CN110084774A - A kind of method of the gradient transmitting and minimum total variation blending image of enhancing - Google Patents
A kind of method of the gradient transmitting and minimum total variation blending image of enhancing Download PDFInfo
- Publication number
- CN110084774A CN110084774A CN201910288177.9A CN201910288177A CN110084774A CN 110084774 A CN110084774 A CN 110084774A CN 201910288177 A CN201910288177 A CN 201910288177A CN 110084774 A CN110084774 A CN 110084774A
- Authority
- CN
- China
- Prior art keywords
- image
- blending
- gradient
- term
- infrared
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000002156 mixing Methods 0.000 title claims abstract description 72
- 238000000034 method Methods 0.000 title claims abstract description 36
- 230000002708 enhancing effect Effects 0.000 title claims abstract description 9
- 230000004927 fusion Effects 0.000 claims abstract description 38
- 230000005855 radiation Effects 0.000 claims abstract description 6
- 230000009466 transformation Effects 0.000 claims description 5
- 238000012546 transfer Methods 0.000 claims description 2
- 238000006243 chemical reaction Methods 0.000 claims 1
- 230000000007 visual effect Effects 0.000 abstract description 13
- 230000000694 effects Effects 0.000 abstract description 8
- 238000007500 overflow downdraw method Methods 0.000 abstract description 8
- 230000006870 function Effects 0.000 description 8
- 238000005457 optimization Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 239000000654 additive Substances 0.000 description 1
- 230000000996 additive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000007499 fusion processing Methods 0.000 description 1
- 238000001027 hydrothermal synthesis Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 238000011017 operating method Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000013442 quality metrics Methods 0.000 description 1
- 230000035945 sensitivity 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
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—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
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of methods of the gradient of enhancing transmitting and minimum total variation blending image, belong to image co-registration field.Mainly solve the problems, such as that target and background texture information is not detailed when infrared and visual image fusion.There is image pixel intensities similar with infrared image, visible images and gradient similar with infrared image, visible images by the way that blending image to be constrained to.Fusion problem is converted L by we1- TV minimization problem, uses m, λ1And λ2Relationship between three parameter control data fidelity terms and regularization term, to achieve the effect that while keep the heat radiation and appearance information in source images.The present invention can sufficiently integrate infrared and visible images target texture detailed information, and effective protection image detail improves visual effect, greatly improves the quality of blending image compared to traditional fusion method.
Description
Technical field
The invention belongs to image co-registration field, be related to a kind of enhancing gradient transmitting and minimum total variationization fusion it is infrared and
The method of visible images is infrared and a Visual image processing technical field fusion method, in business and military affairs
Have and widely applies.
Background technique
As a research branch in image co-registration field and research emphasis, with the quick hair of heat radiation image technology
Exhibition, infrared and visual image fusion have become current research hotspot both domestic and external.Infrared image can accurately provide target
The information such as the position details of object, it is seen that light image can accurately provide detailed details and background information.Infrared and visible light
Image co-registration can effectively integrate the scene detailed information of infrared image target signature information and visible images, obtain information more
Comprehensive blending image.The complementary information that infrared and visual light imaging sensor provides, contains fused image more
Comprehensively, information abundant, more meet the visual characteristic of people or machine, be more advantageous to image further analysis processing and
Automatic target detection.In pixel image fusion, the first problem to be solved is most important information in determining source images, with
The information of acquisition is converted into change the smallest blending image, be especially distorted or is lost.In order to solve this problem, mistake
It goes decades to propose many methods, including is based on pyramidal method, wavelet transformation, warp wavelet, multiresolution singular value
It decomposes, guiding filtering is multifocal, rarefaction representation etc..It is pixel by pixel simplest strategy to source images averaging.However, this
Direct method generates many undesired effects, such as contrast reduces.In order to address this issue, it has been proposed that based on more
The method of change of scale is related to three basic steps: source images are resolved into more rulers with low frequency and high-frequency information first
Degree indicates；Then multi-scale Representation is merged according to some fusion rules；The inverse transformation of compound multiple dimensioned coefficient is eventually for composition
Blending image.Method based on multi-scale transform is capable of providing better performance, because they are consistent with human visual system, and
The object of real world is usually by the structure composition of different scale.The example of these methods includes laplacian pyramid, discrete
Wavelet transformation, wave transform of not sub sampled contour etc..Method based on multi-scale transform achieves immense success in many cases；
But they use identical expression to different source images, and attempt to retain in identical significant feature such as source images
Side and line.The problem of for infrared and visual image fusion, the thermal radiation information in infrared image are characterized by image pixel intensities, and
And target usually has the intensity bigger than background, therefore can easily detect；And the texture information master in visual picture
It to be characterized by gradient, and there is the detailed information of significantly gradient (such as edge) offer scene.Therefore, in fusion process
The image of both types is indicated to be inappropriate using identical.On the contrary, in order to retain important information as much as possible, it is expected that
Blending image is to keep the primary intensity in infrared image to be distributed and the change of gradient in visual picture.
