CN117173584B - Land small micro water body extraction method and device for fusion of PolSAR and Pan images - Google Patents
Land small micro water body extraction method and device for fusion of PolSAR and Pan imagesInfo
- Publication number
- CN117173584B CN117173584B CN202310963054.7A CN202310963054A CN117173584B CN 117173584 B CN117173584 B CN 117173584B CN 202310963054 A CN202310963054 A CN 202310963054A CN 117173584 B CN117173584 B CN 117173584B
- Authority
- CN
- China
- Prior art keywords
- image
- polsar
- pan
- fusion
- land
- 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.)
- Active
Links
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 118
- 230000004927 fusion Effects 0.000 title claims abstract description 73
- 238000000605 extraction Methods 0.000 title claims abstract description 51
- 238000000034 method Methods 0.000 claims abstract description 32
- 230000011218 segmentation Effects 0.000 claims abstract description 21
- 230000000877 morphologic effect Effects 0.000 claims abstract description 13
- 239000011159 matrix material Substances 0.000 claims description 68
- 230000010287 polarization Effects 0.000 claims description 40
- 238000007781 pre-processing Methods 0.000 claims description 26
- 238000000354 decomposition reaction Methods 0.000 claims description 17
- 238000003809 water extraction Methods 0.000 claims description 12
- 238000001914 filtration Methods 0.000 claims description 11
- 230000001427 coherent effect Effects 0.000 claims description 9
- 238000010606 normalization Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 9
- 238000002347 injection Methods 0.000 claims description 7
- 239000007924 injection Substances 0.000 claims description 7
- 238000005070 sampling Methods 0.000 claims description 7
- 239000002352 surface water Substances 0.000 claims description 7
- 238000005388 cross polarization Methods 0.000 claims description 6
- 238000004422 calculation algorithm Methods 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 4
- 238000012805 post-processing Methods 0.000 claims description 4
- 230000002194 synthesizing effect Effects 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 230000001105 regulatory effect Effects 0.000 claims description 3
- 238000003384 imaging method Methods 0.000 abstract description 6
- 230000009286 beneficial effect Effects 0.000 abstract description 5
- 230000003287 optical effect Effects 0.000 description 7
- 238000012937 correction Methods 0.000 description 5
- 238000011160 research Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 210000000746 body region Anatomy 0.000 description 2
- 230000009977 dual effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000011835 investigation Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000000926 separation method Methods 0.000 description 2
- 239000013589 supplement Substances 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 235000002566 Capsicum Nutrition 0.000 description 1
- 239000006002 Pepper Substances 0.000 description 1
- 235000016761 Piper aduncum Nutrition 0.000 description 1
- 235000017804 Piper guineense Nutrition 0.000 description 1
- 244000203593 Piper nigrum Species 0.000 description 1
- 235000008184 Piper nigrum Nutrition 0.000 description 1
- 238000012952 Resampling Methods 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000002592 echocardiography Methods 0.000 description 1
- 230000003628 erosive effect Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 150000003839 salts Chemical class 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000003746 surface roughness Effects 0.000 description 1
- 238000012876 topography Methods 0.000 description 1
Abstract
The invention relates to a land micro water body extraction method and device for fusion of PolSAR and Pan images, comprising the following steps: the advantages of multi-source data are complemented based on an image fusion mode by utilizing the high-resolution space advantages of Pan and the backscattering characteristics of PolSAR and combining the space scale characteristics of small land micro-water bodies and a remote sensing imaging mechanism, so that the PolSAR land water body index after space enhancement is constructed; then, determining a segmentation threshold value of the water body and the non-water body through clustering, and finally, carrying out morphological post-treatment to realize fine drawing of the small micro water body. The beneficial effects of the invention are as follows: the method can remarkably improve the extraction precision of the land small micro water body, is simple to operate, easy to realize and high in treatment efficiency, and has important wide application prospect.
Description
Technical Field
The invention relates to the field of multisource remote sensing image fusion and ground object recognition, in particular to a land micro water body extraction method and device for fusion of PolSAR and Pan images.
Background
The small micro water body generally refers to grooves, channels, streams, pits and ponds and the like distributed on the land surface, and is characterized by small area, large quantity, poor fluidity and weak self-cleaning capability, and the fine drawing of the small micro water body has important significance. The extraction difficulty of small micro water bodies is high, natural conditions in various places are severe, or investigation staff are hard to reach, such as swamps, wetlands and the like. The remote sensing means is known to have the advantages of large imaging range, short revisit period and the like, and the acquisition of the underlying surface information is less limited by natural conditions. The field mapping period is longer, and the remote sensing imaging can be performed once every month or even a few days, so that the remote sensing technology becomes a key means for extracting the water body gradually.
The water extraction based on the remote sensing technology is usually performed by using only a single optical image, such as normalized water index, surface water index, spectral information or spatial texture features, and common classification methods include expert interpretation, image clustering, object-oriented classification, decision tree, and the like. The data source commonly used for research is a medium-resolution optical remote sensing image, and is usually Landsat and Sentinel series data; besides optical remote sensing images, the satellite-borne Synthetic Aperture Radar (SAR) imaging process is not affected by weather and meteorological conditions, and is increasingly applied to remote sensing identification of complex water bodies due to sensitivity of the SAR imaging process to dielectric constants and surface roughness, and is also an important technical means for water body extraction and water condition monitoring at present. SAR is used as an active microwave detector, a backward scattering signal of a ground object is acquired for imaging, a water body in an SAR image generally presents mirror image scattering characteristics, a backward scattering coefficient received by a receiver is low, and the SAR image is darker black. However, the high resolution remote sensing data is more beneficial to improving the classification accuracy and can realize further extraction of small micro water bodies, so that the advantages of full-color images are more obvious.
