From the Laptop to the Cloud: How to process satellite imagery?
Over the years satellite imagery data size has been constantly receiving upgrades.
If you look back a few years ago, let’s say 2005, the Earth Observing satellites orbiting Earth were measured with two hands. Currently (2018), there are more than 300 operational earth observation satellites. To be more precise, more than one company (planet.com) have managed to image the entire earth’s landmass every day! And several more companies are coming, including video!
The most common spectral range that earth is being digitized is the visible range. Then near infrared and thermal infrared follow. Significant attention is show in the microwave spectrum with operational satellites like ESA’s Copernicus Sentinel 1A and 1B.
Before going to answer the question, just see the big picture:
According to the “Study to examine the socio-economic impact of Copernicus in the EU”, more than 10.000 related jobs will be needed!
So a very big question rises. How are we going to process or better extract information from so “big data”?
Before examining this, think how we (or at least most of us) currently process the satellite imagery.
- We define our Area Of Interest (AOI).
- Then we order the data and we download the data to our laptops/desktop computers. The most common way is to use online platforms like ESA’s Scihub, USGS’s Earth Explorer, Planet or UrthCast. As the terminology suggest, they are platforms to find and download satellite images (either open like Copernicus or Landsat data or proprietary).
- Now we prepare the downloaded data and start using our remote sensing software and scripts. GEO University provides tutorials for some of the most common softwares and tools used for this step, which are ENVI, ESA SNAP and Python/ArcPy (and more are planned to come!)
- After we finish with all our processing steps, it is time to wrap up and deliver the output to our customer.
Just by reading these 4 steps, you can understand that when you want to deliver to a customer an operational service, this approach can be considered either obsolete or highly costly!
How to upgrade our way of processing?
Having in mind the above, the answer is quite simple. Move to a cloud that has all (or at least the ones you need) the data you need close. A commonly known moto:
Move the algorithms to the data and not the data to the algorithms.
CloudSigma offers a combination of raw underlying computing power with an approach and tools that channel computing power in the most effective way. The result is best price/performance on the market and a solution that customers can and do build production services with.
CloudSigma actually replicated the ESA’s archive that contains the EU’s flagship space program, Copernicus. This gives you a perfect opportunity to practice your skills in a professional and secure cloud environment.
So, let’s say you are practicing on ESA’s Sentinel Application Platform (SNAP) software.
After learning how to process Sentinel 1 and 2 with this practical tutorial, it’s time now to move your this software (and your knowledge) in a CloudSigma hosted Virtual Machine with these data laying right next to your VM!
It’s (almost) that simple. Stay tuned, more are coming to make your Earth Observation skill state-of-art with new cloud dedicated courses.
Need more information?
E-mail here: info@geo.university
