CN103971364A - Remote sensing image variation detecting method on basis of weighted Gabor wavelet characteristics and two-stage clusters - Google Patents
Remote sensing image variation detecting method on basis of weighted Gabor wavelet characteristics and two-stage clusters Download PDFInfo
- Publication number
- CN103971364A CN103971364A CN201410134520.1A CN201410134520A CN103971364A CN 103971364 A CN103971364 A CN 103971364A CN 201410134520 A CN201410134520 A CN 201410134520A CN 103971364 A CN103971364 A CN 103971364A
- Authority
- CN
- China
- Prior art keywords
- remote sensing
- class
- gabor wavelet
- change
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a remote sensing image variation detecting method on the basis of weighted Gabor wavelet characteristics and two-stage clusters. The processed objects include optical remote sensing images and SAR (synthetic aperture radar) images and the remote sensing image variation detecting method includes (1) generating difference images according to remote sensing image types; (2) subjecting the difference images to Gabor wavelet transform; (3) extracting multiscale and multidirectional characteristics of the Gabor wavelet transform of the difference images; (4) designing the weighting coefficient and acquiring the weighted Gabor wavelet characteristics; (5) clustering the weight Gabor wavelet characteristics by means of the two-stage cluster strategy; (6) acquiring variation detecting results. By the remote sensing image variation detecting method, loss of marginal information is reduced, stronger, weak and slight varying areas can be detected at the same time, the total mistake pixel number is decreased, more detail information is reserved and variation results can be effectively extracted.
Description
Technical field
The invention belongs to digital image processing techniques field, relate generally to Multitemporal Remote Sensing Images and change detection research direction, the specifically method for detecting change of remote sensing image based on weighting Gabor wavelet character and two-stage cluster, handling object has comprised multidate remote sensing image and synthetic-aperture radar (SyntheticAperture Radar, SAR) image simultaneously.The method can be applicable in many practical problemss of the earth observations such as forest inventory investigation, soil covering/land utilization dynamic monitors, environmental hazard estimation, city planning and layout, Hitting Effect Evaluation, particularly Natural Disaster monitoring and evaluation.
Background technology
Remote sensing technology is from the electromagnetic wave of remote perception target reflection or self radiation, visible ray, infrared ray etc., target is carried out to the technology of detecting and identifying, have and can obtain data information on a large scale, the feature such as the speed of obtaining information is fast, the cycle is short and contain much information.According to the difference of obtain manner, can be divided into remote optical sensing and microwave remote sensing two classes.Wherein, the equipment that remote optical sensing is corresponding is simple, and institute's image space geometric resolution that obtains is high, and image is easy to decipher, but can only use by day, affected by weather condition more serious; Because microwave has good penetration capacity, microwave remote sensing (SAR) can round-the-clock, round-the-clock imaging over the ground, but coherent imaging characteristic makes its image have the property taken advantage of coherent speckle noise.Both respectively have feature, mutually supplement, and are just bringing into play more and more important effect in national economy and national defense construction.
Along with the development of sensor technology, space technology and computer technology, numerous machine/space remote sensing platforms put into effect, novel, the high resolution sensor carrying can carry out uninterruptedly earth surface, continue observation, obtained a large amount of wide cuts, high-definition remote sensing data.But in the face of growing acquisition capability, unbefitting is with it that the research and development of remote sensing image processing and Interpretation Technology are relatively lagged behind, and can not meet the active demand of practical application.Wherein, Multitemporal Remote Sensing Images variation detects as one of the basis of processing and decipher and gordian technique, refer to that two width or several remote sensing images by the same area, different times are obtained compare analysis, and then obtain the change information of interested atural object, scene or target according to difference between image, being limited by the movable aggravation of the mankind and disaster takes place frequently, its research and development are more and more subject to people and pay attention to, and become gradually the study hotspot in remote sensing field.
In recent years, Chinese scholars has proposed a lot of effectively change detecting methods, and summing up to be divided into has supervision to change detection and change and detect without supervision.Be limited to and have the required ground of supervision variation detection real change classification sample to be difficult to obtain, work at present mainly concentrates on without supervision change detection class, and it is roughly divided into: (1) Multitemporal Remote Sensing Images variation detection based on Cluster Distribution Divergence; (2) Multitemporal Remote Sensing Images of analyzing based on differential image changes detection; And the Multitemporal Remote Sensing Images that merge based on Markov (3) changes detection.Particularly general to the research of Equations of The Second Kind algorithm, core concept is the binary classification/segmentation problem that variation test problems is considered as to image, can be subdivided into again the strategies such as cluster analysis, intelligent optimization, Threshold segmentation, Finite mixture model, markov random file, active profile and level set.Wherein, cluster analysis simple by it, effectively and generally approved.
