As Drone Navigation Advances, SlantRange Focuses on Farm Analytics
(Page 2 of 2)
the U.S. military. According to Ritter, the SlantRange team has an advantage over most drone-related startups because they have a good grasp of the practical limitations of airborne surveillance from their work on sensors and other “mission systems” for the Predator.
“Data acquisition is extremely perishable,” said Ritter, who spent a decade at GA Aeronautical Systems. “A pest infestation can get out of control quite quickly and take over a crop. So making sure the data is accurate and immediately available makes it extremely valuable” to agronomists and farmers. Some rivals address this need for speed by designing their system to upload imaging data to the cloud. But SlantRange took a counter-intuitive approach.
The company developed a new technique for data compression that allows much of the data-crunching analytics to occur in SlantRange’s on-board sensor module and the tablet computer used to control the drone.
“Access to computing sources around the world do not line up well with agricultural lands,” Ritter explained, and bandwidth is a rare commodity. “The volume of data produced by these systems is just so large—maybe 200 to 500 megabytes per acre—that the data overwhelm the uplink capability” in most sparsely populated rural areas. “You need a [high-capacity] T1 connection to upload the data.”
In a field trial for a potential customer in the wheat-growing region of South Africa’s Northern Cape, Ritter said an Israeli-based competitor had to drive several hours to Johannesburg to transmit its data to Israel for analysis. It took several days to get the results to the customer, and SlantRange won the business, Ritter said.
In the meantime, Ritter said the FAA rules for operating radio-controlled drones by visual line-of-sight shouldn’t hinder SlantRange. Keeping a drone in sight usually means the operating range is “something just short of a mile for the types of systems we use.”
Ritter said American farms are usually comprised of multiple non-contiguous parcels spread over a large area. “Our typical Midwest customer is a 10,000 to 12,000-acre farm, which is comprised of perhaps 75 fields spread over several hundred square miles,” he said. Such jobs require the drone service provider to move from parcel to parcel anyway, so flying a drone within view usually isn’t a problem.
Ritter also maintains that low-altitude, slow-flying drones also can deliver very high-resolution images at a much lower cost than a farmer would get from manned aircraft or satellite imagery.
“By very high resolution, we mean resolution whereby the pixel size is smaller than individual leaves on the plant,” Ritter said. “Once you’re in that regime, you can begin invoking techniques in artificial intelligence and computer vision to deliver new types of information that are not possible from lower resolution manned aircraft or satellites. And that is where the real value is in multiple respects for delivering more specific and actionable information to growers.”