臺灣博碩士論文加值系統

在過去十年，視訊監控技術正走向數位化和遠程監控的方法。為了使攝影機的視野增加，需要改進攝影鏡頭監控效率。而普通的固定攝影機並不適合廣大的空曠區域如校園、公園以及交通不便的地方，但如果將攝影經結合無線遠程監控技術，搭載在飛行載具上，並能夠自動偵測異常裝況，將能大大提高監控效率。
本論文嘗試開發一個移動載具利用ArduPilot Mega (APM2.6)飛行控制板為主要控制器，製作兩架多旋翼空拍載具，用於空拍以及感興趣物體追蹤，載具上搭載一台智慧型手機，載具會將影像以及飛航資訊例如:高度、飛行時間、電池電壓、陀螺儀感測數據以及加速度計數值等傳送至地面站。製作影像回傳系統透過智慧型手機的WI-FI回傳影像至地面站，並對回傳影像使用TLD追蹤演算法對感興趣之物體進行追蹤。
本研究使用智慧型手機鏡頭拍攝的影像經由Android手機程式做為發送端，透過WIFI傳輸，影像接收端則是用筆電或是平板電腦來接收，利用此系統來進行即時影像的傳輸，並掌握所要拍攝物體的狀況和飛機運作時所拍攝的畫面。並用回傳之影像做初步的物體追蹤。

The progresses in information technologies and the security demand make the surveillance industry gain the massive growth in the past decade. The video surveillance technology is now moving toward to the digital communication and remote monitoring approaches. The need for improved camera surveillance efficiency has prompted the development of multiple-angle cameras. However, the surveillance system is not suitable for campuses, parks or inaccessible areas. The application will become more flexible if the monitoring cameras are able to detect incidents of unusual objects, which are mounted on a mobile platform such as unmanned aerial vehicle that support path planning and different ways of alerting.
In this thesis, we try to develop a mobile platform, i.e. a remote controlled multicopter, for video surveillance purposes. We have built the multicopter with the ArduPilot Mega (APM) flight control module and develop the flight control console software. An android mobile device is used as the bridge between the multicopter and the control console via Wi-Fi connection. The camera of the android mobile device is also utilized for capturing the monitoring video. The flight information console software and the video surveillance system are developed on the android device to provide the necessary information such as the flight route, flight time, altimeter information, sensor data, and the object tracked by the surveillance system, etc. We have developed an object of interest detection method based on the Tracking-Learning-Detection algorithm to locate the target in the aerial images. The flight vehicle and the video surveillance console developed in this thesis own the advantages of theoretical research and industrial application values.
In this thesis, we try to develop a mobile platform, i.e. a remote controlled multicopter, for video surveillance purposes. We have built the multicopter with the ArduPilot Mega (APM) flight control module and develop the flight control console software. An Android mobile device is used as the bridge between the multicopter and the control console via Wi-Fi connection. The camera of the Android mobile device is also utilized for capturing the monitoring video.

摘　　要 1
ABSTRACT 3
致　　謝 2
目　　次 0
表目錄 2
圖目錄 3
第一章 緒論 6
1.1 背景與動機 6
1.2 研究目的 6
1.3 論文架構 8
第二章 文獻探討 9
2.1多旋翼飛行載具 9
2.2影像處理系統 13
2.2.1 TLD工作框架 14
1. 追蹤器(Tracking) 15
2. 學習過程(Learning) 15
3. 檢測器(Detector) 15
第三章 系統架構 17
3.1 飛行載具製作 18
3.2 影像處理 37
3.3 影像傳輸 42
第四章 實驗結果 45
4.1整合系統介紹 45
4.2傳輸距離測試 49
4.3影像處理 51
第五章 結論 54
參考文獻 55

