Automation of beam straightness measurement

Abstract

This work deals with the development of an image processing methodology to be employed in an inspection machine for measurement of straightness of moving steel beams. The method consists of developing specific software to perform three main tasks: reading the position of the beam at several locations; correlating the data monitored by the cameras; and reporting the quality of the beam straightness. Computer vision technology as a non contact measuring system is used by the inspection machine. From the production process the beams roll onto a conveyor therefore several video cameras are needed to log information about the lateral position and also the height of the steel beams. The number of cameras and the distance between them can be defined by the user. The technique is developed in such a way that if one camera fails or is switched off the measuring system is not affected as long as the system has a minimum of three cameras viewing acceptable points. The system outputs is the deviations from a perfectly straight beam; compensating factors for lateral or angular movements in the horizontal measuring plane of the beam and the beam profile. The output is displayed on a computer screen, and is based on both linear and least squares curve fitting routines.