Application Of Support Vector Machine Optimized By Particle Swarm Optimization Algorithm In High Accuracy Urban Land Cover Classification

In this paper, a method for high accuracy urban land cover classification is proposed by using support vector machine (SVM) based on particle swarm optimization (PSO). SVM has the outstanding advantages in solving small sample classification problems. However, the land cover classification is not a small sample problem, its data are high dimensional data. SVM can solve the defects of high dimensional representations, it is reasonable to apply SVM to land cover classification. The selection of parameters can affect the classification results greatly. PSO has high efficient global search ability. Through PSO to determine the optimal penalty parameter C and kernel function parameter gamma, the classification accuracy rate can reach optimum. And comparison with the classification accuracy rates of different parameter optimization methods and different classification methods, the experiment results indicate that PSO-SVM has the high accuracy rate and is a preferred method for urban land cover classification.