Physically-accurate synthetic images for machine vision design

In machine vision applications, accuracy of the image far outweighs image a ppearance. This paper presents physically-accurate image synthesis as a fle xible, practical tool for examining a large number of hardware/software con figuration combinations for a wide range of parts. Synthetic images cart ef ficiently be used to study the effects of vision system design parameters o n image accuracy: providing insight into the accuracy and efficiency of ima ge-processing algorithms in determining part location and orientation for s pecific applications, as well as reducing the number of hardware prototype configuration to be built and evaluated. We present results illustrating that physically accurate, rather than photo -realistic, synthesis methods are necessary to sufficiently simulate captur ed image gray-scale values. The usefulness of physically-accurate synthetic images in evaluating the effect of conditions in the manufacturing environ ment on captured images is also investigated The prevalent factors investig ated in this study are the effects of illumination, the sensor non-linearit y and the finite-size pinhole on the captured image of retroreflective visi on sensing and therefore, on camera calibration was shown; if not fully und erstood, these effects can introduce apparent error in calibration results. While synthetic images cannot Sully compensate for the real environment, t hey can be efficiently used to study the effects of ambient lighting and ot her important parameters, such as true part and environment reflectance, on image accuracy. We conclude with an evaluation of results and recommendati ons for improving the accuracy of the synthesis methodology.