摘要： 针对星球探测任务中的岩石采样问题，利用双目立体视觉装置采集不同岩石图像，采用立体视觉算法获得岩石表面的三维点云数据。以三角剖分后生成的三角片法向量夹角为基本元素，提出基于C均值聚类的岩石表面评估算法，对岩石平整三维表面进行评估计算，并用不同颜色加以区分。仿真结果证明了该算法的有效性。

关键词: 星球探测, 立体视觉, C均值聚类, 岩石评估

Abstract: According to the rock sample problem in planetary exploration task, binocular vision system is applied to sample different rock images and the rocks’ surface 3D point cloud data are acquired by stereo vision algorithms in this paper. The relative plane on the rock is automatically evaluated by the proposed method based on C-means clustering between the triangulation pieces’ normal vector after triangulation, and the different plane is signed by different color. Simulation results show the validity of the algorithm.

Key words: planetary exploration, stereo vision, C-means clustering, rock evaluation

中图分类号: 

• TP391