engleski

Automatic Coastline Extraction Tool (CET) Based on Very-High Resolution Worldview Satellite Imagery

Exact determination of coastline length and characteristics is necessary for studies of coastal processes and consequently for management and protection of coastal areas. Very-high resolution (VHR) satellite imagery has great potential for coastline extraction, however noises in spectral data (e.g. shadows or sea reflections) can cause significant extraction errors. To overcome this problem, we present newly developed Coastal Extraction Tool (CET) that allows accurate and time-efficient automated coastline extraction based on combination of VHR multispectral satellite imagery and stereo-pair derived digital surface model (DSM). Automated coastline extraction is performed and tested with CET on Iž-Rava island group, situated within Northern Dalmatian archipelago (Croatia). CET extracted coastline accuracy was validated at chosen locations through comparison with the coastline extracted manually from centimetre accuracy reference data collected by in-situ terrestrial LiDAR scanning and UAV photogrammetry. Accuracy of CET extracted coastline was further analysed and quantified through 25 ground control points (GCP) collected along true coastline with high- accuracy RTK-GPS. Validation has shown that CET is highly accurate and that it successfully overcomes spectral induced errors. As a result of analysis two small islets are detected within study area that haven`t been mapped earlier within official state data. Considering its accuracy and ease of use we suggest that CET can be applied for automated coastline extraction in other large and indented coastal areas. Additionally, we suggest that it could be potentially applied in longitudinal geomorphological coastal erosion studies for detection of spatio-temporal coastline displacement.

Worldview imagery, coastline extraction, LiDAR, very-high resolution, CET

nije evidentirano