Visual Input Amplification for Inspecting Specular Surfaces

This paper presents a method to simplify the inspection of specular surfaces by means of amplification the system input. It is based on sensor amplification by means of environmental conditioning and on the use of knowledge bases that store capture conditions. The model is intended as an advance in the conception of realism in vision involving surfaces that make the process difficult due to reflections and shine. This method offers general solutions as opposed to the use of heuristics relating to the characteristics of the objects. Experimentation has been carried out using a realistic simulator and the results prove the success of the amplification of the sensor by means of environmental conditioning for different scales, angles and lighting conditions.