CN104751185A - SAR image change detection method based on mean shift genetic clustering - Google Patents

Abstract

The invention provides an SAR image change detection method based on mean shift genetic clustering. The method includes 1, image inputting; 2, differential image construction; 3, mean shift filtering; 4, genetic fuzzy clustering; 5, differential image segmentation; 6, result outputting. The method has the advantages that the influence on synthetic aperture radar (SAR) image change detection results by the region between the change class and non-change class can be reduced effectively, the inherent noise of synthetic aperture radar (SAR) images can be suppressed, the mean shift filtering, fuzzy clustering local optimality and genetic algorithm global optimizing capability are combined, the convergence speed of the algorithm is increased, the detection missing information of detection results is reduced, and the high change detection accuracy is provided.

Description

Based on the SAR image change detection of average drifting genetic cluster

Technical field

The invention belongs to technical field of image processing, further relate to a kind of synthetic-aperture radar based on average drifting genetic cluster (Synthetic Aperture Radar, the SAR) image change detection method in Image Change Detection technical field.The present invention can be applicable to the detection of dynamic of lake level, the field such as detection of dynamic, city planning, military surveillance of crop growth state, detects the change that atural object occurs in time.

Background technology

It is that image is divided into region of variation and invariant region by the gray-scale value of utilization variance image, detects the information that this area's atural object changes in time by analyzing areal not several SAR image in the same time that SAR image change detects.Synthetic-aperture radar SAR has feature that is round-the-clock, round-the-clock, and not by weather effect, and having certain penetration capacity, is good change-detection images source, and research SAR image change detection techniques has boundless application prospect.

Fuzzy C-means clustering is a kind of change detecting method based on cluster be most widely used, and in recent years, a series of method improved based on fuzzy C-means clustering is suggested.Maoguo Gong, Zhiqiang Zhou, and JingjingMa is at paper " Change Detection in Synthetic Aperture Radar Images based on Image Fusionand Fuzzy Clustering " (IEEE Transactions on Image Processing, 2012, 21 (4): 2141-2151) a kind of FLICM (the reformulated FLICM based on improving is proposed in, RFLICM) SAR image change detection, compared with the existing change detecting method based on fuzzy C-means clustering, the method solves the change test problems of image more accurately, but, RFLICM still haves much room for improvement in degree of accuracy and arithmetic speed.First, RFLICM random selecting initial cluster center, result in the method very responsive to cluster initial center point, and RFLICM carries out cluster using objective function as reference point, is therefore easily absorbed in local optimum.

A kind of hereditary Kernel fuzzy clustering SAR image change detection is disclosed in patent " a kind of SAR image change detection based on hereditary Kernel fuzzy clustering " (number of patent application: 201410497802.8, publication number: CN104268574A) that Xian Electronics Science and Technology University applies at it.The method asks disparity map and gray matrix to two width SAR image, uses Genetic-fuzzy cluster to obtain population, according to population computed segmentation threshold value, and completes the segmentation to disparity map according to segmentation threshold, obtains changing the result detected.The method combines the ability of searching optimum of genetic algorithm and the local search ability of Kernel fuzzy clustering algorithm, accelerates convergence of algorithm speed speed, effectively reduces the arithmetic speed of algorithm.But the weak point that the method still exists is, is difficult to select suitable kernel function, and well can not removes the intrinsic noise in SAR image, responsive to noise spot, reduce the precision that change detects.

Disclose a kind of based on non-mean filter SAR image change detection in the patent " SAR image change detection based on non-local mean " (number of patent application: 201310529323.5, publication number: CN103927737A) that Wang Haoran applies at it.The two width SAR image that the method comprises same region different time obtains carry out pre-service; After utilizing pre-service, two width SAR image construct ratio difference striographs; The each pixel of traversal ratio difference striograph, calculates the Smoothness Index matrix of each pixel; Non-local mean filtering ratio figure is obtained to doing ratio computing after two width SAR image carry out non-local mean filtering respectively after pre-service; With Smoothness Index as weight, ratio difference striograph and the summation of non-local mean filtering ratio images are obtained final difference image figure; Use fuzzy Local C means Method to split this final difference image figure to obtain changing testing result figure.Although the method effectively inhibits noise, show real change information better, improve change testing result degree of accuracy.But the weak point still existed is, completing Similarity Measure between different pixels point and search can computing time of at substantial, time complexity is high, and can not effectively reduce be in change class and non-changing class between region on change testing result impact, reduce change detect precision.

Summary of the invention

The object of the invention is to the deficiency overcoming above-mentioned prior art, propose a kind of SAR image change detection based on average drifting genetic cluster.Combine average drifting filtering, the local optimum of fuzzy clustering algorithm and the global optimizing ability of genetic algorithm, remove the intrinsic noise in synthetic-aperture radar SAR image well, decrease the undetected information in testing result, there is higher change accuracy of detection, accelerate convergence of algorithm speed.

