基于机器视觉的印刷电路板检测及分拣系统

PCB detection and sorting system based on machine vision

用机器视觉代替人眼能够进行缺陷检测;用机器人代替人力进行分拣,可保证分拣的效率、稳定性和可重复性。文中结合二者优势,提出一种机器视觉的印刷电路板检测和分拣系统,以达到实时检测和快速分拣的目的。该系统利用智能相机配合光源得到检测对象的图像信息。针对印刷电路板的缺件、漏件、排针弯曲、焊脚不规整等缺陷,分别采用不同的检测算法和图像预处理措施,编写检测程序来快速识别检测对象是否存在缺陷,并将缺陷信息和印刷电路板坐标提供给机器人以实现印刷电路板的分拣,保证缺陷检测的准确性。另外,在图像定位上采用软件配准以节约人力和时间。实验结果表明,所提系统可以满足印刷电路板的实时检测要求,达到较好的检测效果。

The machine vision can perform defect detection instead of human eye,and robots can replace manpower for sorting to ensure the sorting efficiency,stability and repeatability.By combing the advantages of the machine vision and robot,the PCB(printed circuit board)inspection and sorting system based on machine vision is proposed to achieve the real-time inspection and rapid sorting.In this system,the smart camera is used to get the image information of the detected object in conjunction with light source.In allusion to the defects of PCB such as missing parts,bent row pins,and irregular weld leg,different detection algorithms and image preprocessing measures are adopted to compile the detection program to quickly identify whether the inspection object has defects,and provide the defect information and its coordinates on PCB to the robots,so as to realize the PCB sorting and ensure the accuracy of the defect detection.The software registration is adapted in image positioning to save manpower and time.The experimental results show that the proposed system can meet the real-time detection requirements of PCB,and get better detection results.