基于GA的遥感图像目标SVM自动识别

为了高效合理地确定支持矢量机(SVM)的参数,使其对复杂的二值遥感图像目标进行自动识别,采用实值编码遗传算法来实现SVM模型参数的自动选择.与穷举搜索的留一法及随机试凑法相比,采用遗传算法的SVM模型参数选择更简单、更易于实现,并使SVM具有更好的推广能力.二值遥感图像目标的分类识别结果表明,该方法不但可以提高分类识别率,而且显著地缩短了SVM的训练时间.

Genetic Algorithm-based SVM for Automatic Target Classification of Remote Sensing Images

Support vector machine(SVM) has recently been proposed as a new effective learning machine for classification of remote sensing images.However,SVM often requires expensive design phases to choose adequate model parameters to attain high classification accuracy.A real-coded genetic algorithm(RGA) is used to automatically determine the model parameters for SVM, aiming at expediting the model selection process in SVM design with optimal generalization performance.Compared with the commonly used trial-and-error method,the proposed method is easier to implement.Furthermore,the generalization of the RGA-based SVM is much improved.Experimental tests conducted on targets classification of 2-value remote sensing images demonstrate that the proposed approach can conduct automatic model selection with low error while providing significant savings in time.