切膜机生产线上,需将膜切成各种规格的产品。但是在具体的操作过程中,其表面将产生诸多类型的缺陷,因此必须加以检测,例如,污渍、划痕等问题的检测。本文根据过去人工法存在效率相对较低的现实问题,阐明基于计算机视觉的自动检测设计方案,并且阐明优化的背景差法,这种方法对检测非常适用。经过验证,该技术非常有效、算法可行,能够快速检测其中存在的缺陷,在很大程度上改善了生产率。在明显缩短剪切时间的基础上,降低了生产原料的损失,极大地迎合市场需求。同时通过新的检测方法,以及均值滤波得到目标的背景,研究发现,其可以明显改善传统方法的不足。 Cut film machine production line, the film needs to be cut into a variety of products. However, in the specific operation, the surface will produce many types of defects, it must be detected, for example, stains, scratches and other issues detected. According to the fact that the artificial method has relatively low efficiency, this paper expounds the automatic detection design scheme based on computer vision and clarifies the optimized background subtraction method. This method is very suitable for the detection. Proven, this technique is very effective, the algorithm is feasible, can quickly detect the flaw in which, to a large extent, improve the productivity. On the basis of obviously shortening the cutting time, the loss of raw materials for production is reduced and the demand of the market is greatly met. At the same time, through the new detection method and the average filtering, the background of the target is obtained, and the research finds that it can obviously improve the deficiencies of the traditional method.