This week we‚Äôre looking at this year‚Äôs key trends in Earth Observation (EO) that you need to know.
Rise of the Data Buckets!
EO data is big! Anyone who has tried to process EO data knows the issues of downloading and storing large files, and as more and more data becomes available these challenges will grow. Amazon recognised this issue and set up Amazon Web Services which automatically downloads all freely available data such as Copernicus and Landsat, offering people who want to process data a platform where they don‚Äôt have to download the data ‚Äď for a price!
The European Commission also picked up on this and awarded four commercial contracts at the end of last year to establish Copernicus Data and Information Access Services (DIAS) which will offer scalable processing platforms for the development of value-added products and services.
The four successful DIAS consortiums are led by Serco Europe, Creotech Instruments, ATOS Integration & Airbus Defence and Space respectively, and a fifth DIAS is planned to be established by EUMETSAT. It‚Äôs hoped this will kick-start the greater use and exploitation of Copernicus data.
Continued Growth of Data
There are some exciting EO launches planned this year continuing to increase the amount of data available. Earlier this week China launched the last two satellites of the high resolution optical SuperView constellation. In addition, some of the key larger satellites going into orbit this year include:
- ESA‚Äôs Sentinel-3B and its Aeolus wind mission.
- NASA‚Äôs Gravity Recovery and Climate Experiment Follow-on (GRACE-FO) and the Ice, Cloud and land Elevation Satellite (ICESat-2).
- Japan‚Äôs Advanced Satellite with New system Architecture for Observation (ASNARO 2) which is x-band SAR radar satellite with a 1 m ground resolution.
- NOAA‚Äôs GOES-S is the second of four upgraded weather observatories.
In addition, as we described last week, cubesats will continue to have regular launches. We are still a long way from the high watershed of EO data!
SaaS Will Become The Norm
The rise of the data buckets will encourage the Software-as-a-service (SaaS) approach to EO to become the norm. Companies will develop products and services and offer them to customers on a platform via the internet, rather than the historic bespoke application approach. For companies this will be a more effective way of using their resources and will allow them to better leverage products and services. For the customers, it will enable them greater use EO and geospatial data without the need for expert knowledge.
Pixalytics is due to launch its own Product Portal at the Data.Space 2018 conference at the end of this month.
Artificial Intelligence (AI)
AI is becoming more and more important to EO. Part of this is the natural development of AI, however certain EO tasks are far more suited to AI. For example, change detection, identification of new artefacts in imagery, etc. These aspects have a base image and looking for differences, computers can do this much quicker than any human researcher. Although, it‚Äôs also true that humans can see artefacts much more easily than you can program a computer to identify them. Therefore, these AI applications are strongly dependent on training datasets created by humans.
However, things are now moving beyond these simple AI tasks and it‚Äôs becoming an integral part of EO products and services. For example, last year Microsoft launched their AI for Earth programme, support by a $50 m investment, which will deploy their cloud computing, AI and other technology to researchers around the world to help develop new solutions for the agriculture, biodiversity, climate change, and water challenges on the planet.
These are a snapshot of our view of the key trends. What do you think? Have we missed anything? Let us know.