The Variation Model of initial removal additive noise is divided into regularization term and data fidelity term in the model, just in total
Then changing item is to play the role of inhibiting noise, and data fidelity term is to maintain the similitude of image and observed image after denoising, keeps
The edge feature of image.It is had also been proposed later based on gradient transmitting and minimum total variationization is infrared and visual picture fusion, referred to as
Gradient transmitting fusion (GTF).Fusion is expressed as optimization problem, wherein objective function is by data fidelity item and regularization term group
At.Data fidelity item constraint blending image should have image pixel intensities similar with given infrared image, and regularization term ensures
Gradient distribution in visual picture can be for delivery in blending image, L1Norm is used to promote the sparsity of gradient, then can be with
Pass through existing L1- TV minimizes technology to solve optimization problem.In GTF, although it can preferably capture background information,
Target is not prominent enough, and blending image contrast is lower.For this purpose, the invention proposes a kind of transmitting of modified hydrothermal process gradient and total change
Difference, which minimizes, merges infrared and visible images method.It remains infrared and visible images gradients simultaneously and pixel is strong
Degree, and introduce three parameters and carry out the relationship appropriate adjusted between regularization term and fidelity term, to preferably be melted
Close effect
In order to improve the performance of blending image, the selection of fusion rule is equally most important.It selects in the present invention based on ladder
The fusion rule of degree transmitting and minimum total variation, has preferably kept the gradient and image pixel intensities of image, by retaining simultaneously
Infrared and visible images gradients and image pixel intensities, so that the image object of fusion is more accurate, background detail is clear, improves
The quality of blending image.
Summary of the invention
The purpose of the present invention is in view of the above shortcomings of the prior art, propose a kind of gradient transmitting of enhancing with total variance most
Smallization merges infrared and visible images method, solves blending image mesh obtained by existing infrared and visible light image fusion method
Loss in detail and the unsharp problem of background texture are marked, and sufficiently integrates the infrared structural information with visible images of different modalities
And functional information, effective protection image detail, enhancing image detail, texture and edge contour improve its visual effect, raising is melted
Close the quality of image.
The technical solution adopted by the present invention to solve the technical problems is as follows:
A kind of method of the gradient transmitting and minimum total variation blending image of enhancing, steps are as follows:
(1) gradient transfer transformation is carried out to infrared image to be fused and visible images, obtains corresponding gradient and picture
Plain intensity；
(2) data fidelity term and regularization term are established, wherein data fidelity term for Constraint fusion image should have and to
Determine image pixel intensities similar in infrared image and visible images, and regularization term is for ensuring in infrared image and visible images
Gradient distribution can be transmitted in blending image；
(2.1) data fidelity term
Wherein, x indicates blending image；ε1(x) fidelity term of blending image is indicated；The norm of p expression fidelity term；U is indicated
Infrared image；V indicates visible images；Wherein the size of x, u and v image is a × b；Infrared image is usually strong with pixel
Degree is characterized, similarly in visible images there is also image pixel intensities differences between target and background.Due to target and
Image pixel intensities difference between background, target is high-visible usually in infrared image, and the mesh in some visible images
The texture for marking object is also very clear.So data fidelity term is characterized by image pixel intensities, in order to avoid it is strong to lose some pixels
Spend information.
