CN110533581A - A kind of raindrop removing method, system and storage medium based on raindrop probability graph - Google Patents
A kind of raindrop removing method, system and storage medium based on raindrop probability graph Download PDFInfo
- Publication number
- CN110533581A CN110533581A CN201910737935.0A CN201910737935A CN110533581A CN 110533581 A CN110533581 A CN 110533581A CN 201910737935 A CN201910737935 A CN 201910737935A CN 110533581 A CN110533581 A CN 110533581A
- Authority
- CN
- China
- Prior art keywords
- raindrop
- image
- probability graph
- layer
- removing method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 35
- 238000012804 iterative process Methods 0.000 claims abstract description 10
- 239000000203 mixture Substances 0.000 claims description 9
- 238000009472 formulation Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000009434 installation Methods 0.000 claims description 6
- 230000009466 transformation Effects 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 2
- 230000002708 enhancing effect Effects 0.000 abstract description 3
- 238000003384 imaging method Methods 0.000 abstract 1
- 238000012545 processing Methods 0.000 description 9
- 230000008030 elimination Effects 0.000 description 6
- 238000003379 elimination reaction Methods 0.000 description 6
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000008439 repair process Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000010191 image analysis Methods 0.000 description 2
- 235000020061 kirsch Nutrition 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000003014 reinforcing effect Effects 0.000 description 2
- 230000019771 cognition Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000000926 separation method Methods 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
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/04—Context-preserving transformations, e.g. by using an importance map
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/77—Retouching; Inpainting; Scratch removal
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of raindrop removing method, system and storage medium based on raindrop probability graph, which comprises receives a raindrop image；Raindrop probability graph is constructed according to the prior information of raindrop；The background layer and raindrop layer of the raindrop image are constrained by a priori assumption of natural image, to construct optimal model；The optimal model is solved based on the raindrop probability graph, updates the raindrop probability graph simultaneously in the iterative process for solving the optimal model；Export the image without raindrop.The present invention is by carrying out imaging model hypothesis to the raindrop image of shooting, estimate the distribution of raindrop probability graph in model, to constructing optimal model accurately to eliminate the raindrop in image, construct high quality without raindrop image, and can realize effective enhancing to raindrop image.
Description
Technical field
The present invention relates to technical field of image processing more particularly to a kind of raindrop removing method based on raindrop probability graph,
System and storage medium.
Background technique
The important goal of computer vision system is the image or video data for acquiring real world, by handling and dividing
Cognition and understanding to vision data is realized in analysis, as to have become one, computer vision field very typical for video monitoring
With important application.The accuracy of vision system initial data collected brings very big shadow to subsequent processing and analysis
It rings, collected data are more accurate, then its information is more complete, and the performance of vision system is also higher.However current outdoor vision
System usually will receive the influence of boisterous influence, especially rain, therefore eliminate the presence of raindrop image moderate rain, to raising
The robustness of outdoor vision system and the validity of video monitoring system have great importance.
With the development of deep learning, from a regular rain line image obtain one completely the image without rain have become opposite
It is easy.However current educational circles is to the processing of raindrop in image or fewer and fewer.In fact, adopting in rainy day video image data
Remaining raindrop are had during collection, on the camera lens of camera, usually so as to cause the presence for having raindrop in the image of shooting.
Raindrop image deals with difficulty than common rain line image can be bigger, and the direction of the latter's rain is generally more regular, can be preferable
Ground constructs model, and there is two big obstacles for raindrop image, first, being different from common rain line image, in common rain line image
The distribution of rain is regular, global, and the rain in raindrop image is likely to be present in any position of image, and shape size is remote
The line that do not rain is so regular, and there are difficulty in detection；Second, the rain line image for being different from obtaining in common rain line image is back
The superposition of scape image and rain, in raindrop image, the information of the image background covered by raindrop may be to lack completely, this
Recovery operation for raindrop image further increases difficulty.
The raindrop removing method majority of existing raindrop image is first to detect to raindrop, then carry out to the raindrop detected
Repairing, to repair the pixel region damaged in raindrop image by raindrop.However in the image after repairing by existing manner still
It there may be raindrop residual, the raindrop in raindrop image cannot be effectively removed, repairing quality is not high.
Summary of the invention
In order to solve at least one above-mentioned technical problem, the invention proposes a kind of, and the raindrop based on raindrop probability graph are eliminated
Method, system and storage medium.