The full-color image is a collective name of black-and-white images with the whole visible light range of 0.4-0.7 mu m obtained by a remote sensing technology, the incident energy is large, and the image spatial resolution is relatively high. The full-polarization synthetic aperture radar (PolSAR) image has four polarization modes of VV, VH, HH and HV, contains abundant polarization scattering information, and can more fully reflect the difference of backscattering coefficients of water and other ground objects. At present, the commonly used data sources in the small micro water body extraction research are mid-resolution images such as multispectral data, dual-polarization SAR data and the like, and high-resolution Pan data or full-polarization SAR data are rarely utilized. The optical image is larger by the weather image, and the land water body and the ocean are harder to distinguish on the multispectral image. The SAR data alone cannot well distinguish water boundaries or other aquatic vegetation, the space detail expression capability of the optical image or the dual-polarized SAR image with medium resolution is poor, the false separation and the wrong separation are easy to cause, and the existing land water extraction method cannot accurately reflect the real space distribution condition of small micro water.
Disclosure of Invention
The invention aims at overcoming the defects of the prior art, and provides a land micro water body extraction method and device for fusion of PolSAR and Pan images.
In a first aspect, a land micro-water extraction method for fusion of a PolSAR and a Pan image is provided, which comprises the following steps:
step 1, acquiring a PolSAR image and preprocessing;
step 2, calculating a polarization coherence matrix and a covariance matrix of the preprocessed PolSAR image, and then carrying out incoherent Freeman-Durden polarization decomposition on the PolSAR to obtain a mirror image single backward scattering component;
Step 3, acquiring a Pan image, preprocessing, and performing geometric registration operation on the PolSAR image and the Pan image;
step 4, performing heterogeneous image fusion on the PolSAR image and the high-resolution Pan image subjected to geometric registration in a mode of space detail injection and back scattering information compensation to obtain a high-spatial-resolution PolSAR and Pan fusion image;
Step 5, denoising a sliding window with a certain scale on the PolSAR and Pan fusion image according to actual conditions;
step 6, constructing a high-resolution PolSAR land water index PLSWI based on the PolSAR and Pan fusion image;
Step 7, performing image clustering on the calculated PolSAR land water index PLSWI images, determining a clustering center, and presetting a segmentation threshold value according to the clustering center to obtain a land water primary extraction result;
And 8, carrying out morphological post-treatment on the preliminary land water body extraction result to obtain an accurate land micro water body extraction result.
Preferably, step 2 includes:
Step 2.1, performing polarization denoising operation on the PolSAR image matrix T 1 through polarization REFINED LEE filtering to obtain T 2;
Step 2.2, processing T 2 components among poling modes of the PolSAR images to obtain a plurality of coherence matrixes T 3 after polarization decomposition;
step 2.3, decomposing the denoised polarized PolSAR wave band by using an incoherent Freeman-Durden algorithm:
C＝Cs+Cd+Cv
Under the condition that three components are mutually incoherent, the Freeman-Durden polarization decomposition algorithm assumes that the total backscattering model is C; wherein, C s、Cd and C v are a single scattering covariance matrix, a secondary scattering covariance matrix and a bulk scattering average covariance matrix respectively; f s、fd and f v are the single, secondary and bulk scatter components, respectively; alpha is the surface related parameter of the plane producing the single scattering and beta is the ratio related parameter of the complex dielectric constants of the two planes producing the secondary scattering.
Preferably, step4 includes:
Step 4.1, performing image normalization operation on the preprocessed PolSAR image and the Pan image, wherein the image normalization operation with the range of 0 to 1 is expressed as follows:
Wherein A n represents the nth pixel, and the image matrix has N pixels; a min is the minimum value in the image pixel, a max is the maximum value in the image pixel, The normalized result image is completed;
step 4.2, carrying out histogram matching on the Pan image by taking the PolSAR image as a reference image; the standard deviation σ A of all pixels of the Pan image is expressed as:
Wherein, Representing the nth pixel element of the picture,Representing the average value of pixels after normalization of the Pan image matrix, wherein the Pan image matrix has N pixels;
step 4.3, injecting the spatial details of the high-resolution Pan image into the PolSAR image after up-sampling to obtain a fusion image;
And 4.4, obtaining a compensation matrix of the backscattering coefficient by calculating the difference between the original low-resolution PolSAR image and the fusion image, and then carrying out compensation operation of backscattering information on the primary fusion result.
Preferably, in step 4.2, the histogram is matchedThe image matrix is obtained by the following formula:
Wherein sigma A is the standard deviation of all pixels of the Pan image, sigma S is the standard deviation of all pixels of the PolSAR image, And (5) representing the average value of the normalized PolSAR image matrix.
Preferably, in step 4.3, the spatial information and the texture details are obtained by subtracting the high-frequency components of the high-resolution Pan image and the PolSAR image, and the high-frequency components of the PolSAR image are obtained by spatial filtering or band linear combination, and the above process is described as follows:
Wherein Y is a fusion result image, f up (S) is a PolSAR image up-sampling operation, I h is a high-frequency space component obtained by the PolSAR image through space filtering or wave band linear combination, And (5) representing the high-resolution Pan image after histogram matching, and w representing the fusion weight coefficient.
