Dae-Yeon Won* and Min-Jea Tahk**
International Journal of Aeronautical and Space Sicences, vol. 9, no. 1, pp.76-84, 2008

Abstract : This paper is mainly concerned with the vision-based attitude stabilization of a quad-rotor UAV. The methods for attitude control rely on computing the roll and pitch angles of the vehicle from a two-camera vision system. One camera is attached to the body-fixed x-axis and the other to the body-fixed y-axis. The attitude computation for the quad-rotor UAV is performed by image processing consisting of Canny edge and Hough line detection. A proportional and integral controller is employed for the attitude hold autopilot. In this paper, the quad-rotor UAV is modeled by 6-DOF nonlinear equations of motion that includes rotor aerodynamics with blade element theory. The performance of the proposed method is evaluated through 3D environmental numerical simulations.

Keyword : Quad-Rotor UAV, Attitude stabilization, Vision-based control