value of candidate sky areas is 1;
Morphological scale-space is carried out to the image after segmentation, mist traffic image final close shot region and day dead zone is obtained
Domain.
It is further used as preferred embodiment, the average intensity value for taking mist traffic image sky areas is as big
Gas backscatter extinction logarithmic ratio, and the step for dark channel diagram of mist traffic image calculates scene air light transmission is combined with, it includes:
Calculating has the air backscatter extinction logarithmic ratio of mist traffic image sky areas, the atmosphere light for having a mist traffic image sky areas
Coefficient A expression formula is:Wherein, Ig(v) gray-scale map of foggy image is represented, Ψ (v) indicates mist
The sky areas of traffic image, mean is used for the average value for asking for all pixels point;
The dark for having mist traffic image is calculated based on dark primary priori theoretical, wherein, dark primary priori theoretical will be any
Width image J dark Jdark(x) it is defined as:Wherein, c is image J
{ r, g, b } in one of Color Channel, JcJ Color Channel c components are represented, Ω (x) represents one piece of side centered on x
Shape region, y is any pixel point in region Ω (x), Jc(y) it is image JcPixel value at pixel y,For
The minimum value in r, g, tri- passages of b is sought,For minimum filtering device;
Calculating has the scene air light transmission of mist traffic image, the scene air light transmission for having a mist traffic image
T (x) calculation formula is:Wherein, Ic(y) it is to have mist communication chart
As I (y) c channel components, AcFor A c channel components, ω to characterize the constant of defogging degree, ω ∈ (0,1].
It is further used as preferred embodiment, the air light transmission of sky areas in the air light transmission to scene
Rate carries out adaptive correction, and the degradation model for combining air backscatter extinction logarithmic ratio and having mist traffic image obtains the traffic image after defogging
The step for, it includes:
Adaptive correction is carried out to the air light transmission of sky areas in scene air light transmission, adaptively repaiied
Scene air light transmission after just;
The scene air light transmission after adaptive correction is optimized using the method for Steerable filter, obtained after optimization
Scene air light transmission;
According to the scene air light transmission after air backscatter extinction logarithmic ratio and optimization, the degradation model for being combined with mist traffic image is obtained
Traffic image after to defogging.
It is further used as preferred embodiment, the air light transmission of sky areas in the air light transmission to scene
Rate progress adaptive correction, the step for obtaining the scene air light transmission after adaptive correction, it includes:
The air light transmission t (v) of sky areas is obtained from scene air light transmission t (x);
Adaptive correction is carried out to the air light transmission t (v) of sky areas, the sky areas after adaptive correction is obtained
Air light transmission t ' (v), the expression formula of the transmissivity t ' (v) is:T ' (v)=min (| λ × t (v) |, μ), wherein, v ∈
Ψ (x), λ are that, for keeping the successional constant of sky areas transmissivity, μ is the thresholding of transmissivity amendment;
The scene air after adaptive correction is obtained according to the sky areas air light transmission t ' (v) after adaptive correction
Light transmission t ' (x).
It is further used as preferred embodiment, the scene air light transmission according to after air backscatter extinction logarithmic ratio and optimization
Rate, the step for degradation model for being combined with mist traffic image obtains the traffic image after defogging, it is specially:
According to the scene air light transmission t " (x) after air backscatter extinction logarithmic ratio A and optimization, moving back for mist traffic image I (x) is combined with
Change model and obtain the expression formula of the traffic image J (x) after the traffic image after defogging, the defogging and be:
Reference picture 1, a kind of traffic image defogging system split based on dark and image, including:
Image segmentation module, for there is mist traffic image to split, obtain mist traffic image close shot region and
Sky areas;
Scene atmosphere light transmittance calculation module, for taking the average intensity value of mist traffic image sky areas as big
Gas backscatter extinction logarithmic ratio, and it is combined with the dark channel diagram calculating scene air light transmission of mist traffic image;
Adaptive correction and defogging module, enter for the air light transmission to sky areas in scene air light transmission
Row adaptive correction, and the degradation model for combining air backscatter extinction logarithmic ratio and having mist traffic image obtains the traffic image after defogging.
