Position and orientation estimation for unmanned helicopter landing based on ellipse feature

This paper proposed an algorithm of ellipse detection based on Mean Shift, followed by position and orientation estimation. Setting a specific color ellipse mark as the landing target, the algorithm captures the target section by Mean Shift. The ADM edge point extraction and direct least square method are limited in the section to establish the ellipse equation. The problem of estimating UAV´s position and orientation based on ellipse feature is turned into solving an equation of twelfth order. The method removes the background interference and reduces the complexity of image processing. Experimental results on Quanser platform verify the accuracy and real-time performance of the method.