CN106780385B - A kind of fog-degraded image clarification method based on turbulent flow infra-red radiation model - Google Patents
A kind of fog-degraded image clarification method based on turbulent flow infra-red radiation model Download PDFInfo
- Publication number
- CN106780385B CN106780385B CN201611167482.5A CN201611167482A CN106780385B CN 106780385 B CN106780385 B CN 106780385B CN 201611167482 A CN201611167482 A CN 201611167482A CN 106780385 B CN106780385 B CN 106780385B
- Authority
- CN
- China
- Prior art keywords
- image
- frequency domain
- transfer function
- atmospheriacally
- indicate
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 49
- 230000005855 radiation Effects 0.000 title claims abstract description 13
- 238000005352 clarification Methods 0.000 title claims abstract description 6
- 238000012546 transfer Methods 0.000 claims abstract description 40
- 238000012545 processing Methods 0.000 claims abstract description 8
- 230000009466 transformation Effects 0.000 claims abstract description 7
- 230000015556 catabolic process Effects 0.000 claims description 16
- 238000006731 degradation reaction Methods 0.000 claims description 16
- 230000005540 biological transmission Effects 0.000 claims description 11
- 230000001052 transient effect Effects 0.000 claims description 6
- 230000000007 visual effect Effects 0.000 claims description 5
- 230000007850 degeneration Effects 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 101001031492 Methanothermococcus thermolithotrophicus Short-type peptidyl-prolyl cis-trans isomerase Proteins 0.000 claims description 3
- 230000003287 optical effect Effects 0.000 claims description 3
- 238000003384 imaging method Methods 0.000 claims description 2
- 238000005259 measurement Methods 0.000 claims 1
- 230000001869 rapid Effects 0.000 claims 1
- 230000008569 process Effects 0.000 abstract description 15
- 238000011084 recovery Methods 0.000 description 7
- 238000012360 testing method Methods 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 5
- 239000003595 mist Substances 0.000 description 5
- 238000001514 detection method Methods 0.000 description 4
- 238000001914 filtration Methods 0.000 description 4
- 238000003707 image sharpening Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 239000002245 particle Substances 0.000 description 2
- 230000002265 prevention Effects 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 238000012935 Averaging Methods 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000011109 contamination Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000004438 eyesight Effects 0.000 description 1
- 238000003331 infrared imaging Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000004297 night vision Effects 0.000 description 1
- 230000035515 penetration Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G06T5/73—
-
- 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
- G06T2207/10048—Infrared 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
- G06T2207/20048—Transform domain processing
- G06T2207/20056—Discrete and fast Fourier transform, [DFT, FFT]
Landscapes
- Image Processing (AREA)
Abstract
The present invention relates to a kind of fog-degraded image clarification methods based on turbulent flow infra-red radiation model, it include: that Fourier transformation is carried out to image, the degraded image in frequency domain is obtained, according to the degraded image in frequency domain and then obtains atmospheriacally modulation transfer function (MTF)；Calculate refractive index structures coefficient unknown in atmospheriacally modulation transfer function；Degraded image in frequency domain is filtered out into atmospheriacally modulation transfer function (MTF) in frequency domain；The frequency domain image after atmospheriacally modulation transfer function will be filtered out and carry out inverse Fourier transform, obtain the image of sharpening processing.The present invention is not under the premise of increasing hardware cost, it is analyzed from the inverse process for the IR image enhancement that degrades, restores atmospheric scattering and act on the influence that decay to picture contrast, to solve the problems, such as that the degrade sharpening of infrared image of greasy weather is handled, calculating is simpler, and efficiency is higher.
Description
Technical field
The invention belongs to field of image processings, and in particular to a kind of fog-degraded image based on turbulent flow infra-red radiation model
Clarification method.
