Hyperspectral Features And Image Classification Based On Robust Extreme Learning Machine

With the development of sensor technology,hyperspectral images with high spatial resolution have been used more and more,such as face recognition,image processing and so on.However,the hyperspectral image has rich information,but there are also problems such as too much data and strong correlation,which greatly improves the difficulty of hyperspectral image classification.The traditional hyperspectral image classification system uses the spectral features of the extracted image to classify the image,but it ignores the spatial features that contain important information in the image.In recent years,people have great interest in using spatial features to improve the performance of hyperspectral image classification.In this paper,the feature extraction and classification methods of hyperspectral images are systematically studied,and the spectral features and spatial features are combined,and the hyperspectral image classification algorithm based on extreme learning machine classifier is studied.The main contents of this paper are as follows:(1)This paper studies the spatial feature extraction of hyperspectral image based on local binary pattern.This method has significant advantages in spatial feature extraction,and combines the extracted spatial feature vector with spectral feature vector to get a new spatial spectral joint feature vector,so as to solve the problem of insufficient feature extraction and unsatisfactory classification effect caused by single feature extraction in the process of image classification.(2)The algorithm of hyperspectral image classification based on kernel extreme learning machine is proposed.The kernel extreme learning machine is used to classify hyperspectral image instead of the traditional extreme learning machine,so as to avoid the problem that the traditional extreme learning machine does not have a good classification effect due to randomly assigning the input weight and deviation.