The rapid progress in LiDAR technology opens up ways for real time instant 3D (i3D) point cloud processing. With the access and maintenance of detailed 3D city maps through mobile laser scanning (MLS), autonomous vehicles are able to efficiently analyze urban environments. In my Thesis I introduce algorithms for automatic point cloud registration and change detection based on available heterogeneous MLS and instant 3D scans and for tracking dynamic objects from a moving platform. The introduced workflow consists of point cloud segmentation, object detection, a Hough transform based approach for registration, a Markov Random Field (MRF) based change detection technique in the range image domain and a Kalman filter based tracking algorithm. Experiments show qualitative and quantitative evaluations of the algorithms. The proposed workflow can provide centimeter level Global Positioning System (GPS) correction for dense urban areas and effective dynamic object analysis for autonomous vehicles.