目的 为解决传统人工分拣作业效率低、速度慢、质量难以保证的问题。方法 设计一种基于机器视觉的包装物料自动筛选控制系统，适用于多种输送线物料分拣的方案。结合多种图像处理算法，以提高物料筛选精度。改进中值滤波可实现图像预处理；形态学滤波可实现包装物料边缘检测；基于边缘匹配可完成物料识别。最后进行试验研究。结果 试验结果表明，对于圆形、方形、三角形物料，所述包装物料筛选系统的分拣精度在99.5%以上，最大抓取速率能够达到150 次/min。结论 所述图像处理算法能够有效地提取物料边缘并实现物料识别，具有比较高的检测精度和稳定性，可以满足包装需求。

The work aims to solve the problems such as low efficiency, slow speed and poor quality of traditional manual sorting. An automatic screening control system of packaging materials based on machine vision was designed, which was suitable for sorting materials of various conveying lines. Multiple image processing algorithms were combined to improve the precision of material screening. The improved median filter could realize image prepossessing. Morphological filtering could realize edge detection of packaging materials. Material identification could be completed based on edge matching. Finally, experimental research was carried out. The test results showed that, for circular, square and triangular materials, the sorting accuracy of the packaging material screening system was more than 99.5%, and the maximum grasping rate could reach 150 times /min. The proposed image processing algorithm can effectively extract the material edge and realize material identification, and has relatively high detection accuracy and stability, which can meet the packaging requirements.