The main purpose of this paper is to propose a solution to remove errors due to spurious reflections which can occur during the scanning of shiny surfaces. As glass objects and semi transparent objects, shiny surfaces are one of the main problems in 3D scanners based on triangulation. Different techniques which deal with the same problem are investigated. One of these methods was chosen in order to be implemented and tested. This technique uses a 3D data cloud density. The data cloud density of a 3D point is all the points that remain in the close vicinity of that point. Furthermore, the basic 3D laser scanner system is described in order to understand the theory around the data cloud density method.
An implementation and performance tests were done to prove the relevance of this method in the detection and the suppression of errors in cloud of points created by 3D laser scanners. There is no doubt about the advantage this method has to find false points in a set of measured points. However, the velocity of this technique can be problematic in certain area.
A better implementation has to be found to conclude if this method can be used in industrial area or not.