【作者】 王浩然； 王文光； 张玉玺；

【Author】 WANG Haoran;WANG Wenguang;ZHANG Yuxi;School of Electronic and Information Engineering, Beihang University;

【机构】 北京航空航天大学电子信息工程学院；

【摘要】 激光雷达对同一目标进行扫描时,扫描角度、设备等因素会直接影响点云的具体分布,导致同一目标的点云的坐标范围、点云密度有所不同,进而影响后续的检测、识别等流程。针对这一问题,结合点云的三角网格模型和法向量匹配对激光点云进行有效区域的提取,并以三角形间的法向量夹角与内心间的欧式距离作为权值,通过复杂网络的动态演化机制提取有效区域的多尺度特征向量,进一步得到两激光点云的相似度。最后通过实测数据验证了该衡量方法的有效性。更多还原

【Abstract】 When the Lidar scans the same target, the scanning angle, equipment and other factors will directly affect the specific distribution of the point cloud, resulting in different coordinate range and density of the point cloud of the same target, and then affect the subsequent detection, recognition and other processes. To solve this problem, this paper combines the triangular mesh model and normal vector matching of point cloud to extract the effective region of laser point cloud, and takes the angle between the normal vectors and the Euclidean distance between the inner heart as the weight, extracts the multi-scale feature vector of the effective region through the dynamic evolution mechanism of complex network, and further obtains the similarity of two laser point clouds. Finally, the effectiveness of the measurement method is verified by the measured data.更多还原