Automatic Band Selection of Hyperspectral Remote Sensing Image Classification Using Particle Swarm Optimization

It is proposed an automatic band selection and SVM parameter optimization method based on a novel PSO-BSSVM model,which is used to classify the hyperspectral remote sensing images.Comparing with K-nearest neighbors classifier(K-NN),radial basis function-neural network(RBF-NN)classifier and standard SVM classifier,the empirical results have demonstrated that PSO-BSSVM can automatically select appropriate hyperspectral bands and optimize SVM parameters,and the classification accuracy of hyperspectral remote sensing images is improved significantly.