MULTI-SENSORY DATA FUSION FOR MÉTRO INSPECTION

本研究项目的主要目的是探讨利用遥测设备对蒙特勒尔运输银行(Sociétéde Transport de Montréal，STM)的混凝土结构进行检测的可行性。为了实现这一目标，将探索多传感器数据来源。具体地说，对探地雷达系统的剖面进行了分析，得到了结构内部的劣化情况。此外，还对热红外彩色图像进行了处理，以检测结构外观劣化。通过激光扫描将内部和外部劣化条件投影到三维点云中，生成具有详细和准确劣化信息的三维(3D)结构状态图。这样，STM工程师就可以依靠这样的状态图来制定适当的维护策略。本研究项目中提出的方法将在MatLab中实现，并在蒙特利尔市地铁混凝土结构的现场试验中进行评估。这项研究有望减少目前人工检测地铁混凝土结构所需的时间、成本和工作量。因此，Métro工程师可以做出明智和明智的维护决策，在Métro网络结构缺陷或功能过时之前，保持Métro网络健康向公众服务。

The main objective of this research project is to investigate the feasibility of utilizing remote sensingequipment to facilitate the inspection of Métro concrete structures for Société de Transport de Montréal (STM).In order to achieve this objective, multi-sensory data sources will be explored. Specifically, the profiles fromthe ground penetrating radar system are analyzed to obtain the structure interior deterioration. Also, the thermaland color images are processed to detect the structure exterior deterioration. Both interior and exteriordeterioration conditions are projected to the 3D point cloud from laser scanning to produce a three-dimensional(3D) structure condition map with the detailed and accurate deterioration information. This way, the STMengineers could rely on such condition maps to make an appropriate maintenance strategy. The proposedmethod in this research project will be implemented in Matlab and evaluated in the field trials of inspectingMétro concrete structures in the city of Montreal. The research is expected to reduce the time, cost, and effortrequired in the current process of manually inspecting Métro concrete structures. As a result, the Métroengineers could make wise and informed maintenance decisions to keep the Métro network healthy in serviceto the public, before they are structurally deficient or functionally obsolete.