In this research, we propose a methodology to improve the performance of scan matching and point cloud segmentation for 3D mapping of urban river environments. We also focus on the integration of depth image-based scan matching and spatial segmentation using streaming LiDAR data embedded in GNSS/LiDAR-SLAM. Moreover, we conduct experiments using a waterborne mobile mapping system to verify that our methodology can improve the stability and scalability of point cloud processing and achieve high-speed processing even in measured environments that cause SLAM degeneration problems. In addition, we propose a fast object classification based on rule-based segmentation using streaming point clouds.