It is further used as preferred embodiment, described image segmentation module includes:
Probability calculation unit, for using sky areas as background, to there is the area in mist traffic image beyond sky areas
Domain is used as close shot region, it is assumed that t is any segmentation threshold for having mist traffic image, and the probability ω in close shot region is calculated respectivelytWith
The probability ω of sky areasB;
Average gray value computing unit, the average gray value μ for calculating close shot regiontWith the average gray of sky areas
Value μB;
Overall average gray value and inter-class variance computing unit, have the overall average gray value of mist traffic image for calculating view picture
μrWith inter-class variance σ2, wherein, μrAnd σ2Calculation formula be respectively:μr=ωt×μt+ωB×μB, σ2=ωt×(μr-μt)2+
ωB×(μr-μB)2;
Optimal segmenting threshold asks for unit, asks for making inter-class variance σ for the method using traversal2Correspondence when taking maximum
T values as sky areas and close shot region optimal segmenting threshold T;
Cutting unit, for, to there is mist traffic image to split, being partitioned into candidate's close shot area according to optimal segmenting threshold T
Domain and candidate sky areas, wherein, the gray value in candidate's close shot region is 0, and the gray value of candidate sky areas is 1;
Morphological scale-space unit, for carrying out Morphological scale-space to the image after segmentation, has obtained mist traffic image final
Close shot region and sky areas.
It is further used as preferred embodiment, the adaptive correction includes with defogging module:
Adaptive correction unit, is carried out adaptive for the air light transmission to sky areas in scene air light transmission
It should correct, obtain the scene air light transmission after adaptive correction;
Optimize unit, the scene air light transmission after adaptive correction is carried out for the method using Steerable filter excellent
Change, the scene air light transmission after being optimized;
Defogging unit, for according to the scene air light transmission after air backscatter extinction logarithmic ratio and optimization, being combined with mist communication chart
The degradation model of picture obtains the traffic image after defogging.
It is further used as preferred embodiment, the adaptive correction unit includes:
First obtains subelement, the air light transmission for obtaining sky areas from scene air light transmission t (x)
t(v);
Sky areas adaptive correction subelement, adaptive correction is carried out to the air light transmission t (v) of sky areas,
The sky areas air light transmission t ' (v) after adaptive correction is obtained, the expression formula of the transmissivity t ' (v) is:T ' (v)=
Min (| λ × t (v) |, μ), wherein, v ∈ Ψ (x), λ is that, for keeping the successional constant of sky areas transmissivity, μ is transmission
The thresholding of rate amendment;
Second obtains subelement, for being obtained certainly according to the sky areas air light transmission t ' (v) after adaptive correction
Adapt to revised scene air light transmission t ' (x).
The present invention is further explained and illustrated with reference to Figure of description and specific embodiment.
Embodiment one
Effective defogging of image is difficult to for conventional images defogging algorithm, be not suitable for handling this kind of bag of traffic image
Image containing large area sky areas and it is not reasonable the problems such as, the present invention proposes a kind of new based on dark and image
The traffic image defogging method of segmentation.As shown in Fig. 2 this method is mainly comprised the following steps:Pass through improved Otsu algorithm pair first
Traffic image is split, and obtains the close shot region and sky areas of traffic image;Then by the average intensity value of sky areas
As air backscatter extinction logarithmic ratio, and it is combined with the dark channel diagram calculating scene air light transmission of mist traffic image;Then to day dead zone
The air light transmission in domain carries out adaptive correction, and using Steerable filter to scene air light transmission and sky areas air
Light transmission is optimized;The degradation model for being finally combined with mist traffic image realizes the defogging of image.
Now widely used Misty Image degradation model is deformed by atmospherical scattering model, the atmospherical scattering model
Expression is:
I (x)=J (x) t (x)+A (1-t (x))
In above formula, what I (x) was represented is observed image (image for treating defogging), and J (x) is the fogless image to be recovered
(image i.e. after defogging), A is air backscatter extinction logarithmic ratio, and t (x) is air light transmission, and t (x) is object of the description from arrival camera
Or the medium transmission function of the percentage of the atmosphere light of scene transmitting.The target of defogging is exactly to estimate air backscatter extinction logarithmic ratio A and saturating
Penetrate rate t (x), and the recovery J (x) from I (x).The difficult point of image defogging based on the physical model is, if input is one
The image for having mist is opened, then image defogging is exactly an ill-conditioning problem for lacking bound term.Therefore, defogging method of the invention is overall
It is divided into the progress of following four step:
(1) the close shot area of mist traffic image is obtained to there is mist traffic image to split by improved Otsu algorithm
Domain and sky areas.
Image segmentation is the classical problem of digital image arts.In recent years, using neutral net or point of average drifting
Cut algorithm can segmentation figure picture well, but its is computationally intensive, spent time length and convergence rate is slow, high in requirement of real-time
Field of traffic can not be applied well.Compared with these algorithms, Otsu algorithm (OTSU) has clear advantage, it
By the gamma characteristic of image, background and prospect two parts are divided the image into.Otsu algorithm is simple because of its calculating, not by brightness of image
With the influence of contrast, there is obvious advantage in the high traffic scene of requirement of real-time.
In traffic scene, influenceed by mist, building of zebra stripes, vehicle body and white etc. can have to be close with sky
Brightness, if split using using traditional Otsu algorithm to traffic image, then the above region also can be wrong
Sky areas is divided into by mistake.Therefore, carrying out image segmentation using improved Otsu algorithm herein.
