A clear view of actual and past processes affecting coastal areas at local and regional scale is the essential starting point to identify trends, issue forecasts and plan any protective measure. When this is not achieved, local interventions (eg breakwaters, groynes) may simply transfer the problem to adjacent areas, calling for further remedial measures and additional costs: this has happened since the 90’s along the Abruzzo coasts (Italy, Adriatic sea). In low coastal areas, the simplest parameter which can be monitored as proxy of the very complex dynamic equilibrium between mean sea level height, tide/wave energy and amount of incoming sediment is the shoreline, defined as the instantaneous divide between water and land. It has to be noted that even small tidal variation in such low relief areas can affect by several meters the position of the water divide: for cartographic purposes tidal and geomorphologic information is needed to correct for such effect. The COSMOCoast project, carried within the framework of an ASI contract (I/067/09/0), is applying standard image processing, object-oriented approaches and neural nets for boundary detection along the Abruzzo coast from high to very high resolution actual satellite data (eg COSMO-SkyMed, Formosat-2, IKONOS, Kompsat 2, Prism, Quickbird). The temporal observation window is extended backwards by a decade by exploiting older data at a much lower resolution (ERS). Validation data are provided by a kinematic GPS as well as by a ground lidar campaign carried out simultaneously to satellite overpass. Boundary-extraction tests run on COSMO-SkyMed spotlight acquisitions taken with different polarisations and incidence angles have been carried out: the NN outperformed the traditional techniques (based on Sobel’s and Robert’s operators) both in terms of accuracy detection and of computational burden. Segmentation of optical data acquired by different satellites also provide very good results. A second acquisition campaign is foreseen for 2011, whilst all processing results are being ingested in a GIS for quality and accuracy checking, validation, analysis of changes and definition of indicators.
Palazzo, F., Baiocchi, V., DEL FRATE, F., Giannone, F., Dominici, D., Latini, D., et al. (2011). Remote sensing as a tool to monitor and analyse Abruzzo coastal changes: Preliminary results from the ASI COSMOCoast project. In Proceedings 5th EARSeL Workshop on Remote Sensing of the Coastal Zone.
Remote sensing as a tool to monitor and analyse Abruzzo coastal changes: Preliminary results from the ASI COSMOCoast project
DEL FRATE, FABIO;
2011-01-01
Abstract
A clear view of actual and past processes affecting coastal areas at local and regional scale is the essential starting point to identify trends, issue forecasts and plan any protective measure. When this is not achieved, local interventions (eg breakwaters, groynes) may simply transfer the problem to adjacent areas, calling for further remedial measures and additional costs: this has happened since the 90’s along the Abruzzo coasts (Italy, Adriatic sea). In low coastal areas, the simplest parameter which can be monitored as proxy of the very complex dynamic equilibrium between mean sea level height, tide/wave energy and amount of incoming sediment is the shoreline, defined as the instantaneous divide between water and land. It has to be noted that even small tidal variation in such low relief areas can affect by several meters the position of the water divide: for cartographic purposes tidal and geomorphologic information is needed to correct for such effect. The COSMOCoast project, carried within the framework of an ASI contract (I/067/09/0), is applying standard image processing, object-oriented approaches and neural nets for boundary detection along the Abruzzo coast from high to very high resolution actual satellite data (eg COSMO-SkyMed, Formosat-2, IKONOS, Kompsat 2, Prism, Quickbird). The temporal observation window is extended backwards by a decade by exploiting older data at a much lower resolution (ERS). Validation data are provided by a kinematic GPS as well as by a ground lidar campaign carried out simultaneously to satellite overpass. Boundary-extraction tests run on COSMO-SkyMed spotlight acquisitions taken with different polarisations and incidence angles have been carried out: the NN outperformed the traditional techniques (based on Sobel’s and Robert’s operators) both in terms of accuracy detection and of computational burden. Segmentation of optical data acquired by different satellites also provide very good results. A second acquisition campaign is foreseen for 2011, whilst all processing results are being ingested in a GIS for quality and accuracy checking, validation, analysis of changes and definition of indicators.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
