Article March 2000

Reducing Weight, Size and Costs

Space Remote Sensing from Micro- and Mini Satellites

Surrey Satellite Technology Ltd. (SSTL) is a world leader in the design and manufacture of small satellites and is located at the University of Surrey's main campus in Guildford, a town 30 miles (48km.) south-west of London. The company is wholly owned by the University and was set up in 1985 to carry out research and to develop and build advanced micro- and mini-satellites for customers outside the University on a commercial basis. This followed on from the successful development, launch and operation of the experimental UoSAT-1 and UoSAT-2 micro-satellites in the early 1980s. Since then, more than a dozen small satellites have been built and launched with several more scheduled to be launched over the next year or two. Although quite a number of SSTL's satellites have been developed for communications and space science applications, recently several examples have been designed and built specifically for Earth Observation (EO) and remote sensing purposes.

 

By Prof. Gordon Petrie

Micro- & Mini-Satellites

In current general usage, a micro-satellite is one that weighs between 10 and 100 kg.; a mini-satellite has a mass of between 100 and 500kg.; small satellites weigh between 500 and 1,000 kg,, while large satellites weigh over 1,000kg. Most remote sensing satellites fall into the last of these classes - e.g. the weights of current operational satellites at launch were for SPOT, 1,900kg.; for Landsat-7, 2,200kg.; for the European ERS-2 satellite, 2,500kg.; and for NASA's new Terra (or EOS AM-1) satellite, 5,200kg. At the top of the size and weight scale is the forthcoming Envisat from the European Space Agency which will be over 10m in height and is scheduled to weigh 8,140kg., including its payload of 2,145kg. Of course, all of these large satellites carry a number of imaging and other sensors - in the case of Envisat, an advanced SAR, spectrometer, microwave radiometer, radar altimeter, atmospheric sounder, etc. - rather than the small camera systems carried by the SSTL micro-satellites. Nevertheless, the contrast in size and weight is striking - e.g. the tiny SSTL-built Thai-Puhtt micro-satellite (or TM-Sat) launched in July 1998 and producing image data of a quality roughly equivalent to that of the Landsat MSS imager, weighs only 50kg. and has dimensions of 69 x 36 x 36cm (roughly the size of a small refrigerator). The largest SSTL mini-satellite, UoSAT-12, launched in April 1999, has a weight of 315kg. In turn, these relatively light weights mean that these satellites can be launched comparatively cheaply, either piggy-back on larger satellites or using relatively inexpensive lightweight launchers such as those converted from surplus ballistic missiles that are available from CIS (former Soviet Union) countries.

Technology Transfer

The differences in cost of the actual satellites are also rather striking. Each large satellite intended for remote sensing costs US$250 million or often much more, whereas SSTL's micro-satellites cost around US$3 million each with up to US$14 million needed for an enhanced micro-satellite or a mini-satellite. These weight and cost differences have resulted in SSTL building several micro-satellites having a remote sensing capability for emerging space countries such as Korea, Portugal, Chile, Thailand, Malaysia and China under collaborative technology transfer and training programmes. Another order from Turkey for the supply of an enhanced micro-satellite by SSTL on such a basis has just been announced. Some of the SSTL satellites that have been launched or built recently (since 1998) under these programmes are discussed below.

TMSat (Thailand)

The Thai-Puhtt or TMSat micro-satellite is the result of a collaborative programme carried out by SSTL in cooperation with the Thai Micro-Satellite Company (TMSC) Ltd. and Mahanakom University of Technology in Thailand. Two camera systems are mounted on board this satellite. The first is a wide-angle camera (WAC) which utilizes a small-format (368 x 560 pixel) CCD areal array made by the English Electric Valve (EEV) company in the U.K. This camera is fitted with a very short focal length (f = 4.8mm) ultra-wide-angle lens which covers an area of 1,500 x 1,050km with a ground pixel size of over 2km from the satellite's orbital height of 800km. The camera is fitted with an optical filter to allow it to collect its images at wavelengths (l) between 810 and 890nm in the near-IR part of the electro-magnetic (EM) spectrum. Essentially this produces low-resolution synoptic images of the cloud pattern over the Earth's surface for weather forecasting or meteorological purposes.