The thinking that the present invention realizes above-mentioned purpose is: after having constructed differential image, first average drifting filtering is carried out to differential image, obtain differential image after denoising, secondly Fuzzy Genetic Algorithm cluster is carried out to differential image after denoising, obtain the cluster centre of differential image after denoising, then after utilizing denoising, the cluster centre of differential image obtains segmentation threshold, splits differential image after denoising, obtains the change testing result figure of synthetic-aperture radar SAR image.

To achieve these goals, specific implementation step of the present invention is as follows:

Based on a SAR image change detection for average drifting genetic cluster, comprise the steps:

(1) image is imported:

Importing areal, the synthetic-aperture radar SAR image that the two width sizes do not obtained in the same time are identical;

(2) structural differences image:

(2a) the neighborhood error image of two width synthetic-aperture radar SAR image is calculated;

(2b) the neighborhood ratio images of two width synthetic-aperture radar SAR image is calculated;

(2c) the normalization differential image after the neighborhood error image of two width synthetic-aperture radar SAR image and the neighborhood ratio images fusion of two width synthetic-aperture radar SAR image is calculated;

(3) average drifting filtering:

(3a) utilize Density Estimator method, obtain the cuclear density value of each pixel of differential image after the neighborhood error image of two width synthetic-aperture radar SAR image and the neighborhood ratio images fusion of two width synthetic-aperture radar SAR image;

(3b) gray-scale value of each pixel of normalization differential image after fusion and step (3a) are obtained the cuclear density value subtraction of each pixel of the differential image after merging, and subtraction result is taken absolute value;

(3c) judge whether absolute value is less than the threshold value of setting, if so, then perform step (3d); Otherwise, perform step (3a);

(3d) using the pixel gray-scale value of absolute value as differential image, differential image after denoising is obtained;

(4) Genetic-fuzzy cluster:

(4a) first generation population of differential image after initialization denoising;

(4b) first generation population overall situation Classification Index of differential image after denoising according to the following formula, is calculated:

$J\left(t\right)=\underset{i=1}{\overset{c}{\mathrm{\Σ}}}\underset{j=0}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\μ}}_{\mathrm{ij}}^m\left(t\right){d_{\mathrm{ij}}}^{2}H\left(j\right);$

Wherein, J (t) represent differential image after denoising first generation population overall situation Classification Index, t represents the evolution number of times of the first generation population of differential image after denoising, Σ represents sum operation, i represents the sequence number of i-th class in the cluster centre of differential image after denoising, and c represents the cluster number of differential image after denoising, and j represents the sequence number of a jth pixel of differential image after denoising, n represents the total pixel number of differential image after denoising, μ _ijrepresent that the jth pixel of differential image after denoising is under the jurisdiction of the degree of membership of the i-th class in the cluster centre of differential image after denoising, μ _ijspan is [0,1], and must meet constraint condition, m represents the Fuzzy Exponential factor, and m is the positive number that value is greater than 1, d _ij ²represent the distance of the i-th class in the cluster centre of differential image after a jth pixel to denoising of differential image after denoising, H (j) represents the gray-scale value of a jth pixel of the normalization differential image after fusion;

(4c) the ideal adaptation degree of the first generation population of differential image after denoising according to the following formula, is calculated:

$f\left(t\right)=\frac{1}{1+J\left(t\right)};$

Wherein, f (t) represents the ideal adaptation degree of the first generation population of differential image after denoising, and t represents the evolution number of times of the first generation population of differential image after denoising, and J (t) represents the overall Classification Index of differential image after denoising;

(4d) genetic manipulation is adopted to obtain the population of future generation of differential image after denoising;

(4e) according to the fitness of each individuality in the population of future generation of differential image after the method calculating denoising in step (4c);

(4f) after judging denoising, whether the population of future generation of differential image stablizes, and if so, then performs step (4g); Otherwise, perform step (4d);

(4g) the fitness maximal value of the population of future generation of differential image after denoising is calculated;

(4h) using the cluster centre of the individuality of the fitness maximal value of the population of future generation of differential image after denoising as differential image after denoising;

(5) differential image is split:

(5a) element of the subordinated-degree matrix of differential image after denoising according to the following formula, is calculated:

${\mathrm{\μ}}_{\mathrm{ij}}={\left(\underset{k=1}{\overset{c}{\mathrm{\Σ}}}{\left(\frac{{\left|\right|Y_j-V_i\left|\right|}^2}{{d_{\mathrm{kj}}}^2}\right)}^{\frac{1}{m-1}}\right)}^{-1};$

Wherein, μ _ijthe element of the subordinated-degree matrix of differential image after expression denoising, i represents the sequence number of i-th class in the cluster centre of differential image after denoising, j represents the sequence number of a jth pixel of differential image after denoising, Σ represents sum operation, k represents the sequence number of a kth class in differential image cluster centre after denoising, c represents the cluster number of differential image after denoising, || || represent and ask Euclidean distance to operate, Y _ja jth pixel gray-scale value of differential image after expression denoising, V _ii-th cluster centre of differential image after expression denoising, d _kj ²represent the cluster centre distance of differential image kth class after a jth pixel to denoising of differential image after denoising, m represents the Fuzzy Exponential factor, and m is the positive number that value is greater than 1;

