Heuristic Genetic Algorithm-Based Support Vector Classifier for Recognition of Remote Sensing Images

Chunhong Zheng, Guiwen Zheng, Licheng Jiao

Abstract: A heuristic genetic algorithm (GA)-based support vector classifier (SVC) for recognition of remote sensing images is presented in this paper. The model parameters of SVC are automatic selected by a heuristic GA to obtain the better performance with high efficiency. Compared with the leave-one-out (loo) method and the trial and error method, this GA-based model parameters selection is simpler and easier to implement. Furthermore, the generalization of the obtained SVC is much improved. Comparative tests conducted on a 2-value remote sensing images demonstrate the better result of the proposed classifier.