http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
류한성,탁영봉,정태영 慶尙大學校 工科大學 自動化및컴퓨터應用技術硏究所 1996 自動化 및 컴퓨터應用技術 Vol.3 No.1
Alpha-numeric characters representing sizes, patterns and production dates etc. are engraved in dented and raised appearances in automobile tires. The existing methods for tire classification comprise the recognition of sizes, widths and tread patterns by human eyes or the reading of expensive bar-code labels for high temperature use. On the contrary, this work will realize the automatic classification by recognizing the engraved characters using the computer vision technique. Since the characters and backgrounds in tire images have similar grey-level values, it is not easy to recognize the characters. Since the one side of the character string is raised and the other dented, the captured images are changed on the angles of camera and illumination. Therefore it is requested to find out the input conditions for obtaining the optimum images. The input images are smoothed by the fuzzy MIN/MAX operators, and then the character string is searched by using the stochastic line histogram obtained. The existing operators are applied to the detected character string and the comparison among them determines the optimum edge detection operator, which enables the segmentation of the string into the discrete characters. The mask patterns is devised to connect missing pixels in the discrete characters. And then We will extract the features from the retriered discrete characters, and apply them to the input of a neural network for the recognition of the tire characters.