The drones delivered high resolution fully-georeferenced Google Maps

Technology

HS2 drones deliver mind-blowing measurements in weeks

6 December 2018 | By Stephen Cousins

One of the largest drone surveying jobs in the world, covering 230km of the High Speed 2 (HS2) route from London to Birmingham, was able to capture an eye-watering 18.2 billion data measurement points in just three weeks.

The controversial £53bn HS2 project will comprise more than 530km of new track. Phase one is scheduled to open in 2026 and phase two, with additional stations in Manchester, near Nottingham and in Leeds, in 2033.

Digital topographical surveys were required to plan enabling works for phase one, but using traditional static 3D laser scanners would have taken several months, requiring workers to walk and climb across hazardous terrain. The client also wanted to avoid the need for constant landowner engagement to perform the surveys, which had already caused hold ups.

SenSat was uniquely positioned to map the route because it has a permission from the Department for Transport to fly drones up to 12km away from a pilot, where legislation normally limits this to 500m, or “within line of sight”.

“The special permission was a result of a decision by the overnment to allow Amazon to test drone deliveries in Britain under the Department for Transport's Drone Pathfinder Programme,” explains James Dean, CEO of SenSat. “Pathfinder was set up to identify ways to fly beyond visual line of sight of the pilot and SenSat was appointed as the infrastructure partner.”

This made it possible to capture data much more efficiently, covering over 23km per day to a width of 200m and an accuracy of 30-50mm.

SenSat’s SUAVs are miniature unmanned planes and once in the air they are controlled using automated commands from a laptop. Real-time kinematics (RTK) positioning technology makes it possible to precisely locate them to about 15mm accuracy.

The drones are fitted with high resolution cameras that can capture thousands of photos that are subsequently stitched together to create full-colour point clouds.

“The accuracy of the data we are generating is RICS Band E, ie good enough for the early stages of a project control framework, but not good enough for detailed engineering design. The absolute accuracy of the HS2 dataset was 30-50mm, whereas you would need 10-20mm for engineering design works,” says Dean.

A second project deliverable was a set of 2D orthomosaics, essentially high resolution fully-georeferenced Google Maps, from which ancillary datasets for the Digital Train Model and other CAD outputs could be produced.

The vast linear nature of the project meant the curvature of the earth had to be reflected in the survey output. A snake grid projection system converted the data into the the correct format, an internal team at SenSat, led by an ex-GIS professor, worked on quality assurance to ensure the corrections were accurate and met with survey requirements.

The point cloud dataset was so big that SenSat had to build a dedicated online platform to host the model and enable users to interact with it. “Even though it’s online it is high fidelity data – you can zoom in and see a coke can on the ground,” says Dean.

Looking ahead, HS2 is considering the use of drone-captured data to calculate the volume of earthworks faster and more cheaply than traditional methods where surveyors have to walk the route. They may also be deployed to make accurate comparisons of “as-designed” versus “as-built” structures to optimise project control and minimise cost and time overruns. 

SenSat was recently ranked as Europe’s number one artificial intelligence company by data provider Crunchbase. The company is developing a ‘simulated reality’ system that involves capturing the physical environment, using various laser scanners and drones, then reconstructing it in 3D and embedding real-time information, such as locations of workers and machines, the temperature, CCTV feeds and more.

“We are working to teach the computer code to understand certain things, such as identify a digger and know that it is digging a hole, to help it understand the world in the same way as humans. It’s a very difficult task, but civil infrastructure gives us a fixed environment with known variables so we can test and validate those things,” Dean concludes.