PhotoMonitoring: a new low-cost technique for landslide monitoring using multispectral data

Landslides are widespread Natural hazards with significant social and economic impact, needing advanced monitoring and analysis technologies. In recent years, innovative and non-intrusive PhotoMonitoring (PM) techniques have attracted attention for their potential in landslide monitoring. These techniques offer the opportunity to enhance our understanding of landslide processes, contribute to risk mitigation, and provide new perspectives for geotechnical asset management. The research questions guiding this study include the applicability of non-invasive techniques like Digital Image Correlation (DIC) and Change Detection (CD) in monitoring rapid landslides, the additional information that PM techniques can offer compared to conventional methods, and the accuracy and reliability of measurements obtained through PM techniques, along with the development of new analysis methodologies to further enhance their potential. Throughout the thesis, the research explores PM techniques, focusing on specific case studies. Chapter 2 illustrates the application of PM in combining various ground-based remote sensing tools for rockfall monitoring. The results show that PM techniques, particularly Gigapixel images combined with PM analysis, provide a unique view of deformations preceding rockfalls, thus improving the rockfall risk assessment. Chapter 3 extends this investigation to the Pietrafitta Earth Flow, examining the integration of PM techniques using satellite images with other monitoring techniques such as T-InSAR, R-TS, and TLS. The results expose the suitability of PM techniques for improving the understanding of Earth Flow processes, bridging knowledge gaps left by conventional methods. Chapter 4 delves deeper into the potential of PM techniques, expanding the discussion of DIC techniques to assess knowledge, control, and emergency monitoring of Earth Flows. The results suggest that PM analyses can be used in a range of ways and offer a huge amount of information for a variety of monitoring purposes. In Chapter 5, the focus shifts to optical and thermal image processing for monitoring shallow landslides triggered by rainfall. The experimental application of PM techniques reveals their sensitivity in observing precursor signals of slope failure, highlighting their potential for Early Warning Systems. Overall, all these sections demonstrate the high potential of PM techniques in various contexts, with different objectives, and using various data acquisition platforms, all aimed at monitoring rapid geological phenomena such as Rockfall and slow but impulsive processes such as Earth Flow and Shallow landslide. The concluding section of this thesis addresses the research questions introduced in the introduction, providing answers based on the study's findings. It is evident that non-invasive PM techniques contribute significantly to the monitoring of rapid or slow but impulsive landslides and offer unique information not obtainable through conventional methods. The accuracy and reliability of PM measurements are highlighted, particularly when using high-resolution images. However, some limitations still exist, and these are thoroughly discussed. Furthermore, the broader implications and prospects are explored, emphasizing the adaptability, cost-effectiveness, and versatility of PM techniques, particularly in understanding and monitoring landslide processes. This opens new opportunities for research and practical implementation. Finally, the increasing use of optical sensors and smartphones is discussed, further expanding the potential scope of data collection. These implications have led to the creation of "IntelligEarth," a start-up of Sapienza aimed at realizing the vision outlined in this thesis.