In the improved Otsu algorithm of the present invention, background is that the region beyond sky areas, sky areas is all close shot region,
Its specific algorithm steps is as follows:
1) any segmentation threshold that t is image is assumed, the probability that close shot region is calculated respectively (belongs to the pixel of close shot
Number accounts for the ratio of entire image) ωtWith the probability of sky areas (pixel number for belonging to background accounts for the ratio of entire image)
ωB;
2) the average gray value μ in close shot region is calculatedtWith the average gray value μ of sky areasB;
3) the overall average gray value μ of entire image is calculatedrWith inter-class variance σ2;
4) the corresponding t of inter-class variance value for asking for maximum is used as the optimal segmenting threshold T of sky areas and close shot region;
5) image is split according to threshold value T, nearly scene area is set to 0, sky areas is set to 1;
6) Morphological scale-space is carried out to the image after segmentation, obtains image final close shot region and sky areas.
Result after being split using improved Otsu algorithm to the traffic image for having mist is as shown in figure 3, in Fig. 3 (a)
It is to there is the figure that mist traffic image is obtained after splitting using traditional Otsu algorithm to there is (b) in mist traffic image, Fig. 3
Picture;(c) is to there is the image that mist traffic image is obtained after splitting using improved Otsu algorithm of the invention in Fig. 3.From Fig. 3
As can be seen that improved Otsu algorithm of the invention carries out further morphology by the region being partitioned into traditional Otsu algorithm
Processing, it is to avoid the situation that close shot region is mistakenly divided into sky areas occurs, and precision is higher.
(2) to having after mist traffic image splits, the average intensity value of sky areas is taken as air backscatter extinction logarithmic ratio A, and
The dark channel diagram for being combined with mist traffic image calculates scene air light transmission.
This process can be further subdivided into following steps:
(1) air backscatter extinction logarithmic ratio A calculating
During image defogging, the value for solving air backscatter extinction logarithmic ratio A in scene is a crucial step, how more accurate
Ground estimation A is also a difficult point of image defogging, and whether accurate its estimation is can largely influence final image to answer
Former effect quality.Before existing image defogging algorithm is extracted generally from the dark channel diagram of foggy image according to the size of brightness
0.1% pixel, then takes the maximum in original image corresponding to these pixels as air backscatter extinction logarithmic ratio, so doing to go out
Now the air backscatter extinction logarithmic ratio of each passage causes the image after defogging part occur all close to the situation of max pixel value 255
Color spot and color offset phenomenon.
Preferably to estimate atmosphere light A value, present invention employs the strategy of " equalization ", improved big Tianjin will be used
The average intensity value for the sky areas that algorithm is partitioned into has as A estimate:
In above formula, Ig(v) gray-scale map of mist traffic image is indicated, Ψ (x) represents sky areas, and mean is operated for inciting somebody to action
All pixels point is summed and averaged.
(2) dark channel diagram of mist traffic image is obtained based on dark channel prior theory
Dark primary priori theoretical is to carry out statistics sight to outdoor a large amount of fog free images by Tang Xiao gulls team of Hong Kong Chinese University
Examine what is put forward afterwards:In regional area of the overwhelming majority not comprising sky areas, exist at least one cool colour passage
Some brightness very low pixel, or even level off to zero, gone to zero equivalent to minimum luminance value in region.Therefore, to any one
Width image J, its dark is defined:
In above formula, JcJ some Color Channel component is represented, Ω (x) represents one piece of square region centered on x.
To image JcCarry out minimum Value Operations twice and obtain a dark output,It is used to handle each pixel,
It is minimum filtering device.The two minimum Value Operations can be exchanged with each other when calculating.Therefore, the concept based on dark, works as J
During for the outdoor fog free images of a width, sky areas, J dark primary J are removeddarkBrightness it is very low and level off to 0, that is, have:
Jdark→0
(3) calculating of scene air light transmission
Air light transmission t (x) is that description is saturating from the medium of the percentage of the light for object or the scene transmitting for reaching camera
Penetrate function.It is theoretical by obtained air backscatter extinction logarithmic ratio A and dark channel prior, with reference to the degradation model of Misty Image, it can obtain field
Scape air light transmission t (x) is:
Wherein, ω ∈ (0,1], ω size affects the residual condition of mist in image after defogging.In order that after defogging
Image seems more natural, is necessary to retain a certain degree of mist in defogging, therefore, and the value that ω can be set is 0.95.
(3) calculate after scene air light transmission, first the air light transmission of sky areas is adaptively repaiied
Just, scene air light transmission and sky air light transmission are optimized using the method for Steerable filter afterwards.
