Aerospace and Electronic Systems Magazine July 2017 Tutorial XI - 13

Acito et al.

Figure 1.

Processing chain for ACD applications.

rsipg.dii.unipi.it/, [6]). It includes three multitemporal images acquired over a suburban scenario where several different targets were
deployed together with reference tarps for in-scene calibration. Various scenarios were accurately planned and set up in order to enable
hyperspectral target detection (anomaly detection, signature based
target detection, and ACD) experiments. Ground truth was rigorously carried out including target and calibration tarps localization
and spectra collection, as well as accurate recording of movements
and changes occurred in-between subsequent acquisitions.
The images considered in this paper are part of a data collection acquired during a measurement campaign organized by CISAM
(Centro Interforze Applicazioni e Studi Militari, San Piero a Grado,
Pisa, Italy) with the support of CSSN-ITE (Centro Supporto Sperimentazioni Navali - Istituto per le Telecomunicazioni e l'Elettronica,
Livorno, Italy), Selex-ES (now Leonardo-Finmeccanica), and University of Pisa. The campaign was held on May 8-9, 2013 in Viareggio, Italy, a town located on the northern coast of Tuscany.
The Selex ES airborne hyperspectral sensor SIM.GA was
mounted on an ultralight aircraft to collect images over different

Table I.

SIM.GA Sensor Specifications
Characteristics

SIMGA
VNIR

Imaging

SIM.GA
SWIR
Pushbroom

Spectral Range (nm)

400 - 1000

1000 - 2500

Spectral Sampling
(nm)

≈ 1.2

≈ 5.8

# Spectral Bands

512

256

# Spatial Pixels

1024

320

NOMINAL IFOV
(mrad)

0.499

1.330

Field of View (FOV) (°)

± 15

JULY 2017, Part II of II

±12

sites in the morning and afternoon of two subsequent days. SIM.
GA is a modular avionic hyperspectral system composed of an
electro-optical head with two sensors collecting data in the Visible
and NearInfrared (VNIR) and the Short Wave InfraRed (SWIR)
spectral ranges. The main specifications of the SIM.GA are reported in Table I. The ground spatial resolution at 1200 m is about 0.6
m and 1.6 m in the VNIR and SWIR, respectively. When flying at
450 m, SWIR ground spatial resolution is about 0.6 m.
The three images considered in the following experiments were
acquired by the VNIR camera of the SIM.GA sensor flying at 1200
m. The main characteristics of these images are listed in Table II.
The images were coregistered by using the input geometry metadata
acquired by the on-board instrumentation. Furthermore, they are
equipped with the change reference maps (CRMs) including all recorded changes that occurred in the scene across the various acquisitions. For a given pair of images, two CRMs are provided where the

Table II.

Hyperspectral images characteristics.
Characteristics

D1F12H1

D1F12H2

D2F22H2

Day

May 8,
2013

May 9,
2014

Time (GMT)

14.18

14.33

12.64

Latitude

43° 51′
25" N

Longitude

10° 14′
51" E

Sky Conditions

Cloud
Cover

Clear

# Samples

450

# lines

375

# Bands

511

Spectral Range

VNIR

IEEE A&E SYSTEMS MAGAZINE

http://rsipg.dii.unipi.it/

Table of Contents for the Digital Edition of Aerospace and Electronic Systems Magazine July 2017 Tutorial XI