本論文主要探討利用立體視覺量測微型飛行器(MAVs) 的位置和姿態。由於立體視覺可以獲得三維空間中的位置資訊,使得立體視覺在科學領域中日漸重要。目前微飛行器的感測器發展還不成熟,使得量測微型飛行器的姿態比量測普通飛機還來得困難,因此我們利用外部觀測平台來獲取微型飛行器的位置和姿態。在本文中,我們在飛行器上標註若干記號,再利用立體視覺的原理找出各點於空間中的絕對座標,進而換算成飛行器的姿態。為了降低成本,我們使用市售便宜的網路攝影機,並設計簡單的校正實驗,以減少誤差。此外,我們亦引入四軸平台,用以精確掌握飛行器於每一個時刻的姿態角,以便做為定量分析之用。雖然本文中的探討的主要標的物為一拍翼式微飛行器,但所提之方法可應用於任何形式之微飛行器。
This thesis focuses on the measurement of position and attitude of micro-aerial vehicles (MAVs) using stereo-vision technology. Since stereo-vision can obtain three dimensional (3D) coordinate, it has been widely applied to various fields. Currently, micro-class sensors or nano-class sensors are not mature enough to serve for measurements of MAVs. It is more difficult for us to obtain the attitude of an MAV than that of a regular aerial vehicle. We propose the do the measurements using an outside, non-intrusive equipment. In this thesis, we make some marks on the fuselage or wings of a MAV, and use the stereo-vision to identify their 3D positions. This information can then be converted to the attitude. In order to lower down the cost, commercial webcams are employed in the experiments, and simple calibrations are designed to increase the precision. A four axes platform is also introduce to offer the accurate position and attitude of the MAV, so that quantitative analysis is possible. Though the main investigated target in this thesis is a flapping-wing MAV, the methodology and algorithms proposed in this thesis can be applied to any other type of MAVs.