Background technique
Infrared imaging has outstanding atmosphere penetration capacity and night vision function, very sensitive to the variation of temperature, thus in army
Thing national defence, remote sensing, Public Hygienic Prevention, Public Hygienic Prevention, chemical substance detection and identification etc. obtain extensively
Application.However suspended particles radius is larger in greasy weather atmosphere, the scattering process of atmospheric particles make on infrared image originally compared with
Low gray value is reinforced, and higher gray value is weakened, so that the grey level distribution of image is excessively concentrated, is caused pair
It is more poor than degree；On the other hand, in the mapping process in image from three-dimensional space to two-dimensional surface, it is lost depth information, is caused
The edge contour that greasy weather acquires image has ambiguity.
The infrared image of greasy weather shooting, the comparision contents of water vapour are big in atmosphere and system hardware and software sensitivity etc. is asked
Topic, the degenerate problem that the infrared image photographed is commonly present noise, fuzzy and grain details are lost etc., to accurately identify, with
Track target object etc. is permitted various applications and causes difficulty.Generally speaking, solve the problems, such as that the greasy weather degrades the sharpening of infrared image,
Three kinds of methods can be taken:
(1) from the angle of hardware, letter is realized using the method that Multi-sensor Fusion and infrared focal plane array correct
The complementation of breath and correction, such as can in conjunction with common two point correction and Supplements using radar, CCD and laser sensing equipment
To realize higher target detection rate and lower error rate.However, the laser or ccd sensor of view-based access control model are in this weather
Under the conditions of effective monitoring distance can greatly shorten, can not provide effective confirmation information for radar sensor, and based on infrared
Sensor thermal signal can only be realized and detect, effective detection can not be realized to the non-radiated signal in scene.Therefore, this
There are the limitations in some applications in the case where there is mist weather condition for strategy, and increase expensive hardware cost.
(2) concrete reason of image deterioration is not considered, and merely from the angle of image procossing, the greasy weather degrades infrared figure
The problem of being exactly picture superposition on the sharpening question essence of picture；
(3) the inverse process analysis generated from degraded image is then the decaying shadow by atmospheric scattering effect to picture contrast
The process of sound is modeled, and finally solves the problems, such as that the sharpening of degraded image restores.
Since image enchancing method does not account for the physical process that Misty Image degrades, the reason of cannot degenerating for image
Try to compensate, therefore visual effect can only be improved to a certain extent.In recent years, some scholars both domestic and external are based on atmosphere
Scattering theory carries out deep analysis to the image degradation mechanism under severe weather conditions, proposes some based on degradation model
Fog-degraded image sharpening restored method.The research of these methods is concentrated mainly on the following aspects:
(1) assume restored method known to scene each point depth information.Although above-mentioned such methods do not need precognition weather
Information is believed however, but needing to obtain accurate scene depth using hardware devices such as expensive radar or range sensors
Breath, therefore limit the extensive use of algorithm in practice.
(2) image auxiliary information is combined to realize the restored method of depth extraction.Some algorithms pass through user's given scenario
Depth capacity and minimum-depth are obtained the depth information of scene each point using the method for linear interpolation, in some cases may be used
Feedback can not be made to scene depth abrupt information, in addition, needing user continuously to change after determining the depth information of each point
Atmospheric scattering coefficient determines the best restoration result of a visual effect, the excessive subjectivity for relying on people.
(3) restored method based on classical degradation model point spread function.Since premise with this method is desirable
Known point spread function, image contamination process and mechanism due to caused by mist in practice are sufficiently complex, the greasy weather of various concentration
Gas process is difficult to be expressed with unified point diffusion model, can not generally predict the point spread function of image degeneration, therefore also limits
The practical application of this method.
In addition, there are also the sharpening restored method of some images, such as the greasy weather visibility for combining international lighting conference to provide
With the threshold value relationship of human eye vision, the contrast of the visibility on greasy weather road surface and image is set up and is contacted, realized and atmosphere is dissipated
Seeking for coefficient is penetrated, and scene is calculated by the Euclidean distance function that the depth value of scene each point is modeled as on the plane of delineation
The depth of point, realizes the contrast real-time recovery and target detection of dynamic scene.Poisson's equation is constructed based on atmospherical scattering model,
The fusion that the degraded image of mist is seen as to prospect (mist) and background (clear image) is attempted, solution greasy weather scene recovery is inquired into and asks
The new way of topic.These methods have dependence to the depth information of image scene, and part needs excessive manual intervention.
