Investigating the efficacy of OpenMV cameras for use in pose estimation

One of the present research focuses within the MechatronicSystems.Group is unmanned aerial vehicles (UAVs). To this end, our group is currently assessing the feasibility of low-cost cameras for use in pose estimation and control.

This project will investigate using multiple OpenMV (or some other equivalent) cameras for use in estimating the pose of a rigid body, given known markers in the world. Specifically, April Tag identification, as well as arbitrary feature locking will form part of the project.

This project will require the student(s) to research, develop, and implement software that can allow for pose estimation using multiple OpenMV cameras. This will include:

• Researching current solutions
• Develop code to assess the viability of April Tag detection and pose estimation using the abovementioned cameras
• Establish an experimental procedure to evaluate the viability of the pose estimation scheme
• Perform tests and quantify the accuracy using a single camera
• Incorporate additional cameras and quantify the potential performance increase

The ideal candidate(s) will have strong programming ability, with a predominant focus on MATLAB/Simulink development and python.

Supervisor: Arnold Pretorius