输电走廊场景激光点云LOD生成方法研究

Study on LOD Generation of LiDAR Point Cloud in Transmission Corridor Scenes

为适应输电走廊海量点云高效实时三维显示需求,提出一种改进的多层次细节结构高效生成方法。首先对点云划分格网并计数,利用格网特征识别关键要素,将点云快速分成小块文件以利于并行处理;然后采用自适应采样方法对电力线塔等关键要素块进行泊松碟采样,对非电力要素块进行随机格网采样,实现多层次细节高效生成。实验结果表明,该方法在显示电力走廊关键要素时表现出了优秀的蓝噪声视觉效果,对比全局泊松碟采样速度提高了20%以上,满足海量点云电力巡检中实时高效显示需求。

In order to meet the demands of the massive points cloud display of transmission corridors efficiently and timely,an adjusted multi-level of detail(LOD)structure of fast generation method is proposed.Firstly,the points cloud is divided into grids and are counted.Key elements are identified by using grid features,and point clouds are quickly divided into small chunks to facilitate parallel processing.Then,the adaptive sampling method is used to sample the Poisson dish of key element blocks such as power line towers,and at the same time,random grid sampling is performed on non-power element blocks,which realizes the efficient generation of multi-level details.Experimental results show that the proposed method shows excellent blue noise visual effects when displaying the key elements of the power corridor,and the sampling speed is increased by more than 20%compared with the global Poisson disk,which meets the real-time and efficient display requirements in the power inspection of massive point clouds.