Remote sensing: efficiently searching the seas

Building on my post from the 27 February, it seems that 2014 certainly is a year where remote sensing is hitting the mainstream. The reasons for this may not be happy ones (floods, missing aircraft, landslides) but the use of satellite imagery (and satellite technologies in general e.g. the signal processing and modelling undertaken by Inmarsat) has been crucial to the response taken to many global emergencies. This has, of course, been the case for many years but the profile of the use of satellite imagery has really been raised in recent months. The response to the current, and many other, disasters and emergencies is incredible. Image data are captured, downloaded, processed and disseminated (to enthusiastic volunteers as well as professionals) across very short timescales. Complex information is broken down into only that which is required. (Extremely) Large spatial extents can be covered and data from different time periods can be collected. Using different types of systems and different wavebands, data extraction can be filtered in a way that helps the analyst.

International rescue

The tasks that we ask our satellite imagery to help us with are not insignificant. Take the current case of Malaysia Airlines Flight MH370, flying between Kuala Lumpur and Beijing and carrying 239 people (12 crew and 227 passengers), which lost radar contact with Vietnamese air traffic controllers on 07 March 2014.  A search was instantly started for the aircraft. On 11 March the search area expanded to the Malacca Strait, west of Malaysia and on 15 March it grew to include the Indian Ocean. It’s difficult to comprehend the scale of the task at hand, but one commentator on the news said it best with “imagine you are trying to find a single bus in continental USA, and you don’t know where to start”.  Since the start of the search things have improved –  the authorities are now just trying to find the bus in an area the size of Alaska! The absolute numbers may not be exactly correct, but the impression of the massive scale of the search is well made. There is an article here that demonstrates how an individual satellite image fits into such an area.

The same article also shows how the different wavebands of a satellite imaging system can be used to enhance the possibility of spotting a metal target against a background of water. Different imagery have been collected across different parts of the search area. Part of the reason for this is that not all satellites collect from all regions of the globe, especially the open ocean and extreme latitudes.

Numerous systems

Some of the satellite systems that have been supplying imagery to the search are listed below (there are likely to be others that I haven’t found reference to yet):

  • TerraSAR-X (Imagery here);
  • Landsat-8 (Imagery here);
  • Terra Aster (Imagery here);
  • RapidEye (Imagery here);
  • Thailand Earth Observation Satellite (THEOS / Thaichote) (Imagery here);
  • LandSat -7;
  • SPOT 5 and 6;
  • WorldView-2 (see below);
  • Pléiades 1A and 1B (see here);
  • EO-1 multispectral and hyperspectral imagery; and
  • Photography from the International Space Station.

The following image has been created by Mapbox and is hosted on the Washington Post site. It shows a number of different satellite paths and the footprints of images collected by the satellites mentioned. One thing to notice is how small many of the individual images are, and therefore how many will need to be collected and analysed to cover such a large search area (and this only shows the original area around Malaysia). Click on the image to go to the original article.

Satellite-search-2

Power of the crowd

One of the most high-profile, and successful satellite imaging platforms was WorldView-2.  The high resolution, multi-spectral imagery collected aboard Worldview-2 has been used in the Tomnod crowd-sourced feature tagging web-application. Over three million people have helped tag the imagery so far. You can also help out by following this link: http://www.tomnod.com/nod/challenge/mh370_indian_ocean Another company, Mapbox, is doing something similar using RapidEye satellite imagery. Satellite image data are also being disseminated using the Hazards Data Distribution System (http://hdds.usgs.gov/hdds2/) provided by the USGS.

Multi-national, multi-agency, multi-use

Other satellite systems have also been alluded to in the media. For example,  China is utilising ten of its satellites to image the search area, and French radar imagery has also been mentioned as a source of information, identifying 122 objects on the ocean surface for further investigation . The names and specifications of these satellite systems have not been named in the media at the time of posting (although the French systems may be those listed above) and could be military grade platforms. It is also probable that other classified systems are being employed to locate the aircraft.

But satellite remote sensing systems are not the only platforms being utilised.  Vessels and aircraft from at least ten different countries are participating in the search. Many, if not all, of those will include radar, thermal and other imaging systems to help the spotters on-board.

Remote sensing is truly becoming integrated into the way that we do things – how can it help you be more effective?