Preprint volume - SIBM

Pre-print Volume – Posters
Topic 3: INTEGRATED COASTAL ZONE MANAGEMENT
L. NICOLETTI, E. VALENTINI, M. TARGUSI, P. LA VALLE, A. FORNARI, A. TARAMELLI
ISPRA, Istituto Superiore per la Protezione e la Ricerca Ambientale, Via di Casalotti ,300 - 00166 Roma, Italia.
luisa.nicoletti@isprambiente.it
CHARACTERISATION OF POSIDONIA OCEANICA MEADOWS USING
BOTH HYPERSPECTRAL AND LIDAR DATA: A NEW APPROACH
CARATTERIZZAZIONE DELLE PRATERIE DI POSIDONIA OCEANICA
ATTRAVERSO L’USO DI DATI IPERSPETTRALI E LIDAR:
UN NUOVO APPROCCIO METODOLOGICO
Abstract – ISPRA has started an environmental characterization study in two Latium coastal areas using
both hyperspectral and LIDAR data. This study regards emerged and submerged portion of the coastal
zone. In this paper we describe a new approach aimed at mapping the Posidonia oceanica meadows.
Key-words: Posidonia oceanica meadow, spectra mixing analysis, LiDAR, data fusion, Tyrrhenian Sea.
Introduction – Integrated costal zone management is calling to develop and apply data
fusion techniques to combine active and passive remote sensing data for mapping
shallow-water and coastal environments. The present research is developed in the
framework of the Regione Lazio funding program and regards both emerged and
submerged portion of the coastal zone. The aim of this paper is to combine
hyperspectral MIVIS (Multispectral Infrared and Visible Imaging Spectrometer)
resolution data with LiDAR (Ligth Detection and Ranging) detection data in order to
characterize the Posidonia oceanica meadows present in two studied areas of the
Latium coast.
Materials and methods – The northern study area extends from the regional boundary
of Tuscany and S. Agostino (VT) and the southern one between Capo Portiere (LT)
and San Felice Circeo (LT). Airborne remotely sensed survey, carried out in May
2009, has provided two stretches for a total of 100 Km 2 hyperspectral (MIVIS) and
LiDAR (Hakw-eye) data. The MIVIS dataset ground resolution at nadir for an average
flight height of 1500 m is of 3 m. The LiDAR dataset acquired in slant range for an
average flight height of 500 m is of 2 m. Analysis of the shallow water and coastal
environment is done in the context of the spectral mixing space defined by the low
order principal components of the spectra (Small, 2004). Basically the works focuses
primarily on nonlinear mixing arising from intimate mixtures, then to test the validity
of the spectral mixing model we use the low dimensional parameter space defined by
the Principal Components in order to select the representative P. oceanica endmebers.
Our analysis combines the use of the first 20 MIVIS bands (atmospherically
normalized), so that radiance were converted into reflectance. A mask of land area was
applied to minimized the variability in digital numbers, increasing the contrast in the
study area. The contemporary acquisition of LiDAR data highlights bathymetry
morphological characterization and intensity data (Red and Near-InfraRed) to detect
the presence of P. oceanica meadows, rocks and sandy substrata.
Results – MIVIS hyperspectral sensor reached 10 m depht in both of areas and signal
accurrancy decrease with depth and turbidity increase. Normalization of sensor
41 st S.I.B.M. CONGRESS Rapallo (GE), 7-11 June 2010
180