基于视觉的涂胶质量检测方法

Quality Inspection Method for Glue Spreading Based on Vision

传统涂胶质量检测方法需要从标准涂胶图片中获取数据作为质量检测的参照,标准涂胶图片的质量会对检测的精度造成很大影响。针对该问题,文章提出了一种无需标准涂胶图片的视觉涂胶质量检测方法。通过对视觉系统的标定以及对工件的处理和分析获得世界坐标系、图像坐标系和工件坐标系的转换关系;利用工程图中的涂胶轨迹数据,生成了图像坐标系下的涂胶轨迹,并作为标准涂胶轨迹;在标准涂胶轨迹上选取采样点,提出了采样点处胶线宽度和胶线位置的计算方法。实验结果表明,该检测方法不需要从标准涂胶图片中获取数据,利用该方法进行检测的胶线宽度误差小于0.06mm,胶线位置误差小于0.22mm,能够满足工业现场的检测需求。

In traditional glue spreading quality inspection method, it is essential to obtain data from the standard gluing image as a reference, the quality of the standard image will have a great influence on the inspection accuracy. To solve this problem, a no standard image inspection method for glue spreading quality based on machine vision is proposed. Through the calibration of the vision system and the processing and analysis of the workpiece, the transformation relation between the world coordinate system, the image coordinate system and the workpiece coordinate system are obtained. By transforming glue track data in engineering drawing to the image coordinate system, the standard glue track is generated;the algorithm of the glue line width and position at the sampling points which are selected on the standard glue track is proposed. The experimental results show that the inspection method doesn′t need to obtain data from the standard glue image. The width error inspected is less than 0.06 mm, and the position error inspected is less than 0.22 mm, which can meet the inspection requirements of industrial sites.