To achieve the goals above, first aspect present invention proposes a kind of raindrop elimination side based on raindrop probability graph
Method, which comprises
Receive a raindrop image；
Raindrop probability graph is constructed according to the prior information of raindrop；
The background layer and raindrop layer of the raindrop image are constrained by a priori assumption of natural image, to construct most
Optimized model；
The optimal model is solved based on the raindrop probability graph, in the iterative process for solving the optimal model
The raindrop probability graph is updated simultaneously；
Export the image without raindrop.
In the present solution, constructing raindrop probability graph according to the prior information of raindrop, specifically include:
Raindrop probability graph, calculation formula are calculated using the shape smoothness index S of the raindrop image are as follows: α=eS-2π。
In the present solution, the algorithm of the optimal model are as follows:
Wherein, first itemFor fidelity term, the background solved for ensuring algorithm
The combination of layer and raindrop layer meets raindrop image and generates model, Section 2 λ1Φ (I) and Section 3 λ2Ψ (R) is respectively constraints graph
The regularization term of picture, I indicate that background layer, R indicate raindrop layer, λ1And λ2For regularization parameter.
In the present solution, a priori assumption by natural image carries out about the background layer and raindrop layer of the raindrop image
Beam specifically includes:
Sparsity using the background layer of raindrop image described in Tight wavelet frames operator constraint in Transformation Domain, constraint formulations
Are as follows: Φ (I)=| | WI | |1, wherein W is small wave operator.
In the present solution, a priori assumption by natural image carries out about the background layer and raindrop layer of the raindrop image
Beam, specifically further include:
Constrain energy of the raindrop layer of the raindrop image in gradient field, constraint formulations are as follows: Its
InFor gradient operator.
In the present solution, solving the optimal model based on the raindrop probability graph, specifically include:
Initialize installation, presetting the raindrop probability graph is α0, background layer I0=J, raindrop layer R0=J, wherein J is described
Raindrop image；
The optimal model is iteratively solved based on the raindrop probability graph；
When the algorithmic statement of the optimal model, final background tomographic image is exported, the as described raindrop image disappears
Except the image after raindrop.
Further, the optimal model is iteratively solved based on the raindrop probability graph, specifically included:
{ the α according to known to previous step in iterative processk, Ik, update Rk+1, more new formula are as follows:
According to updated { αk, Rk+1, update Ik+1, more new formula are as follows:
According to updated Rk+1, the smoothness S of the raindrop image shape is recalculated, and update raindrop probability graph
αk+1。
Second aspect of the present invention also proposes that a kind of raindrop based on raindrop probability graph eliminate system, described to be based on raindrop probability
It includes: memory and processor that the raindrop of figure, which eliminate system, includes a kind of raindrop based on raindrop probability graph in the memory
Removing method program realizes following step when the raindrop removing method program based on raindrop probability graph is executed by the processor
It is rapid:
Receive a raindrop image；
Raindrop probability graph is constructed according to the prior information of raindrop；
The background layer and raindrop layer of the raindrop image are constrained by a priori assumption of natural image, to construct most
Optimized model；
The optimal model is solved based on the raindrop probability graph, in the iterative process for solving the optimal model
The raindrop probability graph is updated simultaneously；
Export the image without raindrop.
In the present solution, solving the optimal model based on the raindrop probability graph, specifically include:
Initialize installation, presetting the raindrop probability graph is α0, background layer I0=J, raindrop layer R0=J, wherein J is described
Raindrop image；
The optimal model is iteratively solved based on the raindrop probability graph；
When the algorithmic statement of the optimal model, final background tomographic image is exported, the as described raindrop image disappears
Except the image after raindrop.
Third aspect present invention also proposes a kind of computer readable storage medium, wraps in the computer readable storage medium
Include a kind of raindrop removing method program based on raindrop probability graph, the raindrop removing method program quilt based on raindrop probability graph
When processor executes, realize such as the step of a kind of above-mentioned raindrop removing method based on raindrop probability graph.
Raindrop removing method, system and storage medium proposed by the present invention based on raindrop probability graph, first using raindrop
Style characteristic estimates raindrop probability graph, then by the prior-constrained building optimal model to background layer and raindrop layer, with most
Image after the elimination of the background tomographic image obtained eventually, as raindrop.It is more in the existing raindrop minimizing technology in raindrop image
It is detected using to raindrop, then to the mode that the raindrop detected are repaired, to repair in image by raindrop damage
Pixel region.Compared with prior art, the present invention can more precisely eliminate the raindrop in image, construct high quality without rain
Image is dripped, and can realize effective enhancing to raindrop image.
Additional aspect and advantage of the invention will provide in following description section, will partially become from the following description
Obviously, or practice through the invention is recognized.
