CN104867150A - Wave band correction change detection method of remote sensing image fuzzy clustering and system thereof - Google Patents

Abstract

The invention provides a wave band correction change detection method of remote sensing image fuzzy clustering and a system thereof. The method comprises the following steps: step1, carrying out pretreatment of a multi-band remote sensing image, which means filtering and enhancement processing of the multi-band remote sensing image; step 2, carrying out single band separation on the multi-band remote sensing image after the pretreatment and acquiring a single-band remote sensing image; step 3, constructing a single band difference image of two time-phase single-band remote sensing images; step 4, constructing a multi-band combined image based on two time-phase multi-band remote sensing images after the pretreatment and acquiring a mutual neighborhood information content of each pixel of the multi-band combined image; step 5, correcting the mutual neighborhood information content of the pixels of the multi-band combined image and acquiring a correction image; step 6, using a fuzzy C-means method to carry out change detection on the correction image. By using the method and the system of the invention, detection precision is high, an anti-noise interference ability is good and an automation degree is high.

Description

The wave band correction change detecting method of remote sensing image fuzzy clustering and system

Technical field

The invention belongs to Photogrammetry and Remote Sensing image applied technical field, be specifically related to a kind of wave band correction change detecting method and system of remote sensing image fuzzy clustering.

Background technology

Remote sensing image change detection is the remote sensing images by analyzing the different phase in areal, detects the time dependent information of atural object of this area.Along with the develop rapidly of aeronautical and space technology, remote sensing observations data are with in real time, fast, the feature such as wide coverage, high-spatial and temporal resolution obtains applying more and more widely.How effectively extract the change information in mass data, and the prevention of various disasteies for environment, agricultural, the ecosystem and facing mankind, become the hot issue in current remote sensing application research.In recent years, lot of domestic and foreign scholar has carried out theoretical research and the IT system development effortsIT of many change detections in succession, and obtains a series of achievement.

At present, conventional remote sensing image variation detection method has direct comparison method and classification and predicting method, and whether the former speed simple to operate is fast, but can only change quantitative description target area, is difficult to determine qualitative change.The latter can provide change type information, but accuracy of detection is subject to the error propagation effect of classification separately.There is the phenomenon of mixed pixel in Moderate-High Spatial Resolution Remote Sensing Image, higher precision cannot be obtained by tradition " firmly " classification, and fuzzy C-means clustering (FCM) is a kind of soft clustering algorithm, degree of membership is utilized to make between class and class, do not have obvious boundary, be used for processing mixed pixel effective, but existence too relies on initial cluster center, classification number is difficult to automatically determine, to defects such as isolated point noise data sensitivities.

Following list of references is related in literary composition:

[1]Song C,Woodcock CE,Seto KC,Lenney MP,Macomber SA.Classification andchange detection using Landsat TM data:when and how to correct atmospheric effects[J].Remote sensing of Environment.2001,75(2):230-44.

[2]Mas J-F.Monitoring land-cover changes:a comparison of change detectiontechniques[J].International journal of remote sensing.1999,20(1):139-52.

[3]Jensen JR.Introductory digital image processing:a remote sensing perspective:Prentice-Hall Inc.1996.

[4] Shen Shaohong, relies archosaur, Wan Youchuan. detect [J] based on the high-resolution remote sensing image change of merging. and mapping circular .2009 (3): 16-9.

[5]Jensen JR.Introductory digital image processing:a remote sensing perspective:Prentice-Hall Inc.；1996.

[6]Xie XL,Beni G.A validity measure for fuzzy clustering[J].IEEE Transactions onpattern analysis and machine intelligence.1991,13(8):841-7.

Summary of the invention

For problems of the prior art, the invention provides the wave band correction change detecting method of remote sensing image fuzzy clustering of a kind of high real-time, high measurement accuracy, high automation degree.

The thought of fuzzy clustering is introduced in remote sensing image change detection by the present invention, improves FCM carry out on classification process basis in utilization, adopts a kind of single band to the wave band revised law of specific entropy power correction multiband associating mutual information to improve change accuracy of detection.

The present invention mainly comprised for two megastages, the first stage: single band revises multiband associating mutual information to specific entropy power; Subordinate phase: utilize the fuzzy C-means clustering improved to carry out change to correction image and detect.

First stage, because ratioing technigue conventional in direct comparison method can strengthen change information, Background suppression information, but too exaggerate part change sometimes; Differential technique directly and convenient operation, but complete reaction can not go out the change of atural object emittance.Consider ratioing technigue and differential technique advantage separately and the limitation of existence, the present invention, on the basis that difference and ratio image product of transformation merge, adopts a kind of new composite algorithm structural differences image, provides data source for follow-up change detects.This composite algorithm introduces deviation factors, has corrective action, can strengthen amplitude and the structural information of difference image to grey-scale, is more conducive to change and detects.In addition, due to improving constantly of remote sensing image resolution, there is the sensitivity differences between wave band at the different passages of multiband image in atural object, brings empty inspection, the impact such as undetected to change detection.The present invention considers the contribution degree that each wave band detects change, adopts neighborhood entropy assessment to carry out the process of tune power to single band difference image.Information entropy measures the quantity of information that certain random occurrence comprises, and the neighborhood information amount of this wave band pixel is larger or uncertainty is higher, and the weights that this wave band is corresponding are larger, also larger to the contribution degree of change detection.

