The 3D mapping of road and city scenes is useful and necessary for many professional and consumer applications. Active perception systems based on lasers were initially developed for their robustness properties, in applications requiring automation such as robotics. The development of LiDARs has been accelerated by their use on mobile systems, which have fostered the productivity of 3D dense and accurate mapping of environments. Such systems are now widely exploited for the off-line creation of accurate road and city maps, which are stand-alone products of interest. But the development of autonomous vehicles brings back the issue of real-time mapping and processing of data over large areas. This talk will consider several issues regarding 3D road scene perception such as data acquisition, localization and geo-referencing, surface reconstruction, real-time and off-line processing for various applications.