Mapping road surface features using single-camera images acquired by a mobile mapping system

Mapping road surface features, such as manholes, traffic markings, and cracks, is an essential task for transportation facility management. Although, these features can be rapidly surveyed using the latest mobile mapping techniques, a sophisticated sensor system with a complicated post-processing procedure is usually required. In this study, an efficient framework for modeling road surface features is proposed using a single camera system installed on a moving platform. First, the road surface images along a route of interest are acquired and potential objects are identified based on their shapes and recorded spectra in the images. Then, the contour pixels of the identified objects are extracted by the Canny edge detection technique. Finally, the 3D coordinates of the detected features in object space are obtained by integrating the profile-image technique and the instantaneous exterior orientation parameters of the platform. Based on the numerical results from a case study, it has been demonstrated that a fully automatic and reliable extraction of road surface features can be easily achieved by implementing the proposed approach. Consequently, the modeling of road surface features, which essentially contributes to the management of transportation facilities, can be executed in a cost-efficient manner.