建筑全景三维机载雷达点云信息采集方法仿真

Simulation of Point Cloud Information Collection Methodfor Building Panoramic 3 D Airborne Radar

由于机载雷达的测量误差和环境干扰等原因,采集到的点云数据可能存在噪声和缺失,且建筑物的多层结构,使得采集的点云数据之间存在位置偏差和姿态差异,导致建筑全景点云结构信息采集误差较大。为此,提出建筑全景三维机载雷达点云结构信息采集仿真。采用双向布料模拟法提取建筑顶面点云,结合归一化数字地表模型、穿透性分析、形态学开运算、约束生长等方法提取建筑全景点云,利用统计滤波算法滤除离群点噪声,采用体素中心邻近特征点的点云精简算法降低点云冗余度,并对建筑全景点云粗配准和精配准,实现建筑全景三维点云结构信息采集。实验结果表明,所提方法误差较小、完整率和采集质量较高。

Due to measurement errors of airborne radar and environmental interference,the collected point cloud data may contain noise and missing values.In addition,the multi-layer structure of buildings may cause position deviation or attitude differences between point cloud data,resulting in large errors in the collection of panoramic point cloud information.To address this issue,this article designed a simulation of collecting point cloud information of panoramic 3D airborne radar of buildings.Firstly,we adopted the bidirectional cloth simulation method to extract the point cloud of the building’s top surface.Then,we used the normalized digital surface models penetrability analysis,morphological opening operation,and constrained growth method to extract the panoramic point cloud of buildings.Next,we used statistical filtering algorithms to filter outlier noise.Meanwhile,we used a cloud simplification algorithm including voxel centroid adjacent feature points to reduce the redundancy of the point cloud and perform rough and fine registration of the panoramic point cloud.Finally,we achieved the collection of panoramic 3D point cloud information.Experimental results show that the proposed method has smaller errors,higher completeness rates,and higher collection quality.