Hyperspectral image semi-supervised classification method based on comprehensive confidence

Abstract

The invention discloses a hyperspectral image semi-supervised classification method based on comprehensive confidence. The method comprises the following steps: reading a hyperspectral image; Calculating a graph weight matrix; 8, performing adjacent connection on the sparse graph weight matrix; Calculating a normalized graph weight matrix; Obtaining an initial training set and a candidate set; Setting collaborative training iteration times and starting a training process; Training a polynomial logic regression classifier; Obtaining a prediction label of the candidate set sample by using a polynomial logic regression classifier; Obtaining prediction tags of the candidate set samples by using a semi-supervised graph classification method; Selecting two candidate samples with consistent prediction tags and corresponding prediction tags to form a protocol set, and forming a comprehensive confidence set by corresponding confidence coefficients; Screening out a protocol set sample with a comprehensive confidence coefficient higher than 99% and a corresponding prediction label, and forming an amplification set and adding the amplification set into a training set; Removing an amplificationset sample in the candidate set; And judging whether the training reaches a set number of times, if not, continuing iteration, and if yes, stopping iteration, and classifying the hyperspectral imagesby using the semi-supervised graph.

Description

technical field [0001] The invention relates to the technical field of high-dimensional image processing, in particular to a semi-supervised classification method for hyperspectral images based on comprehensive confidence. Background technique [0002] Hyperspectral images are remote sensing images of surface features acquired by hyperspectral sensors with a spectral resolution of nanometers. The spectral data of each pixel of the hyperspectral image comes from the reflectance of light of different wavelengths at the pixel position where the object is located, and the spectral features can be used to identify the object category to which the pixel belongs. Hyperspectral images have the following characteristics: a large number of pixels and high-dimensional spectral features. At the same time, the "map-spectrum integration" feature of hyperspectral images also reflects that the spatial information in the images is relatively rich, and the spectral features are distinguishab...

AI Technical Summary

Problems solved by technology

[0004] Although many supervised classification methods such as multinomial logistic regression can have good results in many classification problems, in semi-supervised classification problems, due to the small number of training samples, only rough classification can be obtained for hyperspectral images with large noise. Results; Although the graph-based semi-supervised classification method can use the spatial smoothness information in the hyperspectral image to achieve relatively good classification results, the class boundary discrimination ability of this type of method is very weak

Therefore, many existing supervised and semi-supervised learning methods need to improve the classification ability of hyperspectral images in the case of fewer training samples.

Images