This process can be further subdivided into:
(1) transmissivity to sky areas carries out adaptive correction
The transmissivity of the sky areas gone out using dark channel prior theoretical calculation is often relatively low, even negative value.Transmission
Rate it is relatively low mainly because being approximately infinite by the depth of field of sky caused by, and it is then because air spectrum that transmissivity, which is changed into negative value,
Number A estimation is small compared with actual value and causes.For preferably restored image, it is to avoid the image after defogging is excessively strengthened or produced
The phenomenon of raw transmissivity negative value, for the transmission plot of sky areas, present invention employs the mode of self-adaptive processing to day dead zone
The air light transmission t (v) in domain is modified, and obtains the sky areas air light transmission t ' (v) after adaptive correction:
T ' (v)=min (| λ × t (v) |, μ), v ∈ Ψ (x)
In above formula, λ represents a constant, and λ is used for the continuity for keeping sky areas transmissivity;μ represents transmissivity amendment
Thresholding, μ be used for prevent that transmissivity is excessive and causes image exposure excessive.Found by a series of experiments, λ value is set to
10, μ value is set to 0.5, can obtain good defog effect.
(2) transmissivity is optimized using Steerable filter.
The transmissivity of close shot region and sky areas is often very coarse, the phenomenon that there is bulk, and this will cause after defogging
There is halation phenomenon in image or mist hovers in the phenomenon of object edge, is that this traditional image defogging method is generally scratched using soft
The method of figure optimizes transmissivity.But the mode of soft pick figure is computationally intensive, elapsed time is long, and Steerable filter then can be smooth
Image detail and the marginal information for keeping image, the biggest advantage is that calculating speed is fast, therefore the present invention utilizes Steerable filter
Method is optimized to the scene air light transmission t ' (x) after adaptive correction, the scene air light transmission after being optimized
Rate t " (x).When carrying out Steerable filter, guiding figure is designated as I using the minimum value passage for having mist traffic image.
And the air light transmission of sky areas is modified, and using Steerable filter method to overall optimize after
Air light transmission it is as shown in Figure 4.(b) is in the scene air light transmission that (a) calculates for step (2) in Fig. 4, Fig. 4
Scene air light transmission after adaptive correction;(c) is the scene air light transmission after Steerable filter optimizes in Fig. 4.
From fig. 4, it can be seen that after scene transmission plot is modified and optimized, sky areas is not in not only transmissivity
Too low the problem of, and the regional transmission in close shot region becomes more careful, it is more natural in visual effect.
(4) degradation model for being combined with mist traffic image realizes the defogging of traffic image.
Scene air light transmission t " after (x) after air backscatter extinction logarithmic ratio A and optimization is obtained, according to there is mist traffic image
Degradation model, its final traffic image restores formula and is:
Using the present invention image defogging method carry out defogging before and after effect contrast figure as shown in figure 5, in Fig. 5 (a) and
In Fig. 5 (b) be respectively before defogging and defogging after traffic image schematic diagram.
A kind of traffic image defogging method split based on dark and image of the present invention and system, it is improved by combining
Otsu algorithm is partitioned into more accurately sky areas, and further calculates by way of taking strength mean value more reasonably air
Backscatter extinction logarithmic ratio;Pass through adaptive correction is carried out to the air light transmission of sky areas, it is to avoid the sky areas after defogging is by mistake
Degree enhancing or the phenomenon for producing transmissivity negative value;When the air light transmission to close shot region and sky areas is optimized,
The method for employing Steerable filter, not only increases the speed of transmissivity optimization, also efficiently solves after image defogging and occurs
The problems such as halation and color distortion.The proposed by the invention traffic image defogging method split based on dark and image and it is
System, can provide a kind of effective theoretical foundation and technical support for traffic monitoring.
Above is the preferable implementation to the present invention is illustrated, but the present invention is not limited to the embodiment, ripe
A variety of equivalent variations or replacement can also be made on the premise of without prejudice to spirit of the invention by knowing those skilled in the art, this
Equivalent deformation or replacement are all contained in the application claim limited range a bit.
Claims (10)
1. a kind of traffic image defogging method split based on dark and image, it is characterised in that:Comprise the following steps:
To there is mist traffic image to split, the close shot region and sky areas of mist traffic image are obtained;
The average intensity value of mist traffic image sky areas is taken as air backscatter extinction logarithmic ratio, and is combined with helping secretly for mist traffic image
Road figure calculates scene air light transmission;
Adaptive correction is carried out to the air light transmission of sky areas in scene air light transmission, and combines air backscatter extinction logarithmic ratio
The traffic image after defogging is obtained with the degradation model for having mist traffic image.
2. a kind of traffic image defogging method split based on dark and image according to claim 1, its feature is existed
In:Described pair has mist traffic image to split, the step for obtaining the close shot region and sky areas of mist traffic image, its
Including:
Using sky areas as background, to there is the region in mist traffic image beyond sky areas as close shot region, it is assumed that t is
There is any segmentation threshold of mist traffic image, the probability ω in close shot region is calculated respectivelytWith the probability ω of sky areasB;
Calculate the average gray value μ in close shot regiontWith the average gray value μ of sky areasB;
Calculating view picture has the overall average gray value μ of mist traffic imagerWith inter-class variance σ2, wherein, μrAnd σ2Calculation formula difference
For:μr=ωt×μt+ωB×μB, σ2=ωt×(μr-μt)2+ωB×(μr-μB)2;
Ask for making inter-class variance σ using the method for traversal2Corresponding t values are used as sky areas and close shot region when taking maximum
Optimal segmenting threshold T;
According to optimal segmenting threshold T to there is mist traffic image to split, candidate's close shot region and candidate sky areas are partitioned into,
Wherein, the gray value in candidate's close shot region is 0, and the gray value of candidate sky areas is 1;
Morphological scale-space is carried out to the image after segmentation, mist traffic image final close shot region and sky areas is obtained.
3. a kind of traffic image defogging method split based on dark and image according to claim 1, its feature is existed
In:The average intensity value for taking mist traffic image sky areas is combined with mist traffic image as air backscatter extinction logarithmic ratio
The step for dark channel diagram calculates scene air light transmission, it includes:
Calculating has the air backscatter extinction logarithmic ratio of mist traffic image sky areas, the air backscatter extinction logarithmic ratio for having a mist traffic image sky areas
A expression formula is:Wherein, Ig(v) gray-scale map of foggy image is represented, Ψ (v) indicates mist traffic
The sky areas of image, mean is used for the average value for asking for all pixels point;
Being calculated based on dark primary priori theoretical has the dark of mist traffic image, wherein, dark primary priori theoretical is by any width figure
As J dark Jdark(x) it is defined as:Wherein, c for image J r,
G, b } in one of Color Channel, JcJ Color Channel c components are represented, Ω (x) represents one piece of squared region centered on x
Domain, y is any pixel point in region Ω (x), Jc(y) it is image JcPixel value at pixel y,For seeking r,
Minimum value in tri- passages of g, b,For minimum filtering device;
Calculating has the scene air light transmission of mist traffic image, the scene air light transmission t (x) for having a mist traffic image
Calculation formula be:Wherein, Ic(y) it is to have mist traffic image I
(y) c channel components, AcFor A c channel components, ω to characterize the constant of defogging degree, ω ∈ (0,1].
4. a kind of traffic image defogging method split based on dark and image according to claim 3, its feature is existed
In:The air light transmission of sky areas carries out adaptive correction in the air light transmission to scene, and combines atmosphere light
The step for coefficient and the degradation model for having mist traffic image obtain the traffic image after defogging, it includes:
Adaptive correction is carried out to the air light transmission of sky areas in scene air light transmission, obtained after adaptive correction
Scene air light transmission;
The scene air light transmission after adaptive correction is optimized using the method for Steerable filter, the field after being optimized
Scape air light transmission;
According to the scene air light transmission after air backscatter extinction logarithmic ratio and optimization, the degradation model for being combined with mist traffic image is gone
Traffic image after mist.
5. a kind of traffic image defogging method split based on dark and image according to claim 4, its feature is existed
In:The air light transmission of sky areas carries out adaptive correction in the air light transmission to scene, is adaptively repaiied
The step for scene air light transmission after just, it includes:
The air light transmission t (v) of sky areas is obtained from scene air light transmission t (x);
Adaptive correction is carried out to the air light transmission t (v) of sky areas, the sky areas air after adaptive correction is obtained
Light transmission t ' (v), the expression formula of the transmissivity t ' (v) is:T ' (v)=min (| λ × t (v) |, μ), wherein, v ∈ Ψ
(x), λ is that, for keeping the successional constant of sky areas transmissivity, μ is the thresholding of transmissivity amendment;
The scene atmosphere light after adaptive correction is obtained according to the sky areas air light transmission t ' (v) after adaptive correction saturating
Penetrate rate t ' (x).
6. a kind of traffic image defogging method split based on dark and image according to claim 4 or 5, its feature
It is:The scene air light transmission according to after air backscatter extinction logarithmic ratio and optimization, is combined with the degradation model of mist traffic image
The step for obtaining the traffic image after defogging, it is specially:
According to the scene air light transmission t " (x) after air backscatter extinction logarithmic ratio A and optimization, mist traffic image I (x) degeneration is combined with
The expression formula that model obtains the traffic image J (x) after the traffic image after defogging, the defogging is:
7. a kind of traffic image defogging system split based on dark and image, it is characterised in that:Including:
Image segmentation module, for there is mist traffic image to split, obtaining close shot region and the sky of mist traffic image
Region;
Scene atmosphere light transmittance calculation module, for taking the average intensity value of mist traffic image sky areas as atmosphere light
Coefficient, and it is combined with the dark channel diagram calculating scene air light transmission of mist traffic image;
Adaptive correction and defogging module, are carried out certainly for the air light transmission to sky areas in scene air light transmission
Amendment is adapted to, and the degradation model for combining air backscatter extinction logarithmic ratio and having mist traffic image obtains the traffic image after defogging.
8. a kind of traffic image defogging system split based on dark and image according to claim 7, its feature is existed
In:Described image segmentation module includes:
Probability calculation unit, for using sky areas as background, to there is the region in mist traffic image beyond sky areas to make
For close shot region, it is assumed that t is any segmentation threshold for having mist traffic image, the probability ω in close shot region is calculated respectivelytAnd sky
The probability ω in regionB;
Average gray value computing unit, for calculating the average gray value μ t in close shot region and the average gray value μ of sky areasB;
Overall average gray value and inter-class variance computing unit, have the overall average gray value μ of mist traffic image for calculating view picturerWith
Inter-class variance σ2, wherein, μ r and σ2Calculation formula be respectively:μr=ωt×μt+ωB×μB, σ2=ωt×(μr-μt)2+ωB
×(μr-μB)2;
Optimal segmenting threshold asks for unit, asks for making inter-class variance σ for the method using traversal2Corresponding t values when taking maximum
It is used as the optimal segmenting threshold T of sky areas and close shot region;
Cutting unit, for there is mist traffic image to split, be partitioned into according to optimal segmenting threshold T candidate's close shot region and
Candidate sky areas, wherein, the gray value in candidate's close shot region is 0, and the gray value of candidate sky areas is 1;
Morphological scale-space unit, for carrying out Morphological scale-space to the image after segmentation, obtains final near of mist traffic image
Scene area and sky areas.
9. a kind of traffic image defogging system split based on dark and image according to claim 8, its feature is existed
In:The adaptive correction includes with defogging module:
Adaptive correction unit, is adaptively repaiied for the air light transmission to sky areas in scene air light transmission
Just, the scene air light transmission after adaptive correction is obtained;
Optimize unit, for being optimized using the method for Steerable filter to the scene air light transmission after adaptive correction,
Scene air light transmission after being optimized;
Defogging unit, for according to the scene air light transmission after air backscatter extinction logarithmic ratio and optimization, being combined with mist traffic image
Degradation model obtains the traffic image after defogging.
10. a kind of traffic image defogging system split based on dark and image according to claim 9, its feature is existed
In:The adaptive correction unit includes:
First obtains subelement, the air light transmission t for obtaining sky areas from scene air light transmission t (x)
(v);
Sky areas adaptive correction subelement, carries out adaptive correction to the air light transmission t (v) of sky areas, obtains
Sky areas air light transmission t ' (v) after adaptive correction, the expression formula of the transmissivity t ' (v) is:T ' (v)=min
(| λ × t (v) |, μ), wherein, v ∈ Ψ (x), λ is that μ repaiies for transmissivity for keeping sky areas transmissivity successional constant
Positive thresholding;
Second obtains subelement, for being obtained adaptively according to the sky areas air light transmission t ' (v) after adaptive correction
Revised scene air light transmission t ' (x).
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710330823.4A CN107301623B (en) | 2017-05-11 | 2017-05-11 | Traffic image defogging method and system based on dark channel and image segmentation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710330823.4A CN107301623B (en) | 2017-05-11 | 2017-05-11 | Traffic image defogging method and system based on dark channel and image segmentation |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107301623A true CN107301623A (en) | 2017-10-27 |
CN107301623B CN107301623B (en) | 2020-02-14 |
Family
ID=60137852
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710330823.4A Active CN107301623B (en) | 2017-05-11 | 2017-05-11 | Traffic image defogging method and system based on dark channel and image segmentation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107301623B (en) |
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108335307A (en) * | 2018-04-19 | 2018-07-27 | 云南佳叶现代农业发展有限公司 | Adaptive tobacco leaf picture segmentation method and system based on dark primary |
CN108876743A (en) * | 2018-06-26 | 2018-11-23 | 中山大学 | A kind of image rapid defogging method, system, terminal and storage medium |
CN109118546A (en) * | 2018-08-30 | 2019-01-01 | 东南大学 | A kind of depth of field hierarchical estimation method based on single-frame images |