Summary of the invention
In order to improve the quality of degraded image under the premise of not increasing hardware cost, from the inverse process of degraded image generation
Analysis, establishes atmospheric radiation transmission and corresponding atmospheriacally modulation transfer function (MTF), by filtering out atmosphere modulation in frequency domain
Transmission function, complete the greasy weather degrade infrared image sharpening processing.
Technical solution proposed by the present invention is as follows:
A kind of fog-degraded image clarification method based on turbulent flow infra-red radiation model, the method includes walking as follows
It is rapid:
Step 1: greasy weather degeneration infrared image is obtained；
Step 2: Fourier transformation is carried out to the described image in step 1, obtains the degraded image in frequency domain, according to frequency
Degraded image in domain obtains atmospheriacally modulation transfer function in turn；
Step 3: the refractive index structures coefficient of atmospheriacally modulation transfer function described in step 2 is calculated；
Step 4: the degraded image in the frequency domain in step 2 is filtered out into atmospheriacally modulation transfer function in frequency domain；
Step 5: the degraded image in the frequency domain after atmospheriacally modulation transfer function will be filtered out in step 4, and to carry out Fourier inverse
Transformation obtains the image of sharpening processing；
Step 6: output sharpening treated image.
Further, the step 2: Fourier transformation is carried out to the described image in step 1, obtains moving back in frequency domain
Change image, according to the degraded image in frequency domain and then obtains atmospheriacally modulation transfer function；Specifically:
On the basis of turbulent atmosphere model, the degradation model of image in airspace is analyzed, and then obtains modulation transfer function；
Wherein, the degradation model of image may be expressed as: in airspace
G (x, y)=f (x, y) * h (x, y)+n (x, y) (1)
Wherein, g (x, y) is degraded image, and f (x, y) is original image, and h (x, y) is turbulent flow transient state point spread function, n
(x, y) is noise item.
The degradation model of image may be expressed as: in a frequency domain accordingly
G (u, v)=F (u, v) * H (u, v)+N (u, v) (2)
Wherein, G (u, v) is the degraded image in frequency domain, and F (u, v) is the original image in frequency domain, and H (u, v) is in frequency domain
Turbulent flow transient state point spread function, N (u, v) be frequency domain in noise item.
Further, H (u, v) is the frequency-domain expression that h (x, y) is obtained by Fourier transform, also referred to as optical delivery
Function can indicate are as follows: H=| H | eiΦ, wherein | H | indicate amplitude, Φ is phase, to amplitude | H | make normalized, makes
The amplitude for obtaining H (0,0) at zero frequency is 1, then the amplitude after this normalization is referred to as atmospheriacally modulation transfer function MTF, it may be assumed that
MTF=| H |/K (3)
Wherein: K is amplitude of the H in zero frequency；
Image degradation model based on MTF is obtained by formula (1) (2) (3) are as follows:
G (u, v)=F (u, v) MTFKeiΦ+N(u,v) (4)
Assuming that atmospheriacally modulation transfer function be it is isotropic, then can ignore the influence of phase Φ, even Φ=0, then
F (u, v)=(G (u, v)-N (u, v))/(MTFK) (5)
Further, expression formula of MTF under the conditions of short exposure are as follows:
Atmospheriacally modulation transfer function in the case of short exposure may be expressed as:
In formula, ν indicates angle spatial frequency,Indicate that refractive index structures coefficient, λ indicate that beam wavelength, R indicate transmission distance
It is empirical coefficient from, μ, μ=1 when short distance, μ=0.5 when remote, D indicate imager bore dia.
Further, expression formula of the MTF under long conditions of exposure are as follows:
Atmospheriacally modulation transfer function under long exposure status may be expressed as:
In formula, ν indicates angle spatial frequency,Indicate that refractive index structures coefficient, λ indicate that beam wavelength, R indicate transmission distance
From.
