2016 – A year of progress in Earth Observation

For me 2016 marked a point when the term ‘Remote Sensing’ was used less frequently and ‘Earth Observation’ significantly more. Make of that what you will. With the impending deluge of data (not sure at what point we move from impending to actual) the race is now on to derive information from the pixels.

You could argue that there are 2 areas of interest in EO today.

  1. The actual imagery “pixels” – basemaps, classification, band ratios, NDVI etc. In this world the output is almost always a raster or graphs based on the information contained in that raster.
  2. Analytics based on feature extraction – counting objects, tying into existing known data that is non EO based, detailed models & metrics on change detection. In this world the output can be anything and the user might not know/care that Earth Observation data is providing the information.

Being able to derive value from Earth Observation data is the challenge. It is a challenge across all parts of the industry – whether through the utilisation of the Copernicus program or justification for launching another high resolution satellite. – How many can the market take?

2017 looks to be an exciting year.

Here are some 6 amazing things being done with Earth Observation I have found this year – though they may have been around longer :).

Machine Learning + Satellite data = mapping poverty


Forecasting food crises

Oil supplies from space

Counting ships in port

Terrapattern is like a search engine for satellite imagery

Google Earth Engine Timelapse

I anticipate even more of these use cases/projects/valuable business metrics in 2017. It is going to be fun.