Detailed description of the invention
Fig. 1 shows a kind of flow chart of the raindrop removing method based on raindrop probability graph of the present invention；
Fig. 2 shows a kind of raindrop elimination frame diagrams based on raindrop probability graph of the invention；
Fig. 3 shows the block diagram that a kind of raindrop based on raindrop probability graph of the present invention eliminate system.
Specific embodiment
To better understand the objects, features and advantages of the present invention, with reference to the accompanying drawing and specific real
Applying mode, the present invention is further described in detail.It should be noted that in the absence of conflict, the implementation of the application
Feature in example and embodiment can be combined with each other.
In the following description, numerous specific details are set forth in order to facilitate a full understanding of the present invention, still, the present invention may be used also
To be implemented using other than the one described here other modes, therefore, protection scope of the present invention is not by described below
Specific embodiment limitation.
Fig. 1 shows a kind of flow chart of the raindrop removing method based on raindrop probability graph of the present invention.
As shown in Figure 1, first aspect present invention proposes a kind of raindrop removing method based on raindrop probability graph, the method
Include:
S102 receives a raindrop image；
S104 constructs raindrop probability graph according to the prior information of raindrop；
S106 constrains the background layer and raindrop layer of the raindrop image by a priori assumption of natural image, with
Construct optimal model；
S108 solves the optimal model based on the raindrop probability graph, in the iteration for solving the optimal model
Update the raindrop probability graph simultaneously in the process；
S110 exports the image without raindrop.
It should be noted that technical solution of the present invention can be achieved in the terminal devices such as PC, mobile phone, PAD.
It should be noted that the prior information can be the characteristics such as physics, optics, the present invention is believed using the priori of raindrop
Breath analyzes the raindrop image, raindrop probability graph is obtained, on this basis, in conjunction with raindrop image background layer and raindrop layer
Sparsity structure constraint, background tomographic image is solved by way of construct optimal model, thus realization to raindrop figure
Raindrop as in are eliminated.
As shown in Fig. 2, in a particular embodiment, inputting a raindrop image first, raindrop are constructed according to the priori of raindrop
Probability graph constrains background layer and raindrop layer by a priori assumption of natural image, is then input to the two together
In optimal model, raindrop probability graph is updated simultaneously during the iterative solution of optimal model, final output is without raindrop
Image realizes the reinforcing effect to image, and can be used for the image analysis of outdoor vision system by enhanced image.
The generating principle of model is generated it is found that raindrop image J can be by background layer I's and raindrop layer R according to raindrop image
Linear superposition forms, relational expression are as follows:
J=(1- α) I+ α R；
Wherein, α illustrates possibility existing for raindrop, as raindrop probability in the raindrop image, in whole picture raindrop figure
The corresponding α of different pixels point is different as in, when α=1, represents the point and is capped completely, and when α=0, then representing the point, there is no rain
Drop.
Further, raindrop probability graph is constructed according to the prior information of raindrop, specifically included:
Raindrop probability graph, calculation formula are calculated using the shape smoothness index S of the raindrop image are as follows: α=eS-2π。
It should be noted that the present invention carries out calculating raindrop probability graph by this index of shape smoothness S.It is described smooth
Spend S be to be integrated by the tangent angle to borderline every bit in raindrop image, when the shape in the region be it is convex, then
Smoothness S is 2 π, and if it is non-convex or jagged, then smoothness S can be greater than 2 π, but since the shape of raindrop is usually convex
, so, when the corresponding smoothness S of certain region shape in raindrop image is 2 π, then illustrate that the point is that the probability α of raindrop connects
Nearly 1.
According to an embodiment of the invention, when carrying out raindrop elimination to raindrop image, it is necessary first to according to a preliminary estimate in image
Raindrop probability graph α.Specifically, can be calculated using the smoothness and satisfactory degree of shape the probability of raindrop.In this base
On plinth, then background layer I and raindrop layer R are constrained according to its a priori assumption, so as to construct the calculation of following optimal model
Method:
Wherein, first itemFor fidelity term, the background solved for ensuring algorithm
The combination of layer and raindrop layer meets raindrop image and generates model, Section 2 λ1Φ (I) and Section 3 λ2Ψ (R) is respectively constraints graph
The regularization term of picture, I indicate that background layer, R indicate raindrop layer, λ1And λ2For regularization parameter.
It should be noted that image can be used in the sparsity constraints in wavelet conversion domain about the Φ in optimal model,
Image can be used in gradient field energy constraint in Ψ, and the mode that alternating iteration can be used in the solution of optimal model carries out.
