Simulation of Specular Surface Imaging Based on Computer Graphics: Application on a Vision Inspection System

Abstract

This work aims at detecting surface defects on reflecting industrial parts. A machine vision system, performing the detection of geometric aspect surface defects, is completely described. The revealing of defects is realized by a particular lighting device. It has been carefully designed to ensure the imaging of defects. The lighting system simplifies a lot the image processing for defect segmentation and so a real-time inspection of reflective products is possible. To bring help in the conception of imaging conditions, a complete simulation is proposed. The simulation, based on computer graphics, enables the rendering of realistic images. Simulation provides here a very efficient way to perform tests compared to the numerous attempts of manual experiments

Simulation of Automated Visual Inspection Systems for Specular Surfaces Quality Control

“…The ratio between the number of points in the difference image (9) and the number of points which would represent defects T(m,e,c) provide the efficiency E of transformation ϒ μ . We use this calculation as a measurement of sensitivity of the system, of the capacity for perceiving defects, and it will be stored in a knowledge base BD δ (11). We have called BD δ sensitivities database.…”

“…Lighting of the scene is a determining factor as it is possible to facilitate the detection of defects in images [8]. Structured lighting techniques are considered reliable and adequate for inspection of the surface shape and are presented in the few systems of the literature [9][10] [11] [12] [13] [14] [15]. In addition, the use of the same sensor to determine reflectance is another argument for this type of technique (regularly it should be acquired in an independent manner).…”

“…Nevertheless, the choice of different patterns has not been considered, nor are other factors taken into account which would affect the effectiveness of the system such as the angle of capture, zoom, focus, etc. [11][21] [22].…”

Active lighting applied to three-dimensional reconstruction of specular metallic surfaces by polarization imaging

“…12(a)]. Previous work carried out by Seulin et al 19 was efficient in discriminating geometric aspect defects thanks to dynamic lighting: when a fringe pattern was projected, defects on a smooth surface were revealed near the transition between dark fringes and bright fringes. Nevertheless, this system is not adapted for inspecting shape defects on the object's decorations.…”

An automatic procedure to identify the areas of interest for the automated inspection of headlamp lenses

“…Then, by computing the object silhouette, it is possible to match the predefined mask on the object shape. This shape fitting could be realized by scaling, translating and rotating the reference mask to make it matching the current object shape [4], [7].…”