基于特征值的激光成像雷达目标识别算法  

Target recognition algorithm of laser imaging radar based on feature value

作　　者：李春苗[1] 高剑波[2] 杨镇源[2] 周昕[1] 张华[2] 冷杰[2] 宋昭[2] 

机构地区：[1]四川大学电子信息学院,四川成都610065 [2]西南技术物理研究所,四川成都610041

出　　处：《激光与红外》2015年第3期335-338,共4页Laser & Infrared

摘　　要：提出了一种基于特征值的激光成像雷达目标识别算法。首先从激光成像雷达中获取目标场景的距离像,将场景中各个目标分离,分别提取各个目标距离像的特征值,然后用网格搜索算法对支持向量机参数进行优化,最后应用支持向量机对分离出来的三种目标进行分类识别。仿真实验表明:该方法简单易于实现,识别效果较好。Target recognition algorithm of laser imaging radar based on feature value was put forward. Firstly range images of the object scene from laser imaging radar were achieved,and each object in the scene was separated,and the feature values were extracted from range images. Secondly the grid search algorithm was used to optimize parameters of support vector machine（ SVM）. At last,SVM was used to classify and recognize the segregated targets. Simulation experiment shows that the method is simple and easy to implement,and has good recognition effect.

关 键 词：特征值 支持向量机 激光雷达 目标识别 

分 类 号：TN958.98[电子电信—信号与信息处理]