中国科学院自动化研究所机构知识库

Pose estimation is an important step in the grasping of workpieces. However, most previous works aim to use the 3D vision system to locate the 3D pose of the object. This paper develops a pose estimation of 3D object with 2D vision system. The proposed method includes two steps: (a) a hierarchy model of 2D views of the object is firstly constructed off-line; (b) the pose of object is then estimated by measuring the similarity of the model and target image. The proposed method is inherently robust against noise and illumination changes, and also efficient in real applications.