Study on a Fast Visual Recognition Method of Workpiece in the Context of CIMS

Abstract

In order to increase the flexibility and automation of manufacturing systems, a fast visual workpiece recognition method is proposed in this paper. The image of a workpiece on a production line is taken by a video camera, and the image is then coded. The coding results, called coding templates, are put into a BP neural network for workpiece recognizing. The network has been trained using coding templates of workpieces calculated directly from their STL files, which are taken from their CAD designs. In this way, the visual recognition method is easily integrated into computer integrated manufacturing system (CIMS). Experiments on recognizing different workpieces using the proposed method have been conducted with satisfactory results.