The Design And Implementation Of The Cigarette Print Online Detection System Based On Machine Vision

Now, in the cigarette industry, cigarette making and plug assembling machineand packing machine mainly uses the infrared electric detection device, the forefrontof machine vision detection equipment are few at present. For cigarettes printdetection, cigarettes appearance detection, cigarettes roundness and girth detectionand cigarettes foil defect detection, precision reliability of photoelectric detectionequipment has seriously affect the quality of cigarette. For high-end brand ofcigarettes are not allowed to have any defect in the cigarette into the market.In the cigarette industry strongly require the use of machine vision detectionsystem at the same times, a problem that does not allow to be ignored. Cigaretteequipment internal structure is complex, and cigarettes are on the internal each axis ofrotation and produced on the components. Want to increase machine vision detection,but due to space size that can’t be available for installing general CCD camera, thus itis so difficult to achieve machine vision detection.In the condition of such a background, this topic has researched and developed asuitable integrator CCD industry camera for The Embedded Machine Vision. Theinterface between camera and controller without using the standard industrial busprotocol, but to design a custom parallel protocol, fully will merge into an organicwhole. It can narrow and change the appearance of the CCD size to infiltrate incigarette equipment, but also save the cost of buying a standard interface CCDcamera.The purpose of this design is to develop a set of custom parallel protocol,moreover, according to specific situations, research and improve the templatematching algorithm to meet the real-time processing requirement of high speed andhigh precision. In the cigarette industry production of cigarettes print detection, usingthe image template matching algorithm. Due to production of7000-10000cigarettesper minute, therefore system should be able to perform high speed detection, thesystem of real-time and accuracy requirement are very high. So in this paper, the mainimage processing algorithm is image template matching strategy based on the grayvalue, according to the actual project situation, some improvements of templatematching algorithm so as to meet the requirement of accuracy and rapidity of thesystem. According to the characteristics of the project, it can achieve high-acuracy andhigh-speed data acquisition by the use of DSP and CCD combination solution. Fastimage matching algorithm was adopted to realize high speed processing, to meetreal-time and accuracy, and complete a high-speed machine vision processing systemand apply to industrial testing industry. In the cigarette industry to realize intelligentmachine vision detection has a strong engineering background, and has importantapplication value and economic value.