ABSTRACT
Road condition data are typically reported for several meter long diagnostic sections. However, in some applications, a higher localization accuracy is necessary. This paper reports on methods and results of finding exact localization of high-resolution road surface images for 100 km of a motorway, paved with 120 thousand slabs. Algorithms of image filtering and pattern recognition were used to find transversal and longitudinal slab joints on the surface pictures. Since the slabs were too wide to be captured in a single run of the measurement vehicle, the data from neighbouring runs were synchronized. Next, they were merged to cover the whole row of slabs and extract images of each of them with a centimetre accuracy. Finally, a geodesic map was used to ensure the correctness of the localization of the slabs. The proposed method can be used for both road and airport pavements.