On the other hand, in order to fully utilize the abundant spectrum of single band and multiband and spatial structural form, and the geography effectively reduced because different sensors or different phase image exist and scattering difference (grey-scale whole departure) cause classification and change the impact detected, the present invention by the pixel point of the multiband image of two phases to based on, calculate neighborhood mutual information as multiband Combined Treatment result.Mutual information is based on the statistical nature of a pair pixel gray-scale value, do not need the textural characteristics of selection reference point or extraction image, and the linear relationship of the gray-scale value in image need not be considered, therefore in the comparison of two image similarities and the detection of change information, there is larger flexibility ratio and accuracy.MI (mutual information) type conventional at present comprises normalized mutual information amount NMI, image gradient mutual information GMI, region mutual information RMI.Although traditional neighborhood mutual information is based on neighborhood, object finds optimum registration transformation parameter, makes the mutual information between two width images maximum, belong to whole figure amplitude range Image registration category.And change detects and often occurs in some pixel or zonule, the right mutual information size of the degree of change and pixel point is closely related, and is embodied in mutual information larger, and the right correlativity of pixel point is larger, and the intensity of variation of correspondence is less.

Subordinate phase, can regard special assorting process as, is change, non-changing two class by wave band correction image cluster.Fuzzy C-Means Cluster Algorithm (FCM) is a kind of typical " soft cluster ", and algorithm is simple, fast convergence rate, and the introduction of degree of membership makes process mixed pixel effective.But this algorithm comes with some shortcomings: as algorithm performance too rely on initial cluster center, classification number can not be determined automatically, ignore spatial information causes local optimum etc.The present invention improves from two aspects FCM algorithm: one, in conjunction with Da-Jin algorithm single threshold segmentation thought, obtains and meets the maximum optimal threshold t of inter-class variance ^*, and under this Threshold segmentation the gray average of two class pixels as initial cluster center.Its two, position and the attribute information of introducing neighborhood pixel obtain Similarity-Weighted degree of membership, and obtain weighted space function further, make fuzzy C-mean algorithm method have stronger anti-noise ability, effectively avoid local optimum, improve classifying quality.Finally utilize the FCM Algorithms of improvement to carry out change to correction image to detect.

Technical scheme of the present invention is as follows:

One, the wave band correction change detecting method of remote sensing image fuzzy clustering, comprises step:

Step 1, the pre-service of multiband remote sensing image, namely the filtering of multiband remote sensing image processes with strengthening;

Step 2, carries out single band separation to pretreated multiband remote sensing image, obtains single band remote sensing image;

Step 3, constructs the single band difference image of two phase single band remote sensing images;

Step 4, based on pretreated two phase multiband remote sensing Image construction multiband associating images, and obtains the neighborhood mutual information of each pixel of multiband associating image;

Step 5, according to formula revise the neighborhood mutual information of multiband associating image picture element, obtain revising image; Wherein: k represents that single band is numbered, k=1,2 ... N, N are wave band number; M _ijrepresent and revise image (i, j) pixel gray-scale value; w _ij-kfor the neighborhood entropy power of (i, j) pixel in the difference image of expression single band k, h _ij-krepresent the neighborhood information entropy of (i, j) pixel in the difference image of single band k; C _ij-krepresent (i, j) pixel gray scale in the difference image of single band k; RMI _ijrepresent the neighborhood mutual information of multiband associating image (i, j) pixel; Mod _ijrepresent the corresponding N number of single-range entropy mould of (i, j) pixel,

Step 6, for sample, makes classification c=2 to revise pixel gray-scale value in image, to revise the gray average of two class pixels after image Threshold segmentation as initial cluster center; Adjust power summation as center pixel degree of membership using neighborhood pixel degree of membership, the neighborhood similarity weight of each neighborhood pixel centering imago unit utilize fuzzy C-mean algorithm method to carry out change to correction image to detect;

In above-mentioned Threshold segmentation, based on the segmentation of Da-Jin algorithm single threshold, using threshold value maximum for the inter-class variance of the backdrop pels gray scale and goal pels gray scale that make correction image as segmentation threshold;

Above-mentioned in, t, r represent neighborhood pixel and the Position Number of center pixel in neighborhood window respectively; x _rthe gray-scale value of expression center pixel r, x _trepresent the gray-scale value of neighborhood pixel t.

Step 2 is specially:

Adopt weighted sum method by multiband remote sensing image greyscale, when the weights of a certain component of R, G, B single band are 1, when all the other are 0, namely obtain R, G, B single band remote sensing image through weighted sum.

In conjunction with differential technique and ratioing technigue structure single band difference image in step 3, be specially:

According to formula construct the single band difference image of two phase single band remote sensing images, wherein:

C _ij-krepresent (i, j) pixel gray-scale value in the difference image of single band k;

R _ij-krepresent (i, j) pixel gray-scale value in the ratio image of single band k;

D _ij-krepresent (i, j) pixel gray-scale value in the Difference image of single band k;

A be all pixels of ratio image in the gray average and Difference image of kth wave band all pixels in the difference of the gray average of kth wave band;

B is all pixels of ratio image in the gray-scale value root mean square of kth wave band and all pixels of the Difference image difference at the gray-scale value root mean square of kth wave band.

Step 6 comprises sub-step further:

6.1 setting classification C=2, rule of thumb set blur level m and convergency value ε, the initial subordinated-degree matrix of random setting, and subordinated-degree matrix is formed by revising the degree of membership of each pixel to cluster centre in image; Make primary iteration step number l=0;

6.2 based on Da-Jin algorithm single threshold split plot design, using make the inter-class variance of the background of correction image and goal pels gray scale maximum threshold value as segmentation threshold, to revise the gray average of two class pixels after image Threshold segmentation as initial cluster center

6.3 by the current degree of membership of neighborhood pixel t weighted sum as center pixel r degree of membership namely the current degree of membership of neighborhood pixel t obtains according to current subordinated-degree matrix U, its weights $w_{\mathrm{rt}}=\frac{2x_r}{x_r+x_t},$

6.4 adopt formula $\left\{\begin{array}{c}{u}_{\mathrm{rk}}^{(l+1)}=\frac{1}{{\mathrm{\Σ}}_{y=1}^C{\left(\frac{u_{\mathrm{rk}}^{\left(l\right)}\×\left|\right|v_k^{\left(l\right)}-x_r\left|\right|}{u_{\mathrm{ry}}^{\left(l\right)}\×\left|\right|v_y^{\left(l\right)}-x_r\left|\right|}\right)}^{\frac{2}{m-1}})}\text{,\∀y,\∀k}\\ v_k^{(l+1)}=\frac{{\mathrm{\Σ}}_{r=1}^n\left[{\left(u_{\mathrm{rk}}^{\left(l\right)}\right)}^m{x}_r\right]}{{\mathrm{\Σ}}_{r=1}^n{\left(u_{\mathrm{rk}}^{\left(l\right)}\right)}^m}\end{array}\right.$ Revise subordinated-degree matrix and cluster centre, wherein, t, r represent neighborhood pixel and the Position Number of center pixel in neighborhood window respectively, and k, y are cluster centre numbering, and C is classification number; with represent in (l+1) and l iteration respectively, center pixel r is to the degree of membership of kth class cluster centre; represent in the l time iteration, center pixel r is to the degree of membership of y class cluster centre; x _rthe gray-scale value of expression center pixel r; with represent kth class and the y class cluster centre of the l time iteration acquisition, y is traversal in [1, C] scope, can be equal with k; represent the kth class cluster centre that (l+1) secondary iteration obtains; N is pixel quantity in neighborhood window;

6.5 compare subordinated-degree matrix norm || U ^(l+1)-U ^(l)||, U ^(l+1), U ^(l)represent the subordinated-degree matrix of (l+1) secondary iteration, the l time iteration acquisition respectively; If || U ^(l+1)-U ^(l)||≤ε, finishing iteration, carries out cluster according to current cluster centre and subordinated-degree matrix; Otherwise, make l=l+1, perform step 6.4.

Two, the wave band correction change detecting system of remote sensing image fuzzy clustering, comprising:

Pretreatment module, is used for the pre-service of multiband remote sensing image, namely multiband remote sensing image filtering and strengthen process;

Single band separation module, is used for carrying out single band separation to pretreated multiband remote sensing image, obtains single band remote sensing image;

Single band difference image constructing module, is used for the single band difference image of structure two phase single band remote sensing image;

Neighborhood mutual information obtains module, is used for based on pretreated two phase multiband remote sensing Image construction multiband associating images, and obtains the neighborhood mutual information of each pixel of multiband associating image;

Revise image and obtain module, be used for according to formula revise the neighborhood mutual information of multiband associating image picture element, obtain revising image; Wherein: k represents that single band is numbered, k=1,2 ... N, N are wave band number; M _ijrepresent and revise image (i, j) pixel gray-scale value; w _ij-kfor the neighborhood entropy power of (i, j) pixel in the difference image of expression single band k, h _ij-krepresent the neighborhood information entropy of (i, j) pixel in the difference image of single band k; C _ij-krepresent (i, j) pixel gray scale in the difference image of single band k; RMI _ijrepresent the neighborhood mutual information of multiband associating image (i, j) pixel; Mod _ijrepresent the corresponding N number of single-range entropy mould of (i, j) pixel,

Fuzzy C-means clustering module, being used for revising pixel gray-scale value in image is sample, makes classification c=2, to revise the gray average of two class pixels after image Threshold segmentation as initial cluster center; Adjust power summation as center pixel degree of membership using neighborhood pixel degree of membership, the neighborhood similarity weight of each neighborhood pixel centering imago unit utilize fuzzy C-mean algorithm method to carry out change to correction image to detect;

In above-mentioned Threshold segmentation, based on the segmentation of Da-Jin algorithm single threshold, using threshold value maximum for the inter-class variance of the backdrop pels gray scale and goal pels gray scale that make correction image as segmentation threshold;

Above-mentioned in, t, r represent neighborhood pixel and the Position Number of center pixel in neighborhood window respectively; x _rthe gray-scale value of expression center pixel r, x _trepresent the gray-scale value of neighborhood pixel t.

The problems such as traditional remote sensing image changes detection method mostly from the angle of single pixel, single decision-making, the wave band susceptibility that " the different spectrum of jljl " easily appears in complicated to space structure, that details is abundant Moderate-High Spatial Resolution Remote Sensing Image, " foreign matter is with spectrum " causes.The change information how extracting Moderate-High Spatial Resolution Remote Sensing Image quickly and accurately then seems most important.

Compare with traditional remote sensing image variation detection method, tool of the present invention has the following advantages and beneficial effect:

1, accuracy of detection is high, and anti-noise jamming is strong.

(1) in single band contrast difference Image construction link, the present invention adopts difference/ratio composite algorithm, on the basis that difference, ratio product merge, introduces deviation factors, adjusts grey-scale.In addition, consider that atural object detects in the sensitivity differences of different-waveband the impact brought to change, adopt neighborhood entropy assessment, neighborhood mutual information method to weigh process to single band difference image and multiband image to carrying out tune respectively, weights represent the contribution degree of this wave band.Through above-mentioned process, can strengthen the change information of remote sensing image and can react the change of atural object emittance, amplitude and the structural information of compound difference image also strengthen, and are more conducive to change and detect.

(2) adopt the fuzzy C-means clustering improved to carry out change to correction image to detect.Wherein, the fuzzy C-means clustering improved is according to the maximum principle of optimality of inter-class variance, determine initial cluster center, and introducing neighborhood space information obtains Similarity-Weighted degree of membership and weighted space function, fuzzy C-means clustering is made to have stronger anti-noise ability, thus effectively avoid local optimum, improve classifying quality, improve change accuracy of detection.

2, automaticity is high, and overall process detects automatically.

Whole remote sensing imagery change detection process, without the need to manual intervention, can reach automatic detection.

Accompanying drawing explanation

Fig. 1 is the idiographic flow schematic diagram of the inventive method;

Fig. 2 is the fuzzy C-means clustering process flow diagram improved.

Embodiment

Major design two parts content of the present invention, namely (1) single band revises multiband associating mutual information and (2) remote sensing imagery change detection based on fuzzy C-means clustering to specific entropy power.Technical solution of the present invention mainly comprises difference/ratio composite algorithm structure single band contrast difference image, single band revises multiband mutual information, remote sensing imagery change detection based on fuzzy C-means clustering to specific entropy power.The present invention can be alleviated the different spectrum of jljl, foreign matter and compose together the single band sensitivity differences that causes and detect empty inspection, undetected impact to change, isolated noise effectively can be suppressed to disturb, avoid local optimum etc., have preferably ageing and precision.

Further illustrate the present invention below in conjunction with the drawings and specific embodiments, concrete steps are as follows:

Step 1: the pre-service of multiband remote sensing image, namely the filtering of multiband remote sensing image processes with strengthening.

Adopt Gaussian filter and histogram equalization to strengthen method carry out filtering to multiband remote sensing image and strengthen process, to increase image contrast, restraint speckle interference, to improve visual effect.

Step 2: after pre-service, the single band of multiband remote sensing image is separated, and obtains single band remote sensing image.

In order to single band can be adapted to, check processing is changed to the wave band correction that specific entropy power revises multiband associating mutual information, in concrete enforcement, adopt weighted sum method by multiband remote sensing image greyscale, when the weights of a certain component of R, G, B single band are 1, when all the other are 0, namely weighted sum realizes R, G, B single band and is separated, and obtains R, G, B single band remote sensing image.

Step 3: adopt difference/ratio composite algorithm structure single band difference image.

For strengthening change information, the Background suppression information of remote sensing image, simultaneous reactions goes out the change of atural object emittance, adopts difference/ratio composite algorithm to construct R, G, B tri-single-range difference image respectively in this concrete enforcement.

In the conventional configurations method of single band difference image, ratioing technigue can strengthen image change information, Background suppression information, but too exaggerate sometimes part change; Differential technique directly and convenient operation, but can not the change of complete reaction atural object emittance.Consider both advantage and limitation, the present invention, on the basis that difference and ratio image product merge, introduces deviation factors, has corrective action, can strengthen amplitude and the structural information of difference image to grey-scale, is more conducive to change and detects.

Difference/ratio compound ratio juris is as follows:

$C_{\mathrm{ij}-k}=a+b\×\sqrt{D_{\mathrm{ij}-k}*R_{\mathrm{ij}-k}}---\left(1\right)$

$R_{\mathrm{ij}-k}=255\×\frac{\min [x_{\mathrm{ij}-k}\left(t_1\right),x_{\mathrm{ij}-k}\left(t_2\right)]}{\max [x_{\mathrm{ij}-k}\left(t_1\right),x_{\mathrm{ij}-k}\left(t_2\right)]}---\left(2\right)$

D _ij-k＝|x _ij-k(t ₂)-x _ij-k(t ₁)| (3)

$a=|u_{D_{\mathrm{ij}-k}}-u_{R_{\mathrm{ij}-k}}|---\left(4\right)$

$b=\frac{{\mathrm{\σ}}_{D_{\mathrm{ij}-k}}}{{\mathrm{\σ}}_{R_{\mathrm{ij}-k}}}---\left(5\right)$

Formula (1) is difference/ratio composite algorithm principle formula.Formula (2) is ratioing technigue principle formula, and formula (3) is differential technique principle formula, ratioing technigue and differential technique all based on the pixel point of multidate single band remote sensing image to processing.

In formula (1) ~ (5):

R _ij-krepresent (i, j) pixel gray-scale value in the ratio image of single band k, its absolute value is in [0,255] scope;

X _ij-k(t ₁), x _ij-k(t ₂) represent T respectively ₁phase, T ₂under phase single band k remote sensing image in the gray-scale value of (i, j) pixel;

D _ij-krepresent (i, j) pixel gray-scale value in the Difference image of single band k, its value is in [0,255] scope;

C _ij-krepresent (i, j) pixel gray-scale value in the difference image of single band k, by ratio image R _ij-k, Difference image D _ij-kproduct root mean square compound after, and introduce deviation factors a, b adjustment obtain;

A is that the pixel gray average of ratio image and Difference image is poor, represent the gray average of all pixels of ratio image at kth wave band, represent the gray average of all pixels of Difference image at kth wave band;

B is the ratio of the pixel gray scale root mean square of ratio image and Difference image, represent the gray-scale value root mean square of all pixels of ratio image at kth wave band, represent the gray-scale value root mean square of all pixels of Difference image at kth wave band.

Step 4: adopt neighborhood entropy assessment to obtain the neighborhood entropy power of single band difference image.

Consider that the sensitivity differences of atural object at different-waveband is to the impact of remote sensing imagery change detection, the present invention adopts neighborhood entropy assessment to carry out the process of tune power to R, G, B tri-single-range difference image, the corresponding each single-range contribution degree of weights.

Because atural object exists sensitivity differences between wave band at multiband remote sensing image different-waveband, empty inspection, the impact such as undetected can be brought to remote sensing image change detection.The present invention considers the contribution degree that each wave band detects change, adopts neighborhood entropy assessment to carry out the process of tune power to single band difference image.Information entropy measures the quantity of information that certain random occurrence comprises, and the neighborhood information amount of this wave band pixel is larger or uncertainty is higher, and the weights that this wave band is corresponding are larger, also larger to the contribution degree of change detection.

Neighborhood entropy power computing formula is as follows:

$\left\{\begin{array}{c}{w}_{\mathrm{ij}-k}=\frac{H_{\mathrm{ij}-k}}{{\mathrm{\Σ}}_{k=1}^N{H}_{\mathrm{ij}-k}}\\ H_{\mathrm{ij}-k}=-{\mathrm{\Σ}}_{t=1}^c{\mathrm{\Σ}}_{s=1}^c\left(p_{\mathrm{ts}}\log p_{\mathrm{ts}}\right)\end{array}\right.---\left(6\right)$

In formula (6):

K represents that single band is numbered, k=1,2 ... N, N are wave band number;

W _ij-krepresent the neighborhood entropy power of (i, j) pixel in the difference image of single band k;

H _ij-krepresent the neighborhood information entropy of (i, j) pixel in the difference image of single band k;

P _tsrepresent (i in the difference image of single band k, j) in pixel neighborhood window each gray-scale value (this gray-scale value span is [0,255] probability) occurred or frequency, t, s represent line number and the row number of pixel in neighborhood window respectively, t, s span [1, c], c × c is neighborhood window size.

Step 5: for pretreated two phase multiband remote sensing Image construction multiband associating images, obtain neighborhood mutual information.

Mutual information is based on the statistical nature of a pair pixel gray-scale value, do not need the textural characteristics of selection reference point or extraction image, and the linear relationship of image greyscale value need not be considered, therefore in the comparison of two image similarities and the detection of change information, there is larger flexibility ratio and accuracy.The mutual information size that degree and the pixel point of change are right is closely related, and is embodied in mutual information larger, and the right correlativity of pixel point is larger, and the intensity of variation of correspondence is less.

Therefore, for making full use of gray scale and the spatial information of image, the geography that effective reduction is existed by different sensors or different phase image and scattering difference (that is, grey-scale whole departure) cause the impact on classification and change detection.The present invention by the pixel point of two phase multiband remote sensing images to based on, construct the gray value vectors P of the neighborhood pixel of current pixel ₁, P ₂, and the neighborhood mutual information calculating current pixel is as multiband Combined Treatment result.

The computing method of neighborhood mutual information are as follows:

$\left\{\begin{array}{c}{\mathrm{RMI}}_{\mathrm{ij}}=H_{\mathrm{ij}}\left(C_{X1}\right)+H_{\mathrm{ij}}\left(C_{X2}\right)-H_{\mathrm{ij}}\left(C_{X12}\right)\\ C_{X1}=\frac{1}{d}{(P_1-m_1)}^{^2},C_{X2}=\frac{1}{d}{(P_2-m_2)}^{^2},C_{X12}=\frac{1}{d}(P_1-m_1)\×(P_2-m_2)\\ m_1=\frac{1}{d}{\mathrm{\Σ}}_d{P}_1,m_2=\frac{1}{d}{\mathrm{\Σ}}_d{P}_2\\ H_{\mathrm{ij}}\left(\mathrm{\Σd}\right)=\log \left({\left(2\mathrm{\πe}\right)}^{d/2}\det {\left(2\mathrm{\πe}\right)}^{d/2}\right)\end{array}\right.---\left(7\right)$

In formula (7):

RMI _ijrepresent the neighborhood mutual information of (i, j) pixel in multiband associating image;

H _ij() represents the computing asking neighborhood information entropy;

P ₁for the neighborhood pixel gray-scale value x of (i, j) pixel in first phase multiband remote sensing image _1iby the vector [x of pixel position sequential combination ₁₁..x _1i..x _1d];

P ₂be the neighborhood pixel gray-scale value x of (i, j) pixel in the second phase remote sensing image _2iby the vector [x of pixel position sequential combination ₂₁..x _2i..x _2d];

C _x1, C _x2be respectively the variance matrix of first phase remote sensing image and the second phase remote sensing image, C _x12for covariance matrix;

M ₁, m ₂vectorial P respectively ₁, P ₂middle element average, d=c × c, is expressed as Size of Neighborhood;

E represents natural logarithm.

Step 6: the neighborhood mutual information adopting single band neighborhood entropy power and single band difference image correction multiband associating image picture element, obtains revising image.

Function correction relation is as follows:

$\left\{\begin{array}{c}{M}_{\mathrm{ij}}=\frac{{\mathrm{RMI}}_{\mathrm{ij}}}{{\mathrm{Mod}}_{\mathrm{ij}}}\×{\mathrm{\Σ}}_{k=1}^N(w_{\mathrm{ij}-k}\×C_{\mathrm{ij}-k})\\ {\mathrm{Mod}}_{\mathrm{ij}}=\sqrt{{\mathrm{\Σ}}_{k=1}^N{\left(H_{\mathrm{ij}-k}\right)}^2}\end{array}\right.---\left(8\right)$

In formula (8):

K represents that single band is numbered, k=1,2 ... N, N are wave band number;

M _ijrepresent and revise image (i, j) pixel gray-scale value;

RMI _ijrepresent the neighborhood mutual information of multiband associating image (i, j) pixel;

C _ij-krepresent (i, j) pixel gray-scale value in the difference image of single band k, can be obtained by formula (1);

W _ij-kfor C _ij-kneighborhood entropy power, represent the contribution degree of single band k;

Mod _ijrepresent the corresponding N number of single-range entropy mould of (i, j) pixel;

H _ij-krepresent the neighborhood information entropy of (i, j) pixel in the difference image of single band k.

Step 7: utilize the fuzzy C-means clustering improved to carry out change to correction image and detect.

Remote sensing image change detects can regard special classification as, is about to revise image and is divided into change, the large class of non-changing two.Fuzzy C-means clustering introduces the concept of degree of membership, is a kind of soft clustering algorithm, has good treatment effect to mixed pixel.

The present invention does 2 improvement to fuzzy C-means clustering, one: in conjunction with Da-Jin algorithm single threshold segmentation thought, obtain and meet the maximum segmentation threshold t of inter-class variance ^*, and according to segmentation threshold t ^*correction image is split, and using the gray average of two class pixels after segmentation as initial cluster center.Its two: introduce the Similarity-Weighted degree of membership that the position of the neighborhood pixel of current pixel and attribute information obtain current pixel, and obtain weighted space function further, make fuzzy C-means clustering have stronger anti-noise ability, effectively avoid local optimum, improve classifying quality.

Threshold value t ^*determined by transfers between divisions function:

$\left\{\begin{array}{c}{w}_0={\mathrm{\Σ}}_{f=0}^t{P}_f,w_1={\mathrm{\Σ}}_{f=t}^{L-1}{P}_f\\ {\mathrm{\μ}}_T={\mathrm{\Σ}}_{f=0}^{L-1}\left(f{P}_{\mathrm{fi}}\right),{\mathrm{\μ}}_t={\mathrm{\Σ}}_{f=0}^t\left(f{P}_f\right)\\ {\mathrm{\μ}}_0=\frac{{\mathrm{\μ}}_t}{w_0},{\mathrm{\μ}}_1=\frac{{\mathrm{\μ}}_T-{\mathrm{\μ}}_t}{w_1},{\mathrm{\σ}}_0^2=\frac{{\mathrm{\Σ}}_{f=0}^t{(f-{\mathrm{\μ}}_0)}^2}{w_0},{\mathrm{\σ}}_1^2=\frac{{\mathrm{\Σ}}_{f=t}^{L-1}{(f-{\mathrm{\μ}}_1)}^2}{w_1}\\ {\mathrm{\σ}}_W^2=w_0{\mathrm{\σ}}_0^2+w_1{\mathrm{\σ}}_1^2\\ {\mathrm{\σ}}_B^2=w_0{({\mathrm{\μ}}_0-{\mathrm{\μ}}_T)}^2+w_1{({\mathrm{\μ}}_1-{\mathrm{\μ}}_T)}^2=w_0{w}_1{({\mathrm{\μ}}_0-{\mathrm{\μ}}_1)}^2\\ {\mathrm{\σ}}_T^2={\mathrm{\σ}}_W^2+{\mathrm{\σ}}_B^2\end{array}\right.---\left(9\right)$

In formula (9):

P _frepresent that revising gray scale in image is the pixel number of f;

W ₀represent that revising gray scale in image is not more than the pixel number of t, w ₁represent that revising gray scale in image is not less than the pixel number of t;

μ _trepresent and revise all pixel gray-scale value sums in image, μ _trepresent that revising gray scale in image is not more than the gray-scale value sum of the pixel of t;

μ ₀, represent that revising gray scale in image is not more than gray average and the gray variance of the pixel of t respectively;

μ ₁, represent that revising gray scale in image is less than gray average and the gray variance of the pixel of t respectively;

represent variance within clusters, inter-class variance, population variance respectively;

L represents grey-scale, generally gets 256.

Make inter-class variance maximum threshold value t and segmentation threshold t ^*.

The Similarity-Weighted degree of membership revising pixel in image is as follows:

$\left\{\begin{array}{c}{w}_{\mathrm{rt}}=1-\frac{x_r-x_t}{x_r+x_t}=\frac{2x_r}{x_r+x_t}\\ u_{\mathrm{rk}}^{\left(l\right)}={\mathrm{\Σ}}_{t=r-\frac{c}{2}}^{r+\frac{c}{2}}(w_{\mathrm{rt}}\×u_{\mathrm{tk}}^{\left(l\right)})\end{array}\right.---\left(10\right)$

In formula (10):

W _rtrepresent the neighborhood similarity weight of neighborhood pixel t centering imago unit r, the center pixel of center pixel r and neighborhood window, pixel in neighborhood pixel t and neighborhood window, its span is t, r represent neighborhood pixel and the Position Number of center pixel in neighborhood window respectively, Position Number namely from neighborhood window the first row first row pixel, according to mode number consecutively from left to right, from the top down, such as, for the pixel of the i-th row jth row, its Position Number is i*c+j;

X _rthe gray-scale value of expression center pixel r, x _trepresent the gray-scale value of neighborhood pixel t;

represent in the l time iteration, center pixel r to the degree of membership of kth class cluster centre, by its neighborhood pixel adjust power summation obtain;

represent in the l time iteration, neighborhood pixel t to the degree of membership of kth class cluster centre, its initial value obtains at random;

K=1,2, neighborhood window size is c × c.

The fuzzy C-mean algorithm method improved is carried out changing the flow process detected to correction image and is seen Fig. 2, comprises step:

(1) classification c=2 is set, respectively corresponding change class and non-changing class; Setting blur level m and convergency value ε, blur level m and convergency value ε is empirical value, in this concrete enforcement, makes m=2, ε=0.01; The initial subordinated-degree matrix U of random setting, subordinated-degree matrix U are by revising in image each pixel to the degree of membership of cluster centre; Setting primary iteration step number l=0.

(2) Ostu thresholding method is utilized to obtain segmentation threshold t ^*, with segmentation threshold t ^*divide the gray average cutting off two class pixels as initial cluster center V=[v ₁, v ₂].

(3) according to formula (10), in acquisition correction image, neighborhood pixel t is to the neighborhood similarity weight w of pixel r _rt, and adjust the degree of membership of pixel r to cluster centre V

(4) formula (11) is adopted to revise subordinated-degree matrix U and cluster centre V:

In formula (11):

T, r represent neighborhood pixel and the Position Number of center pixel in neighborhood window respectively, and k, y are cluster centre numbering, and C is total not number;

with represent in (l+1) and l iteration respectively, center pixel r is to the degree of membership of kth class cluster centre;

represent in the l time iteration, center pixel r is to the degree of membership of y class cluster centre;

X _rthe gray-scale value of expression center pixel r;

with represent kth class and the y class cluster centre of the l time iteration acquisition, y is traversal in [1, C] scope, can be equal with k;

represent the kth class cluster centre that (l+1) secondary iteration obtains;

M represents blur level, and n is pixel quantity in neighborhood window.

(5) subordinated-degree matrix norm is compared || U ^(l+1)-U ^(l)||, U ^(l+1), U ^(l)represent the subordinated-degree matrix under (l+1) secondary iteration, the l time iteration respectively; If || U ^(l+1)-U ^(l)||≤ε, finishing iteration, carries out cluster according to current cluster centre and subordinated-degree matrix; Otherwise, make l=l+1, perform step (4).

Claims (5)

1. the wave band correction change detecting method of remote sensing image fuzzy clustering, is characterized in that, comprise step:

Step 1, the pre-service of multiband remote sensing image, namely the filtering of multiband remote sensing image processes with strengthening;

Step 2, carries out single band separation to pretreated multiband remote sensing image, obtains single band remote sensing image;

Step 3, constructs the single band difference image of two phase single band remote sensing images;

Step 4, based on pretreated two phase multiband remote sensing Image construction multiband associating images, and obtains the neighborhood mutual information of each pixel of multiband associating image;

Step 5, according to formula revise the neighborhood mutual information of multiband associating image picture element, obtain revising image; Wherein: k represents that single band is numbered, k=1,2 ... N, N are wave band number; M _ijrepresent and revise image (i, j) pixel gray-scale value; w _ij-kfor the neighborhood entropy power of (i, j) pixel in the difference image of expression single band k, h _ij-krepresent the neighborhood information entropy of (i, j) pixel in the difference image of single band k; C _ij-krepresent (i, j) pixel gray scale in the difference image of single band k; RMI _ijrepresent the neighborhood mutual information of multiband associating image (i, j) pixel; Mod _ijrepresent the corresponding N number of single-range entropy mould of (i, j) pixel,

Step 6, for sample, makes classification c=2 to revise pixel gray-scale value in image, to revise the gray average of two class pixels after image Threshold segmentation as initial cluster center; Adjust power summation as center pixel degree of membership using neighborhood pixel degree of membership, the neighborhood similarity weight of each neighborhood pixel centering imago unit utilize fuzzy C-mean algorithm method to carry out change to correction image to detect;

In above-mentioned Threshold segmentation, based on the segmentation of Da-Jin algorithm single threshold, using threshold value maximum for the inter-class variance of the backdrop pels gray scale and goal pels gray scale that make correction image as segmentation threshold;

Above-mentioned in, t, r represent neighborhood pixel and the Position Number of center pixel in neighborhood window respectively; x _rthe gray-scale value of expression center pixel r, x _trepresent the gray-scale value of neighborhood pixel t.

2. the wave band correction change detecting method of remote sensing image fuzzy clustering as claimed in claim 1, is characterized in that:

Step 2 is specially:

Adopt weighted sum method by multiband remote sensing image greyscale, when the weights of a certain component of R, G, B single band are 1, when all the other are 0, namely obtain R, G, B single band remote sensing image through weighted sum.

3. the wave band correction change detecting method of remote sensing image fuzzy clustering as claimed in claim 1, is characterized in that:

In conjunction with differential technique and ratioing technigue structure single band difference image in step 3, be specially:

According to formula construct the single band difference image of two phase single band remote sensing images, wherein:

C _ij-krepresent (i, j) pixel gray-scale value in the difference image of single band k;

R _ij-krepresent (i, j) pixel gray-scale value in the ratio image of single band k;

D _ij-krepresent (i, j) pixel gray-scale value in the Difference image of single band k;

A be all pixels of ratio image in the gray average and Difference image of kth wave band all pixels in the difference of the gray average of kth wave band;

B is all pixels of ratio image in the gray-scale value root mean square of kth wave band and all pixels of the Difference image difference at the gray-scale value root mean square of kth wave band.

4. the wave band correction change detecting method of remote sensing image fuzzy clustering as claimed in claim 1, is characterized in that:

Step 6 comprises sub-step further:

6.1 setting classification C=2, rule of thumb set blur level m and convergency value ε, the initial subordinated-degree matrix of random setting, and subordinated-degree matrix is formed by revising the degree of membership of each pixel to cluster centre in image; Make primary iteration step number l=0;

6.2 based on Da-Jin algorithm single threshold split plot design, using make the inter-class variance of the background of correction image and goal pels gray scale maximum threshold value as segmentation threshold, to revise the gray average of two class pixels after image Threshold segmentation as initial cluster center

6.3 by the current degree of membership of neighborhood pixel t weighted sum as center pixel r degree of membership namely the current degree of membership of neighborhood pixel t obtains according to current subordinated-degree matrix U, its weights $w_{\mathrm{rt}}=\frac{2x_r}{x_r+x_t};$

6.4 adopt formula $\left\{\begin{array}{c}{u}_{\mathrm{rk}}^{(l+1)}=\frac{1}{{\mathrm{\Σ}}_{y=1}^C\left(\frac{u_{\mathrm{rk}}^{\left(l\right)}\×\left|\right|v_k^{\left(l\right)-x_r}\left|\right|}{u_{\mathrm{ry}}^{\left(l\right)}\×\left|\right|v_y^{\left(l\right)}-x_r\left|\right|}{)}^{\frac{2}{m-1}}\right)},\∀y,\∀k\\ v_k^{(l+1)}=\frac{{\mathrm{\Σ}}_{r=1}^n\left[{\left(u_{\mathrm{rk}}^{\left(l\right)}\right)}^m{x}_r\right]}{{\mathrm{\Σ}}_{r=1}^n{\left(u_{\mathrm{rk}}^{\left(l\right)}\right)}^m}\end{array}\right.$ Revise subordinated-degree matrix and cluster centre, wherein, t, r represent neighborhood pixel and the Position Number of center pixel in neighborhood window respectively, and k, y are cluster centre numbering, and C is classification number; with represent in (l+1) and l iteration respectively, center pixel r is to the degree of membership of kth class cluster centre; represent in the l time iteration, center pixel r is to the degree of membership of y class cluster centre; x _rthe gray-scale value of expression center pixel r; with represent kth class and the y class cluster centre of the l time iteration acquisition, y is traversal in [1, C] scope, can be equal with k; represent the kth class cluster centre that (l+1) secondary iteration obtains; N is pixel quantity in neighborhood window;

6.5 compare subordinated-degree matrix norm || U ^(l+1)-U ^(l)||, U ^(l+1), U ^(l)represent the subordinated-degree matrix of (l+1) secondary iteration, the l time iteration acquisition respectively; If || U ^(l+1)-U ^(l)||≤ε, finishing iteration, carries out cluster according to current cluster centre and subordinated-degree matrix; Otherwise, make l=l+1, perform step 6.4.

5. the wave band correction change detecting system of remote sensing image fuzzy clustering, is characterized in that, comprising:

Pretreatment module, is used for the pre-service of multiband remote sensing image, namely multiband remote sensing image filtering and strengthen process;

Single band separation module, is used for carrying out single band separation to pretreated multiband remote sensing image, obtains single band remote sensing image;

Single band difference image constructing module, is used for the single band difference image of structure two phase single band remote sensing image;

Neighborhood mutual information obtains module, is used for based on pretreated two phase multiband remote sensing Image construction multiband associating images, and obtains the neighborhood mutual information of each pixel of multiband associating image;

Revise image and obtain module, be used for according to formula revise the neighborhood mutual information of multiband associating image picture element, obtain revising image; Wherein: k represents that single band is numbered, k=1,2 ... N, N are wave band number; M _ijrepresent and revise image (i, j) pixel gray-scale value; w _ij-kfor the neighborhood entropy power of (i, j) pixel in the difference image of expression single band k, h _ij-krepresent the neighborhood information entropy of (i, j) pixel in the difference image of single band k; C _ij-krepresent (i, j) pixel gray scale in the difference image of single band k; RMI _ijrepresent the neighborhood mutual information of multiband associating image (i, j) pixel; Mod _ijrepresent the corresponding N number of single-range entropy mould of (i, j) pixel,

Fuzzy C-means clustering module, being used for revising pixel gray-scale value in image is sample, makes classification c=2, to revise the gray average of two class pixels after image Threshold segmentation as initial cluster center; Adjust power summation as center pixel degree of membership using neighborhood pixel degree of membership, the neighborhood similarity weight of each neighborhood pixel centering imago unit utilize fuzzy C-mean algorithm method to carry out change to correction image to detect;

In above-mentioned Threshold segmentation, based on the segmentation of Da-Jin algorithm single threshold, using threshold value maximum for the inter-class variance of the backdrop pels gray scale and goal pels gray scale that make correction image as segmentation threshold;

Above-mentioned in, t, r represent neighborhood pixel and the Position Number of center pixel in neighborhood window respectively; x _rthe gray-scale value of expression center pixel r, x _trepresent the gray-scale value of neighborhood pixel t.