Deep Forestry is a Swedish startup that focuses on robotics and machine learning. With their focus on forest inspections, they have requested Avular to build a drone that can autonomously navigate in the forest, between the trees, under the canopy.  

For this, Avular used the standard Vertex drone as basis and made specific adjustments to meet Deep Forestry’s requirements allowing for precise forest inspection. This not only includes adding a 3D spinning LiDAR, 4 cameras and a high-power computing platform but also modification to the airframe to allow for carrying the extra weight whilst ensuring minimal impact on the sensors’ field-of-view. 

In essence, the 3D LiDAR is mapping the forest area by creating a so-called point cloud which provides a high-accurate 3D representation, like a map, of the forest. As the LiDAR is unable to ‘see’ colors, the cameras are being used to add real-life colors. The result is a digital representation of the forest area including all kinds of information like tree location, species, diameter, volume and curvature.  

Drone specs: 

• Size 45x45x50cm (LxWxH) (props folded)
• Weight 3.9Kg
• Height LiDAR
• Barometer
• RTK-GPS
• Magnetometer 
• 4x camera (5MP color)
• 3D spinning LiDAR

Compared to the standard Vertex, this drone is much stronger and heavier equipped. The high-power computing platform allows for processing of heigh amounts of data and enables the customer to run their software for autonomous flying at the very same time. Although the drone can be operated by remote control, the typical scenario is for the drone to fly autonomously using RTK-GPS for localization and LiDAR to map (and avoid) objects in the forest.