3D reconstruction system by means of unique camera, structured light and mathematical models

Abstract

3D digital profiles are key information in many areas of knowledge, from medical to agricultural or industry. And in order to achieve reliable digitized surfaces we can have in hand techniques using laser, digital cameras, structured light associated to image processing tools. To choose the best technique we face some challenges based on the application itself, but on the cost of equipment, on the robustness and on the computational cost and programming complexity. This work aimed to address a reliable, robust and accessible methodology using just one camera and a light projector to, with mathematical models, generate digital 3D surfaces. The mathematical model was designed using the geometry of the system (camera, projector and object) with additional routine to adjust the optimum values of the variables involved. We validated the proposed method with known objects and presented a 3D reconstruction of objects that are difficult to achieve manually, such as a callote and the back of human-being. The results showed that the mathematical model created conditions to provide reliable outcomes from a low cost devices (webcam and regular light projector) usage, with correction of warping caused by perspective.