(5b) line number at the subordinated-degree matrix all elements minimum value place of differential image after denoising is asked;

(5c) line number at minimum value place step (5b) obtained is as the segmentation threshold of differential image after denoising;

(5d) after judging denoising, whether the gray-scale value of each pixel of differential image is less than the segmentation threshold of differential image after denoising, if so, performs step (5e); Otherwise, perform step (5f);

(5e) will the segmentation threshold pixel of differential image after denoising be less than, be classified as the non-changing class of differential image after denoising;

(5f) will the segmentation threshold pixel of differential image after denoising be more than or equal to, be classified as the change class of differential image after denoising;

(6) Output rusults:

To the change class of differential image after the non-changing class of differential image after the denoising obtained and denoising, the result figure that the change obtaining two width synthetic-aperture radar SAR image detects.

The present invention compared with prior art has the following advantages:

First, because the present invention is in the process of structural differences image, have employed the method that the neighborhood error image of two width synthetic-aperture radar SAR image and the neighborhood ratio images of two width synthetic-aperture radar SAR image are merged, inhibit background information largely, and inhibit speckle noise, effectively reduce the impact of the region Technologies Against Synthetic Aperture Radar SAR image change testing result be between change class and non-changing class, overcome prior art exists more undetected information shortcoming to the region be between change class and non-changing class, make the precision that invention increases synthetic aperture SAR image change detection.

Second, because the present invention is in the process of average drifting filtering, employing Density Estimator method calculates the cuclear density value of each pixel of differential image after the neighborhood error image of two width synthetic-aperture radar SAR image and the neighborhood ratio images fusion of two width synthetic-aperture radar SAR image, inhibit the intrinsic noise in synthetic aperture SAR image, overcome the shortcoming of prior art to noise spot sensitivity, robustness of the present invention is improved.

3rd, because the present invention is in the process of Genetic-fuzzy cluster, have employed genetic manipulation, accelerate convergence of algorithm speed, overcome the computing time of prior art at substantial and there is the shortcoming of more undetected information, making to invention increases the precision that the change of synthetic-aperture radar SAR image detects.

Accompanying drawing explanation

Fig. 1 is process flow diagram of the present invention;

Fig. 2 is analogous diagram of the present invention.

Embodiment

Below in conjunction with accompanying drawing, step of the present invention is described in further detail.

With reference to accompanying drawing 1, concrete steps of the present invention are as follows.

Step 1, imports image.

Importing areal, the synthetic-aperture radar SAR image that the two width sizes do not obtained in the same time are identical.

Step 2, structural differences image.

According to the following formula, the neighborhood error image of two width synthetic-aperture radar SAR image is calculated:

$S=255-\frac{|\mathrm{\Σ}{X}_1(i,j)-\mathrm{\Σ}{X}_2(i,j)|}{H\×H};$

Wherein, S represents the neighborhood error image of two width synthetic-aperture radar SAR image, || represent and ask absolute value operation, Σ represents sum operation, X ₁(i, j) represents the neighborhood of pixel points set that the first width synthetic-aperture radar SAR image is corresponding on i, j position, X ₂(i, j) represents the neighborhood of pixel points set that the second width synthetic-aperture radar SAR image is corresponding on i, j position, and H represents the length of side of two width synthetic-aperture radar SAR image neighborhoods, and its value is 3.

According to the following formula, the neighborhood ratio images of two width synthetic-aperture radar SAR image is calculated:

$R=255\×\frac{\underset{i=1}{\overset{L\×L}{\mathrm{\Σ}}}\min \{N_1\left(i\right),N_2\left(i\right)\}}{\underset{i=1}{\overset{L\×L}{\mathrm{\Σ}}}\max \{N_1\left(i\right),N_2\left(i\right)\}};$

Wherein, R represents the neighborhood ratio images of two width synthetic-aperture radar SAR image, i represents that two width synthetic-aperture radar SAR image are in the sequence number of i-th pixel together, L represents the length of side of two width synthetic-aperture radar SAR image neighborhoods, L value is 3, Σ represents sum operation, and min represents operation of minimizing, N ₁i () represents the Neighbourhood set of the first width synthetic-aperture radar SAR image i-th pixel, N ₂i () represents the Neighbourhood set of the second width synthetic-aperture radar SAR image i-th pixel, max represents that maximizing operates.

According to the following formula, the differential image after the neighborhood error image of two width synthetic-aperture radar SAR image and the neighborhood ratio images fusion of two width synthetic-aperture radar SAR image is calculated:

$X(x,y)=\frac{\underset{(i,j)\∈M_{x,y}}{\mathrm{\Σ}}\exp (\frac{{|h(i,j)-h(x,y)|}^2}{2{\mathrm{\δ}}^2}+\frac{d^2}{2{\mathrm{\μ}}^2})R(i,j)}{\exp (\frac{{|h(i,j)-h(x,y)|}^2}{2{\mathrm{\δ}}^2}+\frac{d^2}{2{\mathrm{\μ}}^2})};$

Wherein, X (x, y) represents, x, y represent the position of the neighborhood error image cluster centre of two width synthetic-aperture radar SAR image, i, j represent the central pixel point position of the Neighbourhood set of the neighborhood error image of two width synthetic-aperture radar SAR image, and ∈ represents that getting set element operates, M _x,yrepresent the Neighbourhood set of the neighborhood error image of two width synthetic-aperture radar SAR image, the neighborhood length of side size of the neighborhood error image of two width synthetic-aperture radar SAR image is L, L value is 7, Σ represents sum operation, exp represents index operation, || represent and ask absolute value operation, h (i, j) represent that the neighborhood error image of two width synthetic-aperture radar SAR image is at i, pixel gray-scale value corresponding on j position, h (x, y) represent that the neighborhood error image of two width synthetic-aperture radar SAR image is at x, pixel gray-scale value corresponding on y position, δ represents adjustment parameter, its value is 1, d represents i in the neighborhood error image of two width synthetic-aperture radar SAR image, j position is to the neighborhood error image cluster centre position x of two width synthetic-aperture radar SAR image, the Euclidean distance of y, μ represents adjustment parameter, its value is 1, R (i, j) represent that the neighborhood ratio images of two width synthetic-aperture radar SAR image is at i, pixel gray-scale value corresponding on j position.

Utilize linear function method, differential image after the neighborhood error image of two width synthetic-aperture radar SAR image and the neighborhood ratio images of two width synthetic-aperture radar SAR image merge is normalized, obtains the gray-scale value of each pixel of normalization differential image after merging.

Step 3, average drifting filtering.

1st step, utilizes Density Estimator method, obtains the cuclear density value of each pixel of differential image after the neighborhood error image of two width synthetic-aperture radar SAR image and the neighborhood ratio images fusion of two width synthetic-aperture radar SAR image.

2nd step, by the cuclear density value subtraction of each pixel of the differential image after the gray-scale value of each pixel of normalization differential image after fusion and fusion, and takes absolute value to subtraction result.

3rd step, judges whether absolute value is less than the threshold value of setting, if so, then performs the 4th step of this step; Otherwise, perform the 1st step of this step.

4th step, using the pixel gray-scale value of absolute value as differential image, obtains differential image after denoising.

Step 4, Genetic-fuzzy cluster.

1st step, the first generation population of differential image after initialization denoising, is set as 2, population at individual number is set as 30, maximum evolution number of times is set as 100, end condition threshold range is set as 10 by the cluster number of differential image after the cluster centre number of the first generation population of differential image after denoising and denoising ^-8, by the pixel gray-scale value of differential image after Stochastic choice denoising, as the initial cluster center value of differential image after denoising.

2nd step, according to the following formula, the first generation population overall situation Classification Index of differential image after calculating denoising:

$J\left(t\right)=\underset{i=1}{\overset{c}{\mathrm{\Σ}}}\underset{j=0}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\μ}}_{\mathrm{ij}}^m\left(t\right){d_{\mathrm{ij}}}^{2}H\left(j\right);$

Wherein, J (t) represent differential image after denoising first generation population overall situation Classification Index, t represents the evolution number of times of the first generation population of differential image after denoising, Σ represents sum operation, i represents the sequence number of i-th class in the cluster centre of differential image after denoising, and c represents the cluster number of differential image after denoising, and j represents the sequence number of a jth pixel of differential image after denoising, n represents the total pixel number of differential image after denoising, μ _ijrepresent that the jth pixel of differential image after denoising is under the jurisdiction of the degree of membership of the i-th class in the cluster centre of differential image after denoising, μ _ijspan is [0,1], and must meet constraint condition, m represents the Fuzzy Exponential factor, and m value is 2, d _ij ²represent the distance of the i-th class in the cluster centre of differential image after a jth pixel to denoising of differential image after denoising, H (j) represents the gray-scale value of a jth pixel of the normalization differential image after fusion.

3rd step, according to the following formula, the ideal adaptation degree of the first generation population of differential image after calculating denoising:

$f\left(t\right)=\frac{1}{1+J\left(t\right)};$

Wherein, f (t) represents the ideal adaptation degree of the first generation population of differential image after denoising, and t represents the evolution number of times of the first generation population of differential image after denoising, and J (t) represents the overall Classification Index of differential image after denoising.

4th step, adopts the selection of row genetic algorithm, crossover and mutation operation, the population of future generation of differential image after acquisition denoising.

