ReLeaf

Aerial Reforestation Using Autonomous Drones

Context

Desertification in Spain is a significant environmental issue driven by a combination of natural and human factors. Over 74% of Spain’s land is at risk of desertification, particularly in regions like Andalusia, Murcia, Valencia, and Castilla-La Mancha. For Spain, this is a massive problem as it leads to biodiversity loss, the degradation of fertile soils, impacts on water resources, and has economic consequences, all of which ultimately affect people’s quality of life.

This phenomenon is exacerbated by:

• Deforestation
• Wildfires
• Intensive agriculture
• Inadequate water management
• Urbanization and industrial activities

However, this project proposal will focus on the first two. Excessive tree felling or wildfires remove vegetation that protects the soil. Without this cover, the soil becomes exposed to erosion from wind and water.

Here, we can see a graph of the hectares burned in Spain from 1990 to 2023. As you can observe, between 1995 and 1997, there was a significant amount of burned territory, but the trend has generally been decreasing.

The same applies to the number of fires. Between 1990 and 2005, there was a peak, but they have since declined. However, in 2024, more than 5,800 fires still occurred, affecting a total area of 48,550 hectares, which is still an alarming figure. This highlights the ongoing need for prevention and restoration efforts.

For example, the European Union has already launched the EU Forest Strategy for 2030, supported by various funds such as the Rural Development Fund and the Recovery and Resilience Fund. These initiatives include actions like the planting of 3 billion trees by 2030.

With all this in mind, what can be done?

Manual reforestation, through organizations like Reforesta and others, or reforestation using drones, which I will now introduce with a couple of examples.

The first example is Morfo, a Brazilian company primarily dedicated to reforesting areas of the Amazon. The company follows four defined steps in its reforestation process: diagnosis, planning, planting, and monitoring. In three of these stages, drones are used—for analysis, seed distribution, and monitoring growth afterward.

According to Morfo, their system enables a single drone to treat up to 50 hectares per day and plant 180 seed pods per minute.

Another example is Dronecoria, a project previously developed in collaboration with IAAC. This initiative prioritizes DIY and low-budget solutions, allowing anyone to build and deploy it. It also features a seed dispenser designed to release seeds on demand.

Proposal

My proposal lies somewhere between these two examples. I aim to develop an open-source autonomous drone capable of analyzing terrain and dispensing seeds, with the potential for future collaboration with additional drones. I believe scalability is essential, especially in situations like this, where large areas need to be covered.

This is the proposed simple workflow for my system. After the user chooses the site:

• Drone performs SLAM and maps all the area 
• Point cloud and images are then post processed to identify target planting points
• Navigation and dispensing points are planned
• Route is performed and seeds are dispensed

Right now I am mainly focusing on getting the first step designed and done, the SLAM, but have started also planning the rest.

The drone I chose is the Holybro X500 v2, equipped with the Pixhawk PX4 flight controller. I selected it due to its availability (it was accessible at IAAC) and its payload capacity (around 1kg). This is an open-source drone with excellent documentation, allowing me to easily assemble and repair it. This will enable me to focus primarily on the implementation of the application.

This is the drone in question after being repaired.

The Hardware of the Holybro is as follows: we have 4 motors and its ESC connected to the flight controller, as well as the GPS, radio receiver and a telemetry module and everything is powered by a … Lipo battery. On the ground it can be either controlled by a remote controller or using a ground station (like a laptop) and another telemetry module and using software like QGroundControl or ROS2.

So my very next step is to try to fly it with the remote controller and calibrate and fly it with QGroundControl using MAvlink.

MAVLink is the communication protocol used by QGroundControl to connect your laptop with the flight controller on the drone. As you can see, it is also used when controlling the drone via ROS2.

Here is the architecture for an initial test I want to perform. In this setup, I should be able to control the drone using the arrow keys on the keyboard. The command is then sent through MAVROS, and, for example, the GPS position of the drone can be read back.

Here is an example of what I just said, but in simulation. 

I have also planned the next step which would be adding SLAM capabilities to the drone. For this purpose a depth camera and a Jetson Nano will be mounted on board.

I am not going to go very into detail here but just mention that the teleoperation part is the same as before, but while teleoperating, rgb-depth, imu and gps data are feeded into the jetson to perform the slam, using for instance the ORB-SLAM 3 package.

In a future a way of performing autonomous SLAM could also be explored, using algorithms such as Rapidly Exploring Random Tree (RRT)

After completing the SLAM, my plan is to begin working on the seed dispenser and the remaining implementations. Dronecoria provides open-source models for these mechanisms; however, they will need to be adapted to fit the specific size and payload capacity of the drone. I also aim to improve the reliability of the mechanism.

This is a draft table and description of the tests I want to perform in order to determine if the project is a success. As success metrics I will take into account things such as: path efficiency, mapping completeness, operational consistency, correct placement rate and so on.

In order to achieve this, I have done a project plan where I will be working of this first stage until end of January, then will complete the rest by end of April (hopefully) and design and perform a final demonstration by June. Let’s see how that goes.

In the mid to long term, it would also be interesting—at least in simulation—to explore the possibility of adding more drones to the system and creating a swarm, as the application would benefit from it. Recently, there has been an increase in research and practical applications in this area.