The goal of this project is to develop the protocols and algorithms necessary that will particularly suit cost-effective, flexible image dataset collection through drones. They may be particularly advantageous due to their recent increase in accessibility and ability to take pictures in dangerous or difficult to navigate places.
The main requirements to run the code in this repo is Python, ROS, and OpenCV. Additionally, one will need a Parrot AR.Drone Quadricopter. Then, refer to drone_master.py to start flying the drone; the drone will automatically circle around objects, using tape on the floor as guidance. For a demo, please see the youtube video here: