CN109767462A - VideoSAR interframe method for registering based on quantum particle swarm - Google Patents
VideoSAR interframe method for registering based on quantum particle swarm Download PDFInfo
- Publication number
- CN109767462A CN109767462A CN201811652772.8A CN201811652772A CN109767462A CN 109767462 A CN109767462 A CN 109767462A CN 201811652772 A CN201811652772 A CN 201811652772A CN 109767462 A CN109767462 A CN 109767462A
- Authority
- CN
- China
- Prior art keywords
- image
- particle
- registration
- parameter
- reference picture
- 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.)
- Pending
Links
Landscapes
- Image Processing (AREA)
Abstract
The invention belongs to technical field of image processing, the VideoSAR interframe method for registering based on quantum particle swarm is disclosed, this method comprises the following steps: obtaining SAR floating image to be registered and corresponding reference picture；Transformation parameter is sought using quanta particle swarm optimization, floating image is converted；Using the pixel of image as Multidimensional Point Set, point set includes the point of certain single pixel point and its neighborhood, constructs multi-dimensions histogram using point set, calculates the region mutual information between reference picture and changing image；When region mutual information does not change with the number of iterations and reaches maximum, then registration parameter at this time is optimal registration parameter；Image registration is carried out to floating image using optimal registration parameter, the SAR image after being registrated.The present invention can quickly get rid of local optimum to global convergence, the entire search space of covering, to effectively improve registration accuracy and with Quasi velosity.
Description
Technical field
The present invention relates to technical field of image processing, more particularly to the VideoSAR interframe registration side based on quantum particle swarm
Method can effectively solve to be easily trapped into office in traditional method for registering for two width or the high-precision rapid registering of multiple image
Portion's extreme value and it is difficult to the problem of covering entire search space, to effectively improve registration accuracy and with Quasi velosity.
Background technique
Image registration is exactly to carry out two width or multiple image that obtain under different time, different sensors or different condition
The process of matching, superposition.Image registration is in target detection, characteristic matching, the fields such as lesion detection, lesion positioning, geological prospecting
All it is widely used.Lead to the erroneous estimation of transformation parameter since conventional method is easily trapped into local extremum, is in addition also difficult to
Entire search space is covered, therefore studies the high method for registering images of fast convergence rate, registration accuracy and has great importance.
Currently, main method for registering is method for registering based on mutual information and standard particle group's algorithm in the prior art.
Method for registering advantage based on mutual information is not having to consider the significant geometrical characteristic of image data, thus avoids feature extraction band
The extra error come, and can be realized the autoregistration of image.However, this method, which is easily trapped into local extremum, leads to transformation parameter
Erroneous estimation.Standard particle group's algorithm advantage is spy that is easy-to-understand, while assigning each particle self-teaching and summary
Property is with boosting algorithm efficiency.But the flying speed of particle is restricted, and is caused to be difficult to sufficiently to cover entire search space etc. and be asked
Topic.
Summary of the invention
It is a kind of based on quantum particle swarm it is an object of the invention in view of the above shortcomings of the prior art, propose
VideoSAR interframe method for registering can cover entire search space and can quickly jump out locally optimal solution acquisition global optimum
Solution, to effectively improve registration accuracy and with Quasi velosity.
Technical thought of the invention is: being included in adjacent pixels as mutual information statistical regions (region mutual information) with reinforced phase
The robustness measured like property；Entire search space is covered using the randomness of quantum, passes through quantum particle swarm intelligent search algorithm
The globally optimal solution of geometric transformation parameter is obtained with high probability.It is handled by the VideoSAR measured data to the laboratory Sandia
As a result, verifying effectiveness of the invention.
According to above-mentioned thinking, technical solution of the present invention includes the following:
Step 1, SAR floating image to be registered and corresponding reference picture are obtained；
Step 2, transformation parameter is sought using quanta particle swarm optimization, floating image is converted；
Step 3, the region mutual information between changing image obtained in reference picture and step 2 is calculated；
Step 4, when region mutual information does not change with the number of iterations and reaches maximum, then registration parameter at this time is optimal
Registration parameter
Step 5, image registration is carried out using optimal registration parameter, the SAR image after being registrated.
The present invention has the advantage that compared with prior art
1. quantum particle swarm intelligent search algorithm employed in the present invention can quickly get rid of local optimum and receive to the overall situation
It holds back；
2. the randomness using quantum can cover entire search space.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described:
Fig. 1 is the flow diagram of the VideoSAR interframe method for registering based on quantum particle swarm；
Fig. 2 (a) is the reference picture with quasi-experiment；
Fig. 2 (b) is the floating image with quasi-experiment；
Fig. 3 is registration result figure；
Fig. 4 is the Overlapping display figure of registration result image and reference picture；
Optimal Parameters application condition figure of the Fig. 5 based on region mutual information；
Fig. 6 is change curve of the two methods region mutual information with the number of iterations.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Fig. 1 show a kind of VideoSAR interframe method for registering based on quantum particle swarm provided in an embodiment of the present invention
Flow diagram.
As shown in Figure 1, a kind of VideoSAR interframe method for registering based on quantum particle swarm provided in an embodiment of the present invention
The following steps are included:
Step 1, SAR floating image and reference picture to be registered are obtained.
Wherein, reference picture and floating image are as shown in Fig. 2, their size is m × n.
Step 2, transformation parameter is sought using quanta particle swarm optimization, floating image is converted.
Each particle i includes following message: setting problem to be optimized in N-dimensional solution space, group X={ X1,X2,…,XmPacket
Containing m particle, in t moment, coordinate of i-th of particle in solution space is denoted as Xi(t)=[Xi,1(t),Xi,2(t),…,Xi,N
(t)], particle only has location information without velocity information, wherein i=1,2 ..., m.Use Pi(t)=[Pi,1(t),Pi,2
(t),…,Pi,N(t)] the optimum adaptation value of particle individual, i.e. local optimum are indicated；With G (t)=[G1(t),G2
(t),…,GN(t)] the optimum adaptation value of group, i.e. global optimum are indicated.And there is G (t)=Pg(t), wherein g indicates group
There is the corresponding serial number of particle of global desired positions, g ∈ { 1,2 ..., m } in body.
(2a) initializes rotation angle γ, the X-direction translational movement T of particle in search spacex, Y-direction translational movement TyIt constitutes
Solution space X.It is randomly assigned position of the n particle in solution space.Particle number is n, greatest iteration step number itermax。
(2b) calculates cost function value by objective function C (R, T)；
Wherein, xiFor the control point in floating image X, number N；yiFor the series of control point in floating image Y, number
For M；R is the translation transformation in 3 axial directions.
(2c) calculates the average desired positions C (t) of particle, updates local optimum and global optimum.
(2d) calculates random point pi,j(t), the position X of more new particlei,j(t+1)；
Xi,j(t+1)=pi,j(t)±α·|Cj(t)-Xi,j(t)|·ln(1/ui,j(t))
Wherein,And ui,j(t) random number being generally evenly distributed between 0 to 1, ± according to ui,j(t) difference makees phase
The dynamic adjustment answered takes negative sign if random number is greater than 0.5, on the contrary then take positive sign；α is known as expanding contraction factor, plays decision
The effect of particle contraction speed.
Using being carried out with iteration and the control strategy of linear decrease determines α, i.e., as shown by:
α=(α0-α1)×(itermax-k)/itermax+α1
In formula, α0> α1, α0For the initial value of control parameter, α1For the end value of control parameter, itermaxIt changes for maximum
Generation number, k are current iteration number.
(2e) judges whether to meet iterative steps.It is unsatisfactory for returning to (2b)；Satisfaction then terminates.With the transformation parameter sought
(γ,Tx,Ty), spatial alternation is carried out to floating image and obtains changing image.And image space is calculated using bilinear interpolation and is become
Change the gray value of rear every pixel.
Step 3, the region mutual information between changing image obtained in reference picture and step 2 is calculated.By the picture of image
For vegetarian refreshments as Multidimensional Point Set, point set includes the point of certain single pixel point and its neighborhood, constructs multi-dimensions histogram using point set.
(3a) establishes the gray scale value matrix P indicated on changing image and floating image and 8 pixels adjacent around it.
Every group of pixel of reference picture and changing image is to can indicate are as follows:
Q (x, y)=[R (x, y), F (x, y)]
Wherein, Q (x, y) indicates that pixel pair, R (x, y) indicate that reference picture, F (x, y) indicate changing image.
The square window for being r with radius is successively taken the picture in window to the pixel of two images by row respectively to Q (x, y)
Vegetarian refreshments, it is available two (2r+1)2The head and the tail connection of the two column vectors is then generated d=2 (2r+1) by the column vector of dimension2Dimension
Vector pi, then can produce N=(m-2r) (n-2r) a d dimensional vector, form the matrix of a d × N.
P=(p1,p2,...,pN)
(3b) calculates covariance matrix C.
The average value, which is subtracted, using element each in P obtains sequence normalization vector P0, expression formula are as follows:
Utilize sequence normalization vector P0Calculate covariance matrix C, expression formula are as follows:
(3c) calculates combination entropy Hg(C) and edge entropy Hg(CR)、Hg(CF)
Assuming that higher-dimension distribution is similar to normal distribution, since each dimension is independently of each other, can to tie up in distribution and divide from d
D independent one-dimensional distributions are separated out, then normal distribution and the covariance ∑ of d dimension point set entropydRelationship it is as follows:
Then combination entropy Hg(C) and edge entropy Hg(CR)、Hg(CF) be respectively as follows:
Wherein, Hg(CR) it is by calculating the upper left corner covariance matrix CThe edge for the reference image R that matrix obtains
Entropy, Hg(CF) it is by calculating the lower right corner covariance matrix CThe edge entropy for the reference picture F that matrix obtains.
The zoning (3d) mutual information RMI:
RMI=Hg(CR)+Hg(CF)-Hg(C)
Step 4, when region mutual information does not change with the number of iterations and reaches maximum, then registration parameter at this time is optimal
Registration parameter；
Step 5, image registration is carried out using the optimal registration parameter that step 4 obtains, the SAR image after being registrated.
It further, is the correctness for verifying the above method of the present invention, VideoSAR experimental data used derives from
The admission of SandiaNational Laboratories, SAR video is U.S.'s Kirtland Air Force Base a certain gate wagon flow field
Scape.Using VideoSAR framing and selection, wherein for two frames as reference picture and floating image, reference picture is Fig. 2 (a), and float figure
As being Fig. 2 (b), reference picture and floating image are registrated using the above method of the present invention, and is carried out with conventional method
Compare, detailed process is as follows:
(1) using region mutual information as similarity measurement, affine transformation is as geometric transformation model, quanta particle swarm optimization
As search strategy, population population n=40, the number of iterations iter are provided that major parametermax=100, expansion is shunk
Factor-alpha initial value is 1.0, and carries out reducing to 0.5 with iteration.Registration result is as shown in Figure 3.By images after registration and reference picture
Overlapping display it is as shown in Figure 4.
(2) in order to verify superiority of the above method of the present invention compared to conventional method, following comparative test has been carried out.
Experiment is classified into two groups of progress: first group of method for registering (RMI+ for region mutual information combination particle swarm algorithm
PSO)；Second group of method for registering (RMI+QPSO) for region mutual information incorporating quantum particle swarm algorithm.
In experiment, population is set as 20, the number of iterations 100, inertia weight w and the QPSO algorithm of PSO algorithm
Shrinkage expansion factor-alpha initial value is 1.0, and carries out reducing to 0.5 with iteration, actual parameter txIt is set as 4, tyIt is set as 3, θ and is set as 6.
In the above conditions, 50 times are repeated with quasi-experiment to experimental data, acquires the optimization ginseng of two kinds of optimization methods
Number application condition figure is as shown in figure 5, region mutual information is as shown in Figure 6 with the change curve of the number of iterations.XRMSE in Fig. 5,
YRMSE and θ RMSE respectively indicates horizontal direction, vertical direction and the root-mean-square error value for rotating angle.
By above-mentioned simulation result it can be seen that two kinds of method for registering root-mean-square errors are not much different, also all 1 pixel with
It is interior, illustrate that registration has reached sub-pixel, but QPSO convergence speed of the algorithm is still significantly faster than that PSO, can quickly get rid of office
Portion is optimal to global convergence.
