Vision-Based Robotic Solution for Wire Insertion With an Assigned Label Orientation

Abstract

This paper tackles the problem of wire insertion in switchgear assembly according to the current regulations. In particular, the wire connections require that the wire label has to be oriented facing up in order to simplify and speed up testing and maintenance of the switchgear. The proposed approach exploits the a priori knowledge of the scenario with a calibrated RGB camera and a robotic arm to estimate both wire end pose and label position. The procedure combines several techniques (gradient base, trained classifier and stereo vision) to elaborate standard images in order to extract some wire features related to its shape and label. Specific frames are fixed according to estimated features and then used to correctly complete the task by using a robotic system. Experiments are reported to verify the effectiveness of the proposed approach.