Adaptive vision system for high velocity tooling machines

The work we present here is a diagnostic task, which must be solved for high velocity industrial tooling machines URANE-20. Due to environment degraded conditions, direct measurements are not possible, also for rapidity of the machine, human intervention is not possible in case of position fault. Therefore, an oriented vision solution is proposed. Degraded conditions are vibrations, dazzling, water and chips of metal projections. In this case, the once method cannot achieve a diagnostic problem: is it the right piece at the right place? That is why complementary methods presented in this paper are proposed in an adaptive way to solve this diagnostic problem. Image processing methods allow us to find image parameters. After a data analysis, image parameters are reduced. Then, using Bayesian approach and neural approach, it is possible to ensure the diagnostic result. With these two methods, we obtain encouraging results and we show that it is possible to improve the results by combining different classifiers approaches.