I Jornadas de Seguimiento y Transferencia de Investigación Soleres
18th and 19th December 2008, Almeria, Spain
CONFERENCE
Application of advanced methodologies with hyperspectral sensors to the environmental sector.
D. Juan Antonio Ruso Fuentes
Technical Advisor R+D+i Management Unit and Technical Coordinator, EGMASA, Junta de Andalucia.
Summary
Satellite images have played a fundamental role within the set of data sources used for the development of methods for the incorporation of new spatial recognition technologies (remote sensing), especially in aspects concerning the temporal and spatial monitoring of variables subject to intense temporal dynamics.
In recent times, the field of Earth observation has seen a qualitative leap in many different applications. In this regard, hyperspectral sensors deserve special attention, characterized by their ability to measure reflected radiation in a wide range of wavelengths, in hundreds of spectral channels.
The great wealth of information provided by this type of sensors will allow their rapid integration into applications oriented to the analysis of biodiversity in terrestrial ecosystems and the study of global climate change. The advantages of hyperspectral images over the rest are:
- Narrower bandwidths are more sensitive to variations in reflected energy, allowing detection of incipient problems in crops or natural vegetation, probably earlier than with any other method.
- They allow the analysis of a wide combination of bands with a very low spatial resolution, thus optimizing spectral-spatial relationships.
- They make it possible to study transitions and obtain clear spectral signatures from a wide variety of bands to discriminate, for example, characteristic elements in an early prediction system.
Although there is considerable experience in different environmental applications with multispectral sensors, there are few studies in the literature on applications with hyperspectral images. Only in the university field have studies been conducted with various sensors (Held and Jupp, 1994; Blackburn and Milton, 1994; Milton and Rolin, 1994) with the CASI sensor, for the forestry sector, in studies of environmental monitoring and land use classification, considering the phenological behavior of vegetation, or in (Haboudane et al., 2004) using images obtained by CASI on agricultural crops (soybean, corn, and wheat) in various stages of growth and fertilization conditions, for the validation of the vegetation indices obtained.
From the private sector, it is worth mentioning the initiative of the company INDRA, also applying the use of hyperspectral sensors, specifically INTA's AHS for the study of contaminated soils (brownfields) within the framework of the ERA STAR projects in the region of Madrid.
Hyperspectral sensors are currently available on the market:
- Over Aircraft: AISA: 286 bands (0.4 to 0.9 um), AVIRIS: 224 channels (0.4 to 2.5 um), CASI: up to 288 bands (0.4 to 1um), DAIS: 79 bands (optical and thermal), GER: 126 bands (+12 thermal), HyMap: 128 bands (0.4 to 2.5 um), AHS: 80 bands (0.43 to 12.5 um).
- Over Satellite: MODIS: 36 bands (optical and thermal), non-continuous, MERIS: 15 bands, CHRIS: 18 to 63 bands, HYPERION: 220 bands.
- Extra-terrestrial missions: Cassini (Saturn), Galileo (Jupiter), Global Surveyor (Mars).
Egmasa, in its interest to incorporate the most advanced technologies to the object of its activity, started in 2006 an R+D+i project consisting in the development of new methodologies for the treatment and analysis of hyperspectral images obtained from several satellite and airborne sensors (data calibration, identification of spectral signatures, angular classification, linear analysis of spectral mixtures, principal component analysis) in different fields of study, with the intention of obtaining a source of periodic information on the state of two characteristic ecosystems in Andalusia:
- Seagrass meadows: characterization, status and evolution mapping of the main seagrass meadows of the Andalusian coast.
- Forest stands: characterization of forest stands, early prediction of defects due to pests, and (or decay in forest stands and mapping of the evolution of these defects/states).
The conference is one more event within the I Jornadas de Seguimiento y Transferencia de Investigación Soleres (JSTIS2008), organized by the OTRI, Fundación Mediterranea and the research groups Informatica Aplicada (TIC-211) and Informatica y Medio Ambiente (TEP-234), in collaboration with the EPS.
The Conference will take place in the Sala de Grados of CITE-III at 13:30, 18/12/2008.
This R&D project is supported by the EU (FEDER) and the Spanish IT Ministry under the reference TIN2007-61497