| Industry online dimension measurement is one major application direction for Vision system development, and mobile objective and edge detection is two key technical this application. General mobile objective detection use difference arithmetic and for edge detection use differential coefficient zero point judge method. Online dimension measurement require measure product outlook dimension on pipe-line accurate and feedback quickly, with this request these arithmetic are not suitable because can't reach accuracy and also will cost much system source.Industry online measurement has its individual feature: the objective's track is simple and fixed with a stable background, it need high accurate and real-time feedback. Also need precise and real-time edge detection, based on above result, system is able to get dimension result according to calculation on detected edge coordinate.This paper is based on design and setup a real production online vision system project, it introduce project background and request, select key part of vision system and define their parameter. Make analysis on if need complicated mobile objective detection arithmetic, get a result with simple system include a counter and an encoder to control camera system take photo and get system need photo. Late we design a new edge detection method based on statistical knowledge, this new test method will overcome some noise and also combine overall analysis and part analysis, it will find some edge points area and use a optimized-vote method to get a final and closest edge.Final, setup the whole system and integrated all parts for adjustment. Improve real-time feedback system, compare operator measure data to vision system's. Pass feasible phase, summarize the effect on improvement for industry production. |
PDF Full Text Request