目的 针对烟支卷制完成后出现的水松纸质量缺陷等问题，结合生产实际要求，设计基于PLC和机器视觉的水松纸在线检测系统。方法 采用线阵相机对图像进行采集，工业计算机利用Vision-Pro图像处理软件进行图像处理，采用SIFT改进算法简化图像处理过程，实现对水松纸质量的在线检测，同时完成参数设定、状态监控和数据记录等功能。结果 实际运行结果表明，成品烟合格率由92%提高到100%。结论 系统运行可靠，可与现有系统配合实现在线监测，减少了原材料的浪费，提高了成品烟支合格率。

The work aims to design the online detection system for tipping paper based on PLC and machine vision, in combination with the actual production requirements, regarding such problems as quality defect of tipping paper occurring after the cigarette production. Line-array cameras were used to collect images. Industrial computer processed the images with the image processing software Vision-Pro and simplified the image processing course with the improved SIFT algorithm. The quality of tipping papers was detected online. Meanwhile, parameter setting, condition monitoring, data recording and other functions were completed. The actual running results indicated that, the percent of pass of the finished cigarettes was improved from 92% to 100%. The reliably operated system can collaborate with the existing system to achieve the online monitoring, which reduces the waste of raw materials and improves the percent of pass of the finished cigarettes.