Modeling of a mobile setup by networks for object contouring

Abstract

A model of a mobile setup for shape reconstruction is presented. In this technique, the contouring is performed by laser scanning and image processing. The devices of this setup can be moved to examine occluded areas and small details. The modeling of the mobile setup is carried out by an approximation network, which produces an equation that computes the object dimensions from the deformation of a laser line and the camera position. Also, the geometric parameters of the optical system are deduced, based on the network and image processing. This model improves the performance and the accuracy of the reconstruction system, because all steps of the contouring are performed by computer algorithms. In this manner, the errors of the physical measurements are not passed to the reconstruction system. This procedure represents a contribution to the methods of contouring based on laser scanning. To describe the accuracy, a root mean square of error is calculated with respect to a contact method. This technique has been tested with real objects, and experimental results are presented. Also, the processing time is discussed.