基于机器视觉的汽车焊缝检测系统

Research on Automotive Weld Inspection System Based on Machine Vision

针对汽车焊缝工件检测精度低、鲁棒性差和效率低等问题,设计一种基于机器视觉的车辆焊缝检测系统,代替现有人工焊缝检测。利用VS2010开发出界面程序,通过配置工业相机,实现多个相机对汽车焊缝点图像的获取,对当前图像进行一系列预处理,使用标定工件对焊缝图像进行尺寸标定和校正,并采用改进的Hough变换检测算法对图像进行焊缝检测。实验结果表明,系统测量精度为0.006 mm,测量方差量级为10^(-5),系统可以对汽车焊缝进行精准、稳定和快速地测量,并且可以将检测结果导出到对应焊缝检测点的Excel电子表格中保存。

Aiming at the problems of low accuracy,poor robustness and low efficiency of automotive weld work-pieces,a detection system for vehicle seam based on machine vision is designed to replace the existing manual weld inspection.An interface program is developed by using VS2010 through the configuration of industrial cameras to achieve the car weld seam image acquisition from a number of cameras.A series of pre-processing of the current images is realized,the weld size image is calibrated and corrected by using the calibrated work-piece,and the weld inspection of the images is performed through the improved Hough transform detection algorithm.The experimental results show that the system’s measurement accuracy is 0.006mm and the measurement variance is 10-5.The welds of automobiles can be measured accurately,stably,and quickly.The inspection results can be exported to Excel and saved as the electronic table corresponding to the inspection points of the welds.