With the current system of concession roads in Brazil, both concessionaires and regulatory agents have been charged with improving the quality of the pavement on Brazilian highways. This situation requires the search for new tools that can facilitate the survey of pavement conditions in less time and at a lower cost than traditional methods. Recently, an increasing number of high-resolution spatial images have become available on the world market following the development of new remote sensing satellites and airborne sensors. Similarly, multispectral images and even hyperspectral images are now available commercially and for scientific research. The paper presents a new methodology for the identification of asphalted pavement surfaces condition and the classification of the main types of asphalt defects using hyperspectral images from airborne digital sensors. The objective of this study was to generate indexes of pavement conditions from images that can be compared with the indicators of pavement surface conditions already used by national regulatory agencies.
pavements; remote sensing; asphalt; hyperspectral; classification