• Accueil
>
• Bretagne Télédétection
• >
• Objectives
>
• Land

Terrestrial media

Most of the continental observatories are set in zones (test sites) of limited size (a few hundredth of km² with linear structures such as rivers, hedges, catchment area, etc.), due to the instrumentation load. The observations are very often irregularly sampled in time. In order to correctly analyze the behavior of the observatories on the long term, and to follow the land state surface and process at the regional level, the ground measurement sampling becomes a critical point. Spatial Remote Sensing constitutes then, an approach particularly well adapted for identifying the land usage and the settling modes and also for characterizing the land states at different scale level (from a plot of land to a county and to the region eventually). This is done by using sensors with different resolutions and/or sensors with different spectral bands (for a 3D structural description). It is thus necessary to obtain data from high spatial resolution sensors but moderate temporal resolution (week/month) acquired in the optical domain (e.g. SPOT HRV/HRVIR, ASTER, Kompsat, Quickbird, future SENTINEL) and micro-wave (e.g. ERS, ENVISAT-ASAR, ALOS) along with measurements at moderate spatial resolution (300m to 3 km) and high temporal resolution (e.g. SPOT-VEGETATION, MERIS, MODIS, MSG).

Presently the adar images are under exploited, although they have an obvious interest for the land surface characterization in places where the cloud cover is frequent and for applications such as: flood monitoring, wet zones, crop, and identification of agricultural practices like plowings for instance. Furthermore, the coupling of these data with optical images offers very interesting perspectives for a better identification and characterization of soil and vegetation. Developments made by teams specialized in signal processing associated with teams closer to the applications must be encouraged. This will insure a better exploitation of radar images on land and will allow a better evaluation of the radar/optic coupling interest.

We have to note that the new and future high spatial resolution sensors will have a high temporal revisit time, major characteristic being borne, up to now only by spatial mid-resolution sensors. Two challenges must be met:

• exploitation of time series of high resolution images by the development of adapted change detection methods
• registration of series of mid-resolution images (regional scale) with series of high spatial resolution time series (local scale) by the development of multi-scale methods

Again, these developments will mix teams from the signal processing community with the one working on applications.

Combining several sensors is an absolute necessity for answering the many questions ones have about the continental surface behavior and their evolution (terrain correspondence, precision of model outputs, data assimilation):

• Recognition and characterization of surface units (settling and use of land, monitoring of soil surface and vegetation, submerged soils, agricultural practices, …),
• Process estimation (floods, wet zones, vegetation growth, inter-culture implementation, plowings, urban spreading, rural zones, …)
• Regional approach and scale transfer (regional to local, local to regional)

Either for regional analysis or for tests of spatial technologies, the amount of data to store and the processing time for operational use implies the development of infrastructure for hosting the archiving, management and dissemination functions. This Environmental Information System (EIS) will have obviously to be coupled with the remote sensing and in situ data processing chains in order to be able to use them in the models.