A Mechanical Part Sorting System Based on Computer Vision

In order to sort out different parts, computer vision is usually used to recognize parts with features on the sorting production line of mechanical parts. As for the core of machine vision, the design of computer vision software, VC++ program method is usually adopted to develop it. The developers need to encapsulate a special class for image process. And methods of pattern recognition are used to recognize object, whose development efficiency is very low. It is difficult to solve the problem of code optimization and the development efficiency of the method. In this paper the system of combining specialized machine vision software platform (e.g. HDevelop) and current development tool VC++ is proposed to recognize and sort out mechanical parts on the production line, so as to greatly improve development efficiency and running speed of the system. In the end of this paper, an experimental example is given to sort various mechanical parts in image by calculating circularity and area features and using recognition rules.