A robust surface matching technique for coastal geohazard monitoring

摘要

Coastal geohazards, such as landslides, mudflows, and rockfalls, represent a major driver for coastal change in many regions of the world, and often impinge on aspects of the human and natural environment. In such cases, there is a pressing need for the development of more effective monitoring strategies, particularly given the uncertainties associated with the impact of future climate change. Traditional survey approaches tend to suffer from limited spatial resolution, while contemporary techniques are generally unsuitable in isolation, due to the often complex coastal topography. To address these issues, this thesis presents the development and application of a strategy for integrated remote monitoring of coastal geohazards. The monitoring strategy is underpinned by a robust least squares surface matching technique, which has been developed to facilitate change detection through the reliable reconciliation of multi-temporal, multi-sensor datasets in dynamic environments. Specifically, this research has concentrated on integrating the developing techniques of airborne and terrestrial laser-scanning. In addition, archival aerial photography has been incorporated in order to provide a historical context for analysis of geohazard development. Robust surface matching provides a mechanism for reliable registration of DEM surfaces contaminated by regions of difference, which may arise through geohazard activity or vegetation change. The development of this algorithm has been presented, and its potential demonstrated through testing with artificial datasets. The monitoring strategy was applied to the soft-cliff test site of Filey Bay, North Yorkshire. This highlighted the viability of the robust matching algorithm, demonstrating the effectiveness of this technique for absolute orientation of DEMs derived from archival aerial photography. Furthermore, the complementary qualities of airborne and terrestrial laser scanning have been confirmed, particularly in relation to their value for multi-scale terrain monitoring. Issues of transferability were explored through application of the monitoring strategy to the hard rock environment of Whitby East Cliff. Investigations in this challenging environment confirmed the potential of the robust matching algorithm, and highlighted a number of valuable issues in relation to the monitoring techniques. Investigations at both test sites enabled in-depth assessment and quantification of geohazard activity over extended periods of time.