Confidence feedback combined kernel collaborative sparse representation remote sensing image classification method and system

Abstract

The invention provides a confidence feedback combined kernel collaborative sparse representation remote sensing image classification method and system. The method comprises the steps that firstly, spectral features are directly extracted from a remote sensing image, and a spectral kernel isconstructed; secondly, morphological attribute features are extracted from principal components of the remote sensing images subjected to principal component analysis dimension reduction, morphological attribute kernels are constructed, then the two kernels are weighted and combined to obtain a combined kernel, and a collaborative sparse representation classifier is adopted to classify the remote sensing images; finally, before the classification result is output, a confidence feedback mechanism is used to judge whether the classification result accords with a set termination judgment criterion; and if the judgment criterion is not met, the classifier outputs a pre-classification result graph, carries out morphological attribute filtering operation on the pre-classification result graph, stacks the pre-classification result graph to morphological attribute features for iterative classification, and outputs a final classification result until the judgment criterion is met. The method can be used for remote sensing image and hyperspectral image ground object information classification, is rapid and effective, is high in classification precision, and has a good application prospect.

Description

technical field [0001] The invention relates to the technical field of remote sensing image processing, in particular to a remote sensing image classification method and system based on confidence feedback combined kernel cooperative sparse representation classification. Background technique [0002] With the rapid development of spectral imaging technology, it has greatly promoted the development of the field of satellite remote sensing. ", "Different objects with the same spectrum" and other issues, there is still a lot of room for development in the processing technology of spectral remote sensing images, such as classification, unmixing, fusion, denoising, target detection, etc. of remote sensing images. In order to solve these problems, a deep understanding of remote sensing imagery and a proficient grasp of machine learning methods are required. [0003] As the focus of remote sensing image processing, remote sensing image classification has gradually been paid attent...

AI Technical Summary

Problems solved by technology

[0002] With the rapid development of spectral imaging technology, it has greatly promoted the development of the field of satellite remote sensing. ", "Different objects with the same spectrum" and other issues, there is still a lot of room for development in the processing technology of spectral remote sensing images, such as classification, unmixing, fusion, denoising, target detection, etc. of remote sensing images

Images