Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Polarized SAR image classification method based on wishart and svm

A classification method and image technology, applied in character and pattern recognition, instruments, computer parts, etc., to achieve the effect of good universality and generalization, high accuracy, and strong generalization ability

Active Publication Date: 2017-07-28
XIDIAN UNIV
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to overcome the shortcomings of the above-mentioned prior art that polarimetric SAR data cannot achieve higher classification accuracy in the case of relatively few training samples, and propose a high accuracy and strong anti-noise ability based on Wishart and SVM Polarimetric SAR Image Classification Method

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Polarized SAR image classification method based on wishart and svm
  • Polarized SAR image classification method based on wishart and svm
  • Polarized SAR image classification method based on wishart and svm

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0028] The invention is a polarization synthetic aperture radar SAR image classification method based on Wishart and SVM. Refer to attached figure 1 , the specific implementation steps of the present invention are described in further detail:

[0029] Step 1, input image, input an optional polarization synthetic aperture radar SAR image to be classified, specifically figure 2 The L-band multi-look polarimetric SAR image of Flevoland, Netherlands area obtained in 1989 is shown.

[0030] Step 2. Filtering. In the specific simulation experiment, the polarized refined Lee filtering method with the filter window size of 3*3, 5*5, 7*7 and 9*9 is used respectively to filter the polarization synthetic aperture radar SAR image to be classified. filtering to remove coherent speckle noise, to obtain a filtered polarimetric SAR image, and to obtain a coherence matrix of the filtered polarimetric SAR image. In this embodiment, a filter window with a size of 7*7 is selected to remove co...

Embodiment 2

[0048] The polarization synthetic aperture radar SAR image classification method based on Wishart and SVM is the same as embodiment 1, wherein the Cloude decomposition described in step 3 includes the following steps:

[0049] 3.1. Extract the coherence matrix of the filtered polarimetric SAR image;

[0050] 3.2. Decompose the eigenvalue of the coherence matrix to obtain the eigenvalue λ of the coherence matrix 1 ,λ 2 ,λ 3 ;

[0051] 3.3. Using the obtained eigenvalue λ 1 ,λ 2 ,λ 3 , calculate the scattering entropy H, scattering angle alpha and total power span of each pixel according to the following formula,

[0052]

[0053]

[0054] span=λ 1 +λ 2 +λ 3

[0055]

[0056] where H represents the scattering entropy of the polarimetric SAR image, p i Indicates the ratio of the ith eigenvalue of the polarimetric SAR image coherence matrix to the sum of all eigenvalues, alpha indicates the scattering angle, a i Indicates the scattering angle of the polarimetr...

Embodiment 3

[0058] The polarization synthetic aperture radar SAR image classification method based on Wishart and SVM is the same as embodiment 1-2, wherein the similarity matrix W of the computing feature set F described in step 4 F Include the following steps:

[0059] 4.1. Construct feature set F=[H alpha span];

[0060] 4.2. Calculate the similarity matrix W of the feature set F F ;

[0061]

[0062] where d F (F i ,F j )=||F i -Fj || 2 , d F (F i , F j ) represents the i-th feature data F in the feature set F i and the jth feature data F j Euclidean distance, F i and F j Respectively represent two different training samples in the feature set F training sample set of polarization SAR SAR images, ||·|| 2 Denotes a two-norm operation, σ 1 Represents the similarity matrix W of the feature set F F The width of σ in the present invention 1 = 1, since choosing the width of such a similarity matrix can better represent the similarity between training samples, such a simi...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a polarization SAR image classification method based on Wishart and SVM, which mainly solves the technical problems of low classification accuracy and low classification efficiency when the existing polarization SAR classification method has few training samples. The implementation steps are: input image; filter; Cloude decomposition; calculate the similarity matrix of feature set F; calculate the similarity matrix of coherence matrix feature set T; calculate the final similarity matrix; use SVM classification; calculate the accuracy. Classification using the present invention overcomes the problem of many misclassification points caused by noise in the prior art, and can not only avoid crosstalk between polarization channels, but also maintain polarization information and statistical correlation, making polarization synthetic aperture radar The outline and edge of the SAR image are clearer, which improves the image quality after classification, makes the classification accuracy of the polarimetric SAR better, and has stronger adaptability to noise, and can be used for target recognition and identification of the polarimetric SAR image. track.

Description

technical field [0001] The present invention belongs to the technical field of image processing, and further relates to the technical field of machine learning and image classification, specifically a polarimetric SAR image classification method based on Wishart and SVM, which can be applied to the object classification of polarimetric SAR images to realize target recognition with tracking. Background technique [0002] Polarization SAR is a high-resolution active microwave remote sensing imaging radar, which has the advantages of all-day, all-weather, and high resolution. It has been extensively studied in agriculture, forestry, ocean, military and other fields. There are many methods for polarimetric SAR image classification. According to the different classifiers used, they can be divided into statistics, neural networks, support vectors, decision trees, etc.; according to whether spatial information is used, they can be divided into pixel-based and region-based; accordin...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/62G06K9/46
CPCG06F18/2411G06F18/2415
Inventor 焦李成李玲玲党晓婉屈嵘杨淑媛候彪王爽刘红英熊涛马文萍马晶晶
Owner XIDIAN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com