ToF点云强度与位置相关联的目标提取算法  

ToF point cloud intensity and location-related target extraction algorithm

作　　者：韩丰泽 李国栋 韩一菲[1,2] 郑福 孙志斌 HAN Feng-ze;LI Guo-dong;HAN Yi-fei;ZHENG Fu;SUN Zhi-bin(National Space Science Center,Chinese Academy of Sciences,Beijing 100190,China;University of Chinese Academy of Sciences,Beijing 100049,China;Liaoning University,Shenyang 110036,China)

机构地区：[1]中国科学院国家空间科学中心,北京100190 [2]中国科学院大学,北京100049 [3]辽宁大学,辽宁沈阳110036

出　　处：《激光与红外》2020年第12期1521-1528,共8页Laser & Infrared

基　　金：国家自然科学基金项目(No.61274024,No.61474123);中国科学院青年创新促进会项目(No.2013105);国家重大科学仪器设备研发项目(No.2013YQ030595);国家重点研发计划科技部国际合作,中捷合作项目(No.2016YFE0131500)资助。

摘　　要：目前光子飞行时间技术在三维成像中应用广泛,但三维点云目标提取相关研究匮乏,因此提出了一种利用点云位置信息和强度信息相关联的三维点云复杂目标提取方法。该方法首先使用边缘提取及形态学处理对点云强度信息中的目标位置定位,使目标在强度图中被标识。然后将强度信息与点云位置信息相关联,在三维点云中提取出目标所在区域。最后建立K邻近邻域,通过改进邻域大小选取方式,使用高斯均值代替固定的人工选择的邻域。实验结果表明,该方法能有效的滤除离群点并完整的保留真是目标物体点云数据的位置信息和强度信息,同时具有较高的运算速度。At present,time-of-flight(ToF)technology is widely used in 3D imaging;however the method of 3D point cloud target extraction is rarely.Therefore,we present a 3D point cloud complex target extraction method based on point cloud location information and intensity information.Firstly,the edge extraction and morphological processing are used to locate the target position in the intensity information,so that the target can be identified in the intensity image.Then the intensity information is correlated with the position information,and the region of the target is extracted from the 3D point cloud.Finally,by establishing a K-neighboring,and by improving the selection method of the neighborhood size,the Gaussian mean replaces the fixed artificially selected neighborhood.The experimental results show that this method can effectively filter outliers and retain the position and intensity information of the real object point cloud data,and has a high computing speed.

关 键 词：图像处理 光子飞行时间 K邻近 双边滤波 强度信息 

分 类 号：TN249[电子电信—物理电子学] TP391[自动化与计算机技术—计算机应用技术]