Target detection method in polarized sar image based on nsct ladder net model

Abstract

The invention discloses a polarimetric SAR image target detection method based on a non-subsampled contour wave NSCT ladder network model, which mainly solves the problem that the prior art can only rely on scattering information to classify ground objects and does not consider polarimetric SAR images. The problem of low classification accuracy caused by scale features. The concrete steps of the present invention are as follows: (1) carry out Lee filter to the polarization SAR image to be detected; (2) carry out Yamaguchi decomposition with coherence matrix; (3) normalize feature matrix; (4) carry out non-subsampling to feature matrix Contourlet transformation; (5) Constructing a data set; (6) Constructing a ladder net object detection model; (7) Training an object detection model; (8) Obtaining test results. The invention has the advantages of good multi-scale feature extraction for polarimetric SAR images and high target detection precision.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to a polarimetric SAR (Synthetic Aperture Radar) based on a non-subsampled contourlet NSCT (non-subsampled contourlet transform) ladder network model in the technical field of polarimetric synthetic aperture radar image target detection. ) image target detection method. The invention can be applied to accurately detect and identify targets in different regions of the polarimetric SAR image. Background technique [0002] Synthetic Aperture Radar (SAR), as the only radar with all-weather and all-weather remote sensing imaging capabilities among various remote sensing methods, has an irreplaceable role in the field of remote sensing and has been widely used. Polarization synthetic aperture radar is a new type of SAR system radar based on the traditional SAR system, and its appearance greatly broadens the application field of SAR. [0003] With the popularization of pol...

AI Technical Summary

Problems solved by technology

This method belongs to the unsupervised classification method, which can accurately describe the scattering characteristics of ground objects, and can well correspond to the actual scattering situation, and has the advantages of reducing the calculation time of category adjustment. However, the disadvantage of this method is that due to This method belongs to unsupervised classification, and can only rely on scattering information to classify ground objects, which makes the classification accuracy rate low

However, the disadvantage of this method is that since the ladder network model used is based on full connection, the data block needs to be pulled into a vector and then input into the network, which destroys the neighborhood information of the image, and does not take into account the polarization SAR The multi-scale features of the image lead to insufficient utilization of image information, and the edge of the image cannot be detected well, and the result will deviate from the real target.

Images