基于分组LM算法的全波形LiDAR高斯分解

发布时间：2018-03-17 21:22

本文选题：全波形Li　切入点：DAR　出处：《测绘与空间地理信息》2016年07期 　论文类型：期刊论文

【摘要】：LM(Levenberg-Marquardt)算法是全波形机载激光雷达(Li DAR)数据高斯分解中求解模型参数的一种方法。针对其结果在一定程度上依赖初值、雅克比矩阵出现非数值导致无结果等问题,本文提出分组LM算法,以广义高斯混合函数为模型,模型参数初始化后,将参数分组并使用LM算法依次对各组参数进行优化,并生成点云。为验证结果的可靠性,以系统点云为参考,与基于改进的EM(Expectation Maximum)算法全波形分解法做对比。结果表明,本方法不仅得到较高质量的点云,而且得到回波位置和宽度等信息。
[Abstract]:The LMU Levenberg-Marquardt algorithm is a method to solve the model parameters in the decomposition of the data Gao Si of the full waveform airborne lidar / Li DAR. In view of the problem that the results depend on the initial value to some extent, the Jacobian matrix is nonnumerical and has no results. In this paper, a grouping LM algorithm is proposed, in which the generalized Gao Si mixed function is taken as the model. After the model parameters are initialized, the parameters are grouped and optimized by LM algorithm, and point clouds are generated to verify the reliability of the results. The system point cloud is taken as a reference and compared with the full-waveform decomposition method based on the improved EM(Expectation maximum algorithm. The results show that the method not only obtains a high quality point cloud, but also obtains the echo position and width information.
【作者单位】：武汉大学遥感信息工程学院;武汉大学测绘遥感信息工程国家重点实验室;浙江省第二测绘院;
【分类号】：P237

【相似文献】