基于机器视觉的双工位载带检测系统研究

叙述了基于机器视觉的双工位载带检测系统的硬件设计方案和机械结构方案。提出了适合载带检测的图像处理算法,该算法主要包括瑕疵检测算法与尺寸计算算法。将标准合格载带样本图像作为训练集,从这些样本图像中获取检测区域图像的特征量信息作为在线检测时的比较标准。经过实际上线测试验证,载带缺陷检出率均不低于99.6%,重要缺陷漏检率为0%;检测一帧图像时间10 ms,检测速度达到12 m/min。

Research on Duplex Carrier Detection System Based on Machine Vision

The hardware design and mechanical structure of the dual-station carrier tape detection system based on machine vision are described. An image processing algorithm suitable for carrier tape detection is proposed, which mainly includes the defect detection algorithm and the size calculation algorithm. The algorithm uses the standard qualified carrier sample image as the training set, and acquires the characteristics of the detection area image from these standard qualified carrier sample images. The quantity information is used as a comparison standard for online detection. After the actual line test verification, the detection rate of the carrier defect is not less than 99.6%, the critical defect detection rate is 0%. The detection time of one frame of image is 10 ms, and the detection speed reaches 12 m/min.