Hyperspectral Imaging for Autonomous Inspection of Roads Pavement Defects

EasyChair Preprint 1055

Abstract

Autonomous inspection of roads is gaining interest to improve the efficiency of road repair and maintenance. In this paper we will be showing the potential for using Hyper Spectral Cameras, HSC, to identify road defects. The key idea of this paper is that cracks in the road show the interior material of road pavement which have different chemical composition from the surface materials due to surface wear. Material changes of the road surface give rise to a spectral signature that can be easily detected in HSC images. This condition facilitates the detection of cracks and potholes, which can be difficult if working in the visible spectrum domain only. We report on experiments with a HSC to identify the road material changes and their association to cracks and potholes. A new metric is devised to measure the amount of metal oxides and associate its absence to the appearance of cracks. The metric is shown to be more discriminative than previous indicators in the literature.

Keyphrases: Autonomous road inspection, Hyperspectral imaging, Pavement defect inspection, Road Crack Detection, hyperspectral remote sensing, metal oxide, remote sensing, road condition, road pavement

@booklet{EasyChair:1055,
  author    = {Mohamed Abdellatif and Harriet Peel and Anthony G Cohn and Raul Fuentes},
  title     = {Hyperspectral Imaging for Autonomous Inspection of Roads Pavement Defects},
  howpublished = {EasyChair Preprint 1055},
  year      = {EasyChair, 2019}}