According to an embodiment of the invention, solving the optimal model based on the raindrop probability graph, specifically include:
Initialize installation, presetting the raindrop probability graph is α0, background layer I0=J, raindrop layer R0=J, wherein J is described
Raindrop image；
The optimal model is iteratively solved based on the raindrop probability graph；
When the algorithmic statement of the optimal model, final background tomographic image is exported, the as described raindrop image disappears
Except the image after raindrop.
Further, the optimal model is iteratively solved based on the raindrop probability graph, specifically included:
{ the α according to known to previous step in iterative processk, Ik, update Rk+1, more new formula are as follows:
According to updated { αk, Rk+1, update Ik+1, more new formula are as follows:
According to updated Rk+1, the smoothness S of the raindrop image shape is recalculated, and update raindrop probability graph
αk+1。
It should be noted that in the raindrop probability graph α updatedk+1, in conjunction with the I of updatek+1To update Rk+2, then
According to updated { αk+1, Rk+2Come update update Ik+2, then according to updated Rk+2, recalculate the raindrop image shape
The smoothness S of shape, and update raindrop probability graph αk+2, it is cyclically updated out newest raindrop probability graph α, background by this method
Tomographic image I, raindrop tomographic image R.
According to an embodiment of the invention, passing through background layer and raindrop of a priori assumption to the raindrop image of natural image
Layer is constrained, and is specifically included:
Sparsity using the background layer of raindrop image described in Tight wavelet frames operator constraint in Transformation Domain, constraint formulations
Are as follows: Ф (I)=| | WI | |1, wherein W is small wave operator.
According to an embodiment of the invention, passing through background layer and raindrop of a priori assumption to the raindrop image of natural image
Layer is constrained, specifically further include:
Constrain energy of the raindrop layer of the raindrop image in gradient field, constraint formulations are as follows: Its
InFor gradient operator.
It should be noted that the gradient operator can for Sobel operator, Roberts operator, Kirsch operator,
Laplace operator, Piewitt operator, Robinson operator, but not limited to this.
Fig. 3 shows the block diagram that a kind of raindrop based on raindrop probability graph of the present invention eliminate system.
As shown in figure 3, second aspect of the present invention also proposes that a kind of raindrop based on raindrop probability graph eliminate system 3, it is described
It includes: memory 31 and processor 32 that raindrop based on raindrop probability graph, which eliminate system 3, includes a kind of base in the memory 31
In the raindrop removing method program of raindrop probability graph, it is described based on the raindrop removing method program of raindrop probability graph by the processing
Device 32 realizes following steps when executing:
Receive a raindrop image；
Raindrop probability graph is constructed according to the prior information of raindrop；
The background layer and raindrop layer of the raindrop image are constrained by a priori assumption of natural image, to construct most
Optimized model；
The optimal model is solved based on the raindrop probability graph, in the iterative process for solving the optimal model
The raindrop probability graph is updated simultaneously；
Export the image without raindrop.
It should be noted that system of the invention can be operated in the terminal devices such as PC, mobile phone, PAD.
It should be noted that the processor can be central processing unit (Central Processing Unit,
CPU), it can also be other general processors, Digital Signal Processing (Digital Signal Processor, DSP), dedicated collection
At circuit (Application Specific Integrated Circuit, ASIC), ready-made programmable gate array (Field-
Programmable Gate Array, FPGA) either other programmable logic device, discrete gate or transistor logic,
Discrete hardware components etc..General processor can be microprocessor or the processor is also possible to any conventional processor
Deng.
It should be noted that the system can also include display, the display is properly termed as display screen or display
Unit.Display can be light-emitting diode display, liquid crystal display, touch-control liquid crystal display and organic in some embodiments
Light emitting diode (Organic Light-Emitting Diode, OLED) touches device etc..Display is for showing in systems
The information of processing, output without raindrop image and for showing visual working interface.
It should be noted that the prior information can be the characteristics such as physics, optics, the present invention is believed using the priori of raindrop
Breath analyzes the raindrop image, raindrop probability graph is obtained, on this basis, in conjunction with raindrop image background layer and raindrop layer
Sparsity structure constraint, background tomographic image is solved by way of construct optimal model, thus realization to raindrop figure
Raindrop as in are eliminated.
In a particular embodiment, a raindrop image is inputted first, and raindrop probability graph is constructed according to the priori of raindrop, is passed through
The a priori assumption of natural image constrains background layer and raindrop layer, and the two is then input to optimal model together
In, update raindrop probability graph, image of the final output without raindrop, realization pair simultaneously during the iterative solution of optimal model
The reinforcing effect of image, and can be used for by enhanced image the image analysis of outdoor vision system.