Preferably, in step 4.4, the information compensation process is expressed as:
Sy＝Y+fup[S-fdown(fGB(Y))]
Wherein, f GB (·) and f down (·) respectively represent gaussian blur and downsampling operations, f up is an upsampling operation; s is a low spatial resolution PolSAR image, and S y represents the final fusion image.
Preferably, step 6 includes:
step 6.1, summing homopolar VV and HH wave bands;
Step 6.2, calculating a coherence matrix and a covariance matrix to characterize the coherence phenomenon of the backward scattering signal; summing a plurality of coherence matrixes obtained in the step 2.2 and amplifying coherence matrix components with smaller logarithmic values according to a coherence matrix solving principle The above calculation process is expressed as:
Wherein G r is the r-th coherence matrix, r ε [1, R ]; g is the sum of all the coherent matrix bands, Is a regulatory factor;
step 6.3, representing whether the underlying surface is a calm land water body or not by calculating (VH-HV) values, namely, the difference value of cross polarization wave bands;
and 6.4, synthesizing main polarization characteristics of the PolSAR image after spatial enhancement, and constructing a PolSAR land surface water index PLSWI according to the formula:
Wherein G is the sum of each coherent matrix band; VV, HH, VH and HV are four polarization bands of the PolSAR image; c s is the mirror image single scattering component after polarization decomposition; PLSWI is the PolSAR land water index.
Preferably, step 7 includes:
Step 7.1, defining an image domain S= { x 1,…,xn,…xN } to contain N pixel objects, wherein each object takes the index value calculated in step 6.4 as a distinguishing attribute; the distance between any two pixels in the image domain S is denoted as dt (x m,xn):
step 7.2, calculating a distance average value of each pixel from the segmentation center position k, and iteratively solving an optimal center, namely a segmentation point:
wherein t is the iteration number, Is an initialization segmentation center, and the I is L 2 normal form; dt is a preset distance;
step 7.3, obtaining a segmentation center according to the iteration result The value of PLSWI of (2) is scaled appropriately to determine a maximum threshold v max, and if the value is smaller than the threshold, the water body area is determined, otherwise the water body area is not, and the process is expressed as:
0≤RLSWR(i,j)≤vmax
Wherein v max is the upper threshold of the water body, RLSWR (i, j) is the RLSWR value of the position (i, j); if RLSWR (i, j) at the position (i, j) meets the above formula, determining that the position (i, j) is a water body area, and executing morphological processing in the step 8; if RLSWR (i, j) at position (i, j) does not satisfy the above equation, it is determined that the position (i, j) is a non-water region.
In a second aspect, a land micro-water extraction system for fusion of a PolSAR and a Pan image is provided, and the land micro-water extraction method for fusion of a PolSAR and a Pan image in any one of the first aspect is performed, and includes:
The first preprocessing module is used for acquiring a PolSAR image and preprocessing the PolSAR image;
the computing module is used for computing a polarization coherence matrix and a covariance matrix of the preprocessed PolSAR image, and then carrying out incoherent Freeman-Durden polarization decomposition on the PolSAR to obtain a mirror image single backward scattering component;
The second preprocessing module is used for acquiring the Pan image, preprocessing the Pan image and performing geometric registration operation on the PolSAR image and the Pan image;
The fusion module is used for carrying out heterogeneous image fusion on the geometrically registered PolSAR image and the high-resolution Pan image in a mode of space detail injection and back scattering information compensation to obtain a high-spatial-resolution PolSAR and Pan fusion image;
The denoising module is used for denoising the PolSAR and Pan fusion image by a sliding window with a certain scale according to actual conditions;
the construction module is used for constructing a high-resolution PolSAR land water index PLSWI based on the PolSAR and Pan fusion image;
The extraction module is used for carrying out image clustering on the calculated PolSAR land water index PLSWI images, determining a clustering center and presetting a segmentation threshold value according to the clustering center to obtain a land water primary extraction result;
and the processing module is used for carrying out morphological post-processing on the preliminary land water body extraction result to obtain an accurate land micro water body extraction result.
In a third aspect, a computer storage medium having a computer program stored therein is provided; when the computer program runs on a computer, the computer executes the land micro water body extraction method for fusing the PolSAR and the Pan image according to any one of the first aspect.
The beneficial effects of the invention are as follows: according to the invention, the PolSAR image and the high-resolution Pan image are fused, the difference of the land water index increase water body and the non-water body area is constructed, then image clustering is carried out, the binary threshold value of the water body rain and the non-water body is obtained, and finally the false sub-pixel generated by SAR image speckle noise is removed by combining morphological post-processing, so that the fine extraction of the land micro water body is realized. The method and the device are easier to realize and have higher extraction precision, are beneficial supplements to the existing water body extraction method, and can better remove the sea surface water body, thus having important practical application significance.
Drawings
FIG. 1 is a flow chart of a method for extracting a land micro water body by fusing PolSAR and Pan images;
FIG. 2 is a diagram showing the comparison results before and after image fusion;
FIG. 3 is a graph showing the comparison of the water extraction effect of using a medium resolution dual polarized SAR image with an optical image alone;
fig. 4 is a schematic diagram of the extraction result of land micro water.
