Traffic signs are integral elements of any transportation network; however, keeping records of those signs and their condition is a tedious, time-consuming, and labor-intensive process. As a result, many agencies worldwide have been working toward automating the process. One form of automation uses remote sensing techniques to extract traffic sign information. An algorithm is proposed that can automatically extract traffic signs from mobile light detection and ranging data. After the number of signs on a road segment has been determined, the coordinates of those signs are mapped onto the road segment. The sign extraction procedure involves applying multiple filters to the point cloud data and clustering the data into traffic signs. The proposed algorithm was tested on three highways located in different regions of the province of Alberta, Canada. The segments on which the algorithm was tested include a two-lane undivided rural road and four-lane divided highways. The highway geometry varied, as did vegetation and tree density. Success rates ranged from 93% to 100%, and the algorithm performed better on highways without overhead signs. Results indicate that the proposed method is simple but effective for creating an accurate inventory of traffic signs.

No abstract

Horizontal curves are designed to provide a safe and smooth transition between straight segments on a highway network. Although curves are often designed to meet very stringent standards, imperfections during construction and high operating speeds mean that they are still prone to collisions. Therefore, it is essential that attributes of curves are surveyed to ensure they meet design requirements. Moreover, knowledge of the locations of horizontal curves and their attributes is also required to provide drivers with accurate information in advanced curve-warning systems, which are expected to enhance safety. Unfortunately, conventional techniques to obtain information about horizontal alignments are extremely tedious and, in some cases, impractical. This paper proposes a method by which horizontal curves can be automatically detected and their attributes automatically measured on scans of the highways obtained using light detection and ranging (LiDAR) technology. The proposed method is tested on two different highway segments at the Province of Alberta, Canada, where LiDAR data were collected. Moreover, testing was also conducted using virtual highways with curves with known attributes generated in AutoCAD Civil 3D. The results show that the code is successful in detecting all curves on a highway segment; moreover, the attributes of those curves were estimated with a high degree of accuracy (average difference <3%).

No abstract

No abstract