中文摘要：

针对现有LiDAR点云去噪算法难以实现自适应的问题，该文提出了基于空间盒子结构的自适应移动盒子去噪算法。基于经验模型建立盒子自适应准则，较好规避了现有算法的去噪阈值确定问题；采用聚类分析实现噪声点与有效点的精确标示。基于不同类型噪声的点云数据实验结果表明：与现有去噪算法相比，基于自适应移动盒子的去噪算法具有更好的适应性和去噪效果。

英文摘要：

Aiming at the problem that existing LiDAR point cloud denoising algorithm is difficult to re- alize self adapting, an adaptive mobile box denoising algorithm based on spatial box structure was pro- posed. The box adaptive criterion was constructed based on experience model, which was a better solution to the problem of threshold determination for the existing algorithms; using cluster analysis to achieve precise marking of both noise points and efficient points. Experiment results of point cloud data for differ- ent noise types showed that comparing with the existing denosing algorithms, the denoising algorithm based on adaptive mobile box had better adaptability and denoising effect.