無人載具嵌入式即時影像監控及追蹤系統之開發

The development of an embedded real-time image surveillance and tracking system for UAV

張怡如(Yi-Lu Chang)

指導教授 ： 曾百由

國立臺北科技大學/機電學院/機電整合研究所/碩士(2007年)

摘要

本文主題為開發無人飛行載具上的影像系統。一般的影像系統多是建立在可執行高速運算的電腦上，但無人飛行載具無法背負龐大的硬體，因此本文採用嵌入式控制器來完成無人飛行載具上的影像系統，其中包括了基礎及進階DSP處理器。基礎DSP處理器的影像系統中包括兩大功能分別是影像追蹤及環境監控。基礎處理器的影像追蹤系統中，利用物體形心法及物體邊界判別法，配合伺服馬達的定位控制做簡單的物體追蹤；但受限於基礎處理器的硬體及處理速度，故此系統只能在單純環境下追蹤一個單色物體。基礎處理器的環境監控系統中使用影像差異法來判別是否有異物入侵或環境是否有明顯差異。由於基礎處理器系統中能做的影像處理方法有限，因此本文又開發了一套進階DSP處理器來改善系統功能。在進階處理器的影像追蹤系統中，可對多個物體作影像標識法及針對物體形狀做特徵值判別，並針對欲追蹤物配合伺服馬達定位控制，完成即時的物體追蹤。在進階處理器的環境監控系統中針對移動物體使用移動邊緣偵測法偵測物體邊緣，搭配影像追蹤系統有效的追蹤移動物體，完成一個可應用於無人飛行載具之嵌入式影像系統。

關鍵字

影像追蹤 ； 數位訊號處理器 ； 影像標識法 ； 特徵值判別 ； 移動邊緣偵測法

並列摘要

The subject of this study is to develop an image system which can be applied on Unmanned Aerial Vehicles (UAV). Since most image systems are PC-based and the sizes are inappropriate for UAV payload, this study selects embedded controllers as the core of image systems, including basic and advanced Digital Signal Processors (DSP). The image system of basic DSP consists of two parts：Image Tracking and Environment Surveillance. The image tracking system with basic DSP uses Centroid and Boundary Definition integrated with the positioning control of Servo Motors for simple object tracking. Because of the limited hardware capability and processing speed of the basic DSP, this system can only track one monochrome object in a simple environment. On the other hand, the environment surveillance system with the basic DSP uses Frame Differencing to detect if there is foreign object invasion or an obvious change of the environment. Since the capability of image processing in basic DSP are limited, a system with advanced DSP processor is developed to improve the image tracking and environment surveillance performance. The image tracking system with advanced DSP makes real-time object tracking by applying Image Labeling to objects and defining characteristic according to their shapes, and is also integrated with positioning control for tracking. The environment surveillance system with advanced DSP uses Moving Edge Detection to detect moving objects’ edges. Together with the image tracking system, it can track moving objects efficiently and make an embedded image system applicable to UAVs.

並列關鍵字

Image Tracking ； Digital Signal Processors(DSP) ； Image Labeling ； Characteristic Definition ； Moving Edge Detection.

參考文獻

[1] Arun D. Kulkarni, Computer Vision and Fuzzy- Neural Systems, Prentice Hall,Inc., 2001.

[2] C. Cedras and M. Shah, “Motion-based recognition: A survey,” Image and Vision Computing, Vol. 13, pp. 129-154, 1995.

[3] 黃博政，分散式無人載具感測與控制系統之發展，碩士論文，國立臺北科技大學/機電整合研究所，2005。

[7] S. A. El-Azim, I. Ismail, and H. A. El-Latiff, “An efficient object tracking technique using block-matching algorithm,” Proc. of the Nineteenth National, Radio Science Conf., pp.427-433, 2002.

[8] J.A. Leese, C.S. Novak and V.R. Taylor, “The Determination of Cloud Motion Patterns From Geosynchronous Satellite Image Data,” Pattern Recognition, Clustering, Statistics, Grammars, Learning, Vol. 2, No. 5, pp. 272-292, 1970.