Unlock instant, AI-driven research and patent intelligence for your innovation.

PolSAR image multilevel feature extraction and unsupervised classification method

A technology of feature extraction and classification method, applied in the field of PolSAR image interpretation, to achieve the effect of improving classification accuracy and high adaptability

Inactive Publication Date: 2018-08-03
CIVIL AVIATION UNIV OF CHINA
View PDF7 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, these algorithms all require more or less human intervention. When processing PolSAR images of complex and large scenes, their shortcomings of requiring human intervention will be very obvious.

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
  • PolSAR image multilevel feature extraction and unsupervised classification method
  • PolSAR image multilevel feature extraction and unsupervised classification method
  • PolSAR image multilevel feature extraction and unsupervised classification method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0029] The method for multi-level feature extraction and unsupervised classification of PolSAR images provided by the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0030] like figure 1 As shown, the method for multi-level feature extraction and unsupervised classification of PolSAR images provided by the present invention includes the following steps in sequence:

[0031] (3) Preprocess the PolSAR image to obtain a coherence matrix T for deorienting and filtering out speckle noise;

[0032] Different from ordinary optical images, the pixels of PolSAR images are neither like grayscale images with only one grayscale value from 0 to 255, nor like color image pixels with only three values ​​from 0 to 255. A value from 0 to 255, each pixel of the PolSAR image is a coherence matrix T θ , and the coherence matrix T θ The diagonal elements are all real numbers, and the off-diagonal elements are complex nu...

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 PolSAR image multilevel feature extraction and unsupervised classification method. The method comprises the steps of firstly preprocessing a PolSAR image by utilizing a deorientation principle and fine Lee filtration; secondly extracting polarimetric scattering entropy / anisotropy combination features by utilizing Cloude-Pottier decomposition, and performing first-level classification by using the features; thirdly extracting surface scattering, secondary scattering and volume scattering power features by utilizing a three-component decomposition method, and performingsecond-level classification by using the features; fourthly extracting total scattering power features to distinguish ground objects same in scattering mechanism and different in scattering power, and performing third-level classification by using the features; and finally performing iterative optimization on classification results by using a Wishart classifier. The method has the following advantages: 1) the classification precision is high and the texture is clear; and 2) the adaptive degree is high, the method can adaptively determine a classification number according to the complexity ofan image scene, and other manual interventions are not required except an iterative frequency.

Description

technical field [0001] The invention belongs to the technical field of PolSAR image interpretation, in particular to a method for multi-level feature extraction and unsupervised classification of PolSAR images. technical background [0002] Polarimetric Synthetic Aperture Radar (PolSAR), as an important remote sensing technology, has been widely used in military and civilian fields due to its all-weather, all-weather, high-resolution and strong penetration characteristics. Compared with the rapid development of PolSAR system development, PolSAR image interpretation research is still relatively lagging behind. Under the trend of massive growth of various remote sensing data, the adaptive interpretation of remote sensing images is particularly important. As an important part of image interpretation, classification can be applied to forest mapping and urban planning as the final result, and can also be applied to image filtering and object detection as an intermediate result. ...

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 Applications(China)
IPC IPC(8): G06K9/40G06K9/62G06T7/40
CPCG06T7/40G06T2207/10044G06V10/30G06F18/24G06F18/214
Inventor 韩萍韩宾宾孙丹丹宋厅华
Owner CIVIL AVIATION UNIV OF CHINA