Image-Based Approach for Road Profile Analyses

摘要

Road profile extraction and analysis are essential tasks in transportation asset management. Current approaches use vehicle-borne laser sensors in order to precisely measure the variations in elevation along a specific route. However, a complicated sensor mechanism, such as in the mobile light detection and ranging (LiDAR) system, is involved and the resulting quality is compromised owing to multiple factors. In this study, an image-based approach for extracting road profiles is proposed. It requires only a single camera sensor and a low-cost laser module and is capable of collecting road profiles along both the longitudinal and transverse directions. A detailed methodology is first presented in this paper, followed by a simulation evaluation and a case study. The case study illustrates that the quality of the extracted profiles based on the proposed approach achieves millimeter accuracy. Consequently, an accurate and cost-efficient road profile analysis becomes possible when the proposed approach is implemented.