GNU/Linux based motion capture platform for fisheye images / 以GNU/Linux建構魚眼影像動態擷取開發平台

碩士 / 明志科技大學 / 電機工程研究所 / 97 / This thesis foresees an advanced research in developing a motion detection apparatus using fish-eye camera (MOCAP-F). The contents are not only establishing a developing environment, integrating a tooling bundle of software but also developing testing software in system performance evaluation. With the advance of technology, the image streaming flow and complicated computing procedures can be easily executed in a common computer set. However, to perfectly combine software and hardware in achieving a fully resource utilization is still an important task.
This research will base on GNU/Linux environment to build up MOCAP-F application. We adapts a great deal of GNU open source software tools, such as the user interface by GTK(GIMP Tool Kit) library, Video4Linux library for accessing the image data, Eclipse IDE (Integrated Development Environment) for soft debug. At last phase, we use Doxygen to document a reference booklet and, nevertheless, CVS (Concurrent Versions System) is responsible for source code managing projects. Additionally, SAMBA is being a transmission to Window OS.
Under the frame of multi-core CPU, 3 programs of similar outcome target but coded with different soft utilities are used to test the MOCAP-F real time achievement. First two of them are public domain GNU/LGPL software. The No. 1 takes multithread POSIX library to handle 4-ch video processing pipeline thread. The No. 2 software is base on idle events of GTK environment. The No. 3 software combines with the distinctions of these two programs to complete it. However, they have very large difference even the three software base on the same platform. Only the No. 3 program writing by us can enhance the efficiency of CPU. These results build up the principles for developing the future MOCAP-F system.

Identiferoai:union.ndltd.org:TW/096MIT00442004
Date January 2009
CreatorsHong-Wen Hang, 黃鴻文
ContributorsChuang-Jan Chang, 張創然
Source SetsNational Digital Library of Theses and Dissertations in Taiwan
Languagezh-TW
Detected LanguageEnglish
Type學位論文 ; thesis
Format90

Page generated in 0.0081 seconds