TMSat Multi-Spectral Images

However the second narrow angle camera (NAC)) system is of a quite different design. It  comprises three vertically pointing cameras that acquire their images of the terrain simultaneously with an identical ground coverage or footprint. Each of the three cameras is equipped with a CCD areal array of 1,000 x 1,000 pixels = 1 megapixel manufactured by Kodak. The size of each array is approximately 9 x 9mm, with each pixel measuring 9.2 x 9.2mm. With each camera having an f = 75mm lens, this combination produces a ground coverage of 100 x 100km in a single exposure from the satellite's orbital height of 800km. with a ground pixel size of 90 to 100m. The three cameras are each fitted with different optical filters allowing transmission in the green (l = 510 to 590nm); red (l = 610 to 690nm) and near I-R (l = 810 to 890nm) bands respectively. Each of the three cameras passes its image in the form of 256 (8-bit) grey level data to a solid-state data recorder. After 12 seconds, the three cameras simultaneously acquire their images again. The solid-state recorder can accept four sets of these three-band images. Each set of three individual band images can then be combined together to produce a single false-colour image. The set of four false-colour images can then be mosaiced together to form a composite strip image. Essentially this produces imagery that is very similar in terms of its spectral and ground coverage and resolution to that of the well-known Landsat MSS sensor with its 80m ground pixel.

Ground Stations

The TMSat, together with a second similar micro-satellite, FaSAT-Bravo, equipped with different imaging sensors that had been built by SSTL for Chile, were carried piggy-back as secondary payloads on the Russian RESURS satellite that was placed in orbit by a Zenit vehicle launched from Baikonur Cosmodrome in July 1998. Both satellites have operated very successfully since being placed in orbit and both can be activated and commanded from low-cost PC-based ground stations established in Thailand and Chile by engineers who have been trained by SSTL, as well as from the SSTL Mission Control Centre in Guildford.

TiungSAT-1 (Malaysia)

Currently complete and ready for launch is the TiungSAT-1 micro-satellite that has a similar combination of WAC and multi-spectral NAC cameras to those that have been mounted in the TMSat. This particular satellite has been built by SSTL for a Malaysian government agency ATSB. However since this organisation decided to arrange for the launch itself rather than employ SSTL to do so, some delay has occurred. Now arrangements have been made with the Russian Space Agency (RSA) to launch the satellite in March 2000 using a Zenit rocket. It will placed in a higher (1,020km) orbit than the Thai and Chilean satellites. Thus the WAC will give a 2,250 x 2,250km. ground coverage in a single exposure with a 2.2km ground pixel size, while the NAC multi-spectral camera will cover an area of 120 x 120km with a 120m ground pixel size.

Tsinghua-1 (China)

Under construction and due for launch later this year is the Tsinghua-1 micro-satellite which has been constructed through a collaboration (including the formation of joint venture company) between SSTL and the Tsinghua University in Beijing. Present plans are for the satellite to be orbited at an altitude of 750km to provide 3-band multi-spectral images. with a 35m ground pixel size and an off-nadir pointing capability of 15. The Tsinghua-1 satellite is viewed partly as a demonstrator for a proposed constellation of seven micro-satellites that would provide reasonably high-resolution imagery for the monitoring of disasters (floods, fires, earthquakes, tornados, etc.) covering large land areas. Under this proposal, this constellation would provide a daily re-visit capability on a world-wide basis. Each participating country would have its own micro-satellite, but, if there was a disaster, then all the satellites in the constellation would be available and used to monitor the affected area to help plan rescue and relief operations.

UoSAT-12

The newest and most capable of SSTL's satellites is UoSAT-12 which has been constructed largely on SSTL's own account to validate key mini-satellite bus and payload technologies. This mini-satellite has been placed in an orbit which is inclined at 64 to the Equator: thus it is not in a Sun-synchronous orbit. The position and altitude of the satellite are measured continuously by an on-board GPS set, while attitude sensors supply a tilt value to a resolution of 0.1. The satellite has two cameras that operate side-by-side acquiring images simultaneously using fully electronic shutters. Both cameras are tilted by 1 in opposite directions to give in combination as wide a coverage as possible in the cross-track direction. After processing and mosaicing, a single image is produced from the two component images resulting in the final image having a rectangular 60 x 30km format. Each camera uses the same 1,000 x 1,000 pixel CCD areal array from Kodak used in the TMSat NAC camera described above. The cameras use lenses with the relatively long focal length (f) of 180mm. These were supplied by Leica, having been selected by SSTL from the Leica Company's standard range on the basis of their very good illumination qualities. No attempt was made to carry out a geometric calibration as required for photogrammetric or mapping purposes.

