Research On The Online Labeling System Based On The Machine Vision

The rapid development of modern manufacturing，packaging，logistics and otherindustries， the continuous improvement of the domestic commercial capital andmodernization’s degree，and the increasingly mature of the storage of sales，centraldistribution mechanism， timely delivery， information management throughout themarketing process and real-time information collection，all of them proposed higherrequirements to the industrial automation．The propulsion of the industrial automation hasplayed a decisive role in cost saving and improve efficiency．However，today，thetraditional robot used in the production is becoming more difficult to meet the demand inwork efficiency，accuracy of operation，etc．This prompted a breakthrough with thedevelopment of image processing，artificial intelligence，robotics and other moderntechnology fields．Machine vision technology as an important part has become a researchhotspot．At the same time，labeling system has become more and more important inmodern industrial production．Currently，most of labeling systems are using photoelectricsensors and mechanical motion mechanism complete the labeling action together．Thisneed manual inspection, and it is inefficiency， high cost and its accuracy is notenough．This paper is the study of the theory of machine vision technology and applied itto the online labeling system．The main contents of this paper are as follows：1. Illustrates the background and significance of the research which applies themachine vision to the online labeling system，and analysis the domestic and foreign presentsituation of the machine vision，as well as its existing problems．2. Designed an online labeling system with a CCD array camera．Introduce the visualsystem in the selection，source system design principles，the selection and tectonic of thecontrol module’s movement mechanism and driving device．Has completed the prototypedesign and running stably．3. Illustrates the details of the camera model’s coordinate system，which is constitutedby world coordinate system，camera coordinate system and image coordinate system，andanalyze the relationships and conversion between the three coordinate systems．Describesseveral methods of camera calibration．analyses the details of the camera calibrationtechnique’s mathematical model，and calculate the internal and external parameters of thecamera by Zhang’s calibration method． 4. Introduced several image preprocessing techniques，analyzed each algorithm of thepreprocessing techniques，and described a simple method which is for this system to detectthe target object．5. Implements the system’s every modules Using the image processing methods whichdescribed in this article and based on OpenCV an open source computer visionlibrary．Described and analyzed the running results, and explain the current lack of existingsystem and how to improve it in the future．...