Abstract

The accurate delineation of Shorelines is required to determine the respective tidal heights and monitor coastal erosion. In some coastlines, the difference between the high tide and the low tide could be as small as one meter but the accurate extraction of the respective vertical surfaces is required to understand the pattern of the sea wave energy and to chart tidal heights relative to a specific datum. The manual digitizing of shoreline is usually with high accuracies and flexibilities but could be rigorous and hectic especially for longer shorelines. Automatic extraction is more feasible where the wavelength intensity gradient within a sub-pixel level is at maximum for instantaneous shorelines with small tidal range. This is due to topological inconsistencies that could hinder such accuracies. In this study, an improved mean shift segmentation model extracts shorelines from digital elevation models from LiDAR and a UAV survey. The key element of an accurate DTM is the contouring network capable of delineating between the grid pixels each representing water and land whose elevation is below or above tidal datum. The study locations are the Freshwater and the Perranporth bay in the United Kingdom. The third is the Nigeria Maritime Academy Oron shoreline. The choice for the different study areas is to ascertain the reliability of this technique on different geographies. Comparison between manual digitizing and the mean shift segmentation were both by visual interpretation and by volumetric change analysis for two different years. The results obtained indicate that the mean shift segmentation can delineate tide-coordinated shorelines accurately. The limitation of this methodology is on digital images with poor spectral resolution. Recommendations include the use of this technique on open source GIS software and a practical solution to developing a monitoring infrastructure for coastlines in Africa.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call