Pose estimation of objects of interest by aerial multirotor drone

Aerial drones have been seeing a rise in popularity in a variety of domains such as shipping, search and rescue operations, and photography due to their ability to carry cameras and other packages through trajectories that avoid ground obstacles or offer an advantageous vantage point, all while allowing the operator to remain at a safe distance from any danger. In this project, the objective is to improve the autonomy of aerial drones performing aerial surveillance. The specific goal of this research is to develop a drone-based system that is capable of determining the pose of objects it detects through it’s camera. Research in AI-based image processing has made the detection and recognition of objects possible with high accuracy. This project intends to implement such an object detection and classification algorithm into a drone system where monitoring of large areas can be done from the air and on the move with the drone’s operator. This project will help to develop drone technology and demonstrate their use in human activities. As a project based in Canada, it will also keep establishing Canada as an aerospace technologies hub.

Faculty Supervisor:

James Richard Forbes

Student:

Partner:

ARA Robotique

Discipline:

Engineering

Sector:

Aerospace; Other; Technology

University:

McGill University

Program: