CN106127715A - A kind of image defogging method and system - Google Patents
A kind of image defogging method and system Download PDFInfo
- Publication number
- CN106127715A CN106127715A CN201610748954.XA CN201610748954A CN106127715A CN 106127715 A CN106127715 A CN 106127715A CN 201610748954 A CN201610748954 A CN 201610748954A CN 106127715 A CN106127715 A CN 106127715A
- Authority
- CN
- China
- Prior art keywords
- image
- mist
- sky areas
- absorbance
- value
- 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.)
- Withdrawn
Links
- 238000000034 method Methods 0.000 title claims abstract description 94
- 239000003595 mist Substances 0.000 claims abstract description 200
- 238000002835 absorbance Methods 0.000 claims abstract description 64
- 238000003379 elimination reaction Methods 0.000 claims abstract description 54
- 230000008030 elimination Effects 0.000 claims abstract description 53
- 230000001419 dependent effect Effects 0.000 claims abstract description 9
- 238000002834 transmittance Methods 0.000 claims description 25
- 238000011156 evaluation Methods 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 11
- 238000004364 calculation method Methods 0.000 claims description 6
- 239000003550 marker Substances 0.000 claims description 6
- 238000013386 optimize process Methods 0.000 claims description 6
- 108010001267 Protein Subunits Proteins 0.000 claims description 3
- 230000005540 biological transmission Effects 0.000 claims description 3
- 238000002156 mixing Methods 0.000 claims description 2
- 239000000203 mixture Substances 0.000 claims description 2
- 238000012545 processing Methods 0.000 abstract description 10
- 230000009467 reduction Effects 0.000 abstract description 2
- 230000000694 effects Effects 0.000 description 14
- 238000010586 diagram Methods 0.000 description 12
- 238000001914 filtration Methods 0.000 description 7
- 238000011160 research Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 3
- 230000015556 catabolic process Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000003412 degenerative effect Effects 0.000 description 2
- 238000006731 degradation reaction Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000011084 recovery Methods 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000002146 bilateral effect Effects 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 125000001475 halogen functional group Chemical group 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000036039 immunity Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000000149 penetrating effect Effects 0.000 description 1
- 229920006395 saturated elastomer Polymers 0.000 description 1
- 238000009738 saturating Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/73—Deblurring; Sharpening
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a kind of image defogging method and system.The method includes: obtain comprise sky areas have mist image；The sky areas in mist image is had described in determining；Set up atmospherical scattering model；According to described sky areas account for whole described in have the air light value in atmospherical scattering model described in the ratio-dependent of mist image, described air light value is the atmosphere light intensity of infinite point；According to described atmospherical scattering model determine described in have the absorbance of mist image；The observed strength of mist image is had described in acquisition.The method and system that the present invention provides can carry out good mist elimination process to the sky areas having in mist image, and, there is the feature that reduction precision is high, mist elimination complexity rate is low, processing speed is fast of image.
Description
Technical field
The present invention relates to image domains, particularly relate to a kind of image defogging method and system.
Background technology
Mist is a kind of common natural phenomena, and it can make the visibility of air reduce, and scene image is degenerated, in the greasy weather
The picture material of shooting obscures, and contrast declines, and this will have a strong impact on outdoor image collection and process, makes the work cannot be normal
Carry out.Therefore, the problem unfolded image signal processing of the image quality decrease caused by this natural phenomena is had with research
Universal meaning.On the other hand, due to the fast development of computer technology, the arithmetic speed of computer technology is increasingly faster, image
The price of processing system declines day by day, along with the fast development of computer vision Yu image processing techniques, outdoor visual system
Study and be also skyrocketed through with application.Thus image processing techniques is widely used in science and engineering field, in order to ensure to regard
The round-the-clock normal work of vision system, is necessary for making system adapt to various weather conditions, and Misty Image contrast and color all can
Degenerate, cause these systems normally to work.Therefore, how research is to the Degenerate Graphs obtained under the vile weathers such as cloud
Picture effectively processes, the recovery to Atmospheric Degraded Image, and the enhancing of scenery detailed information suffers from very important existing
Sincere justice.
Both at home and abroad the research having mist image mist elimination algorithm is had been achieved for the biggest progress, at present, haze image at present
Restored method is broadly divided into two classes: Misty Image strengthens and Misty Image is restored.The Enhancement Method of Misty Image does not consider image
Causes for Degradation, the suitability is wide, can be effectively improved the contrast of Misty Image, strengthens the details of image, improves the vision of image
Effect, but the information for ledge is likely to result in certain loss.It is that research Misty Image is degenerated that Misty Image is restored
Physical mechanism, and set up greasy weather degradation model, inverting degenerative process, compensate the distortion that degenerative process causes, so as to obtain without
Disturb the image without mist degenerated or the optimal estimation value without mist image, thus improve the quality of Misty Image.This method for
Property strong, obtain goes fog effect natural, does not the most have information loss, and the key point of process is the estimation of Model Parameter.Right
In each class method, it is summarized as different subclass methods further according to the similarity of defogging method: mist based on image procossing
It image enchancing method is divided into the image enchancing method of the overall situationization and the image enchancing method of localization；Mist based on physical model
It image recovery method then include Misty Image based on partial differential equation restore, Misty Image based on depth relationship restore and
Misty Image based on prior information is restored.
Single width figure mist elimination is that recent researches is hotter, and method emerges in an endless stream, but is substantially at following several classics
The differentiation carried out on algorithm: 08 year Tan et al. is found by statistical study, without mist image relative to there being mist image must have relatively
High contrast, eliminates mist by maximizing local contrast.Assume that regional area ambient light is constant.At Markovian model
Under type framework, constructing the cost function about edge strength, use figure segmentation theory estimates optimum illumination.This algorithm is intended to increase
The contrast of strong image.Although significantly improving the observability of image, yet with not recovering true from physical model
Scene albedo, so the color of the image after Hui Fuing seems the most saturated, and produces serious on the border of depth of field sudden change
Halo effect.09 year doctor He Kaiming proposes a kind of mist elimination algorithm based on dark channel prior, simple and effective, immediately causes
The extensive concern of mist elimination circle, but in place of this method there is also some shortcomings: it uses the thinking of soft matting to put down
Sliding propagation in atmosphere function, but effective complexity is excessive, inapplicable practical operation, and also it can not process sky areas.
Summary of the invention
It is an object of the invention to provide the image defogging method processing sky areas based on dark channel prior of a kind of improvement
And system, it is possible to well realize the mist elimination of sky areas.
For achieving the above object, the invention provides following scheme:
A kind of image defogging method, including:
Obtain comprise sky areas have mist image；
The sky areas in mist image is had described in determining；
Set up atmospherical scattering model；
According to described sky areas account for whole described in have the air in atmospherical scattering model described in the ratio-dependent of mist image
Light value, described air light value is the atmosphere light intensity of infinite point；
According to described atmospherical scattering model determine described in have the absorbance of mist image；
The observed strength of mist image is had described in acquisition；
According to described air light value, described in have the observed strength of mist image and the described transmittance calculation having mist image to obtain
Image after mist elimination.
Optionally, described determine described in have the sky areas in mist image, specifically include:
The Grad of each pixel in mist image is had described in calculating；
Less than the pixel of first threshold, described Grad is labeled as 1, and described Grad is more than or equal to first threshold
Pixel is labeled as 0, obtains bianry image；
With a round die block, described bianry image is corroded, obtain new bianry image；
Utilize region growing algorithm that described new bianry image is processed, obtain described new bianry image all of
UNICOM region；
Calculate each described UNICOM region in the described mean flow rate having pixel corresponding on mist image；
It is sky areas by UNICOM's zone marker that described mean flow rate is more than Second Threshold；
The satisfied pixel imposed a condition in non-sky areas is labeled as sky areas；
Leak in described sky areas is filled with.
Optionally, described according to sky areas account for whole described in have in atmospherical scattering model described in the ratio-dependent of mist image
Air light value, specifically include:
Calculate described sky areas account for whole described in have the ratio of mist image；
Judge that whether described ratio is less than setting ratio；
If it is not, then calculate the meansigma methods of the light intensity of pixel in described sky areas, described meansigma methods is labeled as
Described air light value；
During if it is, have described in Huo Quing that in mist image dark channel, brightness arranges from high to low, arrange the pixel of front 0.1%
Point, and before obtaining described row the pixel of 0.1% in the described light intensity value having in mist image, by described light intensity value
Big light intensity value is labeled as described air light value.
Optionally, described according to described atmospherical scattering model determine described in have the absorbance of mist image, specifically include:
Utilize the absorbance having mist image described in the estimation of dark channel prior algorithm, obtain absorbance evaluation function；
Obtain fixed transmittance rate value；
With described absorbance evaluation function, described fixed transmittance rate value is carried out Alpha mix, obtain rational absorbance；
Utilize Steerable filter algorithm that described rational absorbance is optimized process.
Optionally, described according to described air light value, described in have the observed strength of mist image and described have the saturating of mist image
The rate of penetrating is calculated the image after mist elimination, specifically includes:
According to formulaImage J (x) after mist image mist elimination is had described in being calculated, wherein,
A is described air light value, t (x) be described in have the refractive index of mist image, I (x) be described in have the observed strength of mist image.
A kind of image mist elimination system, described system includes:
Image acquisition unit, for obtain comprise sky areas have mist image；
Sky areas determines unit, has the sky areas in mist image described in determining；
Atmospherical scattering model sets up unit, is used for setting up atmospherical scattering model；
Air light value determines unit, for according to described sky areas account for whole described in have described in the ratio-dependent of mist image
Air light value in atmospherical scattering model, described air light value is the atmosphere light intensity of infinite point；
Absorbance determines unit, has the absorbance of mist image described in determining according to described atmospherical scattering model；
Observed strength acquiring unit, is used for the observed strength having mist image described in obtaining；
Mist elimination image computing unit, for according to described air light value, described in have the observed strength of mist image and described have
The transmittance calculation of mist image obtains the image after mist elimination.
Optionally, described sky areas determines unit, specifically includes:
Grad computation subunit, has the Grad of each pixel in mist image described in calculating；
Bianry image determines subelement, for described Grad is labeled as 1 less than the pixel of first threshold, and described ladder
Angle value is labeled as 0 more than or equal to the pixel of first threshold, obtains bianry image；
New bianry image determines subelement, for using a round die block to corrode described bianry image, obtains new
Bianry image；
UNICOM region obtains subelement, for utilizing region growing algorithm that described new bianry image is processed,
To described new bianry image all of UNICOM region；
Mean flow rate computation subunit, has pixel corresponding on mist image for calculating each described UNICOM region described
The mean flow rate of point；
First sky areas determines subelement, for by described mean flow rate more than UNICOM's zone marker of Second Threshold being
Sky areas；
Second sky areas determines subelement, for being labeled as by the satisfied pixel imposed a condition in non-sky areas
Sky areas；
Holes filling subelement, for being filled with the leak in described sky areas.
Optionally, described air light value determines unit, specifically includes:
Ratio computation subunit, for calculate described sky areas account for whole described in have the ratio of mist image；
Ratio judgment sub-unit, is used for judging that whether described ratio is less than setting ratio；
First air light value computation subunit, for when described ratio is greater than equal to setting ratio, calculates described sky
The meansigma methods of the light intensity of pixel in dummy section, is labeled as described air light value by described meansigma methods；
Second air light value computation subunit, for when described ratio is less than setting ratio, has mist figure described in acquisition
As when brightness arranges from high to low in dark, arrange the pixel of front 0.1%, and the pixel of 0.1% exists before obtaining described row
The described light intensity value having in mist image, is labeled as described air light value by the largest light intensity angle value in described light intensity value.
Optionally, described absorbance determines unit, specifically includes:
Absorbance estimation subelement, for utilizing the absorbance having mist image described in the estimation of dark channel prior algorithm, obtains
Absorbance evaluation function；
Fixed transmittance rate value obtains subelement, is used for obtaining fixed transmittance rate value；
Reasonable transmittance determines subelement, for carrying out described fixed transmittance rate value with described absorbance evaluation function
Alpha mixes, and obtains rational absorbance；
Absorbance optimizes subelement, is used for utilizing Steerable filter algorithm that described rational absorbance is optimized process.
Optionally, described mist elimination image computing unit, specifically include:
Mist elimination image computation subunit, for according to formulaMist figure is had described in being calculated
As image J (x) after mist elimination, wherein, A is described air light value, t (x) be described in have the refractive index of mist image, I (x) is described
There is the observed strength of mist image.
The specific embodiment provided according to the present invention, the invention discloses techniques below effect: the present invention is having mist image
On be partitioned into the sky areas in mist image, exactly because sky areas has been split out, can be to sky areas
Process targetedly, simultaneously according to sky areas account for whole described in have the difference of ratio of mist image, use different sides
Method determines air light value, improves mist elimination precision, well achieves the mist elimination to sky areas, leads additionally, the present invention uses
To filtering, absorbance is optimized, reduces complexity, improve image processing speed.
Accompanying drawing explanation
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to institute in embodiment
The accompanying drawing used is needed to be briefly described, it should be apparent that, the accompanying drawing in describing below is only some enforcements of the present invention
Example, for those of ordinary skill in the art, on the premise of not paying creative work, it is also possible to according to these accompanying drawings
Obtain other accompanying drawing.
Fig. 1 is embodiment of the present invention image defogging method flow chart；
Fig. 2 is that embodiment of the present invention sky areas determines method flow diagram；
Fig. 3 is that embodiment of the present invention air light value determines method flow diagram；
Fig. 4 is that embodiment of the present invention image transmission rate determines method flow diagram；
The effect that Fig. 5 reaches for the method mist elimination using the present invention to provide the image containing sky areas is imitated with prior art
First comparison diagram of fruit；
Effect that Fig. 6 reaches for the method mist elimination providing the employing present invention of the image containing sky areas and prior art
Effect the second comparison diagram；
The effect that the method that Fig. 7 provides for the employing this method without sky areas reaches and prior art Contrast on effect
Figure；
Fig. 8 is embodiment of the present invention image mist elimination system structure schematic diagram.
Detailed description of the invention
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Describe, it is clear that described embodiment is only a part of embodiment of the present invention rather than whole embodiments wholely.Based on
Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under not making creative work premise
Embodiment, broadly falls into the scope of protection of the invention.
It is an object of the invention to provide a kind of can be to the image having the sky areas in mist image well to process
Defogging method and system, and, this defogging method and system have the reduction precision height of image, mist elimination complexity rate is low, process speed
Spend fast feature.
Understandable for enabling the above-mentioned purpose of the present invention, feature and advantage to become apparent from, real with concrete below in conjunction with the accompanying drawings
The present invention is further detailed explanation to execute mode.
Fig. 1 is embodiment of the present invention image defogging method flow chart, as it is shown in figure 1, the image mist elimination side that the present invention provides
Method step is as follows:
Step 101: obtain comprise sky areas have mist image；
Step 102: have the sky areas in mist image described in determining；
Step 103: set up atmospherical scattering model；
Step 104: according to described sky areas account for whole described in have atmospherical scattering model described in the ratio-dependent of mist image
In air light value, described air light value is the atmosphere light intensity of infinite point；
Step 105: according to described atmospherical scattering model determine described in have the absorbance of mist image；
Step 106: have the observed strength of mist image described in acquisition；Image is obtained from having mist image of photographing with camera
In the observed strength of each pixel；
Step 107: according to described air light value, described in have the observed strength of mist image and the described absorbance having mist image
It is calculated the image after mist elimination.
Fig. 2 is that embodiment of the present invention sky areas determines method flow diagram, as in figure 2 it is shown, determine the sky having in mist image
Dummy section specifically includes following steps:
Step 201: calculate and have the Grad of each pixel in mist image；
Step 202: less than the pixel of first threshold, described Grad is labeled as 1, described Grad is more than or equal to the
The pixel of one threshold value is labeled as 0, obtains bianry image；First threshold can take 0.02, and the place that gradient is the least represents herein
The most smooth.The position of sky areas is determined by the size of gradient；
Step 203: corrode described bianry image with a round die block, obtains new bianry image, this circle
Radius is setting value, and this setting value takes 3cm；
Step 204: utilize region growing algorithm that described new bianry image is processed, obtain described new binary map
As all of UNICOM region；
Step 205: calculate each described UNICOM region in the described mean flow rate having pixel corresponding on mist image；
Step 206: be sky areas more than UNICOM's zone marker of Second Threshold by described mean flow rate, Second Threshold can
To go 0.81；
Step 207: the satisfied pixel imposed a condition in non-sky areas is labeled as sky areas, mentioned here
Impose a condition for: 1) take 3cm with the space length of sky areas less than R, R；2) intensity difference with sky areas pixel is less than
θ, θ takes 0.02 here；
Step 208: be filled with the leak in described sky areas, i.e. fills hole.
Fig. 3 is that embodiment of the present invention air light value determines method flow diagram, as it is shown on figure 3, the ginseng in atmospherical scattering model
The method step that specifically determines of number air light value is:
Step 301: calculate described sky areas account for whole described in have the ratio of mist image；
Step 302: judge whether described ratio is less than setting ratio, setting ratio can take 5%；
Step 303: if sky areas account for whole described in have the ratio of mist image more than or equal to 5%, then calculate described
The meansigma methods of the light intensity of pixel in sky areas, is labeled as described air light value by described meansigma methods；
Step 304: if sky areas account for whole described in have the ratio of mist image less than 5%, then obtain described in have mist figure
As when brightness arranges from high to low in dark, arrange the pixel of front 0.1%, and the pixel of 0.1% exists before obtaining described row
The described light intensity value having in mist image, is labeled as described air light value by the largest light intensity angle value in described light intensity value.
In conventional atmosphere light value calculating method, in employing picture, the pixel value of maximum intensity is as air light value, but
Being when there is other light sources in image, such as light, now, the air light value using the method to obtain is significantly larger than the big of reality
Gas light value, brings the biggest error for the process of mist elimination, and the present invention according to sky areas account for whole described in have the ratio of mist image
Example determines air light value obtaining value method, if ratio is less than 5%, then calculates atmosphere light with the method for what traditional bright proposition of happy
Value, i.e. obtain described in when having that in mist image dark channel, brightness arranges from high to low, arrange the pixel of front 0.1%, and described in acquisition
Arrange the pixel of front 0.1% in the described light intensity value having in mist image, by the largest light intensity angle value mark in described light intensity value
It is designated as described air light value；If ratio more than or equal to 5%, then calculates the average of the light intensity of pixel in described sky areas
Value, is labeled as described air light value by described meansigma methods.Consequently, it is possible to the value of air light value is more accurate, be conducive to carrying
The precision of hi-vision mist elimination.
Fig. 4 is that embodiment of the present invention image transmission rate determines method flow diagram, as shown in Figure 4, according to atmospherical scattering model
Expression formula calculate the specifically comprising the following steps that of absorbance
Step 401: utilize the absorbance having mist image described in the estimation of dark channel prior algorithm, obtains absorbance estimation letter
Number；
Channel prior algorithm is that a kind of simply and effectively method of HE proposition in 2009 is to estimate t (x).Dark channel prior
Assuming that in most non-sky areas, the value of some pixels at least Color Channel has the lowest intensity.So,
For without mist image J (x), it is defined as follows:
Here, what Ω (x) represented is the regional area centered by x.JdarkWhat x () represented is the dark of J (x),
Dark at each pixel without mist image (except sky areas) levels off to 0. this priori, in 09 year by what
Happy Ming doctor demonstrate.But having mist image, owing to the addition of atmosphere light, the intensity of dark does not reaccees 0, it is assumed that with x
Centered by the pixel of regional area have identical absorbance t.As follows by greasy weather imaging model equation above:
I (x)=J (x) t (x)+A (1-t (x))
Above formula minimizes operation, and equation becomes as follows:
Take to minimize operation divided by A and at three kinds of Color Channels to above formula two ends, have:
According to dark, Wo Menyou
So obtaining
So we have tried to achieve absorbance t (x) the most roughly.
Step 402: obtain fixed transmittance rate value；
Step 403: with described absorbance evaluation function, described fixed transmittance rate value is carried out Alpha and mixes, obtains rationally
Absorbance；
Absorbance owing to estimating in step 401 is only applicable to non-sky portion, it is understood that dark channel prior is not suitable for
Containing sky areas, so we use fixing absorbance t in empty region, skysky.But in the operation that I am above-mentioned,
The sky areas obtained is a secondary masking-out figure, and certain point not necessarily fully belongs to sky or is not belonging to sky completely, because of
This, the most just can be according to this value tskyValue absorbance t (x) that+dark is tried to achieve carries out Alpha mixing, is more closed
Absorbance t of reason, as follows:
T=(tsky*Sky[Y]+t(x)*(255-Sky[Y]))/255
Wherein tskyBeing a fixing transmittance values, we take 0.3, in above formula, ifI.e. complete
Belong to sky, then the absorbance of this point is fixed value tskyIf,It is not belonging to sky the most completely, the value of calculating formula
Meet dark channel prior, do not affect normal mist elimination.If falling between, also can well obtain a rational absorbance.
Step 404: utilize Steerable filter algorithm that described rational absorbance is optimized process.
Steerable filter is the wave filter needing guiding figure.Filtering comprises guiding figure I, input picture p and (i.e. needs filter
The image of ripple), and output image q.Wherein I Yu p can be same image.The filtering output of each pixel can be expressed as
One average weighted form:
Wherein i, j are pixel index, and W is referred to as filtering core.The filtering core of bilateral filtering is:
The critical assumptions guiding filtering are exactly the Local Linear Model between guiding figure I and output figure q:
A, b are linear coefficient, and are constant in local window k.For determining the linear coefficient in above formula, and meet
The difference making q Yu p is minimum, is converted into optimization problem:
Solving of above formula can utilize linear regression:
Here, μkWithRepresent that I is at local window wkIn average and variance.| ω | is the pixel count in window, pkTable
Show that p is at window wkIn average.As a askedkAnd bkRear:
Wherein,
After absorbance is optimized, according to formulaAfter calculating mist image mist elimination
Image J (x), wherein, A is described air light value, t (x) be described in have the refractive index of mist image, I (x) be described in have mist image
Observed strength.
Picture J (x) after mist elimination is a little dim, and this situation is typically caused by the atmosphere light having mist image, air
Light is insufficient by time of exposure.In order to solve this problem, J (x) image can strengthen in the following manner:
First J (x) is transformed into HSV colour model；The contrast of V component is strengthened with CLAHE algorithm；The HSV figure strengthened
As going back to RGB color space to obtain final restoration result.
The effect that Fig. 5 reaches for the method mist elimination using this method to provide the image containing sky areas is imitated with prior art
Really comparison diagram, as it is shown in figure 5,5 (a) has mist image for shoot, 5 (b) is that the dark mist elimination algorithm using what happy bright obtains
Experimental result, 5 (c) is the experimental result picture that obtains of defogging method using the present invention to provide.
The effect that the method mist elimination that Fig. 6 provides for employing this method that another width contains the image of sky areas reaches is with existing
Having technique effect comparison diagram, as shown in Figure 6,6 (a) has mist image for shooting, and 6 (b) is the dark mist elimination using what happy bright
The experimental result that algorithm obtains, the experimental result picture that the defogging method that 6 (c) provides for using the present invention obtains.
The effect that the method that Fig. 7 provides for the employing this method without sky areas reaches and prior art Contrast on effect
Figure, as it is shown in fig. 7,7 (a) has mist image for shoot, 7 (b) is the experiment that the dark mist elimination algorithm using what happy bright obtains
As a result, the experimental result picture that the defogging method that 7 (c) provides for using the present invention obtains.
It will be seen that the defogging method that the present invention provides is processing sky areas part substantially than what happy from Fig. 5, Fig. 6
Bright algorithm is good, and for the Fig. 7 artwork without sky areas, it can be seen that the method using the present invention to provide obtains experiment knot
The result of fruit and the method for He Kaiming almost one touch as, it is seen then that the defogging method immunity that the present invention provides is higher.
The image defogging method that the present invention provides has been partitioned into the sky areas in mist image, just having on mist image
Because sky areas having been split out, sky areas can be processed targetedly, accounting for according to sky areas simultaneously
There is the difference of the ratio of mist image described in whole, use different methods to determine air light value, improve mist elimination precision, well
Achieve the mist elimination to sky areas, additionally, the present invention uses Steerable filter to be optimized absorbance, reduce complexity,
Improve image processing speed.
For reaching above-mentioned purpose, present invention also offers a kind of image mist elimination system, Fig. 5 is that embodiment of the present invention image goes
Mist system structure schematic diagram, as shown in Figure 8, this system includes:
Image acquisition unit 801, for obtain comprise sky areas have mist image；
Sky areas determines unit 802, has the sky areas in mist image described in determining；
Atmospherical scattering model sets up unit 803, is used for setting up atmospherical scattering model；
Air light value determines unit 804, for according to described sky areas account for whole described in have the ratio-dependent of mist image
Air light value in described atmospherical scattering model, described air light value is the atmosphere light intensity of infinite point；
Absorbance determines unit 805, has the absorbance of mist image described in determining according to described atmospherical scattering model；
Observed strength acquiring unit 806, is used for the observed strength having mist image described in obtaining；
Mist elimination image computing unit 807, for according to described air light value, described in have the observed strength of mist image and described
The transmittance calculation having mist image obtains the image after mist elimination.
Wherein, sky areas determines unit 502, specifically includes:
Grad computation subunit, has the Grad of each pixel in mist image described in calculating；
Bianry image determines subelement, for described Grad is labeled as 1 less than the pixel of first threshold, and described ladder
Angle value is labeled as 0 more than or equal to the pixel of first threshold, obtains bianry image；
New bianry image determines subelement, for using a round die block to corrode described bianry image, obtains new
Bianry image；
UNICOM region obtains subelement, for utilizing region growing algorithm that described new bianry image is processed,
To described new bianry image all of UNICOM region；
Mean flow rate computation subunit, has pixel corresponding on mist image for calculating each described UNICOM region described
The mean flow rate of point；
First sky areas determines subelement, for by described mean flow rate more than UNICOM's zone marker of Second Threshold being
Sky areas；
Second sky areas determines subelement, for being labeled as by the satisfied pixel imposed a condition in non-sky areas
Sky areas；
Holes filling subelement, for being filled with the leak in described sky areas.
Air light value determines unit 504, specifically includes:
Ratio computation subunit, for calculate described sky areas account for whole described in have the ratio of mist image；
Ratio judgment sub-unit, is used for judging that whether described ratio is less than setting ratio；
First air light value computation subunit, for when described ratio is greater than equal to setting ratio, calculates described sky
The meansigma methods of the light intensity of pixel in dummy section, is labeled as described air light value by described meansigma methods；
Second air light value computation subunit, for when described ratio is less than setting ratio, has mist figure described in acquisition
As when brightness arranges from high to low in dark, arrange the pixel of front 0.1%, and the pixel of 0.1% exists before obtaining described row
The described light intensity value having in mist image, is labeled as described air light value by the largest light intensity angle value in described light intensity value.
Absorbance determines unit 505, specifically includes:
Absorbance estimation subelement, for utilizing the absorbance having mist image described in the estimation of dark channel prior algorithm, obtains
Absorbance evaluation function；
Fixed transmittance rate value obtains subelement, is used for obtaining fixed transmittance rate value；
Reasonable transmittance determines subelement, for carrying out described fixed transmittance rate value with described absorbance evaluation function
Alpha mixes, and obtains rational absorbance；
Absorbance optimizes subelement, is used for utilizing Steerable filter algorithm that described rational absorbance is optimized process.
Mist elimination image computing unit 507, specifically includes:
Mist elimination image computation subunit, for according to formulaMist figure is had described in being calculated
As image J (x) after mist elimination, wherein, A is described air light value, t (x) be described in have the refractive index of mist image, I (x) is described
There is the observed strength of mist image.
The image mist elimination system that the present invention provides has been partitioned into the sky areas in mist image, just having on mist image
Because sky areas having been split out, sky areas can be processed targetedly, accounting for according to sky areas simultaneously
There is the difference of the ratio of mist image described in whole, use different methods to determine air light value, improve mist elimination precision, well
Achieve the mist elimination to sky areas, additionally, the present invention uses Steerable filter to be optimized absorbance, reduce complexity,
Improve image processing speed.
In this specification, each embodiment uses the mode gone forward one by one to describe, and what each embodiment stressed is and other
The difference of embodiment, between each embodiment, identical similar portion sees mutually.For system disclosed in embodiment
For, owing to it corresponds to the method disclosed in Example, so describe is fairly simple, relevant part sees method part and says
Bright.
Principle and the embodiment of the present invention are set forth by specific case used herein, saying of above example
Bright method and the core concept thereof being only intended to help to understand the present invention；Simultaneously for one of ordinary skill in the art, foundation
The thought of the present invention, the most all will change.In sum, this specification content is not
It is interpreted as limitation of the present invention.
Claims (10)
1. an image defogging method, it is characterised in that including:
Obtain comprise sky areas have mist image；
The sky areas in mist image is had described in determining；
Set up atmospherical scattering model；
According to described sky areas account for whole described in have the air light value in atmospherical scattering model described in the ratio-dependent of mist image,
Described air light value is the atmosphere light intensity of infinite point；
According to described atmospherical scattering model determine described in have the absorbance of mist image；
The observed strength of mist image is had described in acquisition；
According to described air light value, described in have the observed strength of mist image and the described transmittance calculation having mist image to obtain mist elimination
After image.
Method the most according to claim 1, it is characterised in that described determine described in have the sky areas in mist image, tool
Body includes:
The Grad of each pixel in mist image is had described in calculating；
Less than the pixel of first threshold, described Grad is labeled as 1, and described Grad is more than or equal to the pixel of first threshold
Point is labeled as 0, obtains bianry image；
With a round die block, described bianry image is corroded, obtain new bianry image；
Utilize region growing algorithm that described new bianry image is processed, obtain the described all of UNICOM of new bianry image
Region；
Calculate each described UNICOM region in the described mean flow rate having pixel corresponding on mist image；
It is sky areas by UNICOM's zone marker that described mean flow rate is more than Second Threshold；
The satisfied pixel imposed a condition in non-sky areas is labeled as sky areas；
Leak in described sky areas is filled with.
Method the most according to claim 1, it is characterised in that described according to sky areas account for whole described in have mist image
Air light value in atmospherical scattering model described in ratio-dependent, specifically includes:
Calculate described sky areas account for whole described in have the ratio of mist image；
Judge that whether described ratio is less than setting ratio；
If it is not, then calculate the meansigma methods of the light intensity of pixel in described sky areas, described meansigma methods is labeled as described
Air light value；
During if it is, have described in Huo Quing that in mist image dark channel, brightness arranges from high to low, arrange the pixel of front 0.1%, and
Before obtaining described row, the pixel of 0.1% is in the described light intensity value having in mist image, by the maximum light in described light intensity value
Intensity level is labeled as described air light value.
Method the most according to claim 1, it is characterised in that described according to described atmospherical scattering model determine described in have mist
The absorbance of image, specifically includes:
Utilize the absorbance having mist image described in the estimation of dark channel prior algorithm, obtain absorbance evaluation function；
Obtain fixed transmittance rate value；
With described absorbance evaluation function, described fixed transmittance rate value is carried out Alpha mix, obtain rational absorbance；
Utilize Steerable filter algorithm that described rational absorbance is optimized process.
Method the most according to claim 1, it is characterised in that described according to described air light value, described in have mist image
Observed strength and the described transmittance calculation having mist image obtain the image after mist elimination, specifically include:
According to formulaHaving image J (x) after mist image mist elimination described in being calculated, wherein, A is institute
State air light value, t (x) be described in have the refractive index of mist image, I (x) be described in have the observed strength of mist image.
6. an image mist elimination system, it is characterised in that described system includes:
Image acquisition unit, for obtain comprise sky areas have mist image；
Sky areas determines unit, has the sky areas in mist image described in determining；
Atmospherical scattering model sets up unit, is used for setting up atmospherical scattering model；
Air light value determines unit, for according to described sky areas account for whole described in have air described in the ratio-dependent of mist image
Air light value in scattering model, described air light value is the atmosphere light intensity of infinite point；
Absorbance determines unit, has the absorbance of mist image described in determining according to described atmospherical scattering model；
Observed strength acquiring unit, is used for the observed strength having mist image described in obtaining；
Mist elimination image computing unit, for according to described air light value, described in have the observed strength of mist image and described have mist figure
The transmittance calculation of picture obtains the image after mist elimination.
System the most according to claim 6, it is characterised in that described sky areas determines unit, specifically includes:
Grad computation subunit, has the Grad of each pixel in mist image described in calculating；
Bianry image determines subelement, for described Grad is labeled as 1 less than the pixel of first threshold, and described Grad
It is labeled as 0 more than or equal to the pixel of first threshold, obtains bianry image；
New bianry image determines subelement, for using a round die block to corrode described bianry image, obtains new two
Value image；
UNICOM region obtains subelement, for utilizing region growing algorithm to process described new bianry image, obtains institute
State new bianry image all of UNICOM region；
Mean flow rate computation subunit, has on mist image corresponding pixel for calculating each described UNICOM region described
Mean flow rate；
First sky areas determines subelement, is sky for described mean flow rate is more than UNICOM's zone marker of Second Threshold
Region；
Second sky areas determines subelement, for the satisfied pixel imposed a condition in non-sky areas is labeled as sky
Region；
Holes filling subelement, for being filled with the leak in described sky areas.
System the most according to claim 6, it is characterised in that described air light value determines unit, specifically includes:
Ratio computation subunit, for calculate described sky areas account for whole described in have the ratio of mist image；
Ratio judgment sub-unit, is used for judging that whether described ratio is less than setting ratio；
First air light value computation subunit, for when described ratio is greater than equal to setting ratio, calculates dead zone, described sky
The meansigma methods of the light intensity of pixel in territory, is labeled as described air light value by described meansigma methods；
Second air light value computation subunit, for when described ratio is less than setting ratio, has mist image dark described in acquisition
When brightness arranges from high to low in passage, arrange the pixel of front 0.1%, and obtain before described row the pixel of 0.1% described
There is the light intensity value in mist image, the largest light intensity angle value in described light intensity value is labeled as described air light value.
System the most according to claim 6, it is characterised in that described absorbance determines unit, specifically includes:
Absorbance estimation subelement, for utilizing the absorbance having mist image described in the estimation of dark channel prior algorithm, obtains transmission
Rate evaluation function；
Fixed transmittance rate value obtains subelement, is used for obtaining fixed transmittance rate value；
Reasonable transmittance determines subelement, for described fixed transmittance rate value and described absorbance evaluation function are carried out Alpha
Mixing, obtains rational absorbance；
Absorbance optimizes subelement, is used for utilizing Steerable filter algorithm that described rational absorbance is optimized process.
System the most according to claim 6, it is characterised in that described mist elimination image computing unit, specifically includes:
Mist elimination image computation subunit, for according to formulaMist image mist elimination is had described in being calculated
After image J (x), wherein, A is described air light value, t (x) be described in have the refractive index of mist image, I (x) be described in have mist figure
The observed strength of picture.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610748954.XA CN106127715A (en) | 2016-08-29 | 2016-08-29 | A kind of image defogging method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610748954.XA CN106127715A (en) | 2016-08-29 | 2016-08-29 | A kind of image defogging method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106127715A true CN106127715A (en) | 2016-11-16 |
Family
ID=57272059
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610748954.XA Withdrawn CN106127715A (en) | 2016-08-29 | 2016-08-29 | A kind of image defogging method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106127715A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106780380A (en) * | 2016-12-09 | 2017-05-31 | 电子科技大学 | A kind of image defogging method and system |
| CN107240075A (en) * | 2017-05-27 | 2017-10-10 | 上海斐讯数据通信技术有限公司 | A kind of haze image enhancing processing method and system |
| CN107403421A (en) * | 2017-08-10 | 2017-11-28 | 杭州联吉技术有限公司 | A kind of image defogging method, storage medium and terminal device |
| CN107644407A (en) * | 2017-09-26 | 2018-01-30 | 成都国翼电子技术有限公司 | A kind of thin cloud minimizing technology of Aerial Images based on man-machine interactively |
| CN108257094A (en) * | 2016-12-29 | 2018-07-06 | 广东中科遥感技术有限公司 | The quick minimizing technology of remote sensing image mist based on dark |
| CN108460778A (en) * | 2017-12-26 | 2018-08-28 | 浙江工商大学 | Class sky areas detection towards foggy image and localization method |
| CN108596856A (en) * | 2018-05-07 | 2018-09-28 | 北京环境特性研究所 | A kind of image defogging method and device |
| WO2019019890A1 (en) * | 2017-07-27 | 2019-01-31 | Oppo广东移动通信有限公司 | Image processing method, computer equipment, and computer-readable storage medium |
| CN109523480A (en) * | 2018-11-12 | 2019-03-26 | 上海海事大学 | A kind of defogging method, device, computer storage medium and the terminal of sea fog image |
| CN109978789A (en) * | 2019-03-26 | 2019-07-05 | 电子科技大学 | A kind of image enchancing method based on Retinex algorithm and guiding filtering |
| CN110175967A (en) * | 2019-06-05 | 2019-08-27 | 海南大学 | Image defogging processing method, system, computer equipment and storage medium |
| CN110738624A (en) * | 2019-10-18 | 2020-01-31 | 电子科技大学 | area self-adaptive image defogging system and method |
| CN111145105A (en) * | 2019-12-04 | 2020-05-12 | 广东省新一代通信与网络创新研究院 | Image rapid defogging method and device, terminal and storage medium |
-
2016
- 2016-08-29 CN CN201610748954.XA patent/CN106127715A/en not_active Withdrawn
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106780380A (en) * | 2016-12-09 | 2017-05-31 | 电子科技大学 | A kind of image defogging method and system |
| CN108257094A (en) * | 2016-12-29 | 2018-07-06 | 广东中科遥感技术有限公司 | The quick minimizing technology of remote sensing image mist based on dark |
| CN107240075A (en) * | 2017-05-27 | 2017-10-10 | 上海斐讯数据通信技术有限公司 | A kind of haze image enhancing processing method and system |
| WO2019019890A1 (en) * | 2017-07-27 | 2019-01-31 | Oppo广东移动通信有限公司 | Image processing method, computer equipment, and computer-readable storage medium |
| CN107403421A (en) * | 2017-08-10 | 2017-11-28 | 杭州联吉技术有限公司 | A kind of image defogging method, storage medium and terminal device |
| CN107644407A (en) * | 2017-09-26 | 2018-01-30 | 成都国翼电子技术有限公司 | A kind of thin cloud minimizing technology of Aerial Images based on man-machine interactively |
| CN108460778B (en) * | 2017-12-26 | 2021-07-30 | 浙江工商大学 | Sky-like region detection and positioning method facing foggy image |
| CN108460778A (en) * | 2017-12-26 | 2018-08-28 | 浙江工商大学 | Class sky areas detection towards foggy image and localization method |
| CN108596856A (en) * | 2018-05-07 | 2018-09-28 | 北京环境特性研究所 | A kind of image defogging method and device |
| CN109523480A (en) * | 2018-11-12 | 2019-03-26 | 上海海事大学 | A kind of defogging method, device, computer storage medium and the terminal of sea fog image |
| CN109523480B (en) * | 2018-11-12 | 2022-05-06 | 上海海事大学 | Defogging method and device for sea fog image, computer storage medium and terminal |
| CN109978789A (en) * | 2019-03-26 | 2019-07-05 | 电子科技大学 | A kind of image enchancing method based on Retinex algorithm and guiding filtering |
| CN110175967A (en) * | 2019-06-05 | 2019-08-27 | 海南大学 | Image defogging processing method, system, computer equipment and storage medium |
| CN110738624A (en) * | 2019-10-18 | 2020-01-31 | 电子科技大学 | area self-adaptive image defogging system and method |
| CN110738624B (en) * | 2019-10-18 | 2022-02-01 | 电子科技大学 | Area-adaptive image defogging system and method |
| CN111145105A (en) * | 2019-12-04 | 2020-05-12 | 广东省新一代通信与网络创新研究院 | Image rapid defogging method and device, terminal and storage medium |
| CN111145105B (en) * | 2019-12-04 | 2020-09-01 | 广东省新一代通信与网络创新研究院 | Image rapid defogging method and device, terminal and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106127715A (en) | A kind of image defogging method and system | |
| CN106780380A (en) | A kind of image defogging method and system | |
| CN107301623B (en) | Traffic image defogging method and system based on dark channel and image segmentation | |
| CN103218778B (en) | The disposal route of a kind of image and video and device | |
| CN105354806B (en) | Rapid defogging method and system based on dark | |
| CN105574827B (en) | A kind of method, apparatus of image defogging | |
| CN103198459B (en) | Haze image rapid haze removal method | |
| EP3076367A1 (en) | Method for road detection from one image | |
| CN105023256B (en) | A kind of image defogging method and system | |
| CN104299192B (en) | A kind of single image to the fog method based on atmospheric light scattering physical model | |
| CN106204567A (en) | A kind of natural background video matting method | |
| CN102982513B (en) | A kind of adapting to image defogging method capable based on texture | |
| CN102831591A (en) | Gaussian filter-based real-time defogging method for single image | |
| CN106846263A (en) | The image defogging method being immunized based on fusion passage and to sky | |
| CN103914820B (en) | Image haze removal method and system based on image layer enhancement | |
| CN104867128B (en) | Image blurring detection method and device | |
| CN107451966A (en) | A kind of real-time video defogging method realized using gray-scale map guiding filtering | |
| CN102663694A (en) | Digital fog effect filter method based on dark primary color channel prior principle | |
| CN103020914A (en) | Rapid image defogging method based on spatial continuity principle | |
| CN103578083A (en) | Single image defogging method based on joint mean shift | |
| CN105913390A (en) | Image defogging method and system | |
| CN106157270A (en) | A kind of single image rapid defogging method and system | |
| CN105976337A (en) | Image defogging method based on filtering guiding via medians | |
| CN106657948A (en) | low illumination level Bayer image enhancing method and enhancing device | |
| CN104331867B (en) | The method, device and mobile terminal of image defogging |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C04 | Withdrawal of patent application after publication (patent law 2001) | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20161116 |