UoSAT-12: Multi-Spectral Images

Multi-spectral images are acquired from UoSAT-12 through sequential exposures by the same  camera through four different spectral filters. These are mounted on a rotating wheel that can accommodate six filters. The four that are actually used are blue, green, red and near infra-red. The four individual band images are taken at intervals of 300 milliseconds to give time to rotate the filter wheel so that the next filter is placed in the correct position in front of the lens. In fact, only 100 milliseconds of this interval is needed to read out and store the image in the solid-state data recorder. In this way, the series of four individual band images are exposed sequentially from four successive (and different) positions in space rather than simultaneously from a single exposure station. Processing has then to be carried out at the ground station to co-register the four individual band images to produce a single true-colour or false-colour image. These have a ground pixel size of 32m, which, in resolution terms, is roughly the same as that of the Landsat TM multi-spectral imagery.

UoSAT-12: Pan Images

UOSAT-12 is also equipped with a vertically pointing panchromatic camera which employs the same type of CCD areal array as that used in the camera taking the multi-spectral images, but is equipped with a still longer-focus (f = 560mm) lens, again supplied by Leica. This produces pan imagery with a ground coverage of 10 x 10km and a ground pixel size of 10m - the latter value being the same as that of the SPOT Pan sensor. While numerous examples of the multi-spectral imagery have been acquired from UoSAT-12, rather less use has been made of the pan camera due to a problem with its thermal compensation provision that causes the images to be somewhat less sharp than expected.

NASA's Experience

The considerable risks and the ever escalating costs of developing, launching and operating large satellites in the face of shrinking or, at best, constant financial budgets have led recently to a rather painful re-appraisal of the situation by space agencies. This resulted in NASA stating that "smaller, faster, cheaper satellites" needed to be built that would still have a high capability to execute many space missions. This has had its influence on space remote sensing. Thus NASA instituted its Small Spacecraft Technology Initiative under which the Lewis and Clark mini-satellites were built to form part of its Mission to Planet Earth programme. The Lewis mini-satellite, launched in August 1997, had two complementary hyperspectral imagers mounted on-board the platform. Unfortunately, however, the satellite was lost a short time after launch due to it entering an uncontrollable flat spin from which it could not be recovered. The Clark mini-satellite (weighing 290kg.) was to feature a high-resolution pan sensor producing images with 3m ground pixel size and a three-band multi-spectral imager with a 15m ground pixel (as fitted to the EarthWatch EarlyBird satellite). However, in February 1998, NASA terminated the project due to cost overruns and the delays arising from problems with the non-availability of the launcher.

Future Prospects  

While the NASA experience has had a rather chastening effect on the development of very small satellites, the success of the SSTL micro- and mini-satellites has been much more encouraging. So has that of the lightweight (260kg) Ofeq reconnaissance mini-satellites that been successfully orbited and used operationally by Israel. This last success has resulted in the development of the commercial high-resolution EROS satellites that are to be launched by West Indian Space. In 1995, SSTL also produced the Cerise micro-satellite for the French CNES agency for military applications and has just launched the follow-on Clementine micro-satellite for similar use. SSTL has also built a micro-satellite for the U.S. Air Force called PicoSat that is intended to test advanced technologies for research purposes. This is now awaiting launch. Furthermore, NASA - now recovering from its previous disappointments - is proposing to launch three new Trailblazer micro-satellites in 2003 to test other new technologies that will reduce the weight, size and costs of its future satellites. Undoubtedly all of this activity is having its effect - as witness the current proposal by CNES called 3S (Small SPOT Satellites) that the follow-on to the SPOT-5 satellite should take the form of several mini-satellites, each to carry different imaging sensors, instead of mounting all of these on a single large platform as has been done with SPOT-1 to -5. Undoubtedly this whole area of micro- and mini-satellites as applied to remote sensing is one to watch for the future.

Acknowledgements

The author would like to acknowledge (with many thanks) the information given by Dr. Marc Fouquet (head of Earth Observation activities at SSTL) during a recent visit to Surrey Space Centre. Also he is grateful for all the help given by Miss Audrey Nice (SSTL's Press & Publicity Officer) in the provision of the background information and illustrations used in this article.

 

Professor G. Petrie, Department of Geography & Topographic Science, University of Glasgow, Glasgow, G12 8QQ,Scotland, U.K.