5th step, according to the fitness of each individuality in the population of future generation of differential image after the method calculating denoising of the 3rd step.

6th step, after judging denoising, whether the population of future generation of differential image stablizes, and if so, then performs the 7th step of this step; Otherwise, perform the 4th step of this step.

7th step, the fitness maximal value of the population of future generation of differential image after calculating denoising.

8th step, using the cluster centre of the individuality of the fitness maximal value of the population of future generation of differential image after denoising as differential image after denoising.

Step 5, segmentation differential image.

1st step, according to the following formula, the element of the subordinated-degree matrix of differential image after calculating denoising:

${\mathrm{\μ}}_{\mathrm{ij}}={\left(\underset{k=1}{\overset{c}{\mathrm{\Σ}}}{\left(\frac{{\left|\right|Y_j-V_i\left|\right|}^2}{{d_{\mathrm{kj}}}^2}\right)}^{\frac{1}{m-1}}\right)}^{-1};$

Wherein, μ _ijrepresent the element of the subordinated-degree matrix of differential image after denoising, i represents the sequence number of i-th class in the cluster centre of differential image after denoising, and j represents the sequence number of a jth pixel of differential image after denoising _,Σ represents sum operation, and k represents the sequence number of a kth class in differential image cluster centre after denoising, and c represents the cluster number of differential image after denoising, || || represent and ask Euclidean distance to operate, Y _ja jth pixel gray-scale value of differential image after expression denoising, V _ii-th cluster centre of differential image after expression denoising, d _kj ²represent the cluster centre distance of differential image kth class after a jth pixel to denoising of differential image after denoising, m represents the Fuzzy Exponential factor, and m value is 2.

2nd step, asks the line number at the subordinated-degree matrix all elements minimum value place of differential image after denoising.

3rd step, using the segmentation threshold of the line number at the subordinated-degree matrix all elements minimum value place of differential image after making an uproar as differential image after denoising.

4th step, after judging denoising, whether the gray-scale value of each pixel of differential image is less than the segmentation threshold of differential image after denoising, if so, then performs the 5th step of this step; Otherwise, perform the 6th step of this step.

5th step, will be less than the segmentation threshold pixel of differential image after denoising, is classified as the non-changing class of differential image after denoising.

6th step, will be more than or equal to the segmentation threshold pixel of differential image after denoising, is classified as the change class of differential image after denoising.

Step 6, Output rusults.

To the change class of differential image after the non-changing class of differential image after the denoising obtained and denoising, then binaryzation is carried out to the result that the pixel basis of differential image after denoising is classified, the result figure that the change obtaining two width synthetic-aperture radar SAR image detects.

Below in conjunction with accompanying drawing 2, simulation result of the present invention is further described.

1. simulated environment:

Emulation of the present invention is core 22.26GHZ at allocation of computer, and internal memory 1G, WINDOWS XP system and computer software carry out under being configured to MATLAB 2010 environment.

2. emulate content:

It is two be combined into aperture radar SAR image data set that the present invention emulates data used.First is combined into the synthetic-aperture radar SAR image that aperture SAR image data set is Feltwell village and farmland district of Britain, two width figure sizes are 470 × 335 pixels, and the change occurred between two width images is by be affected by factors such as the simulation Changes in weather of the earth and electromagnetic radiation characteristics and caused by artificial some change informations of embedding.Second group data set is Switzerland Bern area synthetic-aperture radar SAR image, and the size of two width images is 301 × 301 pixels, and the change occurred between two width images causes due to floods near Bern suburb.

Accompanying drawing 2 (a) is the first result figure being combined into that aperture radar SAR image data set adopts fuzzy C-mean algorithm change detecting method.Accompanying drawing 2 (b) is the first result figure being combined into that aperture radar SAR image data set adopts FLICM change detecting method.Accompanying drawing 2 (c) is the first result figure being combined into that aperture radar SAR image data set adopts RFLICM change detecting method.Accompanying drawing 2 (d) is first be combined into aperture radar SAR image data set and adopt result figure of the present invention.Accompanying drawing 2 (e) is the second result figure being combined into that aperture radar SAR image data set adopts fuzzy C-mean algorithm change detecting method.Accompanying drawing 2 (f) is the second result figure being combined into that aperture radar SAR image data set adopts FLICM change detecting method.Accompanying drawing 2 (g) is the second result figure being combined into that aperture radar SAR image data set adopts RFLICM change detecting method.Accompanying drawing 2 (h) is second be combined into aperture radar SAR image data set and adopt result figure of the present invention.White portion in eight width result figure is region of variation, and black region is non-changing region.

3. analysis of simulation result:

Observe the white portion of accompanying drawing 2 (a), the fuzzy C-mean algorithm synthetic-aperture radar SAR image change detection adopting prior art can be found out, more noise is existed to the first change testing result being combined into aperture radar SAR image.

