Drones have evolved at such a rapid rate over the years it is hard to imagine life without it. So much so, now, the inclusion of artificial intelligence within the drone industry is due to change life as we know it.
The incorporation of AI (Artificial Intelligence) allows drone vendors to use data from sensors attached to the Drone in a bid to
collect visual data.
So how has AI influenced the drone industry?
Until recently drones were only able to capture what the
in-built camera saw. But now thanks to AI software, drones can soak up more of
its surroundings, enabling it to track objects, map areas and provide
analytical feedback in real time.
Neurala, based in Boston, is a network that helps drones
sift through crowds to identify a person of interest. The company claims that
with the use of AI software, it can understand an image of an individual in 20
minutes rather than hours or days, when carried out by traditional equipment.
Using this service, is a company called the Lindbergh Foundation who use
Neurala-powered drones to catch elephant poachers in Africa. By using the
company’s image recognition, it can now monitor elephant herds and spot
poachers miles before they reach the elephants.
Another company using AI powered drones is Skycatch that
builds software that captures and analyses data from aerial images. Skycatche’s
software turns images into 3D meshes or thermal images to get an overview of
the land being surveyed. Japanese construction company, Komatsu, uses
Skycatche’s drones on more than 5,000 sites. It now only takes 30 minutes to
process aerial images compared to days for humans to derive the same task.
So how AI powered drones work?
There are two main components for this to be achieved.
One of which is Computer Vision and the other is sensors. So lets start to
explain exactly what Computer Vision is about.
According to reports, Computer Vision is gained through
onboard image processing with a neural network. A neural network is a method
used to install algorithms in machine learning. This enables a drone to carry
out object detection, classification and tracking. So for example if a drone is
used for deliveries, it can avoid collision, locate and track targets.
But to install neural network in drones researchers need to
train the machine learning algorithms to recognise and categorise objects in a
wide variety of contexts. This is created by marked images into the algorithms.
This allows the neural network to differentiate one object from another.
And now we move on to Sensors. Sensors collect data
including visual, positioning and environmental.
This information is fed to machine learning modules to
gain knowledge on how a drone should respond to environmental conditions, where
it is allowed to fly to and avoidable objects.
This data is then fed to machine learning models to
determine how a drone should respond to environmental conditions, what objects
it should prioritize or avoid, and where it can fly to. Often sensor data is
also used after a drone has landed in non-flight related analyses. For example,
to find potential mining locations or to evaluate the water quality of
reservoirs.
Ends
Comments
Post a Comment