CN110136079A (en) * | 2019-05-05 | 2019-08-16 | 长安大学 | Image defogging method based on scene depth segmentation |
CN110458029A (en) * | 2019-07-15 | 2019-11-15 | 佛山科学技术学院 | Vehicle checking method and device in a kind of foggy environment |
CN111091501A (en) * | 2018-10-24 | 2020-05-01 | 天津工业大学 | Parameter estimation method of atmosphere scattering defogging model |
CN111325688A (en) * | 2020-02-18 | 2020-06-23 | 长安大学 | Unmanned aerial vehicle image defogging method fusing morphological clustering and optimizing atmospheric light |
CN111899198A (en) * | 2020-08-06 | 2020-11-06 | 北京科技大学 | Defogging method and device for marine image |
CN112200746A (en) * | 2020-10-16 | 2021-01-08 | 中南大学 | Defogging method and device for traffic scene image in foggy day |
CN112465720A (en) * | 2020-11-27 | 2021-03-09 | 南京邮电大学 | Image defogging method and device based on image sky segmentation and storage medium |
CN112816483A (en) * | 2021-03-05 | 2021-05-18 | 北京文安智能技术股份有限公司 | Group fog recognition early warning method and system based on fog value analysis and electronic equipment |
CN113205469A (en) * | 2021-06-04 | 2021-08-03 | 中国人民解放军国防科技大学 | Single image defogging method based on improved dark channel |
CN113379619A (en) * | 2021-05-12 | 2021-09-10 | 电子科技大学 | Integrated processing method for defogging imaging, visibility extraction and depth of field estimation |
CN115496685A (en) * | 2022-09-21 | 2022-12-20 | 重庆地质矿产研究院 | Rapid cloud thinning method for high-resolution second-order satellite remote sensing image |
CN115937144A (en) * | 2022-12-08 | 2023-04-07 | 郑州大学 | Image processing method and system in thoracoscopy |
CN116630349A (en) * | 2023-07-25 | 2023-08-22 | 山东爱福地生物股份有限公司 | Straw returning area rapid segmentation method based on high-resolution remote sensing image |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103077500A (en) * | 2012-12-30 | 2013-05-01 | 信帧电子技术(北京)有限公司 | Image data defogging method and device |
CN105654440A (en) * | 2015-12-30 | 2016-06-08 | 首都师范大学 | Regression model-based fast single-image defogging algorithm and system |
CN106251301A (en) * | 2016-07-26 | 2016-12-21 | 北京工业大学 | A kind of single image defogging method based on dark primary priori |
CN106339988A (en) * | 2016-01-07 | 2017-01-18 | 深圳艾科创新微电子有限公司 | Image defogging method and device |
CN106530246A (en) * | 2016-10-28 | 2017-03-22 | 大连理工大学 | Image dehazing method and system based on dark channel and non-local prior |
CN106548463A (en) * | 2016-10-28 | 2017-03-29 | 大连理工大学 | Based on dark and the sea fog image automatic defogging method and system of Retinex |
-
2017
- 2017-05-11 CN CN201710330823.4A patent/CN107301623B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103077500A (en) * | 2012-12-30 | 2013-05-01 | 信帧电子技术(北京)有限公司 | Image data defogging method and device |
CN105654440A (en) * | 2015-12-30 | 2016-06-08 | 首都师范大学 | Regression model-based fast single-image defogging algorithm and system |
CN106339988A (en) * | 2016-01-07 | 2017-01-18 | 深圳艾科创新微电子有限公司 | Image defogging method and device |
CN106251301A (en) * | 2016-07-26 | 2016-12-21 | 北京工业大学 | A kind of single image defogging method based on dark primary priori |
CN106530246A (en) * | 2016-10-28 | 2017-03-22 | 大连理工大学 | Image dehazing method and system based on dark channel and non-local prior |
CN106548463A (en) * | 2016-10-28 | 2017-03-29 | 大连理工大学 | Based on dark and the sea fog image automatic defogging method and system of Retinex |
Cited By (23)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108335307A (en) * | 2018-04-19 | 2018-07-27 | 云南佳叶现代农业发展有限公司 | Adaptive tobacco leaf picture segmentation method and system based on dark primary |
CN108876743B (en) * | 2018-06-26 | 2020-12-29 | 中山大学 | Image rapid defogging method, system, terminal and storage medium |
CN108876743A (en) * | 2018-06-26 | 2018-11-23 | 中山大学 | A kind of image rapid defogging method, system, terminal and storage medium |
CN109118546A (en) * | 2018-08-30 | 2019-01-01 | 东南大学 | A kind of depth of field hierarchical estimation method based on single-frame images |
CN111091501A (en) * | 2018-10-24 | 2020-05-01 | 天津工业大学 | Parameter estimation method of atmosphere scattering defogging model |
CN110136079A (en) * | 2019-05-05 | 2019-08-16 | 长安大学 | Image defogging method based on scene depth segmentation |
CN110458029A (en) * | 2019-07-15 | 2019-11-15 | 佛山科学技术学院 | Vehicle checking method and device in a kind of foggy environment |
CN110458029B (en) * | 2019-07-15 | 2022-12-20 | 佛山科学技术学院 | Vehicle detection method and device in foggy environment |
