Catching Wavelength 2017

Remote sensing, like GIS, excels in integrating across disciplines and people. Whilst no one ever said being a multi-disciplinary scientist was going to be easy, for the ‘thirsty’ mind it challenges, cross pollinates ideas and looks at problems with new eyes. A diverse group of people connected by a common thread of spatial and remotely sensed data found themselves doing all these things and more in London last week at the Wavelength 2017 conference.

The talks and posters took us on whirlwind tour through the ever varying landscape of remote sensing. We moved through subject areas ranging from detecting ground ice, vegetation and overall land cover, through to earth surface movement and 3D imaging, and onto agriculture yield and drought. We also covered the different vertical scales from which remotely sensed data is collected, whether from satellites, planes, drones or cameras operated from ground level. On top of this focus we also had some great key note talks, running through the varied career of a remote sensing scientist (Groeger Ltd), as well as in depth data assimilation of remote sensing imagery in models (UCL) and commercial developments in airborne camera work (Geoxphere Ltd).

In parallel, we were taken on a grand tour covering the temperate UK, parts of the Middle East, the tundra in North America, the central belt of Africa, and even onto the moon and Mars! In many cases we heard talks from scientists from these countries (though not the moon or Mars …). Some are based at the universities in the UK, whilst, others came specifically to talk at the conference.

I found myself transfixed by the far flung places. Listening to how the dark side of the moon is being mapped, a place that never sees daylight and is incredibly ‘chilly’ and traps ice in these shadowed lands. I also heard about the CO2 that precipitates out of the atmosphere on Mars as snow and forms a 1m blanket. Working in places like Africa started to feel really quite local and accessible!

Possibly the most intriguing aspect of the conference for me, was the advancements that have been made in photogrammetry and how multiple photos are now being used to produce highly intricate 3D models. We saw this applied to cliff morphology and change detection, as well as the 3D point clouds that are produced when modelling trees and vegetation generally.

The 3D models aren’t totally complete due to line of sight and other issues. The model visualisations look like an impressionist painting to me, with tree leaves without trunks or clumps of green mass suspended in mid air. However, this does not matter when calculating leaf volume and biomass, as these discrepancies can be worked with and lead to some very useful estimates of seasonality and change.

Setting this up is no small feat for the organiser, and PhD student, James O’Connor. He delivered an interesting programme and looked after the delegates well. I can truly say I haven’t been to such a friendly conference before. It was also unique in providing ample time to discuss aspects of material presented, both from talks and posters, and sharing technical know-how. This felt of real value, especially to the PhD students and young professionals this conference is geared towards, but equally myself with experience in only certain fields of remote sensing.

I would highly recommend Wavelength, and look forward to seeing what they are planning for 2018!

Blog written by Caroline Chambers, Pixalytics Ltd.

Satellite is SPOT on!

July 2009 SPOT image of Grand Cayman, data courtesy of ESA / CNES

July 2009 SPOT image of Grand Cayman, data courtesy of ESA / CNES

This week marks the thirtieth anniversary of the launch of the first SPOT satellite, making it one of the longest satellite time series datasets.

The French Space Agency (CNES) established the series of Satellites Pour l’Observation de la Terre, known as SPOT. SPOT-1 was launched on the 22nd February 1986 and was fitted with revolutionary steerable mirrors, meaning the satellite could look in multiple directions enabling it to observe the same point on the Earth every five days.

Since SPOT-1 there have been six subsequent satellite launches, with a strong thread of sensor consistency throughout the series, meaning it is much easy to compare imagery over time. Currently, there are three active satellites in the series – SPOT-5, SPOT-6 and SPOT-7.

Looking at SPOT’s history:

  • SPOT-1, SPOT-2 and SPOT-3 all had identical imaging sensors, namely a panchromatic band with 10 m spatial resolution, and multispectral bands in green, red and near infrared (NIR) with 20 m resolution.
  • SPOT-4 had identical multispectral bands to its predecessors, but it also added a middle IR (MIR) band. The panchromatic band operated at slightly different wavelengths, but with the same resolution. In addition, SPOT-4 also carried the first Vegetation instrument with blue, red, NIR and MIR bands at a 1 km resolution that effectively gave daily global coverage.
  • SPOT-5, launched in 2002, offered a step change in spatial resolution. The multispectral green, red and NIR bands had an improved resolution of 10 m. The panchromatic band was at 5m, and returned to its original wavelengths. The vegetation sensor was identical to that flown on SPOT-4.
  • The latest incarnations of the series, SPOT-6 and SPOT-7, launched in 2011 and 2014 respectively, operate as a tandem constellation and again offer an improvement in resolution. The panchromatic band is down to 1.5m, and the multispectral green, red and NIR bands are down to 6m. They are expected to provide data through to 2024.

A time-series continuity of the Vegetation Sensor has been provided by the Belgian Proba-V satellite that launched in 2013, and will be carried on into the future by the OLCI sensor on the recently launched Sentinel-3A mission.

Unlike the oldest time series data, Landsat, SPOT is still categorised as a commercial dataset and its imagery has to be purchased.

The US Geological Survey does have a contract to provide historical data from SPOT-4 and SPOT-5 for North America for some US government staff, and ESA provide limited datasets for approved projects. There was an announcement by the French Government in 2014 that SPOT satellite data, over 5 years old, would be free of charge for non-commercial users – although we’ve struggled to find it!

SPOT’s applications have included exploring for gas, oil and minerals including routing pipelines; mapping the planet including forestry, topographical maps and urban planning; agriculture data to combat drought and support farmers decision making; emergency rapid response information for disaster relief and urban planning. In addition, from SPOT-5 onwards DEM’s can be created using photogrammetry techniques, because of the instruments produce stereoscopic images allowing minute changes on the Earth to measured; and the global coverage of Vegetation sensor has also contributed towards climate change research.

SPOT-1 provided imagery ten days after the Chernobyl disaster, and also picked up photosynthesis in the area in 1988 using its NIR sensor, and was recently used to look at camps for Syrian Refuges in recent times.

The SPOT series of satellites have made a huge contribution the development of remote sensing time series datasets, and that’s worth celebrating.