Projects

Designed the vision-based single object tracking based on colour or contrast (C++).

Colour-based tracking originally developed by Donghan Lee.

Tested and verified through flight test.

Will be extended to multi-target tracking problem.

Designed the vision-based position measurement software based on deep learning (C++).

Possible to classify and measure the LLA position of the objects on the ground.

Tested and verified through flight test and cooperative mission test using UAV-UGV team.

Journal publication had published as an output of this project.

“Development of a Vision-based Recognition and Position Measurement System for Cooperative Missions of Multiple Heterogeneous Unmanned Vehicles,” published in IJASS.

Developed vision-based relative navigation software based on deep learning (C++).

Possible to estimate the relative pose (position + attitude) of the leader aircraft.

Tested and verified through flight test using light sport aircraft (LSA) with other institutes.

Journal publication had published as an output of this project.

“Vision-based Relative Pose Estimation of UAV using YOLO and PnP,” published in JAIS.