Method for selecting hyperspectral image bands based on extraction of all kinds of important bands

Abstract

The invention discloses a method for selecting hyperspectral image bands based on extraction of all kinds of important bands. The hyperspectral data of each training sample of the bands are used as a time sequence; smooth denoising processing is carried out on each time sequence through wavelet transformation; the important band sets of the training samples are ensured by extracting the important points of the smoothed time sequences, wherein the important points correspond to the important bands respectively; the important band sets of the training samples are collected to form an initial band set; the final band combination is selected on the basis of the initial band set through a branch and bound method. According to the method, conditional mutual information grouping is introduced in the branch and bound method as constraint conditions, and compared with the search efficiency of the existing fast branch-and-bound search algorithm, search efficiency is improved by about one half.

Description

technical field [0001] The invention belongs to the field of hyperspectral image processing, and more specifically relates to a hyperspectral image band selection method based on various important band extraction and branch and bound methods. Background technique [0002] The emergence of hyperspectral remote sensing technology (hyperspectral remote sensing) is a revolution in the field of remote sensing, providing humans with a new technical means of observing the external world. With the rapid development of aerospace technology and remote sensing science, the available hyperspectral remote sensing data is increasing at an alarming rate. How to effectively process it and make full use of the rich information contained in it is the current research topic in the field of remote sensing and pattern recognition. One of the hot topics. Hyperspectral images are often composed of hundreds of bands. The large number of bands, high correlation and redundancy between bands bring a ...

AI Technical Summary

Problems solved by technology

[0006] In the prior art, for example, in the patent application whose application number is 2010101529156, and the title of the invention is "Method for selecting bands of remote sensing hyperspectral images based on conditional mutual information", it is recorded that the band conditional mutual information grouping information is used in genetic algorithm search, and obtained In the application number 201010195127.5, the title of the invention is "A hyperspectral remote sensing image band selection method based on the analysis of important points in time series", A band selection method based on the analysis and extraction of important points in time series is proposed, but because the typical spectral curve is obtained based on cluster analysis, there is a lot of information loss, and the final classification accuracy is not as good as that using all the original bands.

Images