Image Classification Based on Dempster-Shafer Evidence Theory and Neural Network

Dempster-Shafer Evidence theory was extended from Bayes Decision, and it can combine together the certainty and uncertainty multi-source remote sensing images to effectively identify the images. Taking the results from the training of neural network as evidences, and combining the neural network with evidence theory, we could integrate their advantages to get better classification results. In this paper, we classified the remote sensing image with computer preprocess, and took the panchromatic image with plentiful spatial information into classification decision to reduce uncertainty and improve the classification accuracy based on evidence theory and neural network.