Transforming bridge inspections through digital technologies

桥梁的检测和维护是保证桥梁正常使用和安全的重要环节。目前的现场检查通常是在纸上人工完成的，而纸是相关利益相关者之间交换条件信息的媒介。在损害登记或信息交换之后，对数据进行数字化处理。重复的数字化是一个容易出错的过程，并导致重复的工作。为了解决这些不足，本研究旨在设计一个框架，并提供一个增强现实工具包，用于基于深度学习的语义分割技术对桥梁损伤进行识别和分类。为实现这一目标，将执行以下步骤：i)为选定的桥梁案例研究获取数据；ii)数据处理和过滤；iii)使用深度学习进行图像分割；iv)基于分割的图像，通过代码交叉检查对损坏类型和严重程度进行分类；v)开发一个用户友好的增强现实工具箱，用于根据训练的数据集进行现场损坏识别和分类。这项技术将能够将重建的桥梁状况与数字化的BIM模型联系起来，并支持已建成基础设施的整个生命周期。最后，专家工具包将嵌入专家知识，用于与诊断的桥梁状况的维修和改造需求和替代方案相关的决策支持。

Inspection and maintenance are essential to ensure the serviceability and safety of bridges. Current inspection on site is often done manually on paper, and paper is the medium to exchange condition information between the involved stakeholders. After damage registration or information exchange, the data is processed digitally. The repeated digitalization is an error-prone process and leads to redundant work. To address these shortcomings, this research aims to devise a framework and deliver an augmented reality toolkit for identification and classification of damage in bridges using deep learning-based sematic segmentation techniques. To achieve this goal, the following steps will be carried out: i) data acquisition for selected bridge case studies; ii) data processing and filtering; iii) image segmentation using deep learning; iv) based on the segmented images, classification of damage type and severity by code cross-checking; v) development of a user-friendly augmented reality toolbox for on-site damage identification and classification based on the trained data set. This technology will enable linking the reconstructed bridge condition with a digitized BIM model and to support the complete lifecycle of built infrastructure. Finally, the expert toolkit will embed expert knowledge for decision making support related to repair and retrofitting demands and alternatives for the diagnosed bridge condition.