Method of remote sensing image automatic classification oriented to simulation of 3d infrared terrain scene

Infrared terrain background simulation is of great significance to virtual battlefield environment. The access to the information of material and boundary is lack of efficient method in the infrared terrain background simulation based on the remote sensing images. A method of remote sensing image automatic classification was proposed to solve the problem. The material category was determined and a sample library was built. The classifier with the sample library was trained which extracted features. The input images were classified by the trained classifier. Boundary information and material category information after post-processing were obtained. Experimental results show that the proposed method can provide front-end data for IR simulation of terrain background efficiently.