Artificial Intelligence for the Condition Assessment of Critical Infrastructure

这一合作伙伴关系的目标是提高通常用于监测混凝土结构恶化的工具和技术的测试速度、可重复性和准确性。我们将使用人工智能（AI）和深度学习（DL）开发自动化工具，用于现场视觉裂纹检查和基于实验室的混凝土损伤评估微观技术。我们的解决方案将包括快速可靠的工具，用于i)自动计算裂缝指数（CI），使用智能手机应用程序（“app”）实时分析结构目视检查期间现场拍摄的照片，以及ii)通过应用图像分析（IA）技术对混凝土试件的高分辨率立体显微镜图像自动计算损伤等级指数（DRI）。提出的技术将显著增强两种重要的混凝土监测技术：a) CI的计算，基于结构的目视检查对损伤原因和程度进行初步评估；b) DRI的计算，其中从结构中提取的核心样品在基于实验室的微观协议中进行检查，以诊断受内部膨胀反应（ISR）等痛苦机制影响的混凝土的损伤原因和程度。这两种方法都有重要的局限性：CI需要现场定性调查，虽然作为初步指标有用，但其评估受影响混凝土损伤的能力对于具有多种损伤机制和不同程度的约束和暴露条件的结构仍然不清楚。DRI的计算是一个耗时的、基于专业知识的实验室程序。CI和DRI在本质上都是主观的，并且严重依赖于执行分析的人的技能和经验。该项目将能够快速、可靠地评估加拿大主要基础设施（如桥梁、水坝和建筑物）混凝土损坏的性质和程度。反过来，这将促成及时和具有成本效益的预防性修复战略，从而确保更安全和更可靠的建筑环境。

The goals of this partnership are to improve testing speed, reproducibility, and accuracy of tools and techniques commonly used for monitoring structural deterioration of concrete. We will use artificial intelligence (AI) and deep learning (DL) to develop automated tools for on-site visual crack inspection and for lab-based microscopic techniques for concrete damage assessment. Our solutions will include rapid and reliable tools to i) automate computation of the cracking index (CI), using a smartphone application ("app") to analyze, in real time, pictures taken in-situ during visual inspections of structures, and ii) to automate the Damage Rating Index (DRI) calculation by applying image analysis (IA) techniques to high-resolution stereomicroscope images of concrete specimens. The proposed technology will significantly enhance two important concrete monitoring techniques: a) computation of the CI, a preliminary assessment of cause and extent of damage based on visual inspection of structures, and b) computation of the DRI, in which core samples retrieved from structures are inspected in a lab-based microscopic protocol to diagnose cause and extent of damage in concrete affected by distress mechanisms such as internal swelling reactions (ISR). Both methods have important limitations: CI requires in-situ qualitative investigations and, although useful as a preliminary indicator, its ability to assess damage of affected concrete is still unclear for structures presenting multiple distress mechanisms and under distinct degrees of confinement and exposure conditions. Computation of the DRI is a time-consuming, expertise-based, lab procedure. Both CI and DRI are subjective in nature and rely heavily on the skill and experience of the person performing the analysis. This project will enable rapid and reliable assessments of the nature and extent of concrete damage in critical infrastructure prevalent in Canada, such as bridges, dams, and buildings. This will, in turn, enable timely and cost-effective preventative rehabilitation strategies, thus ensuring a safer and more reliable built environment.