[1]“Quadcopter.” [Online]. Available: http://www.skylarkfpv.com/store/index.php?route=product/category&path=40. [Accessed: 18-Jul-2014].
[2]“MikroKopter - HexaKopter on Vimeo.” [Online]. Available: http://vimeo.com/6194911. [Accessed: 22-Dec-2012].
[3]“FREEFLY - Freedom To Move / Freefly Systems.” [Online]. Available: http://www.freeflysystems.com/. [Accessed: 18-Jul-2014].
[4]“DCView 數位視野 - 文章總覽 - 讓相機飛天空拍的三款飛行器.” [Online]. Available: http://article.dcview.com/newreadarticle.php?id=10154. [Accessed: 22-Dec-2012].
[5]“IEEE Xplore - A Fully Autonomous Indoor Quadrotor.” [Online]. Available: http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6004839&contentType=Journals+%26+Magazines&searchField%3DSearch_All%26queryText%3DA+Fully+Autonomous+Indoor+Quadrotor. [Accessed: 24-Apr-2013].
[6]H. Lim, J. Park, D. Lee, and H. J. Kim, “Build Your Own Quadrotor: Open-Source Projects on Unmanned Aerial Vehicles,” IEEE Robot. Autom. Mag., vol. 19, no. 3, pp. 33–45, Sep. 2012.
[7]“arducopter - The ArduCopter manual has now moved http://copter.ardupilot.com The manuals here are for legacy hardware and will no longer be updated or maintained. - Google Project Hosting.” [Online]. Available: https://code.google.com/p/arducopter/. [Accessed: 18-Jul-2014].
[8]“The APM 2.5 Board Overview | ArduCopter.” [Online]. Available: http://copter.ardupilot.com/wiki/apm25board_overview/. [Accessed: 18-Jul-2014].
[9]“diydrones/MissionPlanner,” GitHub. [Online]. Available: https://github.com/diydrones/MissionPlanner. [Accessed: 18-Jul-2014].
[10]“OpenPilot.org – The Next Generation Open Source UAV Autopilot.” .
[11]“PaparazziUAV.” [Online]. Available: http://wiki.paparazziuav.org/wiki/Main_Page. [Accessed: 18-Jul-2014].
[12]“PIXHAWK Research Project - ETH PIXHAWK: MAV Computer Vision.” [Online]. Available: https://pixhawk.ethz.ch/. [Accessed: 18-Jul-2014].
[13]“MikroKopter • Home.” [Online]. Available: http://www.mikrokopter.de/en/home. [Accessed: 18-Jul-2014].
[14]“kkmulticopter.com.” [Online]. Available: http://www.kkmulticopter.com/. [Accessed: 18-Jul-2014].
[15]“MultiWii,” MultiWii. [Online]. Available: http://www.multiwii.com/. [Accessed: 18-Jul-2014].
[16]“AeroQuad Forum - AeroQuad - The Open Source Multicopter,” AeroQuad Forum. [Online]. Available: http://aeroquad.com. [Accessed: 18-Jul-2014].
[17]K. Yamasaki, T. Takimoto, and K. Yoshino, “Attitude control of quadrotor helicopters with sensor offsets,” in 2011 IEEE/SICE International Symposium on System Integration (SII), 2011, pp. 961 –965.
[18]“TLD | Zdenek Kalal.” [Online]. Available: http://personal.ee.surrey.ac.uk/Personal/Z.Kalal/tld.html. [Accessed: 18-Jul-2014].
[19]“zk00006/OpenTLD,” GitHub. [Online]. Available: https://github.com/zk00006/OpenTLD. [Accessed: 18-Jul-2014].
[20]“An accessible introduction to mean-shift.” .
[21]D. Comaniciu and P. Meer, “Mean shift: a robust approach toward feature space analysis,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 24, no. 5, pp. 603–619, 2002.
[22]“Zhongshan RC SunnySky Model Co., Ltd.” [Online]. Available: http://www.rcsunnysky.com/kygood/en/index.asp. [Accessed: 05-Aug-2014].
[23]Z. Kalal, K. Mikolajczyk, and J. Matas, “Forward-Backward Error: Automatic Detection of Tracking Failures,” in 2010 20th International Conference on Pattern Recognition (ICPR), 2010, pp. 2756–2759.
[24]Z. Kalal, J. Matas, and K. Mikolajczyk, “P-N learning: Bootstrapping binary classifiers by structural constraints,” in 2010 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2010, pp. 49–56.