Detailed Description
The invention is further described below with reference to examples. The following examples are presented only to aid in the understanding of the invention. It should be noted that it will be apparent to those skilled in the art that modifications can be made to the present invention without departing from the principles of the invention, and such modifications and adaptations are intended to be within the scope of the invention as defined in the following claims.
Example 1:
Aiming at the problems of lower drawing resolution and marine water body error and the like in small micro water body extraction research, the invention provides a land small micro water body extraction method based on PolSAR and Pan image fusion. As an embodiment, the full-color image of Radarsat-2 full-polarization SAR data and resource one 02d star is taken as a data source, the coastal wetland of the south shore of the hangzhou is taken as a research area, and the detailed flow is shown in fig. 1:
And step 1, acquiring a PolSAR image and preprocessing.
The preprocessing of the PolSAR image comprises radiometric calibration, thermal noise removal, polarized Lee filtering denoising, doppler topography correction and decibelization. For example, preprocessing such as thermal noise removal, lee filtering salt and pepper noise denoising with a 5×5 window, radiometric calibration, doppler terrain correction, decibelization and the like is performed on Radarsat-2 full-polarization SAR data, wherein the terrain correction is applied to SRTM 1Sec data with 30 meters of spatial resolution, and the resampling grid size after the Radarsat-2SAR image preprocessing is 5 meters of spatial resolution.
And step 2, calculating a polarization coherence matrix and a covariance matrix of the preprocessed PolSAR image, and then carrying out incoherent Freeman-Durden polarization decomposition on the PolSAR to obtain a mirror image single backward scattering component.
The step 2 comprises the following steps:
And 2.1, performing polarization denoising operation on the PolSAR coherent matrix T 1 through polarization REFINED LEE filtering to obtain T 2.
And 2.2, processing the PolSAR image T 2 component to obtain a plurality of coherence matrixes T 3 of each polarization mode.
Step 2.3, performing polarization characteristic decomposition by a Freeman-Durden Decomposition method:
C＝Cs+Cd+Cv
wherein, C s、Cd and C v are a single scattering covariance matrix, a secondary scattering covariance matrix and a bulk scattering average covariance matrix respectively; f s、fd and f v are the single, secondary and bulk scatter components, respectively; alpha is the surface related parameter of the plane producing the single scattering and beta is the ratio related parameter of the complex dielectric constants of the two planes producing the secondary scattering. Wherein, the Freeman-Durden polarization decomposition algorithm assumes that the three components are mutually incoherent, and the total backscattering model is C.
And step 3, acquiring a Pan image, preprocessing, and performing geometric registration operation on the PolSAR image and the Pan image.
The pre-processing of the Pan image comprises the following steps: orthographic correction and radiometric calibration. For example, preprocessing such as orthographic correction, radiometric calibration and the like is carried out on the ZY1-02d full-color image, and the spatial resolution after preprocessing is 2.5 meters; performing geometric registration operation on the PolSAR and the Pan image, wherein the actual registration Error RMS error=0.23.
And 4, performing heterogeneous image fusion on the PolSAR image and the high-resolution Pan image subjected to geometric registration in a mode of space detail injection and back scattering information compensation to obtain a high-spatial-resolution PolSAR and Pan fusion image.
Specifically, a multiple wavelet (ATWT) method is adopted to perform image fusion on the Radarsat-2 full-polarization SAR and the ZY1-02d Pan image, and the fused image is subjected to back scattering signal compensation to obtain a high-resolution PolSAR image.
Step 4 comprises:
Step 4.1, performing image normalization operation on the preprocessed Radarsat-2 full polarization SAR and ZY1-02d Pan images, wherein the image normalization operation with the range of 0 to 1 is expressed as follows:
Wherein A n represents the nth pixel, and the image matrix has N pixels; a min is the minimum value in the image pixel, a max is the maximum value in the image pixel, The normalized result image is completed;
step 4.2, carrying out histogram matching on the Pan image by taking the PolSAR image as a reference image; the standard deviation σ A of all pixels of the Pan image is expressed as:
Wherein, Representing the nth pixel element of the picture,Representing the average value of pixels after normalization of the Pan image matrix, wherein the Pan image matrix has N pixels; after histogram matchingThe image matrix is obtained by the following formula:
Wherein sigma A is the standard deviation of all pixels of the Pan image, sigma S is the standard deviation of all pixels of the PolSAR image, And (5) representing the average value of the normalized PolSAR image matrix.
Step 4.3, injecting the spatial details of the high-resolution Pan image into the PolSAR image after up-sampling to obtain a fusion image;
The spatial information and the texture detail are obtained by taking the difference between the high-frequency components of the high-resolution Pan image and the PolSAR image, the high-frequency components of the PolSAR image are obtained by taking the difference of the high-frequency information of the two images through multiple wavelets (ATWT), namely, the spatial detail of the up-sampling Radarsat-2 full-polarization SAR image is injected. The fusion over for the ATWT-based multi-resolution analysis can be expressed as:
Wherein Y is a fusion result image, f up (S) is a PolSAR image up-sampling operation, I h is a high-frequency space component obtained by the PolSAR image through space filtering or wave band linear combination, And (5) representing the high-resolution Pan image after histogram matching, and w representing the fusion weight coefficient.
And 4.4, obtaining a compensation matrix of the backscattering coefficient by calculating the difference between the original low-resolution PolSAR image and the fusion image, and then carrying out compensation operation of backscattering information on the primary fusion result.
