MEng
Final Group Project

This project was proposed to our university from an Australia agricutural scientist. Before proposing this method, the scientists had to walk through the fields to collect the samples by hand and as Australia's farms are very large, this took them a long time to collect the samples.
I was the sole electronic engineer who knew coding and wiring. So all those tasks left given to me. The pixhawk 4 was used as the autopilot of the drone to follow the predecided path given by the ardupilot ground control software. When the drone was above the place where it was chosen to do a sample collection, the Raspberry Pi took control of the autopilot and used computer vision algorithm I wrote that used video data from a pi camera to figure out where the wheat plants are.
Once the wheat plant was decided, the next algorithm started running which used a library called DroneKit which allowed me to send information to and from the drone to the Raspberry Pi. This allowed me to position the drone directly above the plant with better accuracy than the autopilot on it's own. All the code can be found here.
However, the drone could not be produced due to the pandemic causing the manufacturing facility at the univeristy to be closed. So the drone became theoretical and the only thing that was not theoretical was the circuity I had produced on breadboards to show how the drone would have worked.