针对室外未知的非结构化的三维地形环境，研究基于安装在室外移动机器人平台上的扫描激光对地形环境模型进行构建。采用里程计和GPS的信息对移动机器人平台进行融合定位，并结合机器人的位姿信息将扫描激光的数据转化为点云信息。采用多模板的中值滤波实现对点云中因混合像素干扰等原因造成的孤立点进行滤波，并将扫描的地形环境构建为基于八叉树的空间三维环境描述。文中对校园环境进行扫描实验，实验结果表明该方法可以对未知的三维环境进行准确地描述，同时可大大节省点云数据的存储空间。%For unstructured outdoor unknown three-dimensional terrain environment,research scanning laser installed in the outdoor mo-bile robot platform to construct terrain model. Use odometer and GPS for information fusion location of the mobile robot platform,com-bined robot pose translate scanning laser point cloud into information. Median filtering of multi-template is applied to achieve the outlier point filtering caused by mixed pixels interference in cloud,and scanning the terrain environment is built as octree space-based three-di-mensional environment description. Campus environment scan experimental results show that the method can be carried out on the un-known of three-dimensional environment with an accurate description of point cloud data storage space,at the same time greatly save the storage space for point cloud data.
展开▼
