Unmanned Aerial Vehicles: Embedded Control

Chapter 9

Hover Stabilization of a Quad-Rotor Using a Single Camera ¹

This chapter presents the stabilization and local positioning of a four-rotor rotorcraft using computer vision. Our approach combines the measurements from an IMU (inertial measurement unit) and a vision system composed of a camera used to estimate the orientation and position of the rotorcraft. The vision system provides the position and yaw angle while the IMU gives the pitch and roll angles at a higher rate. Two different techniques are presented to obtain the position from the information provided by the camera. In the first method the camera calibration is carried out using the two-plane calibration approach and the estimation of the position is based on perspective ofn-points method. The second technique combines a camera calibration method that allows us to recover directly their intrinsic parameters that we have named the homogenous transformation method with the plane-based pose approach to estimating the position. These two techniques have been tested in real-time experiments to stabilize and locate the flying machine with respect a very well-known target and results are shown at the end of this chapter.

9.1. Introduction

Most of the applications involving UAVs require an aerial robot capable of performing hover and navigation. These characteristics strongly depend on the applied control strategy and the set of on-board sensors [CAS 05]. The design of UAVs usually involves the trade-off between performance, ...