A vision system capable of imaging and detecting defects on reflective nonp
lanar surfaces in the production line at a high cadence is presented in thi
s paper. Defects are typically dust located under the metallic layer of pac
kaging products used in cosmetic industries. To realize this processing, st
ructured lighting which reveals the defects in the image is proposed. Defec
ts appear clearly in the images like a set of brilliant pixels in dark zone
s. The signature of the defect is then obtained. The size of this signature
does not depend linearly on the size of the defect. It is a function of th
e observation angle. In order to realize a precise and robust process, the
necessity of acquiring several images of the same defect is demonstrated. B
ecause of the acquisition rate, it has been necessary to optimize image pro
cessing time by the means of an original laplacian filter and of high level
techniques of programming. Results obtained using this detection system at
e finally presented. (C) 2001 SPIE and IS&T.