Safety Screening of Auckland's Harbour Bridge Movable Concrete Barrier
MetadataShow full metadata
A moveable concrete barrier along the Auckland Harbour Bridge(AHB) ena-bles two-way traffic flow optimisation and control. The barrier's block seg-ments are interconnected with metal pins, which sometimes can pop-out of their safe position. This thesis aims to use deep learning to aid visual metal pin inspection to improve traffic safety. The thesis proposes near-real-time IoT alerting solutions using mobile and other video sources. Preliminary experi-ments on a small dataset indicated that we could identify unsafe pin positions with high precision and recall. The first part of the proposed network detects and classify the unsafe pins. The second part actively tracks and alert the user of unsafe pin positions. The novel contributions presented in the thesis include: (1) A universal sys-tem globally applicable to similar traffic flow regulation and safety contexts with minimal modifications. (2) A novel technique for obtaining synthetic frame to produce various degrees of unsafe pin positions derived from the original frames. Collectively, synthetic minority-class data boosting, adaptive, incremental, and transfer learning utilising pre-trained neural networks allow a robust approach to data analysis and modelling on initially small and unbal-anced datasets for circumstances where the expected size of the dataset may or may not become available within the expected timeframes (such as during the pandemic lockdowns and added safety requirements). From the presented proof-of-concept, future work is intended to include collaborative user-centred design, where models, software upgrades and analytical platform up-grades will be under the oversight of New Zealand NZ Transport Agency and Auckland System Management.