(2.2) regularization term
Wherein, ε2(x) regularization term of blending image is indicated；The norm of q expression regularization term；Indicate blending image
Gradient；Indicate the gradient of visible images；Indicate the gradient of infrared image；It can be seen that remaining mesh in fidelity term
The pixel intensity distribution of mark and background, and be substantially spy with the gradient of image for target and scene detailed look information
Sign, therefore regularization term, to enhance target and the background complete picture on vision shape, is made using pixel gradient as target
Obtaining blending image has more detailed look information.
(3) fusion problem is expressed as initial target function in conjunction with public formula (I) and (II):
Wherein, ε (x) indicates objective function；m,λ1And λ2It is for adjusting the parameter weighed between x and u, v.First and second
In the case that item constraint blending image x is to have image pixel intensities similar with infrared image u, visible images v, third and fourth
It is required that blending image x and infrared image u, visible images v have similar gradient, m, λ1And λ2Be for adjust x and u, v it
Between the parameter weighed, objective function (III) purpose is so that the image of fusion seems to be more biased towards in vision shape, appearance details
Clean mark, target can protrude and detailed information will not lose, it appear that while there is target to protrude clean mark
Retain and enhance the image of background detail information function.
(4) it is optimized using minimum total variation
For p, q norm in public formula (III), we the problem of in it is desirable that retaining the thermal radiation information of infrared image
The relatively high information with the image pixel intensities in visible images, so p=1 uses L for the purposes of promoting the sparsity of gradient1Model
Number minimizes gradient disparities, i.e. q=1, for the image having a size of a × b, we use y ∈ Rab×1Indicate the column of its image pixel intensities
Vector form, the gray value with range from 0 to 255.Y=x-v i.e. x=y+v is enabled, optimization (III) can rewrite are as follows:
Wherein, y indicates the difference of blending image x and visible images v；T (y) indicates to minimize energy functional；For eachIndicate image gradientPixel i withWithPoint
Not Dui Yingyu a horizontal and vertical scale, i.e.,WithWherein r (i) and b (i) represent the right side
Nearest-neighbors below side and pixel i.In addition, if pixel i is located in last row or column, then r (i) and b (i) are set
For i.Objective function (IV) is convex, therefore has globally optimal solution, and algorithm is by adjusting m, λ1And λ2Value make fidelity term and
Regularization term reaches suitable point, and the image that fusion obtains in this way can retain the thermal radiation information and clearly of two width source images
Appearance texture information.
(5) minimize method using broad sense total variation functional and solve T (y): problem (IV) is the L of standard1- TV minimum is asked
Topic, the thinking of solution are that public formula (IV) is resolved into two parts of fidelity term and regularization term to carry out substep solution, finally will
Two are combined together to obtain following formula:
Wherein: first item is fidelity term, second and third is regularization term, C (y(k)) it is a constant,WithTable respectively
Show the iteration function of fidelity term and regularization term；y(k)Indicate y to iteration k times；
(6) public formula (V) is needed through existing L1- TV minimizes Optimized Iterative to calculate T (y), it may be assumed that circulation k=
0,1 ...；Work as T(k)(y) meeting the condition of convergence, then iteration terminates, and otherwise k=k+1 returns to (5)；
(7) blending image x is determined by following formula*Globally optimal solution: x*=T (y)+v finally obtains final fusion figure
As x*。
Beneficial effects of the present invention:
1, the present invention is transmitted using the gradient based on enhancing and minimum total variationization merges infrared and visible images method,
Fusion in spatial domain can retain the gradient and image pixel intensities of two width source images simultaneously, can adequately merge source images
The minutias such as texture, profile, keep the image definition of fusion higher, information content is richer, better quality.
2, infrared and visible light image fusion method of the invention adjusts the image and source figure of fusion using three parameters
The proportionate relationship of picture can debug out the syncretizing effect figure needed according to demand, with low excellent of flexible structure, computation complexity
Therefore point is more able to satisfy public demand.
Detailed description of the invention
Fig. 1 (a) is the present invention in parameter m=0, λ1=4, λ2=0 blending image.
Fig. 1 (b) is blending image of the GTF invention in parameter lambda=4.
Fig. 1 (c) is the present invention in parameter m=4, λ1=4, λ2=0 blending image.
Fig. 2 (a) is that the present invention keeps parameter m=0, λ in different source images1In=4 constant situations, λ2Value from 0 to 40
Step-length is the display result of 4 incremental blending image SSIM indexs.
Fig. 2 (b) is that the present invention keeps parameter lambda in different source images1=4, λ2In=0 constant situation, the value of m is from 0 to 40
Step-length is the display result of 4 incremental blending image SSIM indexs.
Fig. 3 is that the present invention keeps parameter m=0, λ in different source images1In=4 constant situations, λ2Value from 0 to 40 step
A length of 4 incremental subjective blending images are as the result is shown.Wherein, (a-1) is respectively indicated to (a-5) works as λ2=0,8,12,16,20
When, corresponding subjective Bunker blending image；(b-1) respectively indicates to (b-5) works as λ2It is corresponding subjective when=0,8,12,16,20
Lake blending image；(c-1) respectively indicates to (c-5) works as λ2When=0,8,12,16,20, corresponding subjective Tank merges figure
Picture.
Fig. 4 is that the present invention keeps parameter lambda in different source images1=4, λ2In=0 constant situation, the value of m step from 0 to 40
A length of 4 incremental subjective blending images are as the result is shown.Wherein, (a-1) is respectively indicated to (a-5) works as m=0, when 4,8,16,40,
Corresponding subjective Bunker blending image；(b-1) respectively indicates to (b-5) works as m=0, corresponding subjective when 4,8,16,40
Lake blending image；(c-1) respectively indicates to (c-5) works as m=0, when 4,8,16,40, corresponding subjective Tank blending image.
Fig. 5 is the present invention to the fusion results schematic diagrames of 6 width difference source images, and (a-1) to (f-1) is Bunker, Lake,
The visible images of oneman in front of hous, Sandpath, NATO-camp and Tank；(a-2) to (f-2) is
The infrared image of Bunker, Lake, one man in front of hous, Sandpath, NATO-camp and Tank；From the 3rd
Row starts to be the result figure for merging each pair of visible light/infrared image using different fusion methods, fusion side to the 11st row
Method is from top to bottom successively are as follows: the blending image of the method for the present invention, the blending image based on LP, is based on the blending image based on GTF
The blending image of RP, the blending image based on DTCWT, the blending image based on CVT, is based on the blending image based on Wavelet
The blending image of MSVD and blending image based on LP-SR.
Fig. 6 is the quantitative comparison result of blending image of the invention in MI index.Wherein, fusion side used in each image
Method is from left to right successively are as follows: LP, RP, Wavelet, DTCWT, CVT, MSVD, LP-SR, GTF, this method.
Fig. 7 is the quantitative comparison result of blending image of the invention in EN index.Wherein, fusion side used in each image
Method is from left to right successively are as follows: LP, RP, Wavelet, DTCWT, CVT, MSVD, LP-SR, GTF, this method.
Fig. 8 is the quantitative comparison result of blending image of the invention in Yang index.Wherein, fusion used in each image
Method is from left to right successively are as follows: LP, RP, Wavelet, DTCWT, CVT, MSVD, LP-SR, GTF, this method.
Fig. 9 is the quantitative comparison result of blending image of the invention in Chen index.Wherein, fusion used in each image
Method is from left to right successively are as follows: LP, RP, Wavelet, DTCWT, CVT, MSVD, LP-SR, GTF, this method.