The generating principle of model is generated it is found that raindrop image J can be by background layer I's and raindrop layer R according to raindrop image
Linear superposition forms, relational expression are as follows:
J=(1- α) I+ α R；
Wherein, α illustrates possibility existing for raindrop, as raindrop probability in the raindrop image, in whole picture raindrop figure
The corresponding α of different pixels point is different as in, when α=1, represents the point and is capped completely, and when α=0, then representing the point, there is no rain
Drop.
Further, raindrop probability graph is constructed according to the prior information of raindrop, specifically included:
Raindrop probability graph, calculation formula are calculated using the shape smoothness index S of the raindrop image are as follows: α=eS-2π。
It should be noted that the present invention carries out calculating raindrop probability graph by this index of shape smoothness S.It is described smooth
Spend S be to be integrated by the tangent angle to borderline every bit in raindrop image, when the shape in the region be it is convex, then
Smoothness S is 2 π, and if it is non-convex or jagged, then smoothness S can be greater than 2 π, but since the shape of raindrop is usually convex
, so, when the corresponding smoothness S of certain region shape in raindrop image is 2 π, then illustrate that the point is that the probability α of raindrop connects
Nearly 1.
According to an embodiment of the invention, when carrying out raindrop elimination to raindrop image, it is necessary first to according to a preliminary estimate in image
Raindrop probability graph α.Specifically, can be calculated using the smoothness and satisfactory degree of shape the probability of raindrop.In this base
On plinth, then background layer I and raindrop layer R are constrained according to its a priori assumption, so as to construct the calculation of following optimal model
Method:
Wherein, first itemFor fidelity term, the background solved for ensuring algorithm
The combination of layer and raindrop layer meets raindrop image and generates model, Section 2 λ1Ф (I) and Section 3 λ2Ψ (R) is respectively constraints graph
The regularization term of picture, I indicate that background layer, R indicate raindrop layer, λ1And λ2For regularization parameter.
It should be noted that image can be used in the sparsity constraints in wavelet conversion domain about the Φ in optimal model,
Image can be used in gradient field energy constraint in Ψ, and the mode that alternating iteration can be used in the solution of optimal model carries out.
According to an embodiment of the invention, solving the optimal model based on the raindrop probability graph, specifically include:
Initialize installation, presetting the raindrop probability graph is α0, background layer I0=J, raindrop layer R0=J, wherein J is described
Raindrop image；
The optimal model is iteratively solved based on the raindrop probability graph；
When the algorithmic statement of the optimal model, final background tomographic image is exported, the as described raindrop image disappears
Except the image after raindrop.
Further, the optimal model is iteratively solved based on the raindrop probability graph, specifically included:
{ the α according to known to previous step in iterative processk, Ik, update Rk+1, more new formula are as follows:
According to updated { αk, Rk+1, update Ik+1, more new formula are as follows:
According to updated Rk+1, the smoothness S of the raindrop image shape is recalculated, and update raindrop probability graph
αk+1。
It should be noted that in the raindrop probability graph α updatedk+1, in conjunction with the I of updatek+1To update Rk+2, then
According to updated { αk+1, Rk+2Come update update Ik+2, then according to updated Rk+2, recalculate the raindrop image shape
The smoothness S of shape, and update raindrop probability graph αk+2, it is cyclically updated out newest raindrop probability graph α, background by this method
Tomographic image I, raindrop tomographic image R.
According to an embodiment of the invention, passing through background layer and raindrop of a priori assumption to the raindrop image of natural image
Layer is constrained, and is specifically included:
Sparsity using the background layer of raindrop image described in Tight wavelet frames operator constraint in Transformation Domain, constraint formulations
Are as follows: Φ (I)=| | WI | |1, wherein W is small wave operator.
According to an embodiment of the invention, passing through background layer and raindrop of a priori assumption to the raindrop image of natural image
Layer is constrained, specifically further include:
Constrain energy of the raindrop layer of the raindrop image in gradient field, constraint formulations are as follows: Its
InFor gradient operator.
It should be noted that the gradient operator can for Sobel operator, Roberts operator, Kirsch operator,
Laplace operator, Piewitt operator, Robinson operator, but not limited to this.
Third aspect present invention also proposes a kind of computer readable storage medium, wraps in the computer readable storage medium
Include a kind of raindrop removing method program based on raindrop probability graph, the raindrop removing method program quilt based on raindrop probability graph
When processor executes, realize such as the step of a kind of above-mentioned raindrop removing method based on raindrop probability graph.