The step can ensure the robustness of the basic properties of SAR images when the obtained fusion images are constructed in an index, and the information compensation process is expressed as follows:
Sy＝Y+fup[S-fdown(fGB(Y))]
Wherein, f GB (·) and f down (·) respectively represent gaussian blur and downsampling operations, f up is an upsampling operation; s is a low spatial resolution PolSAR image, sy represents the final fusion image.
And 5, carrying out Gaussian filtering denoising with the sliding window size of 5 multiplied by 5 on the fused high-resolution Radarsat-2 image.
And 6, constructing a high-resolution PolSAR land water index PLSWI based on the PolSAR and Pan fusion image.
And 7, performing image clustering on the calculated PolSAR land water index PLSWI images, determining a clustering center, and presetting a segmentation threshold value according to the clustering center to obtain a land water primary extraction result.
And 8, carrying out morphological post-treatment on the preliminary land water body extraction result to obtain an accurate land micro water body extraction result.
Example 2:
on the basis of the embodiment 1, the embodiment 2 of the application provides a more specific land micro-water extraction method for fusion of PolSAR and Pan images, which comprises the following steps:
And step 1, acquiring a PolSAR image and preprocessing.
And step 2, calculating a polarization coherence matrix and a covariance matrix of the preprocessed PolSAR image, and then carrying out incoherent Freeman-Durden polarization decomposition on the PolSAR to obtain a mirror image single backward scattering component.
Step 3, acquiring a Pan image, preprocessing, and performing geometric registration operation on the PolSAR image and the Pan image;
and 4, performing heterogeneous image fusion on the PolSAR image and the high-resolution Pan image subjected to geometric registration in a mode of space detail injection and back scattering information compensation to obtain a high-spatial-resolution PolSAR and Pan fusion image.
And 5, denoising the PolSAR and Pan fusion image by a sliding window of a certain scale according to the actual situation.
The actual situation refers to the relevant situation of project application or specific images.
And 6, constructing a high-resolution PolSAR land water index PLSWI based on the PolSAR and Pan fusion image.
The step 6 comprises the following steps:
step 6.1, summing homopolar VV and HH bands.
It should be noted that homopolar VV and HH bands are the most obvious ways to distinguish mirror scattering and other scattering, and the backscattering coefficient of the water body in the homopolar band is far smaller than that of other land features, so that the boundary of the water body can be clearly depicted. Therefore, the homopolar band summation (vv+hh) can significantly increase the difference, so that the water body region pixel value is smaller, and the non-water body region pixel value is larger.
And 6.2, calculating a coherence matrix and a covariance matrix to effectively represent the coherence phenomenon of the back scattering signal, wherein the water body part, particularly the continuous water surface back scattering signal, has weak echoes and relatively consistent directions, and the coherence matrix is a smooth area with relatively low values. Summing a plurality of coherence matrixes obtained in the step 2 and amplifying coherence matrix components with smaller values according to a coherence matrix solving principleThe above calculation process can be expressed as:
Wherein G r is the r-th coherence matrix, r ε [1, R ]; g is the sum of all the coherent matrix bands, Is a regulatory factor.
For example, the 9 coherent matrices obtained in step 2 are summed and amplified by 3 times according to the T3 matrix solving principle of SNAP software, where the value of the coherent matrix is smaller, and the above process can be expressed as:
And 6.3, representing whether the underlying surface is a calm land surface water body or not by calculating a (VH-HV) value of the high-resolution Radarsat-2 image, namely a difference value of cross polarization bands.
In step 6.3, the land surface calm water body on the PolSAR image is represented in a dark area with a cross polarization wave band, the difference is small, and other land surface vegetation, buildings, sea surfaces with stormy waves and other ground features have large differences in the cross polarization wave band, and whether the underlying surface is calm land surface water body can be well represented by calculating (VH-HV) values, namely, the difference of the cross polarization wave bands.
And 6.4, synthesizing main polarization characteristics of the PolSAR image after spatial enhancement, and constructing a PolSAR land surface water index PLSWI according to the formula:
Wherein G is the sum of each coherent matrix band; VV, HH, VH and HV are four polarization bands of the PolSAR image; c s is the mirror image single scattering component after polarization decomposition; PLSWI is the PolSAR land water index.
And 7, carrying out peak Density (DP) clustering on the calculated PolSAR land water index PLSWI image, determining a clustering center, and presetting a segmentation threshold value according to the clustering center to obtain a land water primary extraction result.
In step 7, binary segmentation is carried out on the calculated PolSAR land water index PLSWI image in an image clustering mode, so as to obtain water and non-water. The step 7 comprises the following steps:
Step 7.1, defining an image domain S= { x 1,…,xn,…xN } to contain N pixel objects, wherein each object takes the index value calculated in step 6.4 as a distinguishing attribute; the distance between any two pixels in the image domain S is denoted as dt (x m,xn):
step 7.2, calculating a distance average value of each pixel from the segmentation center position k, and iteratively solving an optimal center, namely a segmentation point:
wherein t is the iteration number, Is an initialization segmentation center, and the I is L 2 normal form; dt is a preset distance;
step 7.3, obtaining a segmentation center according to the iteration result The value of PLSWI of (2) is scaled appropriately to determine a maximum threshold v max, and if the value is smaller than the threshold, the water body area is determined, otherwise the water body area is not, and the process is expressed as:
0≤RLSWR(i,j)≤vmax
Wherein v max is the upper threshold of the water body, RLSWR (i, j) is the RLSWR value of the position (i, j); if RLSWR (i, j) at the position (i, j) meets the above formula, determining that the position (i, j) is a water body area, and executing morphological processing in the step 8; if RLSWR (i, j) at position (i, j) does not satisfy the above equation, it is determined that the position (i, j) is a non-water region.
