Structured Light 3D scanners have released a stampede of successful applications in robotics. This opens up the possibility to identify subtle deficiencies of actual sensors and propose solutions. In this paper we present the design of a novel Structured Light scanner that defies the conventional and simplistic static cameras-projector configuration. We found that by adding some degrees of freedom to the cameras it is possible to improve the capabilities of the system, for example reducing the minimum operating distance and increasing the depth estimation accuracy. This multidisciplinary project involves the fields of mechanics, electronics, control, optics and computer engineering, and aims to design a powerful tool, for academic and industrial applications.

No abstract

In this paper we present the design of a biologically inspired active vision system to estimate the depth of objects in a scene. Our design considers geometric characteristics similar to that of human eyes and uses an algorithm that fovealizes the stereo images at points of interest or previous known objects. The proposed algorithm greatly reduces the computational time. We present preliminary simulated results.