The foregoing is merely a prefered embodiment of the invention, is not intended to limit the invention, all in the spirit and principles in the present invention
Within, any modification, equivalent replacement, improvement and so on should all be included in the protection scope of the present invention.
Claims (3)
1. a kind of VideoSAR interframe method for registering based on quantum particle swarm, which comprises the following steps:
Step 1, SAR floating image to be registered and corresponding reference picture are obtained；
Step 2, transformation parameter is sought using quanta particle swarm optimization, floating image is converted；
Step 3, the region mutual information between changing image obtained in reference picture and step 2 is calculated；
Step 4, when region mutual information does not change with the number of iterations and reaches maximum, then registration parameter at this time is optimal registration
Parameter；
Step 5, image registration is carried out using optimal registration parameter, the SAR image after being registrated.
2. the method according to claim 1, wherein step 2 specifically includes following sub-step:
(2a) initializes rotation angle γ, the X-direction translational movement T of particle in search spacex, Y-direction translational movement TyIt is empty to constitute solution
Between X, be randomly assigned position of the n particle in solution space, particle number n, greatest iteration step number itermax；
(2b) calculates cost function value by objective function C (R, T)；
Wherein, xiFor the control point in floating image X, number N；yiFor the series of control point in floating image Y, number M；R
For the translation transformation in 3 axial directions；
(2c) calculates the average desired positions C (t) of particle, updates local optimum and global optimum；
(2d) utilizes (2c) obtained particle be averaged desired positions C (t) calculating random point pi,j(t), the position X of more new particlei,j
(t+1)；
Xi,j(t+1)=pi,j(t)±α·|Cj(t)-Xi,j(t)|·ln(1/ui,j(t))
Wherein,And ui,j(t) random number being generally evenly distributed between 0 to 1, ± according to ui,j(t) different works are corresponding
Dynamic adjusts, if random number is greater than 0.5, takes negative sign, on the contrary then take positive sign；α is known as expanding contraction factor, plays decision particle
The effect of contraction speed；
(2e) judges whether to meet iterative steps, is unsatisfactory for returning to (2b), satisfaction then terminates；With transformation parameter (γ, the T soughtx,
Ty), spatial alternation is carried out to floating image and obtains changing image, and is calculated using bilinear interpolation every after Image space transformation
The gray value of pixel.
3. the method according to claim 1, wherein step 3 specifically includes following sub-step:
(3a) establishes the gray scale value matrix P indicated on changing image and floating image and 8 pixels adjacent around it；
P=(p1,p2,...,pN)
(3b) calculates covariance matrix C；
(3b1) utilizes each element in P obtained by (3a) to subtract average value acquisition sequence normalization vector P0, expression formula are as follows:
(3b1) utilizes normalized vector P obtained by (3b1)0Calculate covariance matrix C, expression formula are as follows:
(3c) calculates combination entropy Hg(C) and edge entropy Hg(CR)、Hg(CF)；
Wherein, Hg(CR) it is by calculating the upper left corner covariance matrix CThe edge entropy for the reference image R that matrix obtains, Hg
(CF) it is by calculating the lower right corner covariance matrix CThe edge entropy for the reference picture F that matrix obtains；
The zoning (3d) mutual information RMI:
RMI=Hg(CR)+Hg(CF)-Hg(C)。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811652772.8A CN109767462A (en) | 2018-12-29 | 2018-12-29 | VideoSAR interframe method for registering based on quantum particle swarm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811652772.8A CN109767462A (en) | 2018-12-29 | 2018-12-29 | VideoSAR interframe method for registering based on quantum particle swarm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109767462A true CN109767462A (en) | 2019-05-17 |
Family
ID=66453155
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811652772.8A Pending CN109767462A (en) | 2018-12-29 | 2018-12-29 | VideoSAR interframe method for registering based on quantum particle swarm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109767462A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111583311A (en) * | 2020-05-14 | 2020-08-25 | 重庆理工大学 | PCBA rapid image matching method |
