This paper investigates automatic digitization with complete coverage of large and complex environments using a TLS or a mobile scanner. We propose an adaptive multi-objective view-planner that can operate in an unknown environment to provide guidance for the human operator and ease the scanning task or by a mobile robot for an automatic exploration of the environment. The proposed view-planner is adapted to environments where the sensor is operating on a flat surface such as office spaces, urban areas, open fields or in some cultural heritage applications. First, we propose an adaptive gap-based method to extract occluded areas in a point cloud, which is completely automated and does not require extensive computations in a large environment such as ray-tracing or level-set methods. Then, we introduce a novel exploration strategy that uses specific regions of the environment called “Conservative-Cells” to drastically reduce the number of sensing positions to achieve complete digitization of the environment.

Both methods were validated with simulated and real point clouds. The proposed approach has been applied to a scanner carried by a mobile robot, then to data acquired by a TLS used by a human operator in a large, complex environment. Experimental results on both TLS and mobile robot show that our view-planning approach is effective in finding a sequence of positions that leads to a complete reconstruction of the environment. Moreover, the proposed approach shows efficient performance in terms of coverage rate and computational time compared to others view-planning approaches as well as the results of an experienced human operator in a large, complex environment.

本文研究了使用 TLS 或移动扫描仪完全覆盖大型复杂环境的自动数字化。我们提出了一种自适应多目标视图规划器，它可以在未知环境中运行，为人类操作员提供指导并减轻扫描任务，或者由移动机器人自动探索环境。建议的视图规划器适用于传感器在平坦表面上运行的环境，例如办公空间、城市地区、开阔地或某些文化遗产应用。首先，我们提出了一种基于间隙的自适应方法来提取点云中的遮挡区域，该方法是完全自动化的，不需要在大型环境中进行大量计算，例如光线追踪或水平集方法。然后，

这两种方法都通过模拟和真实点云进行了验证。所提出的方法已应用于移动机器人携带的扫描仪，然后应用于人类操作员在大型复杂环境中使用的 TLS 获取的数据。在 TLS 和移动机器人上的实验结果表明，我们的视图规划方法可以有效地找到导致完全重建环境的位置序列。此外，与其他视图规划方法以及经验丰富的操作员在大型复杂环境中的结果相比，所提出的方法在覆盖率和计算时间方面显示出高效的性能。