基于图像识别的生产线零件分拣系统

摘要

针对工业领域的生产线零件分拣系统进行研究，其零件识别采用图像识别方法。主要针对工件图像的预处理进行了研究，使用加权平均的滤波方法对工件图像进行去噪处理，使用基于双峰法的阈值分割方法对去噪图像进行分割，并使用凯西算子提取图像中的工件边缘。同时对工件位置的定位方法进行了研究，为控制机械手抓取零件提供依据，用三菱公司的RV⁃4FL工业机器人进行工件分拣实验，使用上位机的Matlab软件实现工件零件的图像处理、识别和定位。%The sorting system of parts in the production line in the field of industry is researched in this paper,in which the image recognition technology is used to recognize the parts in production line. The study on pretreatment of workpiece images is the important point of the paper. The method of weight average filtering is adopted to achieve on the denoising processing of workpiece images. The threshold segmentation method based on the two⁃peak method is used to segment the denoised image. The Casey operator is employed to extract the workpiece edge in the image. The positioning method of workpiece location is studied to provide basis for controlling the manipulator to grasp parts. PC Matlab software is used to realize image processing,identifica⁃tion and location of workpieces. The workpiece sorting experiment was conducted with sorting robot(Mitsubishi RV⁃4FL industri⁃al robot).