Deep sparse main component analysis-based polarimetric SAR image classification method

Abstract

The invention discloses a deep sparse main component analysis-based polarimetric SAR (Synthetic Aperture Radar) image classification method, and aims to mainly solve the problems of complex process and low classification accuracy of polarimetric SAR image classification in the prior art. The method is implemented by the following steps: (1) inputting a coherence matrix of a polarimetric SAR image to be classified, and filtering the coherence matrix to obtain a de-noised coherence matrix; (2) taking an element of each column vector in the de-noised coherence matrix as a sample, and randomly selecting a training sample set and a test sample set from the samples; (3) performing deep characteristic learning on the training sample set of the polarimetric SAR image, and training a two-layer sparse main component analysis network; (4) classifying the test sample set of the polarimetric SAR image to finish the classification of the polarimetric SAR image by utilizing the trained deep sparse main component analysis network. The method is high in classification accuracy, and can be used for ground object classification and target identification.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a polarimetric SAR image classification method, which can be used for ground object classification and target recognition. Background technique [0002] In recent years, with the rapid development of computer technology, a large amount of text, voice, image, video and other data is increasing day by day. Therefore, how to mine valuable information from massive data has become a problem worthy of attention. The core principle of data mining is to extract its essential characteristics from a large amount of complex data through advanced computer technology, so that valuable data information can be fully utilized. Data mining includes many methods, one of which is very important is classification, which can extract models describing important data from databases with rich content and a large amount of information for intelligent decision-making, so it is widely...

AI Technical Summary

Problems solved by technology

Although this method can cluster on a sample space of any shape and converge to the global optimal solution, it still has the disadvantage that when using a Gaussian function to construct a similarity matrix, the scale parameter has a great influence on the classification structure, and it is difficult to obtain the optimal solution. parameters, leading to unreasonable feature extraction, affecting the stability of image segmentation, resulting in a decrease in classification accuracy

Images