Autonomous drones show promise for shooing birds away from crops

Birds regularly eat large quantities of crops, and often become accustomed to stationary devices designed to scare them away. That’s why scientists are now looking at using autonomous drones to do the job.

In a study recently conducted by a team from Washington State University, cameras were installed around small plots of land, accompanied by customized multicopter drones. The live output of the cameras was continuously analyzed by a machine vision algorithm, which had been trained to recognize the wing-flapping movement of birds.

Since it would have been difficult to have a good supply of birds ready to go whenever needed, the researchers instead had field-roaming volunteers mimic the flapping motion with their hands. The system was able to identify those movements approximately 92 percent of the time, responding by deploying a drone to hover at the relevant location.

Of course, the human volunteers wouldn’t have been scared of the aircraft. In previous studies conducted by the same scientists, however, it was found that manually piloted drones were very effective at chasing off actual birds.

In one study, use of drones resulted in a four-fold reduction in bird counts in a given area. In another, fields in which bird-scaring drones had been used had a 50-percent reduction in fruit damage. Additionally, an unrelated study announced earlier this year showed that an automatically deployed drone was effective at chasing pigeons off a rooftop.

That said, the researchers state that more research still needs to be conducted, particularly to see if birds just get used to the drones over time. If they do, additional measures will need to be taken.

“We could make drones look like predators, or have reflective propellers that are really shiny,” said the lead scientist, Assoc. Prof. Manoj Karkee. “All of these working together would likely keep birds away from those vineyards and fields. We need to research that over multiple years to make sure.”

A paper on the study was recently published in the journal Computers and Electronics in Agriculture.

Source: Washington State University via EurekAlert

Birds regularly eat large quantities of crops, and often become accustomed to stationary devices designed to scare them away. That’s why scientists are now looking at using autonomous drones to do the job.

In a study recently conducted by a team from Washington State University, cameras were installed around small plots of land, accompanied by customized multicopter drones. The live output of the cameras was continuously analyzed by a machine vision algorithm, which had been trained to recognize the wing-flapping movement of birds.

Since it would have been difficult to have a good supply of birds ready to go whenever needed, the researchers instead had field-roaming volunteers mimic the flapping motion with their hands. The system was able to identify those movements approximately 92 percent of the time, responding by deploying a drone to hover at the relevant location.

Of course, the human volunteers wouldn’t have been scared of the aircraft. In previous studies conducted by the same scientists, however, it was found that manually piloted drones were very effective at chasing off actual birds.

In one study, use of drones resulted in a four-fold reduction in bird counts in a given area. In another, fields in which bird-scaring drones had been used had a 50-percent reduction in fruit damage. Additionally, an unrelated study announced earlier this year showed that an automatically deployed drone was effective at chasing pigeons off a rooftop.

That said, the researchers state that more research still needs to be conducted, particularly to see if birds just get used to the drones over time. If they do, additional measures will need to be taken.

“We could make drones look like predators, or have reflective propellers that are really shiny,” said the lead scientist, Assoc. Prof. Manoj Karkee. “All of these working together would likely keep birds away from those vineyards and fields. We need to research that over multiple years to make sure.”

A paper on the study was recently published in the journal Computers and Electronics in Agriculture.

Source: Washington State University via EurekAlert