The present invention proposes a kind of raindrop removing method, system and storage medium based on raindrop probability graph, first uses raindrop
Style characteristic estimate raindrop probability graph, then by the prior-constrained building optimal model to background layer and raindrop layer, with
Image after the elimination of finally obtained background tomographic image, as raindrop.In the existing raindrop minimizing technology in raindrop image
It mostly uses and raindrop is detected, then to the mode that the raindrop detected are repaired, damaged to repair in image by raindrop
Pixel region.Compared with prior art, the present invention can more precisely eliminate the raindrop in image, construct the nothing of high quality
Raindrop image, and can realize effective enhancing to raindrop image.
In several embodiments provided herein, it should be understood that disclosed device and method can pass through it
Its mode is realized.Apparatus embodiments described above are merely indicative, for example, the division of the unit, only
A kind of logical function partition, there may be another division manner in actual implementation, such as: multiple units or components can combine, or
It is desirably integrated into another system, or some features can be ignored or not executed.In addition, shown or discussed each composition portion
Mutual coupling or direct-coupling or communication connection is divided to can be through some interfaces, the INDIRECT COUPLING of equipment or unit
Or communication connection, it can be electrical, mechanical or other forms.
Above-mentioned unit as illustrated by the separation member, which can be or may not be, to be physically separated, aobvious as unit
The component shown can be or may not be physical unit；Both it can be located in one place, and may be distributed over multiple network lists
In member；Some or all of units can be selected to achieve the purpose of the solution of this embodiment according to the actual needs.
In addition, each functional unit in various embodiments of the present invention can be fully integrated in one processing unit, it can also
To be each unit individually as a unit, can also be integrated in one unit with two or more units；It is above-mentioned
Integrated unit both can take the form of hardware realization, can also realize in the form of hardware adds SFU software functional unit.
Those of ordinary skill in the art will appreciate that: realize that all or part of the steps of above method embodiment can pass through
The relevant hardware of program instruction is completed, and program above-mentioned can store in computer-readable storage medium, which exists
When execution, step including the steps of the foregoing method embodiments is executed；And storage medium above-mentioned includes: movable storage device, read-only deposits
Reservoir (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or
The various media that can store program code such as CD.
If alternatively, the above-mentioned integrated unit of the present invention is realized in the form of software function module and as independent product
When selling or using, it also can store in a computer readable storage medium.Based on this understanding, the present invention is implemented
Substantially the part that contributes to existing technology can be embodied in the form of software products the technical solution of example in other words,
The computer software product is stored in a storage medium, including some instructions are used so that computer equipment (can be with
It is personal computer, server or network equipment etc.) execute all or part of each embodiment the method for the present invention.
And storage medium above-mentioned includes: that movable storage device, ROM, RAM, magnetic or disk etc. are various can store program code
Medium.
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can easily think of the change or the replacement, and should all contain
Lid is within protection scope of the present invention.Therefore, protection scope of the present invention should be based on the protection scope of the described claims.
Claims (10)
1. a kind of raindrop removing method based on raindrop probability graph, which is characterized in that the described method includes:
Receive a raindrop image；
Raindrop probability graph is constructed according to the prior information of raindrop；
The background layer and raindrop layer of the raindrop image are constrained by a priori assumption of natural image, optimized with constructing
Model；
The optimal model is solved based on the raindrop probability graph, in the iterative process for solving the optimal model simultaneously
Update the raindrop probability graph；
Export the image without raindrop.
2. a kind of raindrop removing method based on raindrop probability graph according to claim 1, which is characterized in that according to raindrop
Prior information construct raindrop probability graph, specifically include:
Raindrop probability graph, calculation formula are calculated using the shape smoothness index S of the raindrop image are as follows: α=eS-2π。
3. a kind of raindrop removing method based on raindrop probability graph according to claim 1, which is characterized in that described optimal
Change the algorithm of model are as follows:
Wherein, first itemFor fidelity term, for ensure algorithm solve the background layer come with
The combination of raindrop layer meets raindrop image and generates model, Section 2 λ1Φ (I) and Section 3 λ2Ψ (R) is respectively to constrain image
Regularization term, I indicate that background layer, R indicate raindrop layer, λ1And λ2For regularization parameter.
4. a kind of raindrop removing method based on raindrop probability graph according to claim 1, which is characterized in that pass through nature
The a priori assumption of image constrains the background layer and raindrop layer of the raindrop image, specifically includes:
Sparsity using the background layer of raindrop image described in Tight wavelet frames operator constraint in Transformation Domain, constraint formulations are as follows: Φ
(I)=‖ WI ‖1, wherein W is small wave operator.