T.Celik is at document [T.Celik, " Unsupervised change detection insatellite images using principal component analysis and K-meansclustering; " IEEE Geoscience and Remote Sensing Letters, vol.6, no.4, the method for detecting change of remote sensing image of cluster analysis is proposed pp.772-776,2009.] the earliest.Particularly, first utilize principal component analysis (Principal Component Analysis, PCA) extract each pixel characteristic of correspondence vector, re-use K-means algorithm proper vector is divided into two classes (change class and do not change class), obtain changing detection figure.Although the method is simply effective, does not make full use of the partial structurtes information of differential image, and be vulnerable to noise.For further improving the validity detecting, [M.Volpi, D.Tuia, the G.Camps-Valls such as M.Volpi, and M.Kanevski, " Unsupervisedchange detection with kernels, " IEEE Geoscience and Remote SensingLetters, vol.9, no.6, pp.1026-1030,2012] shine upon the Nonlinear separability problem in the original input space is converted into the linear separability problem in feature space by core, utilize core K-means algorithm to realize cluster.Simultaneously, for reducing noise effect, applicant [Y.Q.Cheng, H.C.Li, T.Celik, and F.Zhang, " FRFT-based improved algorithm of unsupervised changedetection in SAR images via PCA and K-means clustering ", in Proceedingof IEEE International Geoscience and Remote Sensing Symposium, Melbourne, Australia, and [the Y.G.Zheng such as Y.G.Zheng July2013:1952-1955.], X.R.Zhang, B.Hou, and G.C.Liu, " Using combined differenceimage and K-means clustering for SAR image change detection, " IEEEGeoscience and Remote Sensing Letters, vol.11, no.3, pp.691-695, 2014] utilize respectively low order Fractional Fourier Transform (Fractional Fourier Transform, FRFT), PPB(Probabilistic Patch-Based) wave filter produces the differential image of squelch, carry out again spatial domain cluster and change detection, can obtain satisfied testing result, but all still do not take into full account the partial structurtes information of differential image.In addition along with the raising of remote sensing images resolution, the separability that changes and do not change between class reduces, uncertain and ambiguity strengthens, make existing change detecting method can not effectively detect the change information of high-resolution remote sensing image, there is higher flase drop and false dismissal probability, become restriction high-definition remote sensing and change the bottleneck that detects application, be badly in need of more excellent change detecting method.
Summary of the invention
The present invention utilizes Gabor wavelet transformation and fuzzy C-means clustering (Fuzzy C-means, FCM) to propose to change new detecting method based on the Multitemporal Remote Sensing Images of weighting Gabor wavelet character and two-stage cluster.The method, by the multiple dimensioned and Orientation Features of Gabor wavelet transformation, can effectively be portrayed the partial structurtes information of differential image, and two-stage cluster strategy can reduce noise testing result is affected, and algorithm simply effectively can meet actual needs.
Based on a method for detecting change of remote sensing image for weighting Gabor wavelet character and two-stage cluster, comprise the steps:
Step 1, utilizes and obtains at areal different time that two width sizes are identical, the remote sensing images X of mutual registration
0and X
1to produce differential image X
d;
Step 2, to differential image X
dcarry out Gabor wavelet transformation;
Step 3, the multiple dimensioned and Orientation Features of extraction differential image Gabor wavelet transformation;
Step 4, design weights coefficient, obtains weighting Gabor wavelet character;
Step 5, adopts the two-stage cluster strategy based on FCM algorithm to carry out cluster to weighting Gabor wavelet character, is specifically divided into two steps, that is:
(a) use FCM algorithm that the proper vector of the differential image pixel of previous step acquisition is divided into: change class, do not change class and border class, the cluster centre that obtains changing class and do not change class, uses v
cand v
urepresent respectively to change class ω
cdo not change class ω
ucluster centre, ω
brepresent border class;
(b) computation bound class ω
bin each proper vector with change class and do not change the distance at the class center of birdsing of the same feather flock together, be assigned to cluster centre apart from minimum class; Then the variation class obtaining with (a) step and do not change class combination, obtains new variation class and does not change class;
Step 6, changes class location of pixels and puts 1, does not change class location of pixels and sets to 0, and obtains final two-value and changes testing result.
Wherein: the multiple dimensioned and Orientation Features of the extraction differential image Gabor wavelet transformation described in step 3, be embodied as: the Gabor small echo of selecting 5 yardsticks and 8 directions, v ∈ { 0,, 4}, u ∈ { 0,7}, v represents the scale parameter of Gabor small echo, u has embodied the directional selectivity of Gabor small echo; After the Gabor small echo phase convolution of differential image with 5 yardsticks and 8 directions, obtain 40 characteristic image G
v, u(z), wherein v ∈ 0 ..., 4}, u ∈ 0 ..., 7}; For convenience of representing with following symbol: J
k(z)=G
v, u(z), wherein k=u+8v, v ∈ 0 ..., 4}, u ∈ 0 ..., 7}, the corresponding proper vector of each pixel can be expressed as v (i, j)=[J
0(i, j) J
1(i, j) ... J
39(i, j)], wherein (i, j) is coordinate points, v (i, j) is the pixel Gabor wavelet character vector representation that on image, coordinate is (i, j).
In addition, the design of the weights coefficient to extracted Gabor wavelet character described in step 4, is embodied as: the frequency parameter of different scale can obtain as k
v=k
max/ f
v; If v ∈ 0 ..., 4}, this corresponding frequency parameter is k
0, k
1, k
2, k
3, k
4, calculate weights coefficient and be followed successively by
and be summed to
Then be normalized and be followed successively by
different scale v ∈ 0 ..., the weights coefficient of 4} set gradually into
it is exactly lower directive weights coefficient of each yardstick that normalization weights coefficient is overturn according to yardstick order, and then obtains weighting Gabor wavelet character.
For different image sources, differential image X
dthere is different obtain manners: for remote sensing image, differential image X
dfor input picture X
0and X
1the absolute value directly subtracting each other, X
d=| X
1-X
0|; For synthetic aperture radar (SAR) image, its differential image X
dget input picture X
0and X
1make the absolute value of logarithm ratio,
Beneficial effect
1, the present invention is by Gabor wavelet transformation, the partial structurtes information that can portray fully and effectively differential image by extracting the multiple dimensioned and Orientation Features of differential image, and affect to weaken noise and heterogeneous pixel according to different frequency design weight coefficient.
2, the two-stage cluster strategy based on FCM algorithm proposed by the invention carries out cluster to weighting Gabor wavelet character, can in first order cluster process, obtain the cluster centre that separability better changes class and do not change class, and then in the cluster process of the second level, can more effectively distinguish the classification of edge class pixel.
3, the method can at the stronger region of variation of complete acquisition simultaneously, detect faint, less region of variation, reduces the loss of marginal information.Can take into account false retrieval pixel count and undetected pixel count, total erroneous pixel number is reduced, retain more details information, effectively extract result of variations.
Brief description of the drawings
The Multitemporal Remote Sensing Images change detecting method general frame of Fig. 1 based on weighting Gabor wavelet character and two-stage cluster;
Feature extraction mode after Fig. 2 differential image Gabor wavelet transformation;
First group of remote sensing image data collection Bern that Fig. 3 emulation experiment is used, the ERS2SAR image (a) obtaining in April, 1999, the ERS2SAR image (b) obtaining in May, 1999, (c) changes and detects reference diagram;
Second group of remote sensing image data collection burn that Fig. 4 emulation experiment is used, the remote sensing image (a) obtaining on August 5th, 1986, the remote sensing image (b) obtaining on August 5th, 1992, (c) changes and detects reference diagram;
The 3rd group of remote sensing image data collection lake that Fig. 5 emulation experiment is used, the remote sensing image (a) obtaining on August 5th, 1986, the remote sensing image (b) obtaining on August 5th, 1992;
Tri-groups of differential images corresponding to remote sensing image data collection difference of Fig. 6: (a) Bern data set differential image, (b) burn data set differential image, (c) lake data set differential image;
The variation testing result of Fig. 7 the present invention and control methods emulation SAR view data Bern, (a) the variation testing result of control methods 1, (b) the variation testing result of control methods 2, (c) the variation testing result of control methods 3, (d) variation testing result of the present invention, (e) changes and detects reference diagram;
The variation testing result of Fig. 8 the present invention and control methods simulate optical remote sensing image data burn, (a) the variation testing result of control methods 1, (b) the variation testing result of control methods 2, (c) the variation testing result of control methods 3, (d) variation testing result of the present invention, (e) changes and detects reference diagram;
The variation testing result of Fig. 9 the present invention and control methods simulate optical remote sensing image data lake, (a) the variation testing result of control methods 1, (b) the variation testing result of control methods 2, (c) the variation testing result of control methods 3, (d) variation testing result of the present invention.
Embodiment
Embodiments of the invention are described with reference to the accompanying drawings.Should be appreciated that each embodiment of the present invention described here is only used to explain better principle of the present invention and concept, instead of will limit the present invention.After reading such description, those skilled in the art are easy to construct other amendments or replacement, and such amendment or replacement should be understood to fall into scope of the present invention.
Fig. 1 has provided the flow process framework of the embodiment of the present invention, and concrete enforcement comprises following steps:
(1) utilize Multitemporal Remote Sensing Images to produce differential image.Two width sizes are the remote sensing images X of H × W and mutual registration
0={ x
0(i, j) | 1≤i≤H, 1≤j≤W} and X
1={ x
1(i, j) | 1≤i≤H, 1≤j≤W}, they are at areal different time t
0and t
1the remotely-sensed data obtaining respectively.Generate differential image X according to the type of remote sensing images
d.
For optical imagery (Optical Images), input picture X
0and X
1the absolute value directly subtracting each other is exactly differential image X
d,
X
D=|X
1-X
0|
And for SAR(Synthetic Aperture Radar) image, differential image X
dinput picture X
0and X
1ratio images, in order to improve the pixel value of low amplitude, our correlative value image is carried out a logarithm operation,
Wherein log is natural logarithm operator, can make to change in differential image class ω
cdo not change class ω
udistribute more even.Make X simultaneously
0and X
1the inherent residual property taken advantage of (×) spot disturbs and converts additivity (+) interference to.
(2) the differential image X first step being obtained
dcarry out Gabor wavelet transformation.Two-Dimensional Gabor Wavelets kernel function (wave filter) form can be expressed as,
Wherein
v represents the scale parameter of Gabor small echo, and u has embodied the directional selectivity of Gabor small echo, a pixel coordinate in z=(x, y) correspondence image, || || represent European norm computing.
Gabor wavelet vectors parameter k
v, ucontrol the width of Gaussian window, wavelength and the direction of oscillating part, can be expressed as,
Wherein k
v=k
max/ f
v(v=0,1 ..., V-1),
(u=0,1 ..., U-1), generally
parameter k
maxbe the maximum frequency of Gabor small echo, and f represent the interval factor at frequency domain between adjacent Gabor small echo.
Gray level image I (x, y) is carried out to Gabor wavelet transformation, is equivalent to image and Gabor small echo is made convolution, can be expressed as follows,
G
v，u(z)＝I(z)*g
v，u(z)
Wherein, a pixel coordinate in z=(x, y) correspondence image, " * " represents two-dimensional convolution operation symbol, and v represents the scale parameter of Gabor small echo, and u has embodied the directional selectivity of Gabor small echo, g
v, u(z) be Two-Dimensional Gabor Wavelets function, G
v, u(z) be the Gabor wavelet function convolution results of image and corresponding yardstick v, direction u, i.e. the Gabor Wavelet representation for transient of image.
(3) the multiple dimensioned and Orientation Features of extraction differential image Gabor wavelet transformation.Generally, select the Gabor small echo of 5 yardsticks and 8 directions, v ∈ 0 ..., 4}, u ∈ 0 ..., 7}.After the Gabor small echo phase convolution of differential image with 5 yardsticks and 8 directions, obtain 40 characteristic image G
v, u(z), wherein v ∈ 0 ..., 4}, u ∈ 0 ..., 7}.For below easy to use we represent with following symbol, J
k=G
v, u(z), wherein k=u+8v, v ∈ 0 ..., 4}, u ∈ 0 ..., 7}, 40 characteristic images can be expressed as J (z)={ J
k(z), k=0,1 ..., 39}.Therefore the corresponding proper vector composition of each pixel of image can be expressed as v (i, j)=[J
0(i, j) J
1(i, j ... J
39(i, j)], wherein (i, j) be coordinate points, v (i, j) is the pixel Gabor wavelet character vector representation that on image, coordinate is (i, j), it is partial structurtes information and the multidirectional feature of picture engraving accurately, and rotation, yardstick and illumination variation etc. are had to certain robustness.Characteristic extraction procedure as shown in Figure 2.
(4) to extracted Gabor wavelet character design weights.Generally, select the Gabor small echo of 5 yardsticks and 8 directions, v ∈ 0 ..., 4}, u ∈ 0 ..., 7}.The Gabor feature of lower 8 directions of same yardstick has identical weights, and weights setting is only relevant to yardstick, independent of direction.The frequency parameter of different scale can obtain as k
v=k
max/ f
v.If v ∈ 0 ..., 4}, this corresponding frequency parameter is k
0, k
1, k
2, k
3, k
4, calculate weights coefficient and be followed successively by
and be summed to
Then be normalized and be followed successively by
different scale v ∈ 0 ..., the weights coefficient of 4} set gradually into
explain like this and be convenient to understand.For the different directions under same yardstick, identical weights are set, calculate weights normalization.It is exactly lower directive weights coefficient of each yardstick that normalized weights are overturn according to yardstick order, and then obtains weighting Gabor wavelet character.Can weaken noise and the impact of heterogeneous pixel by weights are set, be conducive to complete extraction change information.
(5) utilize the two-stage cluster strategy based on FCM algorithm that the present invention proposes to carry out cluster to the Gabor feature of weighting.
(5a) use FCM algorithm that the proper vector of the differential image pixel of previous step acquisition is divided into: change class, do not change class and border class, the cluster centre that obtains changing class and do not change class, uses v
cand v
urepresent respectively to change class ω
cdo not change class ω
ucluster centre, ω
brepresent border class.
(5b) computation bound class ω
bin each proper vector with change class and do not change the distance at the class center of birdsing of the same feather flock together, be assigned to cluster centre apart from minimum class.Then the variation class obtaining with (5a) step and do not change class combination, obtains new variation class and does not change class.
(6) obtain and change testing result.It is CM={cm (i that two-value changes testing result, j) | 1≤i≤H, 1≤j≤W}, wherein " 1 " represents that corresponding picture position is (that is to say of changing, belong to and change class), " 0 " represents that corresponding picture position is indeclinable (that is to say, belong to and do not change class).We put 1 by variation class location of pixels, do not change class location of pixels and set to 0, and obtain final variation testing result.
Effect of the present invention can further illustrate by following experiment:
1, emulation experiment condition
The method for detecting change of remote sensing image based on Gabor small echo and two-stage cluster and control methods that the present invention proposes are all to realize on identical experiment simulation platform.Experimental calculation machine is configured to AMD Athlon (tm) 7750 dual core processors, dominant frequency 2.70GHz, RAM is that 4.00GB(3.25GB can use), 32-bit operating system, software platform is MATLAB7.13.
2, emulation experiment data
Emulation experiment of the present invention is used real remotely-sensed data, comprises remote sensing image and SAR image.
First group of remote sensing image data collection Bern.In Fig. 3, be two width SAR view data and an amplitude variation reference diagram.The identical SAR image of this two width size is by ERS2(European Remote Sensing2) the SAR sensor of satellite obtains respectively in April, 1999 and in May, 1999 in the neighbourhood at Switzerland Bern.Piece image is Bern earth's surface information in the neighbourhood before in April, 1999 flood generation, and the second width image is the earth's surface information of the same area after in May, 1999 flood generation, and the size of two width images is 301 × 301, and gray level is 256.The 3rd width image is to change reference diagram, and it is the width bianry image obtaining in conjunction with known view data by the real terrestrial information in locality, and in figure, white pixel represents the situation of change on ground.
Second group of remote sensing image data collection burn, Fig. 4 is two width remote sensing images and an amplitude variation reference diagram.The size of front two width images is 200 × 200, respectively on August 5th, 1986 and on August 5th, 1992 at Lake Tahoe, Reno, Nevada obtains.Change because ground forest fire causes a large amount of regions, the 3rd width image is to change reference diagram.
The 3rd group of remote sensing image data collection lake, Fig. 5 is two width remote sensing images.The size of two width images is 200 × 200, respectively on August 5th, 1986 and on August 5th, 1992 at Lake Tahoe, Reno, Nevada obtains, the variation that mainly shown that lake surface is dry, land use and deforestation causes.
3, evaluation of result index
General in the situation that actual change is known, weighing change detection algorithm performance has following three indexs.
(1) False Alarms (FA): error-detecting pixel count, represents that reality does not change and is detected as the sum of all pixels of variation.
(2) Missed Alarms (MA): loss detection pixel count, represents actual change and be detected as unchanged sum of all pixels.
(3) Total Errors (TE): total erroneous pixel number is error-detecting pixel count (FA) above and the summation of loss detection pixel count (MA).This is to weigh the most important index of change detection algorithm, is the main basis for estimation that proves result quality.
4, contrast simulation experiment
For validity of the present invention is described, we compare simulated effect of the present invention and following three methods:
Control methods 1, document [T.Celik, " Unsupervised change detection insatellite images using principal component analysis and K-meansclustering; " IEEE Geoscience and Remote Sensing Letters, vol.6, no.4, pp.772-776,2009.] variation that utilizes principal component analysis (PCA) (PCA, Principal Component Analysis) and K-means clustering algorithm to realize remote sensing images in detects.First the method utilizes principal component analysis to extract each pixel characteristic of correspondence vector, re-uses K-means algorithm proper vector is divided into two classes (change class and do not change class), obtains changing detection figure.
Control methods 2, document [Y.Q.Cheng, H.C.Li, T.Celik, and F.Zhang, " FRFT-based improved algorithm of unsupervised change detection in SARimages via PCA and K-means clustering ", in Proceeding of IEEEInternational Geoscience and Remote Sensing Symposium, Melbourne, Australia, July2013:1952-1955.] be to utilize the Fractional Fourier Transform of low order to improve the method that document proposes above.
Control methods 3, document [Y.G.Zheng, X.R.Zhang, B.Hou, and G.C.Liu, " Using combined difference image and K-means clustering for SAR imagechange detection, " IEEE Geoscience and Remote Sensing Letters, vol.11, no.3, pp.691-695, 2014] first to carry out PPB(ProbabilisticPatch-Based to Multitemporal Remote Sensing Images) filtering, then utilize two different operators and produce differential image to be combined in conjunction with average and medium filtering respectively, obtain target differential image by linear weighted function again, finally use K-means cluster to obtain changing testing result.
5, emulation experiment content
Test 1 calculated difference image
Emulation experiment of the present invention adopts three groups of remote sensing image data collection, comprises a pair of SAR image and two pairs of remote sensing images.According to different formula, generate differential image X according to the type of remote sensing images
d, respectively as shown in Fig. 6 (a)-(c).
Experiment 2SAR view data Bern
That this emulation experiment is used is first group of remote sensing image data collection Bern.Fig. 7 is the variation testing result of the present invention and control methods emulation SAR view data Bern, wherein Fig. 7 (a) is the variation testing result of control methods 1, Fig. 7 (b) is the variation testing result of control methods 2, Fig. 7 (c) is the variation testing result of control methods 3, Fig. 7 (d) is variation testing result of the present invention, and Fig. 7 (e) changes to detect reference diagram.
Except observing and carry out qualitative contrast according to vision, also carry out quantitative comparison.Table 1 has been listed the accuracy evaluation data of emulation experiment output image.
Table 1 the present invention and control methods emulation Bern data variation testing result accuracy evaluation data
Test 3 remote sensing image data burn
That this emulation experiment is used is first group of remote sensing image data collection burn.Fig. 8 is the variation testing result of the present invention and control methods simulate optical remote sensing image data burn, wherein Fig. 8 (a) is the variation testing result of control methods 1, Fig. 8 (b) is the variation testing result of control methods 2, Fig. 8 (c) is the variation testing result of control methods 3, Fig. 8 (d) is variation testing result of the present invention, and Fig. 8 (e) changes to detect reference diagram.
Except observing and carry out qualitative contrast according to vision, also carry out quantitative comparison.Table 2 has been listed the accuracy evaluation data of emulation experiment output image.
Table 2 the present invention and control methods emulation burn data variation testing result accuracy evaluation data
Test 4 remote sensing image data lake
That this emulation experiment is used is first group of remote sensing image data collection lake.Fig. 9 is the variation testing result of the present invention and control methods simulate optical remote sensing image data lake, wherein Fig. 9 (a) is the variation testing result of control methods 1, Fig. 9 (b) is the variation testing result of control methods 2, Fig. 9 (c) is the variation testing result of control methods 3, and Fig. 9 (d) is variation testing result of the present invention.Detect reference diagram because remote sensing image lake lacks to change, will provide qualitative comparative result.
6, analysis of simulation result
Variation testing result and accuracy evaluation data by analysis the present invention and control methods emulation SAR view data Bern are known, and total erroneous pixel number of simulation result of the present invention is 296, has obtained good result than other control methods.Its error-detecting pixel count is 131 simultaneously, and loss detection pixel count 165, has reached a kind of compromise preferably.
Variation testing result and accuracy evaluation data by analysis the present invention and control methods simulate optical remote sensing image data burn are known, total erroneous pixel number of simulation result of the present invention is 1108, control methods 2 be 1057, these two kinds of methods are all obviously better than other two kinds of control methodss.And in simulation result of the present invention loss detection pixel count minimum be 79, can detect preferably faint region of variation, be conducive to the maintenance of region of variation detailed information.
The variation that known by analyzing variation testing result and the image data set information of the present invention and control methods simulate optical remote sensing image data lake, this view data has mainly shown that lake surface dries up, land use and deforestation cause.Control methods 3 can detect the dry variation causing of lake surface preferably, still cannot effectively detect the variation that land use and deforestation cause, and loses a large amount of faint region of variation.Simulation result of the present invention not only can more intactly detect the dry variation causing of lake surface than other control methods, has retained the variation that land use and deforestation cause simultaneously.
In sum, the present invention is applicable to detect multidate remote sensing image and SAR image simultaneously, utilize Gabor wavelet transformation to extract and portray fully and effectively the multiple dimensioned Orientation Features of differential image partial structurtes information, consider the impact of avoiding noise and heterogeneous pixel, obtain weighting Gabor wavelet character by design weights coefficient and represent.Utilize the two-stage cluster strategy based on FCM algorithm proposed by the invention to carry out cluster to weighting Gabor wavelet character, two-stage cluster can produce two at a distance of farther cluster centre simultaneously, more easily will change class and not change class separately, is better detected effect.By the comparison of quantitative and qualitative analysis, can prove the method for detecting change of remote sensing image based on weighting Gabor wavelet character and two-stage cluster proposed by the invention, can in the stronger region of variation of complete acquisition, faint, less region of variation be detected, reduce the loss of marginal information.And take into account false retrieval pixel count and undetected pixel count, total erroneous pixel number is reduced, retained more details information, effectively extract result of variations.
Those of ordinary skill in the art is obviously clear and understand, the inventive method for above embodiment only for the inventive method is described, and be not limited to the inventive method.Although effectively described the present invention by embodiment, there are many variations and do not depart from spirit of the present invention in the present invention.Without departing from the spirit and substance of the case in the method for the present invention, those skilled in the art are when making various corresponding changes or distortion according to the inventive method, but these corresponding changes or distortion all belong to the protection domain that the inventive method requires.
Claims (4)
1. the method for detecting change of remote sensing image based on weighting Gabor wavelet character and two-stage cluster, comprises the steps:
Step 1, utilizes and obtains at areal different time that two width sizes are identical, the remote sensing images X of mutual registration
0and X
1to produce differential image X
d;
Step 2, to differential image X
dcarry out Gabor wavelet transformation;
Step 3, the multiple dimensioned and Orientation Features of extraction differential image Gabor wavelet transformation;
Step 4, design weights coefficient, obtains weighting Gabor wavelet character;
Step 5, adopts the two-stage cluster strategy based on FCM algorithm to carry out cluster to weighting Gabor wavelet character, is specifically divided into two steps, that is:
(a) use FCM algorithm that the proper vector of the differential image pixel of previous step acquisition is divided into: change class, do not change class and border class, the cluster centre that obtains changing class and do not change class, uses v
cand v
urepresent respectively to change class ω
cdo not change class ω
ucluster centre, ω
brepresent border class;
(b) computation bound class ω
bin each proper vector with change class and do not change the distance at the class center of birdsing of the same feather flock together, be assigned to cluster centre apart from minimum class; Then the variation class obtaining with (a) step and do not change class combination, obtains new variation class and does not change class;
Step 6, changes class location of pixels and puts 1, does not change class location of pixels and sets to 0, and obtains final two-value and changes testing result.
2. the method for detecting change of remote sensing image based on weighting Gabor wavelet character and two-stage cluster according to claim 1, it is characterized in that the multiple dimensioned and Orientation Features of the extraction differential image Gabor wavelet transformation described in step 3, be embodied as: select the Gabor small echo of 5 yardsticks and 8 directions, v ∈ 0 ... 4), u ∈ 0 ..., 7}, v represents the scale parameter of Gabor small echo, and u has embodied the directional selectivity of Gabor small echo; After the Gabor small echo phase convolution of differential image with 5 yardsticks and 8 directions, obtain 40 characteristic image G
v, u(z), wherein v ∈ 0 ..., 4}, u ∈ 0 ..., 7}; For convenience of representing with following symbol: J
k(z)=G
v, u(z), wherein k=u+8v, v ∈ 0 ..., 4}, u ∈ 0 ..., 7}, the corresponding proper vector of each pixel can be expressed as v (i, j)=[J
0(i, j) J
1(i, j) ... J
39(i, j)], wherein (i, j) is coordinate points, v (i, j) is the pixel Gabor wavelet character vector representation that on image, coordinate is (i, j).
3. the method for detecting change of remote sensing image based on weighting Gabor wavelet character and two-stage cluster according to claim 1, it is characterized in that the design of the weights coefficient to extracted Gabor wavelet character described in step 4, be embodied as: the frequency parameter of different scale can obtain as k
v=k
max/ f
v; If v ∈ 0 ..., 4}, this corresponding frequency parameter is k
0, k
1, k
2, k
3, k
4, calculate weights coefficient and be followed successively by
and be summed to
Then be normalized and be followed successively by
different scale v ∈ 0 ..., the weights coefficient of 4} set gradually into
it is exactly lower directive weights coefficient of each yardstick that normalization weights coefficient is overturn according to yardstick order, and then obtains weighting Gabor wavelet character.
4. the method for detecting change of remote sensing image based on weighting Gabor wavelet character and two-stage cluster according to claim 1, is characterized in that described differential image X
d: for remote sensing image, differential image X
dfor input picture X
0and X
1the absolute value directly subtracting each other, X
d=| X
1-X
0|; For synthetic aperture radar (SAR) image, its differential image X
dget input picture X
0and X
1make the absolute value of logarithm ratio,
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410134520.1A CN103971364B (en) | 2014-04-04 | 2014-04-04 | Remote sensing image variation detecting method on basis of weighted Gabor wavelet characteristics and two-stage clusters |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410134520.1A CN103971364B (en) | 2014-04-04 | 2014-04-04 | Remote sensing image variation detecting method on basis of weighted Gabor wavelet characteristics and two-stage clusters |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103971364A true CN103971364A (en) | 2014-08-06 |
| CN103971364B CN103971364B (en) | 2017-02-01 |
Family
ID=51240813
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410134520.1A Expired - Fee Related CN103971364B (en) | 2014-04-04 | 2014-04-04 | Remote sensing image variation detecting method on basis of weighted Gabor wavelet characteristics and two-stage clusters |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103971364B (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104268879A (en) * | 2014-09-28 | 2015-01-07 | 民政部国家减灾中心 | Physical building quantity damage evaluation method based on remote sensing multi-spectral images |
| CN104700411A (en) * | 2015-03-15 | 2015-06-10 | 西安电子科技大学 | Sparse reconstruction-based dual-time phase remote-sensing image change detecting method |
| CN105118065A (en) * | 2015-09-14 | 2015-12-02 | 中国民航大学 | Polari SAR (metric synthetic aperture radar) image variation detection method of wavelet domain polarization distance transformation |
| CN104268875B (en) * | 2014-09-26 | 2017-02-15 | 西安电子科技大学 | SAR image registration method based on specific value correlation function |
| CN107368781A (en) * | 2017-06-09 | 2017-11-21 | 陕西师范大学 | Synthetic Aperture Radar images change detecting method based on Subspace partition |
| CN107886103A (en) * | 2016-09-29 | 2018-04-06 | 日本电气株式会社 | For identifying the method, apparatus and system of behavior pattern |
| CN110119782A (en) * | 2019-05-14 | 2019-08-13 | 西安电子科技大学 | SAR image change detection based on FPGA |
| CN111242864A (en) * | 2020-01-12 | 2020-06-05 | 杭州电子科技大学 | Finger vein image restoration method based on Gabor texture constraint |
| CN111369457A (en) * | 2020-02-28 | 2020-07-03 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Remote sensing image denoising method for sparse discrimination tensor robustness PCA |
| CN111814671A (en) * | 2020-07-08 | 2020-10-23 | 三亚中科遥感研究所 | Forest land dynamic change monitoring method based on remote sensing cooperation |
| CN112131969A (en) * | 2020-09-01 | 2020-12-25 | 河海大学 | Remote sensing image change detection method based on full convolution neural network |
| CN112131968A (en) * | 2020-09-01 | 2020-12-25 | 河海大学 | Double-time-phase remote sensing image change detection method based on DCNN |
| CN113537063A (en) * | 2021-07-16 | 2021-10-22 | 海南省林业科学研究院(海南省红树林研究院) | Forest environment remote sensing monitoring method and terminal based on comprehensive remote sensing technology |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101510310A (en) * | 2009-02-19 | 2009-08-19 | 上海交通大学 | Method for segmentation of high resolution remote sensing image based on veins clustering constrain |
| CN102629380A (en) * | 2012-03-03 | 2012-08-08 | 西安电子科技大学 | Remote sensing image change detection method based on multi-group filtering and dimension reduction |
| CN103218823A (en) * | 2013-05-08 | 2013-07-24 | 西安电子科技大学 | Remote sensing image change detection method based on nuclear transmission |
| CN103366365A (en) * | 2013-06-18 | 2013-10-23 | 西安电子科技大学 | SAR image varying detecting method based on artificial immunity multi-target clustering |
-
2014
- 2014-04-04 CN CN201410134520.1A patent/CN103971364B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101510310A (en) * | 2009-02-19 | 2009-08-19 | 上海交通大学 | Method for segmentation of high resolution remote sensing image based on veins clustering constrain |
| CN102629380A (en) * | 2012-03-03 | 2012-08-08 | 西安电子科技大学 | Remote sensing image change detection method based on multi-group filtering and dimension reduction |
| CN103218823A (en) * | 2013-05-08 | 2013-07-24 | 西安电子科技大学 | Remote sensing image change detection method based on nuclear transmission |
| CN103366365A (en) * | 2013-06-18 | 2013-10-23 | 西安电子科技大学 | SAR image varying detecting method based on artificial immunity multi-target clustering |
Non-Patent Citations (3)
| Title |
|---|
| LINLIN SHEN等: "A review on Gabor wavelets for face recognition", 《PATTERN ANAL APPLIC》 * |
| 李敏: "基于模糊C均值算法的遥感图像变化检测的研究", 《中国优秀硕士学位论文全文数据库(信息科技辑)》 * |
| 赵英男: "Gabor滤波器在车辆检测和车型识别中的应用研究", 《中国博士学位论文全文数据库 信息科技辑》 * |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104268875B (en) * | 2014-09-26 | 2017-02-15 | 西安电子科技大学 | SAR image registration method based on specific value correlation function |
| CN104268879A (en) * | 2014-09-28 | 2015-01-07 | 民政部国家减灾中心 | Physical building quantity damage evaluation method based on remote sensing multi-spectral images |
| CN104268879B (en) * | 2014-09-28 | 2017-03-01 | 民政部国家减灾中心 | Building quantity of goods produced Damage assessment method based on remote sensing multispectral image |
| CN104700411A (en) * | 2015-03-15 | 2015-06-10 | 西安电子科技大学 | Sparse reconstruction-based dual-time phase remote-sensing image change detecting method |
| CN105118065A (en) * | 2015-09-14 | 2015-12-02 | 中国民航大学 | Polari SAR (metric synthetic aperture radar) image variation detection method of wavelet domain polarization distance transformation |
| CN107886103B (en) * | 2016-09-29 | 2023-12-08 | 日本电气株式会社 | Method, device and system for identifying behavior patterns |
| CN107886103A (en) * | 2016-09-29 | 2018-04-06 | 日本电气株式会社 | For identifying the method, apparatus and system of behavior pattern |
| CN107368781B (en) * | 2017-06-09 | 2019-08-20 | 陕西师范大学 | Synthetic Aperture Radar images change detecting method based on Subspace partition |
| CN107368781A (en) * | 2017-06-09 | 2017-11-21 | 陕西师范大学 | Synthetic Aperture Radar images change detecting method based on Subspace partition |
| CN110119782A (en) * | 2019-05-14 | 2019-08-13 | 西安电子科技大学 | SAR image change detection based on FPGA |
| CN111242864A (en) * | 2020-01-12 | 2020-06-05 | 杭州电子科技大学 | Finger vein image restoration method based on Gabor texture constraint |
| CN111242864B (en) * | 2020-01-12 | 2023-05-05 | 杭州电子科技大学 | Finger vein image restoration method based on Gabor texture constraint |
| CN111369457A (en) * | 2020-02-28 | 2020-07-03 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Remote sensing image denoising method for sparse discrimination tensor robustness PCA |
| CN111369457B (en) * | 2020-02-28 | 2022-05-17 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Remote sensing image denoising method for sparse discrimination tensor robustness PCA |
| CN111814671A (en) * | 2020-07-08 | 2020-10-23 | 三亚中科遥感研究所 | Forest land dynamic change monitoring method based on remote sensing cooperation |
| CN112131969A (en) * | 2020-09-01 | 2020-12-25 | 河海大学 | Remote sensing image change detection method based on full convolution neural network |
| CN112131968A (en) * | 2020-09-01 | 2020-12-25 | 河海大学 | Double-time-phase remote sensing image change detection method based on DCNN |
| CN113537063A (en) * | 2021-07-16 | 2021-10-22 | 海南省林业科学研究院(海南省红树林研究院) | Forest environment remote sensing monitoring method and terminal based on comprehensive remote sensing technology |
| CN113537063B (en) * | 2021-07-16 | 2022-09-13 | 海南省林业科学研究院(海南省红树林研究院) | Forest environment remote sensing monitoring method and terminal based on comprehensive remote sensing technology |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103971364B (en) | 2017-02-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103971364B (en) | Remote sensing image variation detecting method on basis of weighted Gabor wavelet characteristics and two-stage clusters | |
| Zhan et al. | Log-based transformation feature learning for change detection in heterogeneous images | |
| Huang et al. | A new building extraction postprocessing framework for high-spatial-resolution remote-sensing imagery | |
| Chen et al. | A practical trial of landslide detection from single-temporal Landsat8 images using contour-based proposals and random forest: A case study of national Nepal | |
| Im et al. | A change detection model based on neighborhood correlation image analysis and decision tree classification | |
| Lu et al. | A survey of image classification methods and techniques for improving classification performance | |
| Nasiri et al. | Infrared small target enhancement based on variance difference | |
| Lu et al. | Land cover classification in a complex urban-rural landscape with QuickBird imagery | |
| CN103955926B (en) | Method for detecting change of remote sensing image based on Semi-NMF | |
| Khazai et al. | An approach for subpixel anomaly detection in hyperspectral images | |
| CN105844279A (en) | Depth learning and SIFT feature-based SAR image change detection method | |
| CN103236063B (en) | Based on the SAR image oil spilling detection method of multiple dimensioned spectral clustering and decision level fusion | |
| CN103578110B (en) | Multiband high-resolution remote sensing image dividing method based on gray level co-occurrence matrixes | |
| Jiang et al. | The potential of more accurate InSAR covariance matrix estimation for land cover mapping | |
| CN102982338B (en) | Classification of Polarimetric SAR Image method based on spectral clustering | |
| CN102819740B (en) | A kind of Single Infrared Image Frame Dim targets detection and localization method | |
| CN102629378B (en) | Remote sensing image change detection method based on multi-feature fusion | |
| CN103871039B (en) | Generation method for difference chart in SAR (Synthetic Aperture Radar) image change detection | |
| CN107403433A (en) | A kind of complicated cloud infrared small target in background detection method | |
| CN105718942A (en) | Hyperspectral image imbalance classification method based on mean value drifting and oversampling | |
| CN103824302B (en) | The SAR image change detection merged based on direction wave area image | |
| CN106156758B (en) | A kind of tidal saltmarsh method in SAR seashore image | |
| CN105160648A (en) | Radar target and shadow segmentation method based on wavelet and constant false alarm rate | |
| CN106295498A (en) | Remote sensing image target area detection apparatus and method | |
| CN107689051A (en) | A kind of multitemporal SAR image change detecting method based on changed factor |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170201 Termination date: 20200404 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |