5 September 1997


Spy-in-the-sky technology is already checking IACS claims and is set to give early yield forecasts to help iron out volatile markets. Edward Long explains

REMOTE sensors and radar antennae on Earth-orbiting satellites are being focused on UK farms in a pan-European project to check IACS subsidy claims, and allow yields to be predicted weeks or months before harvest.

Monitoring agriculture was one of the first civilian applications of satellites. Since Landsat was launched in 1972, successive generations have been equipped with increasingly sophisticated sensors to enable global agriculture to be watched from space.

Farnborough-based company National Remote Sensing Centre has an EU-funded MAFF contract to exploit spy-in-the-sky technology to verify subsidy claims.

It has access to data gathered by Landsat 5, the French SPOT, and the new Indian IRS satellites, which orbit 800km above the Earth. Their sensors can see objects as small as 5m across. They detect reflected light which is both visible and invisible to the human eye.

Shades of grey

A total of 256 different shades of grey, ranging from black to white, can be picked up. Each is converted into a number and a mass of digital information is beamed down to ground receivers then sent to Farnborough on a CD.

Data is beamed to earth and converted into images each covering 60km x 60km. Different crops reflect light at different wavelengths, so a field of flowering rape gives a different signal, or signature, from a field of grass.

Once it is confirmed that a certain field contains a certain crop, an image processing computer highlights areas within the image with a similar signature. The end result resembles a false-colour photograph.

Lush green foliage which reflects a lot of infra-red light shows up dark red, bare ground and buildings which bounce visible light back into space appear blue, grey or white.

This is then warped to match the Earths curvature and existing Ordnance Survey maps and a digitised version of the official maps overlaid to provide field boundaries and other topographical features.

As many field boundaries no longer correspond with those on maps, further data and farm sketch plans are also incorporated into the final version.

Unlike military spy satellites which orbit frequently, civilian ones with sensors come round less often. Usually only three or four cloud-free images can be obtained during a growing season.

One in May gives the best overview of oilseed rape, another in June is good for spotting cereals, and a July image is best for picking out sugar beet, potatoes and linseed.

The sequence provides the continuity which gives the clues to what is being grown. Ground surveys in selected areas provide the ground-truth evidence to verify the interpretation of computer manipulated images.

The final image can see cropped areas as small as 0.3ha, and can even determine the width of set-aside strips. Inconsistencies between the actual image and that claimed alert MAFF to areas for further investigation.

The next generation of satellite imaging will involve microwave radar to back-up information gathered by optical sensors, and to provide data for yield predictions.

Radar penetrates clouds and operates round the clock, so can provide data during the winter to assess cropping intentions.

Microwave scanners also detect roughness of soil, so can show whether it has been ploughed, weathered or cultivated. Moisture also shows up.

Radar interaction

NRSC is investigating how radar interacts with different plants under various moisture conditions. It can already detect grass, and winter and spring crops.

Radar information is already available from the European Remote Sensing satellite, which has a 30m resolution. A sharper image down to just 8m will soon be available from a Canadian satellite.

In a pilot study space-based radar is being developed to predict crop yields weeks or even months before harvest. Each month 60 sites, each 40km x 40 km square, across Europe from Greece to Ireland are checked from both space and the ground to determine whether growth development can be assessed from satellite-based radar. Five of these sites are in the UK, one in the Kings Lynn area of Norfolk. &#42

False colours are used to identify crops in this image from a Landsat satellite. The original image showed infra-red and red reflectance, which indicate plant cell structure and moisture content, both good indications of crop type. © ESA (1992), distributed by Eurimage/NRSC, processed by NRSC.

An EU pilot project is evaluating cloud-piercing radar as an alternative for collecting data. This view of Norfolk uses radar images taken on three occasions last winter. Colours show changes in soil structure associated with cultivations, ploughing and seeding. Grass is dark as around the airfield. © European Space Agency 1997. Data processed by NRSC Ltd. Reproduced with kind permission of European Directorate General VI Agriculture.

See more