基于机器视觉的硅钢钢带表面缺陷自动检测识别研究

Research on surface defects detection and identification for silicon steel strips based on machine vision

硅钢钢带是变压器等工业设备发展的一项重要原材料,其质量的高低直接影响着产品性能的好坏。传统的人工检测具有效率低、精确度差等特点。为此提出基于机器视觉的表面缺陷自动检测研究。该研究采用图像处理及模式匹配的方法,通过对由CCD工业相机采集到的图片进行几何矫正、图像拼接、缺陷处理等过程,实现了硅钢钢带表面缺陷轮廓检测、特征提取、分类等功能,从而完成钢带质量的判定。

Silicon steel strip is an important raw material in the development of transformer and other industrial devices,and its quality of high and low is very important. The traditional artificial detection is low efficiency,poor accuracy,so this paper puts forward a research on automatic surface defect detection based on machine vision. This study adopts the method of image processing and pattern matching,and through geometric correction,image matching,defect processing on images collected by a CCD industrial camera,to realize the silicon steel strip surface defect detection,feature extraction,classification and other functions,so as to complete the determination of the quality of steel strip.