3D digitization has advanced considerably in both hardware and algorithmic development, driven by the growing demand across a wide range of applications. In parallel, increasing attention has been given to methods that accelerate the digitization workflow by reducing computational time while preserving reconstruction accuracy. In this work, we evaluate the in-house handheld multi-camera mobile mapping system (MMS), ATOM-ANT3D, equipped with a V-SLAM-based 3D reconstruction pipeline, in a range of complex real-world environments. The processing pipeline relies on a multi-instance V-SLAM framework that processes synchronized image streams in real time to generate sparse 3D reconstructions and camera poses (i.e., trajectories), which are subsequently refined through the developed near-real-time optimization techniques to improve geometric accuracy of the final outputs. The proposed workflow is assessed across diverse case studies to test its generalizability, including a historical tower composed of rooms and narrow staircases, complex underground tunnels, inner-corridor urban city mapping, and confined infrastructure. The results demonstrate centimeter-level accuracy comparable to photogrammetric reconstruction, while significantly reducing computational time up to ca. 93%. These findings highlight the potential of the proposed system as an efficient and reliable solution for rapid and accurate 3D reconstruction in complex environments.