Figure 10 is the flow chart of the method for the present invention.
Specific embodiment
It elaborates below to one embodiment of the present of invention (infrared and visible images) in conjunction with attached drawing, the present embodiment
It carries out under the premise of the technical scheme of the present invention, detailed embodiment and specific operating procedure are as follows:
Step 1: for public formula (III), parameter m, λ is arranged in we2Equal to 0, so that it may which formula is converted into following format:
λ is enabled for public formula (VI)1Shown in blending image such as Fig. 1 (a) that=4 pairs of images are merged；Then we
In original GTF model, shown in the blending image for enabling λ=4 obtain such as Fig. 1 (b)；
Step 2: similarly for public formula (III), m=4, λ is arranged in we1=4, λ2=0 obtained blending image such as Fig. 1
(c) shown in；
Step 3: we are used to verify λ in public formula (III)2Effect, first set m=0, λ1=4, then adjust λ2Value
Step-length is gradually incremented by from 0 to 40 for 4, (is Bunker, Lake, one man in front of respectively to 6 width source images
Hous, Sandpath, NATO-camp and Tank) fusion results that are merged, it is commented with SSIM objective indicator
Valence, display the result is that with λ2Value increase index and move closer to 1 but be never equal to 1 as shown in Fig. 2 (a), this and institute
The fusion purpose of operation is consistent, wherein having chosen λ2=0,8,12,16,20 corresponding subjective blending images, as a result such as Fig. 3 institute
Show.
Step 4: we are used to verify the effect of m in public formula (III), first set λ1=4, λ2Then=0 adjusts the value of m from 0
Be 4 to be gradually incremented by 40 step-lengths, to 6 width source images (be Bunker, Lake, one man in front of hous respectively,
Sandpath, NATO-camp and Tank) fusion results that are merged, it is evaluated with SSIM objective indicator, is shown
It is showing the result is that with m value increase index move closer to 1 but never be equal to 1, as shown in Fig. 2 (b), this with it is operated
It is consistent to merge purpose, wherein m=0 is had chosen, 4,8,16,40 corresponding subjective blending images, as a result as shown in Figure 4.
Emulation experiment:
In order to verify feasibility and validity of the invention, using to 6 groups of infrared and visible images, such as Fig. 5 the first row
With shown in the second row, fusion experiment is carried out according to the method for the present invention.
In conclusion compared as the fusion results of Fig. 5 it can be seen that blending image obtained by the method for the present invention utmostly
Raw information is loyal on ground, has been preferably kept object, Texture eigenvalue in image to be fused, has been effectively prevented target texture
With the loss of background pixel intensity, thus the contrast of image and clarity are higher, and details is more prominent, and subjective vision effect is most
Good, i.e., fusion results are more preferable.
Fig. 6, Fig. 7, Fig. 8 and Fig. 9, which give, objectively evaluates index using fusion results obtained by various fusion methods.Its
In, histogram show higher to represent this source images evaluation index value obtained by image interfusion method optimal.Have wherein several
It is not highest that width image objectively evaluates index in the method, this is because infrared image gradient and visible light in source images
The weaker obtained image that will lead to fusion of image pixel intensity objectively evaluates index and is not so good as other fusion methods.
By Fig. 6, Fig. 7, Fig. 8 and Fig. 9 data it is found that the method for the present invention blending image obtained is in mutual information (MI), letter
Breath entropy (EN), Yang, Chen etc., which objectively evaluate, is better than other fusion methods in index.MI reflects the fusion that blending algorithm obtains
Mutual information between image and blending image is bigger, and the correlation between blending image and blending image is higher, blending image
Effect is higher.Entropy reflect image carry information content number, entropy is bigger, illustrate that the information content for including is more, merge
Effect is better；Image co-registration quality metric of the Yang based on similarity；Chen index is to utilize contrast sensitivity function class vision
What system tested the perception degree of image space frequency.Picture is carried out piecemeal first by this method, then calculates area
The significance in domain.
