3D spatial modeling of Canada's critical infrastructure using mobile lidar technology

从卫生、通信、能源和公用事业等到交通运输等十个部门的关键基础设施（例如物理设施）的准确和最新的3D空间模型对加拿大人的健康、安全、保障和经济福祉以及加拿大政府的有效运作至关重要。机载和地面激光扫描（或激光雷达）方法的出现已被证明是创建和更新3D城市模型的强大工具。然而，这两种方法都有其局限性，从较大的街道部分的快速和成本效益的三维点云捕获。如果这些部分包括隧道或如果需要建筑物立面的密集点覆盖，则尤其如此。作为一个较好的选择，应使用移动的激光雷达-一种车载激光扫描系统，运行速度最高可达100 km/h。在移动中，它可以扫描基础设施和周围环境，作为一个非常高密度的3D点云。通过将激光扫描仪与一套集成的定位设备（GPS、惯性测量单元和里程计）连接，可以对点进行地理编码。然而，应急管理人员和其他最终用户既没有工具也没有兴趣处理点云-他们想要的是空间对象（例如建筑物）而不是点云。由于城市街道场景的庞大数据量和复杂性，迄今为止，将3D点云转换为3D空间模型是由操作员驱动的，并且需要包括多个处理步骤的工作流程。现有的方法强调各个步骤，没有一个软件系统可以完成从输入到模型创建的所有必要步骤。本研究通过开发一个全面的策略和一个创新的软件系统来解决从点云重建三维建筑物的非常具有挑战性的问题。通过整合移动的激光雷达、机载激光雷达和（静态）地面激光雷达的点云、光学影像、平面图和三维GIS数据，可以对建筑物内外进行三维建模，有效支持应急疏散和救援路线规划。

Accurate and up-to-date 3D spatial models of critical infrastructure (e.g. physical facilities) across the ten sectors ranging from health, communication, energy and utility, etc. to transportation) are critical to the health, safety, security and economic well-being of Canadians and the effective functioning of governments in Canada. The advent of airborne and terrestrial laser scanning (or lidar) methods has proven to be a powerful tool for creating and updating 3D city models. However, both methods have their limitations for rapid and cost-effective capturing of 3D point clouds from larger street sections. This is especially true if these sections include tunnels or if dense point coverage of the building facades is required. As a preferable option, mobile lidar - a vehicle-mounted laser scanning system operating at speeds of up to 100 km/h, should be used. Whilst on the move, it can scan infrastructure and the surrounding environment as a very high density 3D point cloud. The geocoding of points is made possible by interfacing the laser scanners with an integrated suite of positioning devices (GPS, inertial measuring unit, and odometer). However, emergency managers and other end-users have neither the tools nor the interest in processing point clouds - they want spatial objects (e.g. buildings) rather than point clouds. Due to huge data volumes and complexity of urban street scenes, converting 3D point clouds to 3D spatial models to date is operator-driven and requires a workflow comprising several processing steps. Existing methods stress the steps individually and there is no single software system that does all of the necessary steps from input to model creation. This research tackles the very challenging problem of 3D building reconstruction from point clouds by developing a comprehensive strategy and an innovative software system. Through integrating point clouds acquired from mobile lidar with those from airborne lidar and (static) terrestrial lidar, optical images, floor plans and 3D GIS data, both the outside and inside of buildings can be modeled in 3D, which can effectively support emergency evacuation and rescue route planning.