Classification Method of Polarized SAR Image Based on Deep Directional Wave Network

Abstract

The invention discloses a polarimetric SAR image classification method based on a deep directional wave network. The implementation steps of the method are: (1) inputting a polarimetric SAR image; (2) extracting Pauli decomposition features; (3) constructing training sample features (4) Initialize the convolutional neural network; (5) Train the convolutional neural network; (6) Build the test sample feature matrix; (7) Get the class label of the test sample; (8) Calculate the classification accuracy; (9) (10) Output the colored polarimetric SAR image. The present invention uses the direction filter as the filter of the convolutional neural network to classify the polarimetric SAR image, so that the present invention has the advantage of well preserving the direction information of the polarimetric SAR image.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to a polarization synthetic aperture radar (Synthetic Aperture Radar, SAR) image classification method based on a depth directional wave network in the technical field of polarization synthetic aperture radar image classification. The invention can be used to classify the ground objects of the polarimetric SAR image, and can effectively improve the classification accuracy of the polarimetric SAR image. Background technique [0002] Synthetic aperture radar is a high-resolution imaging radar. Because microwaves have penetrating properties and are not affected by light intensity, synthetic aperture radars have all-day and all-weather working capabilities. With the development of technology, synthetic aperture radar is gradually developing in the direction of high resolution, multi-polarization and multi-channel. Compared with traditional single-polarization SAR, multi...

AI Technical Summary

Problems solved by technology

This method is an unsupervised classification method, which has the advantages of being able to accurately describe the scattering of ground objects, and can well correspond to the actual scattering situation, and reduce the calculation time of category adjustment. Supervised classification can only rely on scattering information to classify ground objects, making the classification accuracy rate low

This method uses a layer-by-layer method to train the deep wavelet neural network, which avoids the problem of gradient diffusion when there are many layers. Preserving orientation information of polarimetric SAR images

Images