AI Backed Big Data in Space: Satellites & Drones
|AI enabled analytics on EO data will be competing with drones|
Having closely worked on satellite systems since 2014, analyzed remote sensed data from multi-spectral LANDSAT data for a journal paper [HERE] and worked on a drone startup that aimed at providing time-scaled, AI enabled, 3D extrapolated data in 2017, I can see a future where Earth Observation (EO) Satellite and drones will be either 1) competing in it's own niche market 2) joining forces to provide an all inclusive data set ranging from medium, high, very high and extremely high analyzed resolution imagery that will allow industries in agriculture, urban planning, construction or safety make better, extremely accurate informed decisions. The latter is ideal but there will be competition nevertheless.
AI is great at repetitive, ultra boring pattern finding work. And that's why they are such an effective tool to get information out imagery acquired from satellites. Space News back in January of 2018 reported that "All told, AI-driven EO start ups raised $96 million in 2017 - nearly three times than 2016." Looking at 2017, 2018 is projected to have even greater funds available for exponential, robust growth in the sector. With more satellites in LEO coming online, and with higher temporal and spatial data being readily available in the market, these start-ups will have plenty of fodder to feed on.
Drones, in fact, are already ahead. While AI wasn't implemented on the startup that I worked in, it was clear that extremely high resolution, extremely fast output, 3D enabled AI data-driven analytics that gave construction, repair, mining and gas-oil companies the edge was the intended purpose even if it means it's a pivot from humanitarian tag that the company was trying to sell. By providing a simple web interface like how one would log into facebook, companies now would have access to a wealth of information previously unfathomable.
|Drone startup AngelSwing from Korea had already been looking into NN to help analyze data in early 2017|
Photo Credit [HERE]
It's anything but inevitable that the two will clash to secure the market in the long run. The same article from the Space News magazine reports that "drone companies are ahead" in "reaching target industries" which I can anecdotally confirm too. What drone cannot do compete however, is the scale at which Planet, for instance, have been mapping the world. Scaling reaches out to a wider spectrum of data consumers, on a wider spectrum of actual captured data and can profit from one sector if the other doesn't give those returns. Drones, well, drones are a bit tricky in that the efforts are more focused on a particular sector at any given time.
Funny enough, this could all come down to an interesting time when market is saturated with analyzed data from all parts of the world and that there might not be enough demand to keep up with the supply. A data overload will particularly hit startups who rely on a day-to-day exponential growth. This could well be true for companies like Angelswing (as an example, no worries Peter) if they are not to rely on government funding and have a sustainable business plan. Or they might just collaborate with satellite companies. As the cliche goes, if you can't beat them, join 'em.