Development of a low-cost photogrammetric monitoring system for timely detection of slope instability

The growing frequency of extreme rainfall events increases the need to monitor unstable slopes and warn of impeding failure. Additionally, the increasing number of vulnerable sites means that low-cost solutions are necessary. This project assumed that a system based on modern photogrammetric workflows and terrestrial multi-stereo acquisitions can simplify hazard prediction and mitigation of landslide-related threat impacting vulnerable people and critical infrastructure. Thus, the challenge of this thesis has been to develop an efficient and automated system to monitor slopes with a self-adaptive monitoring frequency and early warnings capabilities, by adopting a low-cost photogrammetric solution. This was achieved using three case studies, including one scaled laboratory test field (camera-object distance of 2 m) and two field sites (a seacliff - distance of ~ 30 m; a quarry face - distance of ~ 700 m). This research shows the monitoring capabilities of a terrestrial photogrammetric system based on just four low-cost digital cameras combined with Structure-from-Motion and Multi-View-Stereo (SfM-MVS) techniques. The photogrammetric outputs (e.g. dense point clouds) were highly comparable to results obtained with more established techniques (e.g. terrestrial laser scanner and drone-based photogrammetry). Tests demonstrate that when a stationary camera configuration is adopted, imperfections in the camera calibration can be tolerated. In response to the challenges encountered aligning 3D photogrammetric outputs, an automated solution to registration is identified, which involves the use of the scale-invariant feature transform (SIFT) algorithm. This approach produced results comparable to those achieved using a manual procedure involving the adoption of the Iterative Closest Point (ICP) algorithm. An additional case study allowed the implementation of an algorithm to autonomously adapt the temporal frequency for image acquisition based on recorded change and the development of a framework for an SfM-MVS-based Early Warning System (EWS). The results obtained in this research, demonstrate that an EWS based upon low-cost SfM photogrammetry is practicable. This is significant as it proves that low-cost EWS based upon SfM photogrammetry can help to monitor and mitigate the effects of the growing number of slope instabilities expected as a consequence of climate change