Knowledge for Development

Remote sensing - an overview

Author: Paul Geerders

Date: 02/01/2005

Introduction:

Remote imaging sensors use electromagnetic radiation, emanating from the earths' surface, either directly (thermal radiation: temperature) or reflected by it. There are two main types; active and passive. In the latter case the source is sunlight (visible light scanners) or an artificial source carried by the sensing platform (radar, lidar). Remote Sensing uses radiation with wavelengths roughly between 400 nm (ultraviolet) and 4 cm (radio waves).


 

These different wavelengths have very different properties

  • 400-800 nm - visible light, detecting the colour of objects
  • 1 to 15 micron - thermal radiation, detecting the temperature of objects
  • 1 mm to 4 cm - microwave radiation, basically detecting the structure of those parts of an object that conduct electricity; in agriculture these are the water holding parts of vegetation.

In case of a cloud cover, visible light and thermal infrared observations are severely hindered. Only microwave radiation (radar) succeeds to penetrate clouds and observe the earth surface beneath.

While for agricultural applications of remote sensing clouds form a serious hindrance, in meteorology the clouds are the main objects, of which the size, type and movement can be easily followed. Also in meteorology use is made of data on the surface temperature of land and oceans, as a driving force for weather and climate.

There are two main types of satellites:

  • geostationary satellites, which always remain above the same location on earth, at an altitude of about 36000 km;
  • polar satellites, following an orbit at an altitude between 400 and 800 km, that brings them periodically over the same location, for instance a few times a day or only once every 2 weeks, depending on the specific characteristics of their orbit.

Earth observation satellites are operated by a limited number of countries: USA, Canada, Russia, Japan, France, China and Brazil, or by regional organisations such as ESA (European Space Agency). Furthermore, there is an increasing number of commercial earth observing satellites, such as Orbcomm and IKONOS [*].

In most cases, remote sensing observations imply images, like photos or TV like imagery. These optical image data can be obtained from a variety of sensors and systems. While NOAA-AHRR data provides geometrically coarse, but temporally detailed information on the presence and health of a vegetation cover, MODIS, Landsat, ENVISAT and the SPOT Vegetation Instruments provide more detailed information on type and distribution of forests and agricultural crops. Recently, radar systems such as JERS and Radarsat, which are not affected by clouds, have demonstrated their potential to determine different types of vegetation and forests.

Processing is required to translate the 'raw satellite data into a form suitable for the user and the application: a dedicated data or information product. This implies a manual or automatic analysis or interpretation of the data, taking into account aspects such as colour, shape, grey tone and texture. Currently several existing software packages allow this process to be highly automated through the application of a certain 'intelligence' (eCognition [*]) leading to an output product aimed at specific users and tuned to their needs.

A summary of phenomena that can be observed through remote sensing includes: land cover and land use, type of soil and vegetation cover, status of crops, type of trees, roads, urbanisation, rivers, lakes, watersheds, topography, geological structure of the landscape, ecosystems, suitable habitats for certain animals, flora and fauna but not directly in case of small species, colour of water leading to certain aspects of water quality and water pollution.

02/01/2005