一种用于机载LiDAR波形数据高斯分解的高斯拐点匹配法

发布时间：2018-07-08 15:34

本文选题：遥感 + 机载激光雷达　；参考：《激光与光电子学进展》2014年10期

【摘要】：针对小光斑全波形机载激光雷达(LiDAR)波形数据高斯分解法的核心问题——高斯分量个数估计,提出一种高斯拐点匹配法。该算法用平面曲线离散点集拐点的快速查找算法检测波形数据中的拐点,计算过检测出的拐点及其右边第一个点的直线的斜率,根据斜率将所有检测出的拐点分为左、右拐点,一个左拐点与其邻近的一个右拐点组成一个高斯分量,据此可以确定波形数据中高斯分量个数。采用高斯拐点匹配法对模拟和实测波形数据进行分解,并与传统的脉冲检测方法(重心法和高斯脉冲拟合法)相比。结果表明,高斯拐点匹配法方法能极大地减小伪拐点的影响,快速、准确地检测并分解出波形数据中高斯分量,提高波形数据分解速度。同时其能分解出更多的高斯分量,从而提高点云密度。
[Abstract]:In view of the core problem of the Gauss decomposition method of the small spot full waveform airborne laser radar (LiDAR) waveform data, the Gauss component number estimation, a Gauss inflection point matching method is proposed. The algorithm uses the fast searching algorithm of the discrete-time point set of the plane curve to detect the inflection point in the waveform data, and calculates the detected inflection point and the first right one on the right. The slope of the point line, according to the slope, divides all the detected inflection points into left, right turn points, a left turning point and a Gauss component adjacent to a right turning point. According to this, the number of Gauss components in the waveform data can be determined. The simulation and actual waveform data are decomposed with the Gauss inflection point matching method and the traditional pulse detection is used. Compared with the measurement method (center of gravity method and Gauss pulse fitting method), the results show that the Gauss inflection point matching method can greatly reduce the influence of the pseudo inflection point, quickly, accurately detect and decompose the Gauss component in the waveform data, and improve the velocity of the waveform data decomposition. At the same time, it can decompose more Gauss components and thus improve the density of the point cloud.
【作者单位】：首都师范大学三维信息获取与应用教育部重点实验室;首都师范大学空间信息技术教育部工程研究中心;中国地震局地震预测研究所;
【基金】：国家科技支撑计划(2012BAH31B01) 北京市自然科学基金重点项目(KZ201310028035)
【分类号】：TN958.98

【参考文献】