應用立體視覺獲取物體姿態

機構典藏 > College of Engineering > Graduate Institute & Department of Aerospace Engineering > Thesis > Item 987654321/35553

Please use this identifier to cite or link to this item: https://tkuir.lib.tku.edu.tw/dspace/handle/987654321/35553

Title:	應用立體視覺獲取物體姿態
Other Titles:	Attitude acquisition using stereo vision
Authors:	陳正霖;Chen, Cheng-lin
Contributors:	淡江大學航空太空工程學系碩士班蕭富元;Hsiao, Fu-yuen
Keywords:	微型飛行器;立體視覺;影像處理;姿態量測;慣性量測元件;Stereo Vision;Image Processing;MAV;IMU;Attitude Measurement
Date:	2009
Issue Date:	2010-01-11 06:46:20 (UTC+8)
Abstract:	本篇論文旨在探討應用立體視覺觀測微型飛行器(MAVs) 姿態的可行性。由於立體視覺能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。此外, 因為微型的感測器發展還不成熟, 使得測量MAV 的姿態比測量普通飛行器還來得困難, 因此, 利用立體視覺的方式來觀測目標姿態, 可成為解決此一問題的方法。本篇論文展示利用影像處理的方法來辨識有無移動物體, 並計算目標物的位置、速度及姿態。為了降低成本, 本文實驗採用市售便宜的網路攝影機, 而此種攝影機常會導致成像扭曲, 所以我們搭配簡單的影像校正方法, 再利用實驗來驗證本文方法的實用性。因此我們目前的成果, 可以應用到日後MAV 飛行中姿態的獲取與探討。 This thesis studies the potential applications of stereo vision to the acquisition of attitude of micro aerial vehicles (MAVs). Stereo vision is of interests among scientists due to its capability of acquiring three dimensional position. On the other hand, to measure the attitude of an MAV is much more difficult than to measure a normal sized one, because the development of micro sensors is not matured enough. Accordingly, to acquire the attitude using stereo vision must be one of the solutions. In this thesis, we demonstrate our ability of computing objects acceleration and attitude using image processing. In order to lower down the cost, a cheap web-cam is selected for the experiment, which in general always causes image distortion. As a result, simple calibration algorithms are applied, and experiments are performed to demonstrate the validity of our algorithms. We also construct a Graphic User Interface for the convenience. Our current achievement is applicable to the future investigation of MAV attitude determination.
Appears in Collections:	[Graduate Institute & Department of Aerospace Engineering] Thesis

Files in This Item:

File	Size	Format
	0Kb	Unknown	334	View/Open

Loading...