Illumination control expert for machine vision: a goal driven approach

摘要: An optimization approach to the automatic sensor and light source positioning for a machine vision task, where geometric measurement and/or object verification is important, proposed. The goal of task assumed be specified in terms measurements related edges. optimal positions are defined as which produce picture from we can obtain data with minimum variance. A method propagating uncertainty edge point position 2D discussed. Experimental evidence presented showing that inversely proportional contrast across edge. In order compute edge, an illumination model based on Torrance-Sparrow reflectance developed. consists two parts, one specular reflection other diffusion. intensity specularly reflected depends degree polarization incident light, because coefficient perpendicularly polarized(S-pol) plane incidence different parallelly polarized(P-pol). Since any wave represented sum orthogonal polarized pieces, may by intensities: S-pol component P-pol component. is, then, products each corresponding coefficient. To demonstrate our feasible acceptable practical use, software system called ICE designed implemented. series experiments, using variant takes into account has been conducted show found really positions.