Further, the step 3: the refractive index structures coefficient of atmospheriacally modulation transfer function described in step 2 is calculated；
Specifically:
Step 3.1: calculating image level gradient IXWith vertical gradient IY；
Step 3.2: each picture element of image (m, n) is calculatedAnd it chooses gradient value and is greater than centainly
The pixel of Grads threshold alternately pixel；
Step 3.3: time intensity variance is calculated to the alternate pixel point
Step 3.4: according to time intensity varianceCalculate refractive index structures coefficient
Further, it the step 3.4: calculatesExpression formula specifically:
Wherein, N is the pixel number with high gradient chosen,For the refractive index structures coefficient estimated.
The utility model has the advantages that
The characteristics of present invention is according to infrared propagation in atmosphere establishes atmosphere spoke by atmospheric turbulance infra-red radiation model recovery
Mode and corresponding atmospheriacally modulation transfer function (MTF) are penetrated, by filtering out atmospheriacally modulation transfer function in frequency domain, completes mist
The sharpening of its infrared image that degrades is handled.
The present invention has a characteristic that (1) under the premise of not increasing hardware cost, from the inverse of the IR image enhancement that degrades
Process analysis procedure analysis restores atmospheric scattering and acts on the influence that decays to picture contrast, degrades the clear of infrared image to solve the greasy weather
Clearization processing problem；(2) algorithm for image enhancement only enhances the contrast of image from visual effect, and the present invention considers image drop
The physical process of matter, compensates for causes for Degradation；(3) present invention does not need to obtain accurate depth information of scene, not yet
Need manual intervention too much；(4) this method does not need accurate apparatus measures atmospheric parameter, but uses in recuperation
The information of sequence image itself estimates the refractive index structures coefficient of atmospheriacally modulation transfer function, calculates simpler；(5) work as image
In there are when apparent edge, this method is more effective.
Figure of description
Fig. 1 flow chart of the method for the present invention
Fig. 2 (a) test image 1 (original greasy weather degrade near-infrared image)
Fig. 2 (b) is to test image 1 with based on atmospheric turbulance infra-red radiation model recovery result
Fig. 3 (a) test image 2 (original greasy weather degrade near-infrared image)
Fig. 3 (b) is to test image 2 with based on atmospheric turbulance infra-red radiation model recovery result
Fig. 4 (a) test image 3 (original greasy weather degrade near-infrared image)
Fig. 4 (b) is to test image 3 with based on atmospheric turbulance infra-red radiation model recovery result
Specific embodiment
Technical solution of the present invention improves the quality of degraded image under the premise of not increasing hardware cost, from degraded image
The inverse process of generation is analyzed, and atmospheric radiation transmission and corresponding atmospheriacally modulation transfer function (MTF) is established, by frequency domain
Filter out atmospheriacally modulation transfer function, completing the greasy weather degrades the sharpening processing of infrared image.
Below with reference to Fig. 1~4, introduce at the fog-degraded image sharpening based on atmospheric turbulance infra-red radiation model recovery
Reason method, specific embodiment are as follows:
Step 1: one width greasy weather degeneration infrared image of input；
Step 2: Fourier transformation is carried out to the image in step 1, the degraded image in frequency domain is obtained, according in frequency domain
Degraded image so that obtain atmospheriacally modulation transfer function (MTF), wherein the atmosphere modulation in long exposure and short exposure
Transmission function (MTF) is different, but all includes unknown refractive index structures coefficient；
Further, step 2 specifically:
On the basis of turbulent atmosphere model, the degradation model of image in airspace is analyzed, and then obtains modulation transfer function
(MTF)；The degradation model of image may be expressed as: in airspace
G (x, y)=f (x, y) * h (x, y)+n (x, y) (1)
Wherein, g (x, y) is degraded image, and f (x, y) is original image, and h (x, y) is turbulent flow transient state point spread function, n
(x, y) is noise item.
The degradation model of image may be expressed as: in a frequency domain accordingly
G (u, v)=F (u, v) * H (u, v)+N (u, v) (2)
Wherein, G (u, v) is the degraded image in frequency domain, and F (u, v) is the original image in frequency domain, and N (u, v) is in frequency domain
Noise item.H (u, v) is the turbulent flow transient state point spread function in frequency domain, and H (u, v) is that h (x, y) is obtained by Fourier transform
Frequency-domain expression, also referred to as optical transfer function can indicate are as follows: H=| H | eiΦ, wherein | H | indicate amplitude, Φ is phase
Position, to amplitude | H | make normalized, so that the amplitude of H (0,0) is 1 at zero frequency, then the amplitude after this normalization is referred to as to adjust
Modulation trnasfer function MTF (it is considered that MTF is | H | normalization indicate), it may be assumed that
MTF=| H |/K (3)
Wherein: K is amplitude of the H in zero frequency；
Image degradation model based on MTF is obtained by formula (1) (2) (3) are as follows:
G (u, v)=F (u, v) MTFKeiΦ+N(u,v) (4)
Assuming that MTF be it is isotropic, then can ignore the influence of phase Φ, even Φ=0, then
F (u, v)=(G (u, v)-N (u, v))/(MTFK) (5)
MTF is respectively in short exposure and the expression formula under long conditions of exposure are as follows:
In the case of short exposure, modulation transfer function be may be expressed as:
In formula, ν indicates angle spatial frequency,Indicate that refractive index structures coefficient, λ indicate that beam wavelength, R indicate transmission distance
It is empirical coefficient from, μ, μ=1 when short distance, μ=0.5 when remote, D indicate imager bore dia；
Under long exposure status, modulation transfer function be may be expressed as:
In formula, ν indicates angle spatial frequency,Indicate that refractive index structures coefficient, λ indicate that beam wavelength, R indicate transmission distance
From.
Step 3: the refractive index structures coefficient of atmospheriacally modulation transfer function in step 2 is calculated；
Each atmospheric parameter can just obtain more accurate in only accurate measuring process twoThis in practice can
It meets difficulty.The present invention provides a kind of new estimationsMethod, do not need accurate apparatus measures atmospheric parameter, and only
It is to be estimated using the information of sequence image itselfAtmospheric turbulance will cause light wave and occur in transmission process in various degree
Refraction, cause wavefront angle of arrival to rise and fall, in turn result in the fluctuating and shake of image, the present invention utilizesAngle of arrival rises
Relationship opening relationships formula between volt, image fluctuating three, to obtain finally
Realize that steps are as follows in detail:
Step 3.1: calculating image level gradient IXWith vertical gradient IY；
Step 3.2: according to horizontal gradient and vertical gradient IX、IYThe each picture element of image (m, n) is calculatedAnd choose pixel alternately pixel of the gradient value greater than certain Grads threshold；
Step 3.3: time intensity variance is calculated to the alternate pixel point
Flating will cause the offset of image, then intensity of the kth frame image at position (m, n) can be formulated
Are as follows:
I (m, n, k)=I0(m+Δxm,n,k,n+Δym,n,k) (8)
In formula, I0() indicates the ideal image under no turbulent flow condition, Δ xm,n,k、Δym,n,kIt respectively indicates horizontal and vertical
Displacement of the histogram to relative ideal image.
First approximation is carried out to formula (8), can be obtained:
I(m,n,k)≈I0(m,n)+[Ix(m,n),Iy(m,n)][Δx,Δy]T (9)
Wherein, I0() indicates the ideal image under no turbulent flow condition, IX、IYRespectively indicate the horizontal gradient of ideal image
And vertical gradient, Δ x and Δ y respectively indicate the offset on horizontally and vertically.
Time intensity variance indicates are as follows:
In formula, it is to be averaged that<>, which indicates to take all k values,For the mean value of I (m, n, k), in conjunction with formula (9),
Time intensity variance can indicate are as follows:
In formula, IX、IYThe horizontal gradient and vertical gradient of ideal image are respectively indicated, Δ x and Δ y respectively indicate level side
To with the offset in vertical direction, since Δ x, Δ y are independent from each other, so<Δ x Δ y>=0, it is assumed that image is along dampening
Square it is respectively to the intensity variance with vertical direction (the space variance that may be considered image), and because scheming
Offset is identical as caused by both direction, it is possible to set the variance of image caused by turbulent flow spatiallyFormula (11) are substituted into obtain:
It can be obtained by formula (10), andImage after being averaging to all image superpositions takes gradient
It is worth, then the variance of image caused by turbulent flow spatiallyIt can be acquired by formula (12).Formula (12) describes intensity ladder
Degree, intensity fluctuation, the relationship between image shift variance three, step 3.4: according to time intensity varianceUse formula
(13), (14) calculate refractive index structures coefficient
Can be obtained from the region with high gradient, why select the pixel of high gradient be because are as follows: in high gradient
Place, image shift amount caused by turbulent flow is larger, and offset relative turbulent caused by other noise factors is smaller, so at this time can be with
Ignore image shift caused by the other factors such as noise, reduces the error of estimation.
Some pixel with high gradient is chosen, following estimated value is calculated:
In formula,PFOV indicates pixel visual field (pixel field of view), L
Indicate that path-length, λ are beam wavelength, D indicates imager aperture, l0Indicate scale in turbulent flow, L0Indicate the outer ruler of turbulent flow
Degree, L0=0.4h, h indicate the height on imaging object distance ground,For time intensity variance.
N number of pixel with high gradient is chosen, to what is estimatedMean value is taken to get arriving finally
Step 4: filtering out atmospheriacally modulation transfer function (MTF) for the degraded image in the frequency domain in step 2 in frequency domain,
This process can be understood as filtering out the process of noise, and the expression formula of noise is atmospheriacally modulation transfer function MTF；
Step 5: the frequency domain image after atmospheriacally modulation transfer function will be filtered out in step 4 and carries out inverse Fourier transform, is obtained
To the image of sharpening processing；
Step 6: output sharpening treated image.
Fig. 2~Fig. 4 lists three groups of test results, and as can be seen from the results, treated that picture quality is improved for sharpening,
Picture contrast is enhanced, and especially image border is enhanced, and as shown in Figure 4, and this method is to there are more in image
It is more effective when edge.
Above-mentioned specific embodiment is only used for explanation and illustration technical solution of the present invention, but can not constitute and want to right
The restriction for the protection scope asked.It will be apparent to those skilled in the art that doing any letter based on the technical solutions of the present invention
New technical solution, will fall under the scope of the present invention obtained from single deformation or replacement.
Claims (5)
1. a kind of fog-degraded image clarification method based on turbulent flow infra-red radiation model, which is characterized in that the method packet
Include following steps:
Step 1: greasy weather degeneration infrared image is obtained；
Step 2: Fourier transformation is carried out to the described image in step 1, the degraded image in frequency domain is obtained, according in frequency domain
Degraded image so that obtain atmospheriacally modulation transfer function；
Step 3: calculating the refractive index structures coefficient of atmospheriacally modulation transfer function described in step 2,
Specifically:
Step 3.1: calculating image level gradient IXWith vertical gradient IY；
Step 3.2: each pixel of image (m, n) is calculatedAnd gradient value is chosen greater than certain gradient
The pixel of threshold value alternately pixel；
Step 3.3: time intensity variance is calculated to the alternate pixel point
Step 3.4: according to time intensity varianceCalculate refractive index structures coefficientIt specifically includes:
Some pixel with high gradient is chosen, following estimated value is calculated:
In formula,PFOV indicates pixel visual field, and L indicates that path-length, λ are ray wave
Long, D indicates imager aperture, l0Indicate scale in turbulent flow, L0Indicate turbulent flow external measurement, L0=0.4h, h indicate imaging object away from
Height from the ground,For time intensity variance；
It calculatesExpression formula be specially
Wherein, N is the pixel number with high gradient chosen,For the refractive index structures coefficient estimated；
Step 4: the degraded image in the frequency domain in step 2 is filtered out into atmospheriacally modulation transfer function in frequency domain；
Step 5: the degraded image in the frequency domain after atmospheriacally modulation transfer function will be filtered out in step 4 and carries out Fourier's inversion
It changes, obtains the image of sharpening processing；
Step 6: output sharpening treated image.
2. the method as described in claim 1, which is characterized in that the step 2: carrying out Fu to the described image in step 1
In leaf transformation, obtain the degraded image in frequency domain, according to the degraded image in frequency domain and then obtain atmospheriacally modulation transfer function；Tool
Body are as follows:
On the basis of turbulent atmosphere model, the degradation model of image in airspace is analyzed, and then obtains modulation transfer function；Its
In, the degradation model of image indicates in airspace are as follows:
G (x, y)=f (x, y) * h (x, y)+n (x, y) (1)
Wherein, g (x, y) is degraded image, and f (x, y) is original image, and h (x, y) is turbulent flow transient state point spread function, n (x, y)
For noise item；
The degradation model of image indicates in a frequency domain accordingly are as follows:
G (u, v)=F (u, v) * H (u, v)+N (u, v) (2)
Wherein, G (u, v) is the degraded image in frequency domain, and F (u, v) is the original image in frequency domain, and H (u, v) is the rapids in frequency domain
Transient state point spread function is flowed, N (u, v) is the noise item in frequency domain.
3. method according to claim 2, which is characterized in that H (u, v) is the frequency domain that h (x, y) is obtained by Fourier transform
Expression formula, also referred to as optical transfer function indicate are as follows: H=| H | eiΦ, wherein | H | indicate amplitude, Φ is phase, to amplitude |
H | make normalized, so that the amplitude of H (0,0) is 1 at zero frequency, then the amplitude after this normalization is referred to as atmosphere modulation transmitting
Function MTF, it may be assumed that
MTF=| H |/K (3)
Wherein: K is amplitude of the H in zero frequency；
Image degradation model based on MTF is obtained by formula (1) (2) (3) are as follows:
G (u, v)=F (u, v) MTFKeiΦ+N(u,v) (4)
Assuming that atmospheriacally modulation transfer function be it is isotropic, then ignore the influence of phase Φ, even Φ=0, then
F (u, v)=(G (u, v)-N (u, v))/(MTFK) (5).
4. method as claimed in claim 3, which is characterized in that expression formula of MTF under the conditions of short exposure are as follows:
Atmospheriacally modulation transfer function in the case of short exposure indicates are as follows:
In formula, ν indicates angle spatial frequency,Indicate that refractive index structures coefficient, λ indicate that beam wavelength, R indicate transmission range, μ is
Empirical coefficient, μ=1 when short distance, μ=0.5 when remote, D indicate imager bore dia.
5. method as claimed in claim 3, which is characterized in that expression formula of the MTF under long conditions of exposure are as follows:
Atmospheriacally modulation transfer function under long exposure status indicates are as follows:
In formula, ν indicates angle spatial frequency,Indicate that refractive index structures coefficient, λ indicate that beam wavelength, R indicate transmission range.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611167482.5A CN106780385B (en) | 2016-12-16 | 2016-12-16 | A kind of fog-degraded image clarification method based on turbulent flow infra-red radiation model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611167482.5A CN106780385B (en) | 2016-12-16 | 2016-12-16 | A kind of fog-degraded image clarification method based on turbulent flow infra-red radiation model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106780385A CN106780385A (en) | 2017-05-31 |
| CN106780385B true CN106780385B (en) | 2019-06-25 |
Family
ID=58892182
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611167482.5A Active CN106780385B (en) | 2016-12-16 | 2016-12-16 | A kind of fog-degraded image clarification method based on turbulent flow infra-red radiation model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106780385B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107680066B (en) * | 2017-09-04 | 2020-05-12 | 北京空间机电研究所 | Dynamic imaging simulation and frequency domain filtering method for space camera |
| CN108765286B (en) * | 2018-05-18 | 2020-06-30 | 浙江大学 | Digital image interpolation algorithm for radiation modulation function fidelity |
| CN109255804A (en) * | 2018-08-10 | 2019-01-22 | 重庆大数据研究院有限公司 | A kind of haze concentration sealing method |
| CN109520969B (en) * | 2018-10-26 | 2021-03-09 | 中国科学院国家空间科学中心 | Distributed scattering imaging method based on atmospheric medium self-modulation |
| CN109949251B (en) * | 2019-04-03 | 2023-06-30 | 天津中航亿达科技有限公司 | Li Ci pyramid-based rapid stabilizing method for atmospheric turbulence degradation video |
| CN110111277B (en) * | 2019-04-29 | 2021-01-29 | 北京理工大学 | Planar thermal image repairing method and device |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1653480A (en) * | 2002-05-14 | 2005-08-10 | 维森盖特有限公司 | Optical projection imaging system and method for automatically detecting cells with molecular marker compartmentalization associated with malignancy and disease |
-
2016
- 2016-12-16 CN CN201611167482.5A patent/CN106780385B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1653480A (en) * | 2002-05-14 | 2005-08-10 | 维森盖特有限公司 | Optical projection imaging system and method for automatically detecting cells with molecular marker compartmentalization associated with malignancy and disease |
Non-Patent Citations (2)
| Title |
|---|
| 基于 MTF 的 γ 辐照 CCD 退化图像复原方法研究;耿小兵;《第十七届全国核电子学与核探测技术学术年会》;20141231;第454-458页 |
| 遥感影像中大气模糊消除恢复算法研究;王振国;《中国博士学位论文全文数据库》;20110715(第7期);正文第1-124页 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106780385A (en) | 2017-05-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106780385B (en) | A kind of fog-degraded image clarification method based on turbulent flow infra-red radiation model | |
| CN105931220B (en) | Traffic haze visibility detecting method based on dark channel prior Yu minimum image entropy | |
| Kang et al. | Pothole detection system using 2D LiDAR and camera | |
| CN106251344B (en) | A kind of multiple dimensioned infrared target self-adapting detecting method of view-based access control model receptive field | |
| US10187617B2 (en) | Automatic detection of moving object by using stereo vision technique | |
| CN101281250B (en) | Method for monitoring on-rail satellite remote sensor modulation transfer function based on image element | |
| CN104899866A (en) | Intelligent infrared small target detection method | |
| CN103617617A (en) | Underwater image quality evaluating and measuring method based on power spectrum description | |
| Li et al. | Lane detection based on spiking neural network and hough transform | |
| CN103996027A (en) | Space-based space target recognizing method | |
| CN103065320A (en) | Synthetic aperture radar (SAR) image change detection method based on constant false alarm threshold value | |
| CN114549642A (en) | Low-contrast infrared weak and small target detection method | |
| Wang et al. | Multi-sensor image decision level fusion detection algorithm based on DS evidence theory | |
| CN107909546A (en) | A kind of fog-degraded image clearness processing method | |
| Cheng et al. | Visibility improvement in turbid water by the fusion technology of Mueller matrix images | |
| Zhang et al. | Dim moving infrared target enhancement based on precise trajectory extraction | |
| Mittal et al. | IoT based image defogging system for road accident control during winters | |
| Rahmani et al. | Grid-edge-depth map building employing sad with sobel edge detector | |
| Liu et al. | Background subtraction and dust storm detection | |
| Hariyanto et al. | Developing of total suspended sediment model Using Landsat-8 satellite image and in-situ data at the Surabaya Coast, East Java, Indonesia | |
| Chung et al. | Applying Second Order Least Square Regression Technology to Estimate Region Area. | |
| Shen et al. | Enhancing visibility of hazy images based on the estimation of the polarization value | |
| Kavya et al. | Face Tracking Using Unscented Kalman Filter | |
| CN116380076A (en) | Underwater positioning method and system based on polarization characteristics and sky image restoration | |
| Arhip et al. | Image processing for acquiring target coordinates with a platform |
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 |