基于语义分割和点云配准的物体检测与位姿估计

A Method for Object Recognition and Pose Estimation Based on the Semantic Segmentation

面向基于视觉的机器人零件抓取任务,提出了一种基于语义分割的物体识别与位姿估计方法。首先通过语义分割模型对彩色图进行分割,然后从对应的深度图中获得物体的深度信息,并转换为点云,再使用法向采样对应点的最近点云迭代算法进行物体位姿精确配准。从而实现在复杂环境下对无纹理、相互间有遮挡的物体的有效位姿估计,为后续抓取提供必要的信息。

In order to accomplish a robot-intelligent-grasping project,a method for object recognition and pose estimation based on the semantic segmentation was proposed,starting with conducting segmentation on color image,following with extracting object’s point-cloud from the depth image,and finishing the work by applying the normal sampling cloud from the nearest point of the corresponding points of iterative algorithm for accurate registration object position.It can effectively estimate the position and pose of texture-less objects with shielding each other in complex environment,providing necessary information for subsequent grasping.