A Synergic Photometric Stereo and Super Resolution Approach for Optical Inspection

Abstract

Three-dimensional reconstruction is a very important technique of Computer Vision that produces a three-dimensional (3D) model of a real scene. For this purpose, we use Photometric Stereo that allows you to shape estimation from several images under different lighting conditions. In the last few years, this technology has been used in many areas both working and scientific. An important sector where Photometric Stereo is used is quality control inside productive cycle. In some case, the model 3D is used to identify defects or to make measurements of the order of hundredths or even thousandths of millimeters. To obtain an accurate 3D model, it is necessary to use an cost-effective reconstruction system with a performance hardware. Alternatively, we propose a method that combine Photometric Stereo with Super Resolution that produces a very accurate model using cheap and lower performance hardware that is suitable for optical inspection. Indeed, the Super Resolution allows to overcome the hardware constraint and produce one high resolution image from a set of low resolution images. With this novel method, we obtain a very accurate 3D model that could be in a decision support system for optical inspection.