CN111325688A (en) * | 2020-02-18 | 2020-06-23 | 长安大学 | Unmanned aerial vehicle image defogging method fusing morphological clustering and optimizing atmospheric light |
CN111325688B (en) * | 2020-02-18 | 2023-05-05 | 西安汇智信息科技有限公司 | Unmanned aerial vehicle image defogging method for optimizing atmosphere light by fusion morphology clustering |
CN111899198A (en) * | 2020-08-06 | 2020-11-06 | 北京科技大学 | Defogging method and device for marine image |
CN112200746A (en) * | 2020-10-16 | 2021-01-08 | 中南大学 | Defogging method and device for traffic scene image in foggy day |
CN112200746B (en) * | 2020-10-16 | 2024-03-08 | 中南大学 | Defogging method and equipment for foggy-day traffic scene image |
CN112465720A (en) * | 2020-11-27 | 2021-03-09 | 南京邮电大学 | Image defogging method and device based on image sky segmentation and storage medium |
CN112465720B (en) * | 2020-11-27 | 2024-02-23 | 南京邮电大学 | Image defogging method and device based on image sky segmentation and storage medium |
CN112816483A (en) * | 2021-03-05 | 2021-05-18 | 北京文安智能技术股份有限公司 | Group fog recognition early warning method and system based on fog value analysis and electronic equipment |
CN113379619A (en) * | 2021-05-12 | 2021-09-10 | 电子科技大学 | Integrated processing method for defogging imaging, visibility extraction and depth of field estimation |
CN113205469A (en) * | 2021-06-04 | 2021-08-03 | 中国人民解放军国防科技大学 | Single image defogging method based on improved dark channel |
CN115496685A (en) * | 2022-09-21 | 2022-12-20 | 重庆地质矿产研究院 | Rapid cloud thinning method for high-resolution second-order satellite remote sensing image |
CN115937144A (en) * | 2022-12-08 | 2023-04-07 | 郑州大学 | Image processing method and system in thoracoscopy |
CN115937144B (en) * | 2022-12-08 | 2023-08-25 | 郑州大学 | Image processing method and system in thoracoscopy |
CN116630349A (en) * | 2023-07-25 | 2023-08-22 | 山东爱福地生物股份有限公司 | Straw returning area rapid segmentation method based on high-resolution remote sensing image |
CN116630349B (en) * | 2023-07-25 | 2023-10-20 | 山东爱福地生物股份有限公司 | Straw returning area rapid segmentation method based on high-resolution remote sensing image |
Also Published As
Publication number | Publication date |
---|---|
CN107301623B (en) | 2020-02-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107301623A (en) | A kind of traffic image defogging method split based on dark and image and system | |
CN106157267B (en) | Image defogging transmissivity optimization method based on dark channel prior | |
CN102831591B (en) | Gaussian filter-based real-time defogging method for single image | |
CN111292258B (en) | Image defogging method based on dark channel prior and bright channel prior | |
CN106548461B (en) | Image defogging method | |
CN107767353A (en) | A kind of adapting to image defogging method based on definition evaluation | |
CN108765342A (en) | A kind of underwater image restoration method based on improvement dark | |
CN108596853A (en) | Underwater picture Enhancement Method based on bias light statistical model and transmission map optimization | |
CN110570360B (en) | Retinex-based robust and comprehensive low-quality illumination image enhancement method | |
CN105023256B (en) | A kind of image defogging method and system | |
CN104318524A (en) | Method, device and system for image enhancement based on YCbCr color space | |
CN106780380A (en) | A kind of image defogging method and system | |
CN110211067B (en) | Defogging method for UUV visible light image on offshore surface | |
CN109255759A (en) | Image defogging method based on sky segmentation and transmissivity adaptive correction | |
CN103020914A (en) | Rapid image defogging method based on spatial continuity principle | |
CN104766286A (en) | Image defogging device and method based on pilotless automobile | |
CN103578083A (en) | Single image defogging method based on joint mean shift | |
CN107360344A (en) | Monitor video rapid defogging method | |
CN107895357A (en) | A kind of real-time water surface thick fog scene image Enhancement Method based on FPGA | |
CN106530240A (en) | Image defogging method based on multi-scale fusion and total variational optimization | |
CN110827221A (en) | Single image defogging method based on double-channel prior and side window guide filtering | |
CN107527325A (en) | A kind of monocular underweater vision Enhancement Method preferential based on dark | |
CN106023108A (en) | Image defogging algorithm based on boundary constraint and context regularization | |
CN107085830B (en) | Single image defogging method based on propagation filtering | |
CN103778605A (en) | Greasy weather image enhancement method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |