Three-dimensional object recognition using pseudo-random color-encoded structured light.

Abstract

3D surface detection and measurement is one of the primary tasks involved in machine vision. Visual sensing of 3D position parameters can be done passively or actively. A method for 3D surface detection and measurement using a projected grid of light is presented. A novel technique using color projected grids is also explained for the detection and extraction of 3D objects in any colored environment. The "point identification" problem is solved using colored pseudo-random multi-valued sequence encoding. A new technique of finding grid line edges instead of its skeleton is also proved to increase the resolution of the extracted object and obtain additional information on its depth from the camera.