MEASUREMENT AND EVALUATION OF CRACKS ON ROAD PAVEMENTS FOR DAILY INSPECTIONS BY LOCAL GOVERNMENTS

Abstract

This study develops a system for measuring and evaluating cracks on road pavements, in order to support their daily inspections in local governments. The system consists of three units. The first one is a measurement unit for recording images of road pavements, together with information of GPS. Characteristic points of the unit are that its cost is relatively low and it can easily be installed/detached to standard cars. The second one is an analysis unit of the recorded road images and it divides a road image into blocks, and computes its crack ratio by evaluating cracks inside those blocks using two neural networks. The third one is a visualization unit of the evaluated results. The unit shows crack ratio of each road image by a color point on the map, whose location is identified from GPS data and, when cricking a color point on the map, it can also show an evaluated road image, which is an origin of the color point.