5. a kind of raindrop removing method based on raindrop probability graph according to claim 1, which is characterized in that pass through nature
The a priori assumption of image constrains the background layer and raindrop layer of the raindrop image, specifically further include:
Constrain energy of the raindrop layer of the raindrop image in gradient field, constraint formulations are as follows: WhereinFor
Gradient operator.
6. a kind of raindrop removing method based on raindrop probability graph according to claim 1, which is characterized in that based on described
Raindrop probability graph solves the optimal model, specifically includes:
Initialize installation, presetting the raindrop probability graph is α0, background layer I0=J, raindrop layer R0=J, wherein J is the raindrop
Image；
The optimal model is iteratively solved based on the raindrop probability graph；
When the algorithmic statement of the optimal model, final background tomographic image is exported, the as described raindrop image eliminates rain
Image after drop.
7. a kind of raindrop removing method based on raindrop probability graph according to claim 6, which is characterized in that based on described
Raindrop probability graph iteratively solves the optimal model, specifically includes:
{ the α according to known to previous step in iterative processk, Ik, update Rk+1, more new formula are as follows:
According to updated { αk, Rk+1, update Ik+1, more new formula are as follows:
According to updated Rk+1, the smoothness S of the raindrop image shape is recalculated, and update raindrop probability graph αk+1。
8. a kind of raindrop based on raindrop probability graph eliminate system, which is characterized in that the raindrop based on raindrop probability graph disappear
It include a kind of raindrop removing method journey based on raindrop probability graph in the memory except system includes: memory and processor
Sequence, the raindrop removing method program based on raindrop probability graph realize following steps when being executed by the processor:
Receive a raindrop image；
Raindrop probability graph is constructed according to the prior information of raindrop；
The background layer and raindrop layer of the raindrop image are constrained by a priori assumption of natural image, optimized with constructing
Model；
The optimal model is solved based on the raindrop probability graph, in the iterative process for solving the optimal model simultaneously
Update the raindrop probability graph；
Export the image without raindrop.
9. a kind of raindrop based on raindrop probability graph according to claim 8 eliminate system, which is characterized in that based on described
Raindrop probability graph solves the optimal model, specifically includes:
Initialize installation, presetting the raindrop probability graph is α0, background layer I0=J, raindrop layer R0=J, wherein J is the raindrop
Image；
The optimal model is iteratively solved based on the raindrop probability graph；
When the algorithmic statement of the optimal model, final background tomographic image is exported, the as described raindrop image eliminates rain
Image after drop.
10. a kind of computer readable storage medium, which is characterized in that be based in the computer readable storage medium including one kind
The raindrop removing method program of raindrop probability graph, the raindrop removing method program based on raindrop probability graph are executed by processor
When, the step of realizing a kind of raindrop removing method based on raindrop probability graph as described in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910737935.0A CN110533581B (en) | 2019-08-12 | 2019-08-12 | Raindrop eliminating method, system and storage medium based on raindrop probability map |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910737935.0A CN110533581B (en) | 2019-08-12 | 2019-08-12 | Raindrop eliminating method, system and storage medium based on raindrop probability map |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110533581A true CN110533581A (en) | 2019-12-03 |
| CN110533581B CN110533581B (en) | 2023-05-02 |
Family
ID=68662954
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910737935.0A Active CN110533581B (en) | 2019-08-12 | 2019-08-12 | Raindrop eliminating method, system and storage medium based on raindrop probability map |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110533581B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112330739A (en) * | 2020-10-27 | 2021-02-05 | 北京航空航天大学 | Satellite detection method based on optical probability map model |
| CN117152000A (en) * | 2023-08-08 | 2023-12-01 | 华中科技大学 | Rainy day image-clear background paired data set manufacturing method and device and application thereof |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102707340A (en) * | 2012-06-06 | 2012-10-03 | 南京大学 | Rainfall measurement method based on video images |
| CN104331865A (en) * | 2014-10-22 | 2015-02-04 | 中国科学院深圳先进技术研究院 | Video raindrop detection and removing method based on naive Bayesian probability model |
| CN105335949A (en) * | 2015-08-28 | 2016-02-17 | 中国科学院深圳先进技术研究院 | Video image rain removal method and system |
