基于支持向量机的遥感图像分类研究

支持向量机(Support Vector Machine,SVM)是一种基于统计学习理论的新型机器学习算法.通过解算最优化问题,在高维特征空间中寻找最优分类超平面,从而解决复杂数据的分类及回归问题.将支持向量机理论应用到遥感图像分类的研究还处在初级阶段,传统分类算法应用于遥感图像分类存在运算速度慢、精度比较低和难以收敛等问题.从支持向量机基本理论出发,建立了一个基于支持向量机的遥感图像分类器.用遥感图像数据进行实验,并将结果与其它方法的结果进行了比较分析.实验结果表明,利用SVM进行遥感图像分类的精度明显优于神经网络算法和最大似然算法分类精度.

Remote Sensing Image Classification Based on Support Vector Machines

Support Vector Machine(SVM) is a new machine learning algorithm based on statistical learning theory.It tries to find the optimal classification hyperplane in high dimensional feature space to handle complicated classification and regression problems by solving optimization problems.The research of the application of SVM to remote sensing image classification is still in the initial stage.Traditional algorithms used in remote sensing image classification have some problems such as low efficiency,low accuracy and much difficulty for convergence.A classification model based on SVM was constructed to classify the content of remote sensing images.Experiments ware done to verify the classification model.Experimental results indicated that SVM classifier had more advantages in the classification of remote sensing images than neural network classifier and maximum likelihood classifier.