Claims (1)
1. a kind of method of the gradient transmitting and minimum total variation blending image of enhancing, which is characterized in that steps are as follows:
(1) gradient transfer transformation is carried out to infrared image to be fused and visible images, obtains corresponding gradient and pixel is strong
Degree；
(2) data fidelity term and regularization term are established, wherein data fidelity term should have and give red for Constraint fusion image
Image pixel intensities similar in outer image and visible images, and regularization term is used to ensure the ladder in infrared image and visible images
Degree distribution can be transmitted in blending image；
(2.1) data fidelity term is established
Wherein, x indicates blending image；ε1(x) fidelity term of blending image is indicated；The norm of p expression fidelity term；U indicates infrared figure
Picture；V indicates visible images；Wherein the size of x, u and v image is a × b；
(2.2) regularization term is established
Wherein, ε2(x) regularization term of blending image is indicated；The norm of q expression regularization term；Indicate the ladder of blending image
Degree；Indicate the gradient of visible images；Indicate the gradient of infrared image；
(3) data fidelity term is combined with regularization term, obtains initial objective function
Wherein, ε (x) indicates objective function；m,λ1And λ2It is for adjusting the parameter weighed between x and u, v；
(4) for the image having a size of a × b, with y ∈ Rab×1The column vector form for indicating its image pixel intensities, with range from 0
To 255 gray value；P=1, q=1 are enabled, y=x-v i.e. x=y+v is enabled, gives initial allowable error ε > 0, by fusion problem conversion
To optimize least energy functional model:
Wherein, y indicates the difference of blending image x and visible images v；T (y) indicates to minimize energy functional；For each x=(x1, x2)∈R2；I indicates pixel；Indicate image gradientIn pixel
I withWithA horizontal and vertical scale is corresponded respectively to, i.e.,WithR (i) and b
(i) nearest-neighbors below the right and pixel i are represented, when during pixel i is located at last line or last is arranged, then r (i) and b
(i) it both is set to i；
Due to objective function (IV) be it is convex, it is all have globally optimal solution, pass through adjust objective function (IV) in m, λ1And λ2's
Value, so that fidelity term and regularization term reach suitable point, the image merged can retain the heat radiation of two width source images
Information and clearly appearance texture information；
(5) method is minimized using broad sense total variation functional solve T (y):
Public formula (IV) is resolved into two parts of fidelity term and regularization term and carries out substep solution, is finally combined together two
Obtain following formula:
Wherein, first itemIt is fidelity term, Section 2And third
?It is regularization term, C (y(k)) it is constant,WithRespectively indicate fidelity term and regularization term repeatedly
For function；y(k)Indicate y to iteration k times；
(6) public formula (V) is needed through existing L1- TV minimizes Optimized Iterative to calculate T (y), it may be assumed that circulation k=0,
1 ...；Work as T(k)(y) meeting the condition of convergence, then iteration terminates, and otherwise k=k+1 returns to (5)；
(7) pass through formula x*=T (y)+v, determines blending image x*Globally optimal solution, finally obtain final blending image x*。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910288177.9A CN110084774B (en) | 2019-04-11 | 2019-04-11 | Method for minimizing fusion image by enhanced gradient transfer and total variation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910288177.9A CN110084774B (en) | 2019-04-11 | 2019-04-11 | Method for minimizing fusion image by enhanced gradient transfer and total variation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110084774A true CN110084774A (en) | 2019-08-02 |
| CN110084774B CN110084774B (en) | 2023-05-05 |
Family
ID=67414771
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910288177.9A Active CN110084774B (en) | 2019-04-11 | 2019-04-11 | Method for minimizing fusion image by enhanced gradient transfer and total variation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110084774B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115082968A (en) * | 2022-08-23 | 2022-09-20 | 天津瑞津智能科技有限公司 | Behavior identification method based on infrared light and visible light fusion and terminal equipment |
