Hyperspectral Image Classification Method and System Based on Stack Width Learning

Abstract

The present invention proposes a hyperspectral image classification method and system based on stack width learning. First, the input image is preprocessed, including normalization, wavelet decomposition and difference; secondly, a small number of samples are selected as training samples, and the stack width is used to The learning model learns the spatial spectral features of the hyperspectral image; then, the training samples are used to train the width learning classifier, and finally, the features of the test samples are input into the trained width learning classifier to obtain the class labels. Using the stack width learning model of the present invention can obtain more abstract spatial spectral features, so the method of the present invention can obtain very accurate classification results. Due to the characteristics of width learning, the present invention has less sample complexity, and only needs a small amount of training samples to obtain better classification results, so it is more conducive to practical application.

Description

technical field [0001] The invention belongs to the technical field of remote sensing impact processing and analysis, and designs a hyperspectral image classification method and system, which can be used for resource management, scene interpretation, precision agriculture, urban planning, disaster prevention and mitigation, and the like. Background technique [0002] As a cutting-edge technology in the field of remote sensing, hyperspectral remote sensing has a wide range of applications in military and civilian fields. The rich spectral information of hyperspectral images not only provides more information for object classification, but also brings a series of new problems of feature extraction and pattern recognition. Researchers have conducted in-depth research on a series of problems in hyperspectral image classification, such as high data dimensionality and few training samples. However, the above-mentioned problems have not been effectively solved so far. [0003] Fi...

AI Technical Summary

Problems solved by technology

[0009] The purpose of the present invention is to address the deficiencies of the above-mentioned prior art, and propose a hyperspectral image classification method based on stack width learning to solve the difficulty in learning effective classification features and high sample complexity in the hyperspectral image classification problem in the prior art. problem, improving the performance of spectral image classification

Images