20120129

Keeping an eye on the planet

Publish Date: 24 January 2012
Prof. Arup Dasgupta
Geospatial World, Managing Editor



With technology advancements in sensors, payload weights and data processing, earth observation as a domain is transforming rapidly. And with its application expanding across the domains, the future looks replete with opportunities. Here’s an exhaustive analysis of the technology trends...


The technology of earth observation has seen many changes over the past few years with four major trends emerging. The first is the government- funded missions for earth observation, using a variety of sensors on large satellites which address mapping as well as scientific studies. The news, however, is dominated by the second trend consisting of commercial imaging satellites with sub-metre spatial resolution for land applications. The third is a shift away from big multi-sensor satellites towards small single-function satellites. The fourth trend is to use small satellites in constellations and swarms. Furthermore, these trends tend to overlap with each other. RapidEye is a commercial constellation of small satellites while Disaster Monitoring Constellation (DMC) is government owned but operated by DMCii.

SATELLITE TRENDS

Large earth observation satellites are being supported by government agencies. India has its IRS series and is perhaps the only country to have such a large commitment to continuing government-funded earth observation satellites and application programmes. Apart from its workhorses, INSAT, RESOURCESAT and CARTOSAT, the Indian programme also involves the piggyback launching of small satellites from different countries and more recently nanosatellites like SRMSat and Jugnu from educational institutions. Joint programmes include Megha-Tropiques and SARAL, in collaboration with CNES, France.

The recently launched Pleiades 1A is the first of a new generation satellites operated by Astrium Services. Pleiades 1A will be followed between 2012 and 2014 by SPOT 6, its twin Pleiades 1B and finally SPOT 7. Built around similar architecture and phased in the same orbit, the constellation of four satellites will ensure better responsiveness and availability of 50 cm to 2 m products through to 2023. Pleiades is a component of the ORFEO programme in which Italy is a partner with its COSMO-Skymed series of satellites.

The US Landsat programme has ended with Landsat 7. NASA has launched the new millennium programme for next generation spacecraft. The first was EO-1, which, among other mission goals, was flown in constellation mode with Landsat 7. EO-1 mission has ended and the NMP has no other satellites planned. The Landsat Data Continuity Mission (LDCM is expected to be launched in late 2012 and will carry two sensors, the operational land imager, OLI and the thermal infrared sensor, TIRS. NASA is also concentrating on their followon to the EOS missions, the earth systematic missions (ESM) programme which will continue to advance understanding of the climate system and climate change. The ESM is a three-tiered programme. Apart from this, there are joint missions with NOAA for weather and climate studies.

Europe has two major programmes, GMES and the Living Planet. The satellites are one off specific mission oriented satellites which form parts of the total programme. The Living Planet contains science and research elements which include the earth explorer missions and an earth watch element, which is designed to facilitate the delivery of earth observation data for use in operational services. Global monitoring for environment and security, GMES includes five sentinel satellites, each unique in its mission. Meteosat third generation satellites, in collaboration with EUMETSAT, will provide continuity of the Meteosat series of meteorological satellites.

José Achache, Director, Group on Earth Observations (GEO) Secretariat points out that "host payloads are a fantastic opportunity. It may be very difficult to handle because space agencies do not like that. They want to build their own satellites and they want to go for cutting edge technology and new developments. But this is an opportunity; it is going to be a new trend". Megha- Tropiques is in fact an Indian bus with hosted payloads from India and France. Another interesting view of José Achache is that "Imagery from GEO will be interesting as well because it provides a revisit time which is of the other minutes that gives an entirely different perspective on a number of highly viable processes". Matthew O'Connell feels that multiple satellite launches is also a good costcutting idea and points out that RapidEye constellation was launched this way.

While the large satellites will continue to be launched, there is a trend towards smaller single mission satellites. At the 8th IAA Symposium on Small Satellites for Earth Observation held in April 2011 in Berlin, Germany, some of the key findings were summarised by Sir Martin Sweeting, Executive Chairman SSTL. In the next 5 to 10 years there will be more constellations of earth observation satellites like RapidEye. Satellites will get fractionated, i.e. each satellite will form a functional part of a total system. There may be separate satellites for different functions like imaging, processing, transmission, etc. These could be through sparse aperture arrays, reconfigurable systems, in orbit assembly of large structures and free-flying swarms. Such satellites create a greater opportunity for participation in space activities by smaller countries as the examples of NigeriaSat and SumbandilaSat have shown. The disaster management constellation is an example of multi-nation cooperation. The challenges are regular, timely and economical launches and a method of removal of space debris that pose a serious risk to small satellites. Matthew O'Connell, on the other hand, feels that small, single-sensor satellite constellations like RapidEye may be a great idea for coverage and such satellite/ sensor combinations will grow but high resolution precision sensors will always be needed by users.

The satellites are also shrinking in size. Small satellites or minisatellites fall in the range of 100 to 500 kg in weight. Satellites in the range of 10-100 kg are called microsatellites; one to 10 kg satellites are called nanosatellites and 100 gm to one kg are called picosatellites. While mini and micro satellites are now operational, nano and pico satellites are research areas. Mini and micro satellites operate in constellations. They are controlled from the ground. Nano and pico satellites will form parts of satellite swarms which are autonomous in their control and may communicate through a 'master' satellite which could be a mini or micro satellite.

SENSORS

There are a plethora of sensors but considering the imaging sensors alone there are three major ones. The first are the tried and tested CCD multispectral and panchromatic imaging sensors. Operational sensors have already reached 40 cm spatial resolution and this could be near the limit from space borne optical sensors, according to Martin Sweeting. Matthew O'Connell of GeoEye has the same view. The next generation GeoEye-2 will have a resolution of 0.33m. This could be stretched to 0.25m but beyond this it would call for significant design change. Kumar Navulur, Director, Next Gen Products, DigitalGlobe feels that design and orbit height, which has a significant effect on satellite life, will decide the resolution limit.

According to him, 0.25m is realisable with the current technology. He also sees the number of bands increasing from four to 20 and beyond, which falls in the definition of hyperspectral sensors. The commercial WorldView 2 satellite of DigitalGlobe has an eight- band sensor. Sentinel 2 has a MSI sensor with 13 bands. NASA's OLI and TIRS data on board LDCM will provide 15m panchromatic and ten-band multispectral data, five 30m resolution in the optical range, three 30m resolution in the near IR and two 100m resolution in the thermal IR ranges. Matthew O'Connell, on the other hand, feels that more bands may only be of academic interest.

Talking of radiometric resolution, Kumar Navulur says that 11 to 12 bit resolution is essential and will increase to about 16 bits but not beyond. Such resolution will compare to aerial sensors but satellite sensors will never replace aerial sensors.

Another area is stereo coverage, which is achieved by fore and aft looking cameras or by using very agile satellites that can be repositioned to image areas at different angles to create stereo pairs. Agile satellites can also image small areas and also sweep large areas. Extreme oblique views, up to 40 to 50 degrees off nadir, are also gaining ground in some applications.

Hypespectral sensors form the second group of sensors. Some are already available on several satellites like EO1 and Aster and will be available on most of the future imaging satellites. The third group of sensors are the synthetic aperture radar, SAR. All major government backed programmes have the SAR as a major component.

There are several synthetic aperture radar satellites in orbit and more on the way. Two of these are ISRO's RISAT- 1 and ESA's Sentinel-1. RISAT (Radar Imaging Satellite)-1 is one of a series of Indian radar imaging reconnaissance satellites being built by ISRO to provide all-weather surveillance using synthetic aperture radars. The synthetic aperture radar onboard RISAT will have the ability to acquire data at C-band in different modes of polarisation, incidence angle and resolution. The Sentinel-1 mission has a C-band SAR instrument which provides three radar imaging modes with dual polarisation capability (HH-HV, VV-VH).

Other Sensors

Weather, environmental and other scientific satellites have a wide range of sensors. They include thermal scanners, optical and microwave radiometers, scatterometers and altimeters. Typical satellites are Sentinel 4 and 5 from ESA and NPOESS from NASA. While at the outset these might look like scientific missions of little use to practical resources management there are also surprises. José Achache points to the outcomes of the GRACE mission as an example. GRACE or gravity recovery and climate experiment was intended to primarily measure the earth's gravity field and its time variability with unprecedented accuracy, but it could also be used for looking at changing of the water table at small scale. Researchers using GRACE data found that there is an alarming depletion of groundwater in north-west India, largely comprising of Punjab and Haryana - a fact well known but not its extent, measure and potential for damage.

Prior to 2009, no commercially available satellite system was able to provide very high temporal resolution with consistent radiometry at a competitive price. To address this issue, the RapidEye optical satellite system, which consisted of five identical satellites, was developed and launched. In the system's design phase, the agricultural industry was identified to have the highest market potential for improved earth observation (EO) imagery and related services. Consequently, the system was designed to meet the following parameters:

•Inclusion of a special red-edge-band for vegetation analysis in addition to 4 more radiometric bands for blue, green, red and near-infrared.

•A high radiometric resolution of 12 bit for improved classification results.

•A native ground sampling distance of 6.5m, orthorectified to 5m pixel size for all five spectral bands.

•A very large imaging capacity of up to 4.2 million sq km per day.

•Automated image processing for rapid delivery and quick turnaround within 24 to 48 hours.

The biggest challenge for any optical satellite system is the desired minimal cloud coverage over an area of interest at any given time and the best way to overcome this principal problem is to increase revisit times, which in turn increase the opportunity to capture favourable weather conditions. The system, with five identical satellites and a wide swath width of 77 km, has five times more opportunity than a single satellite of the same specification. The satellites can also be pointed up to 20 degrees off-nadir. However, with more opportunities for nadir imaging, this lessens any terrain-induced geometric distortions in mountainous areas.

However, the operation of a constellation of five identical satellites does not differ much from the operation of a single satellite, as these satellites too require just one ground station contact per 90-minute orbit. In daily contacts with the control centre in Brandenburg, Germany, the satellites exchange technical status information, which is registered and analysed automatically. Every 30 to 90 days, each satellite must be adjusted back to its initial altitude by firing the onboard rocket engines. While the original design for the constellation was seven years, the onboard fuel tanks actually have enough capacity to maintain the 630 km altitude until 2017 or even longer.

Imaging the earth with a fleet of five satellites requires additional effort in relation to sensor calibration. For the end user, the quality of an image from one satellite must be equal in quality to imagery from any of the other four. In order to maintain and guarantee identical imaging parameters, the system uses a number of calibration sites located in different parts of the world. These are imaged by all satellites regularly and the collected data is used for the calculation of calibration parameters. Such procedures ensure that the satellites are calibrated relatively to each other, while the spectral information itself is not distorted.

In comparison to single satellites, there are some advantages for constellations of multiple identical satellites like better imaging performance, resulting in fast coverage of large areas, especially important for cloudy areas and the system redundancy. The system is designed to counter emergencies and the programme can be fulfilled even if a satellite ceases to function.

Dr. Rene Griesbach, RapidEye Germany

Stefan Oeldenberger

German GeoConsultants Group, Tunisia & Libya

DATA PROCESSING

Earth observation has become an important source of data but it requires processing like geo-referencing and ortho-rectification before they can be integrated into a GIS database. Such pre-processing is largely done automatically today, the only exception being the generation of true ortho-imagery. GIS-ready ortho-imagery is available which typically provides a ninety percent circle of error (CE90%) of 4.8m. It does not come cheap and may cost anywhere up to USD 90 per sq km. GIS-ready imagery is available from many sources like TerraLook, a joint project of the USGS and the NASA jet propulsion laboratory (JPL). There are many private companies that can produce customised datasets as per client needs including integration with aerial data and other sources.

It would make sense to have critical features pre-extracted like transportation and drainage since it is only based on such features that one can query a GIS. At the moment image classification, change detection and feature recognition are possible through image processing tools. The technology is looking at issues like 2D and 3D object extraction, including building reconstruction and 3D city modelling. Techniques being addressed are surface modelling and reconstruction, surveillance and change detection, learning and statistical methods for object extraction, automated sensor orientation and data fusion including information from GIS, BIM or CAD.

SERVICES

High resolution data is also high volume data. Most data users are aware of the hassles of storing petabytes of data. Even though modern data storage media have extremely high capacities but the data deluge can still overpower such capacities. Further, data once used becomes a dead investment. Data as a service is a model that is being explored to overcome such a situation. While Google showed the way initially, there are a number of such data providers like Bing and also very comprehensive sources which enable multiple dataset collection and analysis like Eye on Earth by European Environment AGency (EEA) and World Wind by NASA. India has its Bhuvan, which is touted as India's reply to Google. China too has its own Google type service.

Matthew O'Connell wants to make it easy for nontechnical people to access imagery, to manage imagery and share imagery and for this GeoEye has developed a platform called EyeQ. That then ties into another aspect of Web distribution which is the Cloud. In fact, GeoEye is using the Cloud for an NGA initiated programme called the rapid dissemination of online geospatial information. Every major military effort and every major disaster relief is a coalition effort and imagery has to be shared rapidly. That requires two things, the imagery has to be unclassified and there must be effective and efficient Web distribution so that disaster relief, whether it is in Haiti or in Japan, can get the imagery through the Web and they have to be able to select only the imagery that they need. One thing that GeoEye focussed on in the design of their service was letting people chip out just the image they needed. So if they want to see the Fukushima nuclear plant, they need not have to download too much of data because then the imagery is very large and it is hard to transmit.

DigitalGlobe solves the problem of data volume in three ways. First, scaling is used to reduce data volume by nearly 98 percent. Secondly, very fast processors are used to do the pre-processing on the fly and thirdly, data is stored on the Cloud. This enables users to get data within hours of acquisition on any device. There is also a move to extract layers like bathymetry, land cover, change detection and providing the same as GIS-ready imagery.

While these are great initiatives, it needs to be pointed out that satellite data as a service for real-time data or near-real time data is yet to be established. Real-time or near real-time high or medium data, which is needed to address disasters, requires the setting up of an operational system. The DMC constellation is one approach to address this need. Ultimately, a system which directly downloads imagery to the end user will be required. This exists for meteorologist and oceanographers but is not yet developed for land based applications.

CONCLUSION

The world of earth observation is dynamic and fast changing. Technology requires funds and end users. The utility of space- based earth observation has been proved beyond doubt but without proper downstream use of the information, the technology will remain underutilised. Private players have cutting edge technologies at their command which they have used to address niche markets but have had problems in this market. The benefits of earth observation often cannot always be valued in monetary terms. Take for example the study on groundwater depletion quoted earlier using data from a government sponsored satellite. It will be impossible to monetise this information but it will be foolhardy to ignore it because it cannot be monetised.

soyrce: http://www.geospatialworld.net/

No comments:

Post a Comment