| CN106709926A (en) * | 2016-12-12 | 2017-05-24 | 四川大学 | Quick calculation rain removal algorithm based on dynamic priori knowledge estimation |
| CN108230278A (en) * | 2018-02-24 | 2018-06-29 | 中山大学 | A kind of image based on generation confrontation network goes raindrop method |
| CN110009578A (en) * | 2019-03-13 | 2019-07-12 | 五邑大学 | A kind of image rain removing method and its system, device, storage medium |
| CN110111267A (en) * | 2019-04-17 | 2019-08-09 | 大连理工大学 | A kind of single image based on optimization algorithm combination residual error network removes rain method |
-
2019
- 2019-08-12 CN CN201910737935.0A patent/CN110533581B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102707340A (en) * | 2012-06-06 | 2012-10-03 | 南京大学 | Rainfall measurement method based on video images |
| CN104331865A (en) * | 2014-10-22 | 2015-02-04 | 中国科学院深圳先进技术研究院 | Video raindrop detection and removing method based on naive Bayesian probability model |
| CN105335949A (en) * | 2015-08-28 | 2016-02-17 | 中国科学院深圳先进技术研究院 | Video image rain removal method and system |
| CN106709926A (en) * | 2016-12-12 | 2017-05-24 | 四川大学 | Quick calculation rain removal algorithm based on dynamic priori knowledge estimation |
| CN108230278A (en) * | 2018-02-24 | 2018-06-29 | 中山大学 | A kind of image based on generation confrontation network goes raindrop method |
| CN110009578A (en) * | 2019-03-13 | 2019-07-12 | 五邑大学 | A kind of image rain removing method and its system, device, storage medium |
| CN110111267A (en) * | 2019-04-17 | 2019-08-09 | 大连理工大学 | A kind of single image based on optimization algorithm combination residual error network removes rain method |
Non-Patent Citations (2)
| Title |
|---|
| YU LUO ET AL: "Fast Removal of Rain Streaks From a Single Image via a Shape Prior" * |
| 林向伟等: "基于多细节卷积神经网络的单幅图像去雨方法", 《信号处理》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112330739A (en) * | 2020-10-27 | 2021-02-05 | 北京航空航天大学 | Satellite detection method based on optical probability map model |
| CN112330739B (en) * | 2020-10-27 | 2022-05-27 | 北京航空航天大学 | Satellite detection method based on optical probability map model |
| CN117152000A (en) * | 2023-08-08 | 2023-12-01 | 华中科技大学 | Rainy day image-clear background paired data set manufacturing method and device and application thereof |
| CN117152000B (en) * | 2023-08-08 | 2024-05-14 | 华中科技大学 | Rainy day image-clear background paired data set manufacturing method and device and application thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110533581B (en) | 2023-05-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11200696B2 (en) | Method and apparatus for training 6D pose estimation network based on deep learning iterative matching | |
| US20240070546A1 (en) | System and method for end-to-end differentiable joint image refinement and perception | |
| Dong et al. | Multi-scale boosted dehazing network with dense feature fusion | |
| CN109754417B (en) | System and method for unsupervised learning of geometry from images | |
| US10984556B2 (en) | Method and apparatus for calibrating relative parameters of collector, device and storage medium | |
| Wexler et al. | Space-time completion of video | |
| Michaeli et al. | Nonparametric blind super-resolution | |
| Kuanar et al. | Multi-path dilated convolution network for haze and glow removal in nighttime images | |
| US11676257B2 (en) | Method and device for detecting defect of meal box, server, and storage medium | |
| US11314989B2 (en) | Training a generative model and a discriminative model | |
| US20170178293A1 (en) | Method for performing super-resolution on single images and apparatus for performing super-resolution on single images | |
| EP3721382B1 (en) | Method and system of annotation densification for semantic segmentation | |
| CN115082329A (en) | Generating modified digital images using a deep visual guide patch matching model for image inpainting | |
| CN110992244B (en) | Picture generation method, system, device and storage medium with mole pattern | |
| US20220392025A1 (en) | Restoring degraded digital images through a deep learning framework | |
| CN110533581A (en) | A kind of raindrop removing method, system and storage medium based on raindrop probability graph | |
| CN108648149B (en) | Image splicing method, system, equipment and storage medium based on augmented reality | |
| Liu et al. | Infrared and visible image fusion and denoising via ℓ2− ℓp norm minimization | |
| US11069034B2 (en) | Method and system to enhance quality of digital images | |
| US10521918B2 (en) | Method and device for filtering texture, using patch shift | |
| Wang et al. | Total generalized variation-based Retinex image decomposition | |
| US20160217556A1 (en) | Systems and methods for automatic image enhancement utilizing feedback control | |
| CN110738624B (en) | Area-adaptive image defogging system and method | |
| Wang et al. | Image dehazing based on partitioning reconstruction and entropy-based alternating fast-weighted guided filters | |
| Zheng et al. | Memory-efficient multi-scale residual dense network for single image rain removal |
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 |