A surgical instruments sorting system based on stereo vision and impedance control

Abstract

Surgical instrument sorting is a repetitive work with a high risk of infection, which requires medical staffs to classify them after cleaning. Currently, the surgical instrument sorting is usually accomplished by medical staffs' manual operation. It is a time costing work, which may cause operators' fatigue even classification errors. Meanwhile, surgical instruments are relatively sharp and often stick with infectious substances like bloodstain, which brings a potential threat to the health of medical staffs. The sorting robot system can better avoid these threats when it replaces the medical staffs to complete the tasks accurately. This paper proposes a surgical instrument recognition and location method for sorting robot system based on stereo vision. The Voting Scheme for S2S Features will be applied for surgical instrument recognition, in which hash table searching will be used to boost searching speed. The impedance control and artificial potential field will be applied on robot to grab and place the classified surgical instruments. The experiments are conducted with binocular stereo vision and universal robot, and the results show that the system has a good sorting accuracy.