A new approach for automatic Extraction of Shoreline from Satellite Images

Anna Spinosa, Deltares (NL)

Providing critical information for policy and decision makers, the monitoring of coastal environments, and particularly the investigation of shoreline trends, has an enormous potential to improve the understanding of the coastal environment and of beaches’ state of erosion or accretion. Remote sensing and satellite imagery, coupled with periodic surveys of emerged beaches, provide a systematic view of the Earth, useful to monitor short and long-term changes of coastal zones.

In the contest of the EU H2020 funded ECOPOTENTIAL Project, a new automatic method to retrieve the shoreline position has been developed. Sentinel -1 freely downloadable data by the Copernicus Open Access Hub web portal have been employed and the method has been based on image processing techniques to detect land-water edges. Edge detection techniques, one of the common methods to detect the shoreline from active sensors data, is a process of finding boundaries which separate two different regions. 

The method used consists in two steps, here shortly described: first, a median filtering approach is adopted in order to reduce the speckle effects due to the constructive/destructive interferences of backscattered electromagnetic waves which, appearing as a salt-pepper noise in the image, preclude the recognition of the water/land boundary by intensity values. Then, by means of the Otsu’s threshold method, a binary image is obtained and morphological operations (opening, closing, fill holes) are carried out to remove peaks and noises on the image. At last, the Canny’s edge filter is applied in order to automatically get the shoreline. 

This method has been tested on a total of 11 satellite images over 2017 along Torre Canne beaches, located in the Apulia region (south of Italy) and facing the Adriatic Sea. In order to obtain accuracy and precision metrics, the position of the derived shoreline has been compared with video-monitoring derived shoreline, freely downloadable from the web-portal ( The method shows sub-pixel accuracy.This work has been presented at the International Workshop “IEEE Metrology for the Sea” that has taken place in Bari, Italy, Oct 8-10 2018. The first author Anna Spinosa, a young Italian scientist working at Deltares, has won the “Best Paper Presented by a Young Scientist” award. Other ECOPOTENTIAL co-authors are Alexander Ziemba and Ghada El Serafy from Deltares and Victor Navarro from Starlab.