Research On Grasping The Center And Identification Of Assembly Reference Terminal Of Tubular Workpiece Based On Machine Vision

In the process of assembly，the determination of grasping the center of workpiece and accurate identification of assembly reference terminal are the important preliminary parts of assembly based on machine vision. In order to make the assembly process towards fast, precise and intelligent direction, the packing of tubular workpiece in the complex environment as the object of study is researched. The machine vision system included two parts of gripper position determination of the packing of tubular workpiece and identification of tubular workpiece assembly reference terminal is established; the fitting algorithm of Gaussian fitting and Hough transform is presented according to reflective direct light characteristics reflected on the surface of tubular workpiece by the forward lighting system, to determine the capture center position of each tubular workpiece by fitting and extracting the light stripe center line on the surface of tubular workpiece; a method of identifying assembly reference terminal base on the improved Hu invariant and LM-BP neural network is presented according to the different shape characteristics of both ends, to realize the shape identification of both ends of tubular workpiece and judge the assembly reference terminal. The experimental results show that, the presented method of Gaussian fitting and Hough transform can effectively realize the intelligent achievement of the grasping position of packing tubular workpiece, and the method of improved Hu invariant and LM-BP neural network can have the good recognition ability to the two ends of tubular workpiece, providing a basis for determining the assembly reference terminal. The research has laid certain foundation for the research on the assembly technology based on machine vision.