| CN115620030A (en) * | 2022-12-06 | 2023-01-17 | 浙江正泰智维能源服务有限公司 | Image matching method, device, equipment and medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014078985A1 (en) * | 2012-11-20 | 2014-05-30 | Thomson Licensing | Method and apparatus for image regularization |
| WO2014102458A1 (en) * | 2012-12-31 | 2014-07-03 | Nokia Corporation | Method and apparatus for image fusion |
| CN107945145A (en) * | 2017-11-17 | 2018-04-20 | 西安电子科技大学 | Infrared image fusion Enhancement Method based on gradient confidence Variation Model |
-
2019
- 2019-04-11 CN CN201910288177.9A patent/CN110084774B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014078985A1 (en) * | 2012-11-20 | 2014-05-30 | Thomson Licensing | Method and apparatus for image regularization |
| WO2014102458A1 (en) * | 2012-12-31 | 2014-07-03 | Nokia Corporation | Method and apparatus for image fusion |
| CN107945145A (en) * | 2017-11-17 | 2018-04-20 | 西安电子科技大学 | Infrared image fusion Enhancement Method based on gradient confidence Variation Model |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115082968A (en) * | 2022-08-23 | 2022-09-20 | 天津瑞津智能科技有限公司 | Behavior identification method based on infrared light and visible light fusion and terminal equipment |
| CN115620030A (en) * | 2022-12-06 | 2023-01-17 | 浙江正泰智维能源服务有限公司 | Image matching method, device, equipment and medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110084774B (en) | 2023-05-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | Adaptive image enhancement method for correcting low-illumination images | |
| CN104318569B (en) | Space salient region extraction method based on depth variation model | |
| CN102881011B (en) | Region-segmentation-based portrait illumination transfer method | |
| CN110349247B (en) | Indoor scene CAD three-dimensional reconstruction method based on semantic understanding | |
| CN107451984A (en) | A kind of infrared and visual image fusion algorithm based on mixing multiscale analysis | |
| CN106056564B (en) | Edge clear image interfusion method based on joint sparse model | |
| CN108269244B (en) | Image defogging system based on deep learning and prior constraint | |
| CN109872285A (en) | A kind of Retinex low-luminance color image enchancing method based on variational methods | |
| CN109754377A (en) | A kind of more exposure image fusion methods | |
| CN106960428A (en) | Visible ray and infrared double-waveband image co-registration Enhancement Method | |
| CN103020933B (en) | A kind of multisource image anastomosing method based on bionic visual mechanism | |
| CN103942769B (en) | A kind of satellite remote-sensing image fusion method | |
| Wang et al. | Variational single nighttime image haze removal with a gray haze-line prior | |
| Sánchez-Ferreira et al. | Bio-inspired optimization algorithms for real underwater image restoration | |
| CN109242834A (en) | It is a kind of based on convolutional neural networks without reference stereo image quality evaluation method | |
| CN109493309A (en) | A kind of infrared and visible images variation fusion method keeping conspicuousness information | |
| Luo et al. | Infrared and visible image fusion based on visibility enhancement and hybrid multiscale decomposition | |
| CN105809650B (en) | A kind of image interfusion method based on bidirectional iteration optimization | |
| CN110084774A (en) | A kind of method of the gradient transmitting and minimum total variation blending image of enhancing | |
| CN108154493A (en) | A kind of pseudo- color blending algorithm of the dual-band infrared image based on FPGA | |
| Junwu et al. | An infrared and visible image fusion algorithm based on LSWT-NSST | |
| Huang et al. | Color correction and restoration based on multi-scale recursive network for underwater optical image | |
| CN109255773A (en) | Different resolution ratio based on full variation is infrared with visible light image fusion method and system | |
| Liu et al. | TSE_Fuse: Two stage enhancement method using attention mechanism and feature-linking model for infrared and visible image fusion | |
| CN110211216B (en) | Three-dimensional image spatial domain fusion method based on volume rendering opacity weighting |
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 |