Rectification Strategies for a Binary Coded Structured Light 3D Scanner

Making a computer able to see exactly as a human being does was for many years one of the most interesting and challenging tasks involving lots of experts and pioneers in fields such as Computer Science and Artificial Intelligence. As a result, a whole field called Computer Vision has emerged becoming very soon a part of our daily life. The successful methodologies of this discipline have been applied in countless areas of application and their use is still in continuous expansion. On the other hand, in an increasing number of applications extracting information from simple 2D images is not enough and what is more requested instead is to use three-dimensional imaging techniques in order to reconstruct the 3D shape of the imaged objects and scene. The techniques developed in this context include both active systems, where some form of illumination is projected onto the scene, and passive systems, where the natural illumination of the scene is used. Among the active systems, one of the most reliable approaches for recovering the surface of objects is the use of structured light. This technique is based on projecting a light pattern and viewing the illuminated scene from one or more points of view. Since the pattern is coded, correspondences between image points and points of the projected pattern can be easily found. In particular, the performances of this kind of 3D scanner are determined by two key aspects, the accuracy and the acquisition time. This thesis aims to design and experiment some rectification strategies for a prototype of binary coded structured light 3D scanner. The rectification is a commonly used technique for stereo vision systems which, in case of structured light, facilitates the establishment of correspondences across a projected pattern and an acquired image and reduces the number of pattern images to be projected, resulting finally in a speeding-up of the acquisition times.