| CN112750152A (en) * | 2021-01-18 | 2021-05-04 | 上海联影医疗科技股份有限公司 | Image registration method and device, computer equipment and storage medium |
| CN116152316A (en) * | 2023-04-17 | 2023-05-23 | 山东省工业技术研究院 | Image registration method based on self-adaptive parameter particle swarm algorithm |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101216939A (en) * | 2008-01-04 | 2008-07-09 | 江南大学 | A multi-resolution medical image registration method based on quantum behaviors particle swarm algorithm |
-
2018
- 2018-12-29 CN CN201811652772.8A patent/CN109767462A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101216939A (en) * | 2008-01-04 | 2008-07-09 | 江南大学 | A multi-resolution medical image registration method based on quantum behaviors particle swarm algorithm |
Non-Patent Citations (5)
| Title |
|---|
| 潘婷婷等: "QPSO算法和Powell法结合的多分辨率医学图像配准", 《计算机应用与软件》 * |
| 王小根等: "基于最大互信息和量子粒子群优化算法的医学图像配准研究", 《计算机应用研究》 * |
| 谢振平等: "一种基于混合量子行为粒子群优化算法的图像配准方法", 《计算机应用研究》 * |
| 赵辽英等: "基于尺度不变特征变换和区域互信息优化的多源遥感图像配准", 《物理学报》 * |
| 陈庆芳等: "基于分块互信息和量子粒子群算法的图像配准", 《数据采集与处理》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111583311A (en) * | 2020-05-14 | 2020-08-25 | 重庆理工大学 | PCBA rapid image matching method |
| CN111583311B (en) * | 2020-05-14 | 2023-04-18 | 重庆理工大学 | PCBA rapid image matching method |
| CN112750152A (en) * | 2021-01-18 | 2021-05-04 | 上海联影医疗科技股份有限公司 | Image registration method and device, computer equipment and storage medium |
| CN116152316A (en) * | 2023-04-17 | 2023-05-23 | 山东省工业技术研究院 | Image registration method based on self-adaptive parameter particle swarm algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Bai et al. | Adaptive dilated network with self-correction supervision for counting | |
| CN109767462A (en) | VideoSAR interframe method for registering based on quantum particle swarm | |
| CN105809712B (en) | A kind of efficient big displacement light stream method of estimation | |
| CN110111248B (en) | Image splicing method based on feature points, virtual reality system and camera | |
| CN112001926B (en) | RGBD multi-camera calibration method, system and application based on multi-dimensional semantic mapping | |
| CN107833270A (en) | Real-time object dimensional method for reconstructing based on depth camera | |
| CN105719352B (en) | Face three-dimensional point cloud super-resolution fusion method and apply its data processing equipment | |
| CN110688905B (en) | Three-dimensional object detection and tracking method based on key frame | |
| CN111881773A (en) | Event camera human body posture estimation method and system based on position deviation | |
| CN114862932B (en) | BIM global positioning-based pose correction method and motion distortion correction method | |
| CN106778739B (en) | A kind of curving transmogrified text page-images antidote | |
| CN110110694B (en) | Visual SLAM closed-loop detection method based on target detection | |
| CN107862735B (en) | RGBD three-dimensional scene reconstruction method based on structural information | |
| CN111144213B (en) | Object detection method and related equipment | |
| CN109115184B (en) | Collaborative measurement method and system based on non-cooperative target | |
| CN110097584A (en) | The method for registering images of combining target detection and semantic segmentation | |
| CN105335977B (en) | The localization method of camera system and target object | |
| CN108961385B (en) | SLAM composition method and device | |
| CN107909643B (en) | Mixed scene reconstruction method and device based on model segmentation | |
| CN113361499B (en) | Local object extraction method and device based on two-dimensional texture and three-dimensional attitude fusion | |
| CN107025449A (en) | A kind of inclination image linear feature matching process of unchanged view angle regional area constraint | |
| CN112614169A (en) | 2D/3D spine CT (computed tomography) level registration method based on deep learning network | |
| CN113298947A (en) | Multi-source data fusion-based three-dimensional modeling method medium and system for transformer substation | |
| CN105184789A (en) | Camera positioning system and method based on point cloud reduction | |
| CN114140527A (en) | Dynamic environment binocular vision SLAM method based on semantic segmentation |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190517 |