| Machine vision technology involves many disciplines, such as neurobiology,computer, image processing, etc. With the advantages of high efficiency, high precision,non-contact and easily integrated, it is the foundation of modern integrated manufacturingtechnology. With the application to wheels production, machine vision technology canlargely improve the level of the production process of mechanization and intelligent.Firstly, the article analyzes existing problems for use of on-line recognitiontechnology on wheels production, on the basis of machine vision technology, complexesimage processing, data analysis and pattern recognition, treats different types, materials ofwheels as the research object, on the use of photoelectric switch, motion control card andother hardware to get images, use Hough transform and other algorithms to achieve imageprocess, image segmentation and feature extraction. On basis of Visual C++6.0, the Intelopen source library OpenCV is adopted for image processing in this identification system.Secondly, uses Median Filtering to deal with the noise of images. During the processof casting, there will appear large or small burr at the edge of wheel blanks, which has abad influence on the analysis and identification of wheel images. So mathematicalmorphology is used to smooth edge of wheels and fill gap caused by external interference.Lastly, wheel’s position information is got by calibrated camera, manipulator isguided by motion control card to operate on wheels, so automatic delivery, drilling,handing and stacking is realized.Wheels on-line identification system research can be well adapted to the massproduction of wheels. Realize the fast online recognition positioning, overcome thedrawbacks of manual handling, clamping and sorting in wheels production process.Research on machine vision system for shape recognition in this paper, makes a beneficialattempt and exploration for the realization of the industrial production of modernidentification techniques and methods. |
PDF Full Text Request