| With the improvement of production technology, the traditional manual detection method for medicinal glass-bottle has long been unable to meet the needs of large-scale modernized production. In recent years,researchs based on machine version technology have been a hot topic duing to its huge advantages such as high accuracy, quickly detecting speed and out of contact inspection. This paper first introduces the principle of machine vision technology and the development process in our country and abroad; then with focusing on the characteristics of whole medicinal glass-bottle inspection system, a new system framework is proposed,and the suitable algorithms are exactly chosen and programed to deal with problems; finally,this paper uses DSP technology as a key alternative and acheives some good results.The main contributions of this paper are as follows:1,The structure of the inspecting system was introduced, the hardware choice was analysised and fixed in detail,and the physical experimental system and research ideas were given.2,The algorithms of whole system were studied deeply. Carry out experiments of various image filtering, segmentation and edge detection algorithms with analysis and comparisons on PC platform. Select an feasible and advanced image processing operation flow embracing mathematical morphology filtering, OTUS segmentation, improved morphological edge detection,etc. Calibrate the circle center of top bottle image by Hough transform,and calculate the bottles'size .3,The image processing algorithms based on PC were realized on SEED-VPM642 successfully, and the real-time detection of dynamic images was discussed preliminary. The software design,focusing on the demand of inspection and the characteristic of DSP was accomplished and a detailed description of the program is provided.4,The system GUI was designed, the system error factors and the whole component experimental results are analysized, the instructions of decreasing error for future research are proposed. |
PDF Full Text Request