August 24, 2016

Flying in strong winds is one of the great limiting factors of drones. Foam drones and small quads will get knocked around in wind creating blurry image captures with curvy flight lines. So, when an agronomist out on the Great Plains, unimpressed with poor image quality from light weight drones, called us we jumped at the opportunity to put the Nova to the test by flying in heavy winds.

This agronomist had been doing crop scouting at random to monitor fields. At the time, the fields were growing corn, soybeans, wheat, sunflowers, and—to the west—potatoes. The agronomist had the idea that drones could be more efficient by examining the problem areas directly. He knew, given the amount of acreage he was responsible for, drones could dramatically improve the quality of his work. He gave us a call and we headed out to the Great Plains to fly the Nova.

Planning for the Final Data Product

As with all our projects, we started prepping for the flights before we even arrived. The agronomist, delayed by poor imagery from previous flights, wanted to get to work on the crops quickly to address any problem areas. Our initial meteorological research told us we’d be facing a string of days with winds at ‘comfort-displacing’ levels. But first, we needed to choose the best payload for the final data product.

The agronomist used regular aerial photography to help identify large areas of distress, say where a sprinkler malfunctioned. However, the blurry image captures from his first drone were not helpful. Given the size and scope of the project, we knew the fixed wing Nova would be best suited to fly the fields. He wanted to collect Normal Difference Vegetative Indexes (NDVIs), which meant equipping the Nova with a Fusion MP22 payload with a near-infrared camera. While not as precise as some of the payloads for surveying and construction, this payload is cost-effective for agronomists cover large swaths of land, especially when we know we’ll be flying in heavy winds.

2z

Into the Wind

The first thing our operators do in the morning of a flight is check the Meteorological Aerodrome Report (METAR). METARs are weather reports formatted for pilots and aviation purposes. Thus, they’re highly relevant to drone operations. Using the METAR and a small electronic wind meter we determined we could expect winds around 25mph. While strong, those winds were well within spec for the Nova, and not something we would balk at.

Waiting for the perfect weather wastes time and the weather is just as likely to deteriorate as it is to get better, so we decided to fly that morning. To launch the Nova, we pointed its nose straight into the wind and just…let it go. The climb altitude was normal and the wing level and nose were pitched slightly up. The angle of climb, however, was almost perpendicular to the ground. Once at altitude, the Nova, capable of cruising at 38 mph (17 m/s) had an airspeed of 4 mph (2 m/s). The wind was strong alright.

launch-2-1

Flying in Heavy Winds

Wind speed and direction is never static. Within 15 minutes, our Flare ground control station (GCS) indicated increasing wind speeds. Winds rose to 38 mph (17 m/s) only a few minutes into the flight.

The best practice for flying in heavy winds is to plan the flight lines of your project to be in the same direction of the wind. This means if you’re not nose into the wind it should be at your tail. The idea is to keep cross winds at a minimum. However, winds change, they never stay the same. When a cross wind starts to affect flight lines, the autopilot compensates by turning the nose of the drone into the wind while still flying along the flight line. This is called crabbing. When a plane crabs at an angle greater than 25 degrees, it’s a good indication there’s strong wind. In this case, the Nova was crabbing in excess of 45 degrees, which is pretty extreme.

throwaway8

Crabbing has less effect on data collection than you might think. What will adversely affect data collection is if the plane is knocked around by wind and rolls on its axis. Crabbing provides stability in wind, but only mass and aerodynamics will mitigate gusts. Each degree away from nadir, that is straight top-down orientation, adversely affects data processing. Data becomes difficult to process with distortions and parallax occurring due to the camera’s orientation not being straight top-down.

Despite the strong winds, the Nova only rolled within +/- 2 degrees of nadir for the majority of image captures during the flight. The Nova’s strong structure and mass keep it from rolling on its axis even with wind gusts more than 40mph. Here’s a graph showing the number of images captured and the degree of roll they were captured at in these winds.

throwaway10

Processing The Final Data Product

When the Nova completed the flight lines and returned to land adjacent to the GCS. With the data extracted from the Fusion payload, we headed back to the office to load it into DroneDeploy for overnight processing. The Fusion payload can integrate with almost any data processing software, allowing the agronomist his pick of the bunch. The agronomist previously used DroneDeploy and preferred it for data processing because it was easy to share the data with his crop scouts.

By the following morning, DroneDeploy had finished processing the data. We sat down with the agronomist and examined the results: some beautiful NDVIs.
throwaway6
The agronomist applied a grid and zone filter, exported the data directly as a shapefile, and imported it into SMS, the program he uses every day for field management.

throwaway7-jpg

With a quality data product, despite flying in heavy winds, the agronomist decided to widen the scope of the survey. We wound up flying over 200 flights covering 60,000 acres of nearby farms for the agronomist in the following months. All with that same, wind-resistant Nova drone from our first flight on the windy Great Plains.