针对FPC(柔性印刷电路板)缺陷的特点,设计了基于机器视觉的嵌入式检测系统;该系统以线阵CCD为视觉传感器,DSP为处理器核心,完成图像的采集和存储;基于局部自动阈值分割后的二值化图像,采用多尺度的数学形态学抑制噪声,突显边缘细节;结合线宽评价和局部生长的图像匹配算法,准确实现产品缺陷定位识别功能;实验结果表明:系统能准确识别检测产品中存在的各种缺陷,图像检测速度达到了每秒20帧,具有良好的实用性。
Considering the defect characteristics of flexible printed circuit, this paper develops an embedded architecture visual inspection system based on machine vision, The system uses linear CCD sensor and DSP processor to achieve image acquisition and storage, and uses multi--scale morphology to restrain noises and highlight the image edge. The image defects can be located and identified by using image matching based on local growth and line width evaluation. The results show that: the system can automatically and accurately identify the products defects, with the detecting speed of 20 frames per second, and has a good practicality.