And 8, performing morphological post-treatment on the land water body preliminary extraction result, namely the segmented water body part, so as to obtain an accurate land micro water body extraction result.
Specifically, mathematical morphological operators (corrosion and expansion) are used, the window size is set to be 3 multiplied by 3, adjacent similar classification areas are clustered and combined, fine plaque is removed, and a final land micro water body extraction result diagram is obtained. Wherein, the expansion (Majorit) replaces the category of the center pixel with the category of the pixel which takes the most part (the most number of pixels) in the transformation kernel; erosion (Minority) will replace the class of center pels with the class of the secondary pels in the transformation kernel.
As a result of an embodiment, fig. 2 is a comparison result before and after image fusion according to the present invention, and the left image (a) is a PolSAR image; right panel (b) is a fusion image of Radarsat-2 full polarization SAR and resource one 02d star. Fig. 3 is a graph comparing the water extraction effect of using a medium resolution dual polarized SAR image with that of using an optical image alone. Fig. 4 shows the extraction result of the land micro water body in this embodiment, and black represents the extraction result of the land micro water body, and the land micro water body in the south of the Hangzhou bay is accurately extracted by the method provided by the invention, the extraction precision is 94.7%, and the actual investigation and monitoring requirements are met. The extraction result of the method and the device eliminates most of ocean water bodies (upper right corner areas), and the space distribution conditions of ditches, ponds, rivers and tributaries thereof are more finely described, so that the method and the device are beneficial to monitoring and controlling small and micro water bodies on land.
The land small micro water body extraction method based on the PolSAR and Pan fusion image is a method and a device for synthesizing heterogeneous fusion images and remote sensing indexes, effectively supplements the existing land small micro water body extraction method, particularly in a coastal area, and can effectively remove ocean water bodies with stormy waves on the premise of not using vector masks; according to the method and the device, the remote sensing indexes with complementary advantages of multi-source data are constructed according to the characteristics of small spatial scale of the small micro water body and the like by combining the high spatial resolution Pan and PolSAR backscattering characteristics, and then the segmentation threshold is automatically extracted through clustering, so that the extraction precision of the land small micro water body is remarkably improved.
In this embodiment, the same or similar parts as those in embodiment 1 may be referred to each other, and will not be described in detail in the present disclosure.
Example 3:
On the basis of embodiment 1, embodiment 3 of the present invention provides a land micro-water extraction system for fusion of a PolSAR and Pan images, comprising:
The first preprocessing module is used for acquiring a PolSAR image and preprocessing the PolSAR image;
the computing module is used for computing a polarization coherence matrix and a covariance matrix of the preprocessed PolSAR image, and then carrying out incoherent Freeman-Durden polarization decomposition on the PolSAR to obtain a mirror image single backward scattering component;
The second preprocessing module is used for acquiring the Pan image, preprocessing the Pan image and performing geometric registration operation on the PolSAR image and the Pan image;
The fusion module is used for carrying out heterogeneous image fusion on the geometrically registered PolSAR image and the high-resolution Pan image in a mode of space detail injection and back scattering information compensation to obtain a high-spatial-resolution PolSAR and Pan fusion image;
The denoising module is used for denoising the PolSAR and Pan fusion image by a sliding window with a certain scale according to actual conditions;
the construction module is used for constructing a high-resolution PolSAR land water index PLSWI based on the PolSAR and Pan fusion image;
The extraction module is used for carrying out image clustering on the calculated PolSAR land water index PLSWI images, determining a clustering center and presetting a segmentation threshold value according to the clustering center to obtain a land water primary extraction result;
and the processing module is used for carrying out morphological post-processing on the preliminary land water body extraction result to obtain an accurate land micro water body extraction result.
Specifically, the system provided in this embodiment is a system corresponding to the method provided in embodiment 1, so that the portions in this embodiment that are the same as or similar to those in embodiment 1 may be referred to each other, and will not be described in detail in this disclosure.