Observe the white portion of accompanying drawing 2 (b), the fuzzy local message C average FLICM synthetic-aperture radar SAR image change detection adopting prior art can be found out, the change testing result of aperture radar SAR image is combined into first, comparing accompanying drawing 2 (a) adopts the fuzzy C-mean algorithm synthetic-aperture radar SAR image change detection of prior art to have improvement, but also there is a lot of flase drop information.

Observe the white portion of accompanying drawing 2 (c), the fuzzy local message C average FLICM synthetic-aperture radar SAR image change detection of the improvement adopting prior art can be found out, the change testing result of aperture radar SAR image is combined into first, the synthetic-aperture radar SAR image change testing result comparing accompanying drawing 2 (a) and accompanying drawing (2b) all has improvement, but, also there are some flase drop information in the level and smooth marginal information of region of variation.

Contrast accompanying drawing 2 (d) and the white portion in accompanying drawing 2 (a), accompanying drawing 2 (b), accompanying drawing 2 (c), can find out owing to present invention employs average drifting filtering, effectively inhibit the intrinsic noise of synthetic-aperture radar SAR image, combine again the local optimum of fuzzy clustering and the global optimizing ability of genetic algorithm, accelerate convergence of algorithm speed, reduce the undetected information in testing result, there is the highest change accuracy of detection.

Observe the white portion of accompanying drawing 2 (e), the fuzzy C-mean algorithm synthetic-aperture radar SAR image change detection adopting prior art can be found out, more flase drop information is existed to the second change testing result being combined into aperture radar SAR image.

Observe the white portion of accompanying drawing 2 (f), the fuzzy local message C average FLICM synthetic-aperture radar SAR image change detection adopting prior art can be found out, the change testing result of aperture radar SAR image is combined into second, compare and adopt the fuzzy C-mean algorithm synthetic-aperture radar SAR image change detection of the accompanying drawing 2 (e) of prior art to have improvement, but also there is more flase drop information.

Observe the white portion of accompanying drawing 2 (g), the fuzzy local message C average FLICM synthetic-aperture radar SAR image change detection of the improvement adopting prior art can be found out, the synthetic-aperture radar SAR image change testing result second change testing result being combined into aperture radar SAR image being compared to accompanying drawing 2 (e) and accompanying drawing (2f) all has improvement, but, also there are some undetected information in the level and smooth marginal information of region of variation.

Contrast accompanying drawing 2 (h) and the white portion in accompanying drawing 2 (e), accompanying drawing 2 (f), accompanying drawing 2 (g), can find out owing to present invention employs average drifting filtering, effectively inhibit the intrinsic noise of synthetic-aperture radar SAR image, combine again the local optimum of fuzzy clustering and the global optimizing ability of genetic algorithm, accelerate convergence of algorithm speed, reduce the undetected information in testing result, there is the highest change accuracy of detection.

Be combined on aperture radar SAR image data set two, use the fuzzy local message C average RFLICM synthetic-aperture radar SAR image change detection of the fuzzy C-mean algorithm of the SAR image change detection based on average drifting genetic cluster of the present invention and prior art, fuzzy local message C average FLICM, improvement, the result of aperture radar SAR image change detection is combined into two, calculate the undetected number that two are combined into aperture radar SAR image change detection, flase drop number, total error number and computing time.Wherein, undetected number is that actual there occurs changes but the pixel do not detected, and flase drop number is that reality does not change but is detected as the pixel of conversion, total error number=undetected number+flase drop number.Computing time is the time of using above-mentioned four kinds of methods to obtain synthetic-aperture radar SAR image change testing result figure.

Undetected number in the table 1 that effect of the present invention can be obtained by emulation experiment, flase drop number, total error number and computing time four metrics evaluation change detecting methods quality.

By table 1, contrast prior art fuzzy C-mean algorithm, fuzzy local message C average FLICM, improvement fuzzy local message C average RFLICM synthetic-aperture radar SAR image change detection and the present invention is based on the SAR image change detection of average drifting genetic cluster, can find out, to the two change testing results being combined into aperture radar SAR image data set, all there is minimum flase drop number and total error number, there is less undetected number and less computing time, namely there is the highest change accuracy of detection.

Table 1 liang is combined into aperture radar SAR image change evaluation index

Claims (5)

1., based on a SAR image change detection for average drifting genetic cluster, comprise the steps:

(1) image is imported:

Importing areal, the synthetic-aperture radar SAR image that the two width sizes do not obtained in the same time are identical;

(2) structural differences image:

(2a) the neighborhood error image of two width synthetic-aperture radar SAR image is calculated;

(2b) the neighborhood ratio images of two width synthetic-aperture radar SAR image is calculated;

(2c) the normalization differential image after the neighborhood error image of two width synthetic-aperture radar SAR image and the neighborhood ratio images fusion of two width synthetic-aperture radar SAR image is calculated;

(3) average drifting filtering:

(3a) utilize Density Estimator method, obtain the cuclear density value of each pixel of differential image after the neighborhood error image of two width synthetic-aperture radar SAR image and the neighborhood ratio images fusion of two width synthetic-aperture radar SAR image;

(3b) gray-scale value of each pixel of normalization differential image after fusion and step (3a) are obtained the cuclear density value subtraction of each pixel of the differential image after merging, and subtraction result is taken absolute value;

(3c) judge whether absolute value is less than the threshold value of setting, if so, then perform step (3d); Otherwise, perform step (3a);

(3d) using the pixel gray-scale value of absolute value as differential image, differential image after denoising is obtained;

(4) Genetic-fuzzy cluster:

(4a) first generation population of differential image after initialization denoising;

(4b) first generation population overall situation Classification Index of differential image after denoising according to the following formula, is calculated:

$J\left(t\right)=\underset{i=1}{\overset{c}{\mathrm{\Σ}}}\underset{j=0}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\μ}}_{\mathrm{ij}}^m\left(t\right){d_{\mathrm{ij}}}^{2}H\left(j\right);$

Wherein, J (t) represent differential image after denoising first generation population overall situation Classification Index, t represents the evolution number of times of the first generation population of differential image after denoising, Σ represents sum operation, i represents the sequence number of i-th class in the cluster centre of differential image after denoising, and c represents the cluster number of differential image after denoising, and j represents the sequence number of a jth pixel of differential image after denoising, n represents the total pixel number of differential image after denoising, μ _ijrepresent that the jth pixel of differential image after denoising is under the jurisdiction of the degree of membership of the i-th class in the cluster centre of differential image after denoising, μ _ijspan is [0,1], and must meet constraint condition, m represents the Fuzzy Exponential factor, and m is the positive number that value is greater than 1, d _ij ²represent the distance of the i-th class in the cluster centre of differential image after a jth pixel to denoising of differential image after denoising, H (j) represents the gray-scale value of a jth pixel of the normalization differential image after fusion;

(4c) the ideal adaptation degree of the first generation population of differential image after denoising according to the following formula, is calculated:

$f\left(t\right)=\frac{1}{1+J\left(t\right)};$

Wherein, f (t) represents the ideal adaptation degree of the first generation population of differential image after denoising, and t represents the evolution number of times of the first generation population of differential image after denoising, and J (t) represents the overall Classification Index of differential image after denoising;

(4d) genetic manipulation is adopted to obtain the population of future generation of differential image after denoising;

(4e) according to the fitness of each individuality in the population of future generation of differential image after the method calculating denoising in step (4c);

(4f) after judging denoising, whether the population of future generation of differential image stablizes, and if so, then performs step (4g); Otherwise, perform step (4d);

(4g) the fitness maximal value of the population of future generation of differential image after denoising is calculated;

(4h) using the cluster centre of the individuality of the fitness maximal value of the population of future generation of differential image after denoising as differential image after denoising;

(5) differential image is split:

(5a) element of the subordinated-degree matrix of differential image after denoising according to the following formula, is calculated:

${\mathrm{\μ}}_{\mathrm{ij}}={\left(\underset{k=1}{\overset{c}{\mathrm{\Σ}}}{\left(\frac{{\left|\right|Y_j-V_i\left|\right|}^2}{{d_{\mathrm{kj}}}^2}\right)}^{\frac{1}{m-1}}\right)}^{-1};$

Wherein, μ _ijthe element of the subordinated-degree matrix of differential image after expression denoising, i represents the sequence number of i-th class in the cluster centre of differential image after denoising, j represents the sequence number of a jth pixel of differential image after denoising, Σ represents sum operation, k represents the sequence number of a kth class in differential image cluster centre after denoising, c represents the cluster number of differential image after denoising, || || represent and ask Euclidean distance to operate, Y _ja jth pixel gray-scale value of differential image after expression denoising, V _ii-th cluster centre of differential image after expression denoising, d _kj ²represent the cluster centre distance of differential image kth class after a jth pixel to denoising of differential image after denoising, m represents the Fuzzy Exponential factor, and m is the positive number that value is greater than 1;

(5b) line number at the subordinated-degree matrix all elements minimum value place of differential image after denoising is asked;

(5c) line number at minimum value place step (5b) obtained is as the segmentation threshold of differential image after denoising;

(5d) after judging denoising, whether the gray-scale value of each pixel of differential image is less than the segmentation threshold of differential image after denoising, if so, performs step (5e); Otherwise, perform step (5f);

(5e) will the segmentation threshold pixel of differential image after denoising be less than, be classified as the non-changing class of differential image after denoising;

(5f) will the segmentation threshold pixel of differential image after denoising be more than or equal to, be classified as the change class of differential image after denoising;

(6) Output rusults:

To the change class of differential image after the non-changing class of differential image after the denoising obtained and denoising, the result figure that the change obtaining two width synthetic-aperture radar SAR image detects.

2. the SAR image change detection based on average drifting genetic cluster according to claim 1, is characterized in that: the computing formula of the neighborhood error image of two width synthetic-aperture radar SAR image described in step (2a) is as follows:

$S=255-\frac{|\mathrm{\Σ}{X}_1(i,j)-\mathrm{\Σ}{X}_2(i,j)|}{H\×H};$

Wherein, S represents the neighborhood error image of two width synthetic-aperture radar SAR image, || represent and ask absolute value operation, Σ represents sum operation, X ₁(i, j) represents the neighborhood of pixel points set that the first width synthetic-aperture radar SAR image is corresponding on i, j position, X ₂(i, j) represents the neighborhood of pixel points set that the second width synthetic-aperture radar SAR image is corresponding on i, j position, and H represents the length of side of two width synthetic-aperture radar SAR image neighborhoods, and its span is H ∈ { 3,5}.

3. the SAR image change detection based on average drifting genetic cluster according to claim 1, is characterized in that: the computing formula of the neighborhood ratio images of two width synthetic-aperture radar SAR image described in step (2b) is as follows:

$R=255\×\frac{\underset{i=1}{\overset{L\×L}{\mathrm{\Σ}}}\min \{N_i\left(i\right),N_2\left(i\right)\}}{\underset{i=1}{\overset{L\×L}{\mathrm{\Σ}}}\max \{N_1\left(i\right),N_2\left(i\right)\}};$

Wherein, R represents the neighborhood ratio images of two width synthetic-aperture radar SAR image, i represents that two width synthetic-aperture radar SAR image are in the sequence number of i-th pixel together, L represents the length of side of two width synthetic-aperture radar SAR image neighborhoods, { 3,5}, Σ represent sum operation to L ∈, min represents operation of minimizing, N ₁i () represents the Neighbourhood set of the first width synthetic-aperture radar SAR image i-th pixel, N ₂i () represents the Neighbourhood set of the second width synthetic-aperture radar SAR image i-th pixel, max represents that maximizing operates.

4. the SAR image change detection based on average drifting genetic cluster according to claim 1, is characterized in that: the concrete steps of the normalization differential image after the neighborhood error image of the calculating two width synthetic-aperture radar SAR image described in step (2c) and the neighborhood ratio images of two width synthetic-aperture radar SAR image merge are as follows:

The first step, according to the following formula, calculates the differential image after the neighborhood error image of two width synthetic-aperture radar SAR image and the neighborhood ratio images fusion of two width synthetic-aperture radar SAR image:

$X(x,y)=\frac{\underset{(i,j)\∈M_{x,y}}{\mathrm{\Σ}}\exp (\frac{{|h(i,j)-h(x,y)|}^2}{2{\mathrm{\δ}}^2}+\frac{d^2}{2{\mathrm{\μ}}^2})R(i,j)}{\exp (\frac{{|h(i,j)-h(x,y)|}^2}{2{\mathrm{\δ}}^2}+\frac{d^2}{2{\mathrm{\μ}}^2})};$

Wherein, X (x, y) represents, x, y represent the position of the neighborhood error image cluster centre of two width synthetic-aperture radar SAR image, i, j represent the central pixel point position of the Neighbourhood set of the neighborhood error image of two width synthetic-aperture radar SAR image, and ∈ represents that getting set element operates, M _x,yrepresent the Neighbourhood set of the neighborhood error image of two width synthetic-aperture radar SAR image, the neighborhood length of side size of the neighborhood error image of two width synthetic-aperture radar SAR image is L, L ∈ { 7, 11}, Σ represents sum operation, exp represents index operation, || represent and ask absolute value operation, h (i, j) represent that the neighborhood error image of two width synthetic-aperture radar SAR image is at i, pixel gray-scale value corresponding on j position, h (x, y) represent that the neighborhood error image of two width synthetic-aperture radar SAR image is at x, pixel gray-scale value corresponding on y position, δ represents adjustment parameter, its span is (0, 100], d represents i in the neighborhood error image of two width synthetic-aperture radar SAR image, j position is to the neighborhood error image cluster centre position x of two width synthetic-aperture radar SAR image, the Euclidean distance of y, μ represents adjustment parameter, its span is (0, 100], R (i, j) represent that the neighborhood ratio images of two width synthetic-aperture radar SAR image is at i, pixel gray-scale value corresponding on j position,

Second step, utilize linear function method, differential image after the neighborhood error image of two width synthetic-aperture radar SAR image and the neighborhood ratio images of two width synthetic-aperture radar SAR image merge is normalized, obtains the gray-scale value of each pixel of normalization differential image after merging.

5. the SAR image change detection based on average drifting genetic cluster according to claim 1, it is characterized in that: the initialization described in step (4a) refers to, the cluster centre number of the first generation population of differential image after denoising is set as 2, population at individual number is set as 30, maximum evolution number of times is set as 100, end condition threshold range is set as 10 ^-8, by the pixel gray-scale value of differential image after Stochastic choice denoising, as the initial cluster center value of differential image after denoising.