3D Mobile Mapping Using Artificial Intelligence

关键基础设施，即公用事业、交通和设施相互依存的网络，是加拿大经济和社会的支柱。尽管加拿大是世界第二大国家，拥有世界第十大经济体，但其三分之一的基础设施需要重大更新。定期、严格的监测是必要的，以确保基础设施的安全和高效运行。然而，目前以人为中心的检查做法不足以满足这一需求，而且费力。需要能够持续和准确监测的廉价和更有效的技术。3D移动测绘系统(MMS)是一种新兴的解决方案，它可以收集视觉感知数据，使用一个通用的移动平台高速测量整个基础设施。在获取传感数据以生成三维点云方面取得了重大进展；然而，这些数据集的采集后处理仍然是成功生成三维高保真地图的一个挑战，这些地图在语义上是可解释的，在几何上是详细的，并且以测量等级的精度进行地理参考。Teledyne Optech总部设在多伦多，是彩信领域的世界领先者。与Teledyne Optech合作，NSERC CRD项目将使用一种特定类型的人工智能(AI)开发一种先进的数据处理系统，称为深度神经网络，该网络最近在计算机、机器人视觉和机器学习方面取得了显著成功。这项工作将允许使用MMS数据和关键网络的高质量3D模型自主识别基础设施资产，从而为基础设施管理领域做出贡献，并改善整个城市的可持续性。所开发的技术将能够与以人为本的技术竞争，减少采集后数据处理所需的时间；它们将便利多媒体管理系统在全球定位系统无法使用的环境中运行，并使用户能够根据需要实时调整测量参数。通过该计划培养的HQP将为加拿大的各个行业和人工智能技术、基础设施管理、城市规划和彩信领域做出贡献。

Critical infrastructure, the interdependent networks of utilities, transportation, and facilities, are the backbone of Canada's economy and society. Although Canada is the 2nd largest country in the world, with the world's 10th largest economy, one-third of its infrastructure is in need of a significant update. Regular, rigorous monitoring is necessary to ensure the safe and efficient operation of the infrastructure. However, current human-centric inspection practices are inadequate and laborious to fulfill this need. Inexpensive and more efficient technologies that are capable of continuous and accurate monitoring are necessary. 3D mobile mapping system (MMS) is an emerging solution that can collect visual sensory data to survey entire infrastructure with high speed using a common moving platform. There have been major advancements in acquiring sensory data to produce 3D point clouds; however, the post-acquisition processing of these datasets remains a challenge to successfully generate 3D high-fidelity maps, which are semantically interpretable, geometrically detailed and georeferenced with survey grade accuracy. Teledyne Optech, headquartered in Toronto, is a world leader in the MMS field. In collaboration with Teledyne Optech, this NSERC CRD project will develop an advanced data processing system using a specific type of artificial intelligence (AI), called deep neural network, which has recently achieved remarkable success in computer and robotic vision and machine learning. This work will allow for the autonomous recognition of infrastructure assets using the MMS data and high-quality 3D models of critical networks, thus contributing to the field of infrastructure management and improving urban sustainability as a whole. The technologies developed will be able to compete with human-centric technique and reduce the time required for post-acquisition data processing; they will facilitate the operation of the MMS in GPS-denied environments and allow users to adjust survey parameters as needed in real time. The HQP trained through this program will contribute to various Canadian industries and the fields of AI technologies, infrastructure management, urban planning, and MMS.