The Research On Visual Inspection System For Quality Of Plastic Combination Covers

In the modern automation production, the quality of almost all products needs tobe controlled in the process of producing, so manufacturers are growing investment intesting quality of products when they constantly improve the production efficiency.Yet artificial detection can’t meet the needs of requirements of high speed, highprecision and high stability in the production line, the requirements for high-speed,automation and health indicators are more strict, particularly in wine, beverage,pharmaceutical and so on. So applying machine vision technique to detect quality willbecome the inevitable trend.With plastic combination cover as the object for researching, the article designsintelligent detection system based on machine vision. It amply analysis the machinevision technology in detection for defects of plastic combination cover’s quality fromthe five aspects of mechanical transmission system, electrical control system, visualimaging system, image processing algorithms and software system designing.Firstly, this paper introduces the background and significance of the selectedtopic combine with the application of detecting in industry of machine visiontechnology at home and abroad, and it has made the simple elaboration based on thefeatures, development trend and application fields of the machine vision technology.Secondly, the article has made the detailed introduction in the typical models ofmachine vision system from light source, camera, lens, image processing system.Then, combining with technical difficulties need to be overcome of quality plasticcombination covers’ visual inspection system, the article has made detailed treatmentsfor machine driven system, electric control system, visual imaging system anddefective goods sorting device and so on, and analysis the group process and testingprocess of the system. Practice has proved the system is reliable and practicle.Visual detection algorithm of plastic combination cover quality is the key of thewhole detection system. At first, this paper uses the training-serial classificationdecisions for the target recognition of plastic combination cover; Then the articleadopts different target accurate positioning algorithm according to the differentgeometry and gray level characteristics of the object that need be tested such as cover,rubber plug, inside the cover and finished goods; Finally according to the differentinformation characteristics of object uses the corresponding detection and identification algorithm for defects. Experiment has proved the practicability of thesystem and the effectiveness of the detection algorithm.Finally, the article introduces the software system of the visual inspection systemfor quality of plastic combination covers. First article introduces the characteristics ofthe software system and analysis the various functional modules in detail; then thearticle has made the simple elaboration for multi-threading technology which thesoftware system uses; finally, the article makes a brief explanation for OpenCV whichimage processing algorithm adopts.