Claims (5)
1. The land micro water body extraction method for fusion of PolSAR and Pan images is characterized by comprising the following steps:
step 1, acquiring a PolSAR image and preprocessing;
step 2, calculating a polarization coherence matrix and a covariance matrix of the preprocessed PolSAR image, and then carrying out incoherent Freeman-Durden polarization decomposition on the PolSAR to obtain a mirror image single backward scattering component;
Step 3, acquiring a Pan image, preprocessing, and performing geometric registration operation on the PolSAR image and the Pan image;
step 4, performing heterogeneous image fusion on the PolSAR image and the high-resolution Pan image subjected to geometric registration in a mode of space detail injection and back scattering information compensation to obtain a high-spatial-resolution PolSAR and Pan fusion image;
step 4 comprises:
Step 4.1, performing image normalization operation on the preprocessed PolSAR image and the Pan image, wherein the image normalization operation with the range of 0 to 1 is expressed as follows:
Wherein A n represents the nth pixel, and the image matrix has N pixels; a min is the minimum value in the image pixel, a max is the maximum value in the image pixel, The normalized result image is completed;
step 4.2, carrying out histogram matching on the Pan image by taking the PolSAR image as a reference image; the standard deviation σ A of all pixels of the Pan image is expressed as:
Wherein, Representing the nth pixel element of the picture,Representing the average value of pixels after normalization of the Pan image matrix, wherein the Pan image matrix has N pixels;
in step 4.2, the histogram is matched The image matrix is obtained by the following formula:
Wherein sigma A is the standard deviation of all pixels of the Pan image, sigma S is the standard deviation of all pixels of the PolSAR image, Representing the average value of the normalized PolSAR image matrix;
step 4.3, injecting the spatial details of the high-resolution Pan image into the PolSAR image after up-sampling to obtain a fusion image;
In step 4.3, the spatial information and the texture details are obtained by subtracting the high-frequency components of the high-resolution Pan image and the PolSAR image, and the high-frequency components of the PolSAR image are obtained by spatial filtering or band linear combination, and the above process is described as follows:
Wherein Y is a fusion result image, f up (S) is a PolSAR image up-sampling operation, I h is a high-frequency space component obtained by the PolSAR image through space filtering or wave band linear combination, Representing the high-resolution Pan image after histogram matching, and w represents a fusion weight coefficient;
Step 4.4, obtaining a compensation matrix of the backscattering coefficient by calculating the difference between the original low-resolution PolSAR image and the fusion image, and then carrying out compensation operation of backscattering information on the primary fusion result;
In step 4.4, the information compensation process is expressed as:
Sy＝Y+fup[S-fdown(fGB(Y))]
Wherein, f GB (·) and f down (·) respectively represent gaussian blur and downsampling operations, f up is an upsampling operation; s is a PolSAR image with low spatial resolution, and S y represents a final fusion image;
Step 5, denoising a sliding window with a certain scale on the PolSAR and Pan fusion image according to actual conditions;
step 6, constructing a high-resolution PolSAR land water index PLSWI based on the PolSAR and Pan fusion image;
The step 6 comprises the following steps:
step 6.1, summing homopolar VV and HH wave bands;
Step 6.2, calculating a coherence matrix and a covariance matrix to characterize the coherence phenomenon of the backward scattering signal; summing a plurality of coherence matrixes obtained in the step 2.2 and amplifying coherence matrix components with smaller logarithmic values according to a coherence matrix solving principle The above calculation process is expressed as:
Wherein G r is the r-th coherence matrix, r ε [1, R ]; g is the sum of all the coherent matrix bands, Is a regulatory factor;
step 6.3, representing whether the underlying surface is a calm land water body or not by calculating (VH-HV) values, namely, the difference value of cross polarization wave bands;
and 6.4, synthesizing main polarization characteristics of the PolSAR image after spatial enhancement, and constructing a PolSAR land surface water index PLSWI according to the formula:
wherein G is the sum of each coherent matrix band; VV, HH, VH and HV are four polarization bands of the PolSAR image; c s is the mirror image single scattering component after polarization decomposition; PLSWI is the PolSAR land water index;
Step 7, performing image clustering on the calculated PolSAR land water index PLSWI images, determining a clustering center, and presetting a segmentation threshold value according to the clustering center to obtain a land water primary extraction result;
And 8, carrying out morphological post-treatment on the preliminary land water body extraction result to obtain an accurate land micro water body extraction result.
2. The land micro water extraction method of PolSAR and Pan image fusion according to claim 1, wherein step 2 comprises:
Step 2.1, performing polarization denoising operation on the PolSAR image matrix T 1 through polarization REFINED LEE filtering to obtain T 2;
Step 2.2, processing T 2 components among poling modes of the PolSAR images to obtain a plurality of coherence matrixes T 3 after polarization decomposition;
step 2.3, decomposing the denoised polarized PolSAR wave band by using an incoherent Freeman-Durden algorithm:
C＝Cs+Cd+Cv
Under the condition that three components are mutually incoherent, the Freeman-Durden polarization decomposition algorithm assumes that the total backscattering model is C; wherein, C s、Cd and C v are a single scattering covariance matrix, a secondary scattering covariance matrix and a bulk scattering average covariance matrix respectively; f s、fd and f v are the single, secondary and bulk scatter components, respectively; alpha is the surface related parameter of the plane producing the single scattering and beta is the ratio related parameter of the complex dielectric constants of the two planes producing the secondary scattering.
3. The land micro-water extraction method of PolSAR and Pan image fusion of claim 2, wherein step 7 comprises:
Step 7.1, defining an image domain S= { x 1,…,xn,…xN } to contain N pixel objects, wherein each object takes the index value calculated in step 6.4 as a distinguishing attribute; the distance between any two pixels in the image domain S is denoted as dt (x m,xn):
step 7.2, calculating a distance average value of each pixel from the segmentation center position k, and iteratively solving an optimal center, namely a segmentation point:
wherein t is the iteration number, Is the initialization partition center, the/is L 2 normal form; dt is a preset distance;
step 7.3, obtaining a segmentation center according to the iteration result The value of PLSWI of (2) is scaled appropriately to determine a maximum threshold v max, and if the value is smaller than the threshold, the water body area is determined, otherwise the water body area is not, and the process is expressed as:
0≤RLSWR(i,j)≤vmax
wherein v max is the upper threshold of the water body, RLSWR (i, j) is the RLSER value of the position (i, j); if RLSWR (i, j) at the position (i, j) meets the above formula, determining that the position (i, j) is a water body area, and executing morphological processing in the step 8; if RLSWR (i, j) at position (i, j) does not satisfy the above equation, it is determined that the position (i, j) is a non-water region.
