基于机器视觉的弹簧承载座缺陷检测研究

The study of spring bearing shell defect detection based on machine vision

针对弹簧承载座人工检测效率低、劳动强度大等问题,对弹簧承载座缺陷种类及缺陷特征进行了研究,并对其缺陷检测流程进行了分析归纳,提出了一种基于机器视觉的弹簧承载座缺陷检测方法,实现了缺陷检测自动化。通过中值滤波算法对图像进行预处理,去除图片噪声;提出了一种两步找圆的方法,确定检测区域尺寸,再通过对检测区域进行掩膜操作,实现弹簧承载座焊渣和缺口缺陷检测;通过基于边缘点的模板匹配算法来判断字符的完整性。实验结果表明,以上方法能够准确判断弹簧承载座尺寸、焊渣、缺口以及字符等缺陷,检测正确率达到98%。

Aiming at the problems of low detection efficiency and high labor intensity of the spring bearing shell, the types and defects of the spring bearing shell are studied, and the defect detection process is analyzed and summarized.A spring bearing shell detection algorithm based on machine vision is proposed to realize the automation of defect detection.The image is preprocessed by the median filtering algorithm to remove the picture noise.A two-step circle location method is proposed to determine the size of the detection area, and then the masking operation is performed on the detection area to realize the detection of the slag and notch defects of the spring bearing shell.The character integrity is determined by an edge-based template matching algorithm.The experimental results show that the algorithm above can accurately determine the size of the spring bearing shell, welding slag, notch and characters, and the detection accuracy reaches to 98%.