移動物體即時追蹤研究

Real-Time Tracking of the Moving Object

李文昌(Wen-Chang Lee)

指導教授 ： 黃英哲

元智大學/工程學院/電機工程學系/碩士(2002年)

摘要

本論文針對移動物體即時追蹤提出一些控制方法，其包括目標影像萃取與目標追蹤兩部分。影像萃取法採用區塊平均亮度與變異數差異比對法，配合目標中心定位法與區塊全域搜尋法，找出目標所在的中心位置，而目標追蹤控制方法則分別採用位置型控制法，位置速度型控制法，位置速度加速度控制法和改良式位置速度加速度控制法。最後本論文對此四種控制器加以分析比較，並從模擬分析的結果顯示四種方法都可以達到其追蹤目的。且從實驗結果證明我們所提的方法可以完成在可變背景中對移動目標的即時影像追蹤。

關鍵字

即時追蹤 ； 移動物體 ； 區塊全域搜尋

並列摘要

The theses propose some approaches for real-time tracking of the moving object, including the image extraction and object tracking. The methods of image extraction locate the center of the object by using average and variance brightness of the block pixel. Also, we combine the scheme of object-center-location and block-full-search to focus the central position of the object. Additionally, the control methods of object Tracking make use of P (Position) type, PV (Position -Velocity) type, PVA (Position-Velocity-Acceleration) type, and improved PVA-type controllers to achieve the goal. However, The methods of average and variance brightness of block pixel not only can reduce the sampling error of the camera but also fit the higher contrast environment. The thesis also brings up vertical relationship method to get more correct extract information of the moving object. The four kinds of tracking controllers use different algorithms to estimate the next position of the object based on the error value and the change of error value. Finally, we analyze and compare the performance of all the proposed methods as indicated in simulation results, and prove that they can track the moving object on real-time as shown in experiment results.

並列關鍵字

Real-Time Tracking ； Moving Object ； Block-Full-Search

參考文獻

[2] J.C. Terrillon, M.N. Shirazi, H.Fukamachi, and S. Akamatsu, “Comparative Performance of Different Skin Chrominance Models and Chrominance Space for the Automatic Detection of Human Face in Color Images,” Proc. IEEE Int’l Conf. On face and Gesture Recognition, pp. 54-61, 2000.

[3] Chris Cho-Pin Li and Ludwik Kurz,”A New Approach to the Detection of Moving Objects,” INFORMATION SCIENCES 103, pp.115-134, 1997.

[4] Dae-Sik Jang,Gye-Young Kim,Hyung-II Choi,”Kalman Filter incorporated Model Update for Real-time Tarcking,” IEEE Tenco on Digital Signal Processing Applications, pp. 878-882 ,1996.

[5] Dalong Li, “Moving Objects Detection by Block Comparison,” IEEE Int’l Conf. On Electronics Circuits and System, Vol. 1, pp. 341-344, 2000.

[6] Yachi T, Wada T, Matsuyama T, ”Human Head Tracking using Adaptive Appearance Model with a Fixed-Viewpoint Pan-Tilt-Zoom Camera,” IEEE Int’l Conf. On Automatic Face and Gesture Recognition, pp. 150-155, 2000.