4. The land micro-water extraction system for fusion of PolSAR and Pan images is characterized by comprising the following components:
The first preprocessing module is used for acquiring a PolSAR image and preprocessing the PolSAR image;
the computing module is used for computing a polarization coherence matrix and a covariance matrix of the preprocessed PolSAR image, and then carrying out incoherent Freeman-Durden polarization decomposition on the PolSAR to obtain a mirror image single backward scattering component;
The second preprocessing module is used for acquiring the Pan image, preprocessing the Pan image and performing geometric registration operation on the PolSAR image and the Pan image;
The fusion module is used for carrying out heterogeneous image fusion on the geometrically registered PolSAR image and the high-resolution Pan image in a mode of space detail injection and back scattering information compensation to obtain a high-spatial-resolution PolSAR and Pan fusion image;
The denoising module is used for denoising the PolSAR and Pan fusion image by a sliding window with a certain scale according to actual conditions;
the construction module is used for constructing a high-resolution PolSAR land water index PLSWI based on the PolSAR and Pan fusion image;
The extraction module is used for carrying out image clustering on the calculated PolSAR land water index PLSWI images, determining a clustering center and presetting a segmentation threshold value according to the clustering center to obtain a land water primary extraction result;
and the processing module is used for carrying out morphological post-processing on the preliminary land water body extraction result to obtain an accurate land micro water body extraction result.
5. A computer storage medium, wherein a computer program is stored in the computer storage medium; when the computer program runs on a computer, the computer is caused to execute the land micro water body extraction method for fusing the PolSAR and the Pan images according to any one of claims 1 to 3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310963054.7A CN117173584B (en) | 2023-08-02 | Land small micro water body extraction method and device for fusion of PolSAR and Pan images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310963054.7A CN117173584B (en) | 2023-08-02 | Land small micro water body extraction method and device for fusion of PolSAR and Pan images |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117173584A CN117173584A (en) | 2023-12-05 |
| CN117173584B true CN117173584B (en) | 2024-07-02 |
Family
ID=
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108446716A (en) * | 2018-02-07 | 2018-08-24 | 武汉大学 | Based on FCN the PolSAR image classification methods merged are indicated with sparse-low-rank subspace |
| AU2020100179A4 (en) * | 2020-02-04 | 2020-03-19 | Huang, Shuying DR | Optimization Details-Based Injection Model for Remote Sensing Image Fusion |
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108446716A (en) * | 2018-02-07 | 2018-08-24 | 武汉大学 | Based on FCN the PolSAR image classification methods merged are indicated with sparse-low-rank subspace |
| AU2020100179A4 (en) * | 2020-02-04 | 2020-03-19 | Huang, Shuying DR | Optimization Details-Based Injection Model for Remote Sensing Image Fusion |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kulkarni et al. | Pixel level fusion techniques for SAR and optical images: A review | |
| Alonso et al. | Edge enhancement algorithm based on the wavelet transform for automatic edge detection in SAR images | |
| Zhang et al. | Mapping flood by the object-based method using backscattering coefficient and interference coherence of Sentinel-1 time series | |
| Ferrentino et al. | On the use of multipolarization satellite SAR data for coastline extraction in harsh coastal environments: The case of Solway Firth | |
| CN111046772A (en) | Multi-temporal satellite remote sensing island shore line and development and utilization information extraction method | |
| CN108830808B (en) | On-satellite infrared image stripe noise removing method based on similar line window mean value compensation | |
| Demir et al. | Integrated shoreline extraction approach with use of Rasat MS and SENTINEL-1A SAR Images | |
| Cheng et al. | Integration of hyperspectral imagery and sparse sonar data for shallow water bathymetry mapping | |
| Liu et al. | Mapping China’s offshore mariculture based on dense time-series optical and radar data | |
| Wang et al. | Land contained sea area ship detection using spaceborne image | |
| CN116758049A (en) | Urban flood three-dimensional monitoring method based on active and passive satellite remote sensing | |
| Ping et al. | Applications of deep learning-based super-resolution for sea surface temperature reconstruction | |
| Han et al. | A method for classifying land and ocean area by removing Sentinel-1 speckle noise | |
| CN112989940B (en) | Raft culture area extraction method based on high-resolution third satellite SAR image | |
| Amitrano et al. | Flood detection with SAR: A review of techniques and datasets | |
| Yu et al. | Coastline detection using optical and synthetic aperture radar images | |
| CN117173584B (en) | Land small micro water body extraction method and device for fusion of PolSAR and Pan images | |
| Deng et al. | Mapping bathymetry from multi-source remote sensing images: a case study in the Beilun estuary, Guangxi, China | |
| Yu et al. | Automatic extraction of green tide using dual polarization Chinese GF-3 SAR images | |
| Nagaraj et al. | Extraction of Surface Water Bodies using Optical Remote Sensing Images: A Review | |
| CN117173584A (en) | Land small micro water body extraction method and device for fusion of PolSAR and Pan images | |
| Zhang et al. | A novel saliency-driven oil tank detection method for synthetic aperture radar images | |
| Kulkarni et al. | Application of Taguchi method to improve land use land cover classification using PCA-DWT-based SAR-multispectral image fusion | |
| Tian et al. | Use of intensity and coherence of X-band SAR data to map thermokarst lakes on the Northern Tibetan Plateau | |
| Song et al. | SLWE-Net: An improved lightweight U-Net for Sargassum extraction from GOCI images |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |