Abstract
This paper describes a 3D plane detection method of an intelligent robot to achieve a task grasping an unknown object based on the information of a 3D range sensor. The previous methods of plane detection have problems which are the high computational cost and the low-accuracy of small object detection. To solve the problems, we have proposed a retinal structure-based plane detection method. The proposed method consists of a simple plane detection using region growing, a visual attention model based on the retinal structure and the integration of detected planes. To verify the proposed method, we show that the computational cost of the method is less than that of the previous methods through several experiments.
