Evaluation of road signs using radiometric and geometric data from terrestrial LiDAR

Maintenance works are crucial to reduce the risk of accidents. Road signs appear to be one of the most important elements for safety purposes so their inspection is commonly included in the most extended road management systems. The geometric state of the sign is of great importance, especially interesting is its flatness and the inclination relative to the ground. Road signs are printed with reflective paints to maximize the visibility to the drivers. This coating produces a high reflected radiation that is easily recorded by the photoelectric detector of the laser scanner mechanisms. It allows establishing an intensity based filter in order to perform the 3D classification of the road sign. In this work, a number of road signs are evaluated, under a geometric point of view, using the laser scanner Riegl LMS-Z390i. A Matlab algorithm is developed for all the data processing (3D classification and evaluation of geometric parameters). Results do not show the evidence of folded or abnormally tilted signs. The developed algorithms open the possibility of using the attribute of intensity of laser scanning data for classification purposes, during the automatic evaluation of the condition state of road signs. Such algorithms could increase the productivity and reliability of the inspection works.