The rapid location and extraction of urban municipal manhole covers is one of the key problems to be solved in the component management of smart city, and also the key technical problem to be solved in the realization of intelligent management of urban road maintenance. Therefore, we propose an automatic location and extraction method for manhole covers based on mobile mapping system. The ground point cloud is extracted from the surface roughness, the ground points are generated into the ground intensity feature images, and rapid location and extraction of manhole covers are achieved based on the combination of three methods, such as histogram of oriented gradient algorithm, principal component analysis and support vector machine classifier. We obtain the ownership information of manhole covers with synchronous images. Qualitative and quantitative analysis of the method is carried out based on experimental data, the accuracy of automatic extraction of urban manhole covers is 88.495% and the accuracy is 99%. The results also verify the correctness, feasibility, and robustness of the proposed method. This method further promotes the improvement of intelligent working mode of road maintenance department, which not only reduces the danger coefficient of work, improves the detection accuracy, but also transforms the traditional field test into indoor operation mode, and the degree of automation and intelligence increases greatly.