基于图像识别的面板边缘检测分拣装置的设计

Design of Panel Edge Detection Sorting Device Based on Image Recognition

针对目前大板玻璃切割断裂后的单片面板的边缘质量检测以人工为主,效率低下的现象,设计了一套基于图像识别的面板边缘检测分拣装置。通过设计面板边缘检测分拣装置的整体方案、机械结构、控制系统的功能及控制流程,采用两条单片拾取处理机构A线和B线组成下料分拣检测部分,4台机器人双线程同时对前道工序输出的单片面板进行拾取检测,运行速度快,精度高,稳定性好,提高了工作效率。采用线阵相机实时地拍摄高精度的图像,对目标边缘像素采用最小二乘法拟合比较,建立模板和误差响应函数对面板边缘与拟合结果滤波后统计误差,可有效地检测单片面板的边破和角破,并将不良品分拣处理。实验结果表明该系统能够提高单片面板的边缘检测效率,降低人工成本。

Aiming at the phenomenon that the edge quality detection of the single-panel panel after the large-plate glass cutting is artificial and the efficiency is low, a panel edge detection sorting device based on image recognition is designed. Through the design of the panel edge detection sorting device’s overall scheme, mechanical structure, control system function and control flow, two single-chip picking and processing mechanisms A and B lines are used to form the blank sorting detection part, and the four robots are dual-threaded simultaneously. The single-panel panel output from the previous process is picked up and detected, and the running speed is fast, the precision is high, the stability is good, and the work efficiency is improved. The linear array camera is used to capture high-precision images in real time, and the least squares fitting and comparison are performed on the target edge pixels, and the statistical error of the template and the error response function on the edge of the panel and the fitting result is filtered, which can effectively detect the single-panel panel. Broken edges and corners, and sorted the defective products. The experimental results show that the system can improve the edge detection efficiency of single-panel panels and reduce labor costs.