3D mapping made easy: no drone skills required
Pilkington and his long-time friends Mike Winn (CEO) and Jono Millin co-founded DroneDeploy in 2013 and since then, have raised a total of 31 million US dollars, now boasting thousands of customers spread across 160 countries on all continents.
Partly, the startup owes its success to the fact that it is a drone-agnostic software-only company.
a drone’s flight and collecting real time 3D mapping
information.
“A key difference from other 3D mapping tools offered by drone manufacturers is that we don’t build any hardware, we are just a software company and we don’t build walled-gardens around specific drone models. That means our customers can use the best software, for any drone they choose”, Pilkington told eeNews Europe during a phone interview.
“We can fly any DJI phantom or DJI inspire 100, 200, 211 drones. We can process any sort of aerial imagery uploaded to our cloud-based operation. Customers use us for processing and analysis and because we are cloud-based, we can process more than one job at a time like it would be the case for software running on a person’s desk. Our solution is scalable to large drone operations with many pilots flying at the same time and in different places”.
Although Pilkington acknowledges that DroneDeploy’s target customers are mainly found in agriculture and building, he says the platform is horizontal, meaning it could be used in any vertical market.
“Users can customize anything in the interface so it becomes part of their workflow, whether it be in mining or forestry or agriculture. They can build their own processing and analysis tools.”
video feed.
“Our offer comes in two pieces. One is real time photogrammetry, happening as the drone flies for direct visualization on a phone or tablet. The second part is a cloud-based photogrammetry that gives you really accurate 3D maps (down to a resolution of 1cm/pixel). We try to select the best for our customers, so they don’t have to be experts to use the software. Our motto is: A drone on every job site”.
On its website, DroneDeploy boasts it holds the largest drone data set in the world, covering over 20 million acres of land across 160 countries and all 7 continents. Was the company able to learn anything from that data? We asked.
“We are starting to leverage machine learning to get the most out of data, to tackle the most tedious problems. One nice example is to identify the Ground Control Points (GCPs) used to tag the exact GPS location of given points on a site”.
That information is used to generate precision maps and setting them correctly in the real world.
“Today, it takes about an hour to visually identify and manually process 20 GCPs points. This process could be removed entirely with AI. We have seen hundreds of thousands of GCPs, the way they are exposed in uploaded imagery” says the CTO.
“By knowing the most about the data that is collected, AI can benefit users, it gets to know what the customers want to do with it. For example, if you identify stockpiles, it is likely you will want to perform volumetric measurements. If you identify GCPs, then you’ll want to tag and process them.
Where it gets really exciting is when AI is on the product, with machine learning able to make decisions during the drone’s flight. It can work to guarantee good image quality, routing the drone for a second pass where and when necessary”, Pilkington continues.
“We’ll be able to go beyond today’s 3-step process (fly/process/analyse), all these things are going to be intertwined and integrated, no longer a linear 3-step flow”.
DroneDeploy offers its software services on a monthly or annual subscription basis, starting with a free service (with a limited resolution and support) and scaling up to various support schemes, including the uploading of an unlimited number of video feeds for maps and 3D model processing, together with full support for metrics extraction, contouring and other useful features.
DroneDeploy – www.dronedeploy.com
Related articles:
Parrot takes instant 3D mapping to the sky
Open source drone navigation kit integrates GNSS module
Vision processing and AI set to take-off, literally
Real-time 3D SLAM reaches millimetre-level accuracy
Bat-inspired ultrasound 3D mapping could equip drones
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
