Manhole covers, which are a key element of urban infrastructure management, have a direct impact on travel safety. At present, there is no automatic, safe, and efficient system specially used for the intelligent detection, identification, and assessment of manhole covers. In this work, we developed an automatic detection, identification, and assessment system for manhole covers. First, we developed a sequential exposure system via the addition of multiple cameras in a symmetrical arrangement to realize the joint acquisition of high-precision laser data and ultra-high-resolution ground images. Second, we proposed an improved histogram of an oriented gradient with symmetry features and a support vector machine method to detect manhole covers effectively and accurately, by using the intensity images and ground orthophotos that are derived from the laser points and images, respectively, and apply the graph segmentation and statistical analysis to achieve the detection, identification, and assessment of manhole covers. Qualitative and quantitative analyses are performed using large experimental datasets that were acquired with the modified manhole-cover detection system. The detected results yield an average accuracy of 96.18%, completeness of 94.27%, and F-measure value of 95.22% in manhole cover detection. Defective manhole-cover monitoring and manhole-cover ownership information are achieved from these detection results. The results not only provide strong support for road administration works, such as data acquisition, manhole cover inquiry and inspection, and statistical analysis of resources, but also demonstrate the feasibility and effectiveness of the proposed method, which reduces the risk involved in performing manual inspections, improves the manhole-cover detection accuracy, and serves as a powerful tool in intelligent road administration.

Over the past decade, land-based laser scanning technologies have been actively studied and implemented, in response to the need for detailed three-dimensional (3D) data about our rural and urban environment for topographic mapping, cadastral mapping, and other street-level features, which are difficult and time consuming to measure by other instruments. For rural areas in China, the complex terrain and poor planning limit the applicability of this advanced technology. To improve the efficiency of rural surveys, we present two SSW (Shoushi and SiWei) laser scanning systems for rapid topographic mapping: stop-and-go and electro-tricycle laser scanning systems. The objective of this paper is to evaluate whether laser scanning data collected by the developed SSW systems meet the accuracy requirements for rural homestead mapping. We investigated the performance of the two laser scanning systems on Ma’anshan Village, a small, typical village in Hubei Province, China. To obtain full coverage of the village, we fused the stop-and-go and electro-tricycle laser scanning data. The performance of the developed SSW systems is described by the results of building contours extracted from the fused data against the established building vector map.