Image-based 2D Form-closure Grasping Detection Method

Peitao Hong, Liucun Zhu, Jiyuan Liu, Zhenyu Zhang, Liang Li, Wenhao Bai, Zhanpeng Jing, Shanjun Zhang

Abstract: Generalisability of grasping algorithms, grasping target recognition rates, and real-time grasping judgements in robot grasping tasks have been modern industrial challenges, from analytical heuristics to more recent new deep learning strategies, grasping in complex scenarios is still the aim of several works’ that propose distinctive approaches. In this context, the purpose of this paper is to discuss the real-time robot grasping task, and for the two-dimensional form-closure proposed a real-time grasping point contact resolution method, this method uses traditional image processing, the test proved that the two-dimensional Form-closure point resolution real-time stability of 100fps, for the limited resources of the effectiveness of the high requirements of the scene has a computationally efficient, easy to build the mathematical model, and can be interpreted with strong and other advantages.