應用於監視系統之強健式整合性的即時移動物體偵測與追蹤演算法

A Robust Integral Real-Time Moving Object Detection and Tracking Algorithm for Surveillance Systems

許詠泰(Yung-Tai Hsu)

指導教授 ： 王照明 ； 陳永樹

元智大學/工程學院/機械工程學系/碩士(2003年)

摘要

目前監視系統所能提供的資訊相當有限，主要原因在於攝影機是採用固定式的定點監視，因此有監視範圍的死角，在這樣的因素下，即使拍攝到移動物體，亦很難達到進一步的監視效果。有鑑於此，本文以PTZ (Pan-Tilt-Zoom)攝影機為監視設備，提出一套強健式的整合性即時追蹤演算法則，來提升監視系統的效能。
本文提出一套強健式的整合性即時追蹤演算法則，不但可以在攝影機靜止的狀態下，進行移動物體追蹤，亦可當改變攝影機的位置時，來追蹤移動目標物。其設計原理為當攝影機固定時，利用影像相減方式來偵測出移動物體，以物體為基底(Content-Based)的方式，建立移動物體的特徵結果，再採用前後影片對應(frame-to-frame correspondence)之關係，來追蹤移動目標物。當攝影機改變位置時，其追蹤法則能自動切換到以邊緣為基底(Edge-Based)的方式，來進行移動物的輪廓比對，並以所建立的chamfer漸層擴展圖來解決物體外型變化之比對問題，進而找出移動物體，達到自動追蹤的目的。
本論文所提出的方法，亦能解決物體遮蔽情況下之移動物追蹤，以及適用於多個移動物體之追蹤。最後由實驗結果證實本文所提出的方法確實可行。

關鍵字

PTZ攝影機 ； 移動物偵測與追蹤 ； 攝影機校正 ； 追蹤演算法 ； 距離漸層擴展 ； 比對法則

並列摘要

The information that current surveillance systems can provide is very limited, mainly because they use the fixed-point supervision and controls. Analyzing information on the features of moving objects is difficult. Therefore, this paper addresses the PTZ camera used as monitoring equipment, to develop a robust integral real-time tracking algorithm to improve surveillance systems.
The principle of this system is as follows. When the camera is fixed, the content-based method is used to identify the features of the moving objects, and track the moving objects according to the construction of those features. When the camera changes position, the system follows a tracking rule, based on an edge-based method, to match the outlines of features of the moving objects, and determine the position of the moving objects.
Finally, the result of specify is verified experimentally. The proposed tracking algorithm can adapt to complicated backgrounds while continuing to track moving objects.

並列關鍵字

PTZ camera ； Moving object detection and tracking ； Camera calibration ； Tracking algorithm ； Chamfer distance map ； Matching algorithm

參考文獻

1. A. Baumberg and D. Hogg, “An Efficient Method for Contour Tracking Using Active Shape Models,” In Proceeding of the Workshop on Motion of Nonrigid and Articulated Objects, IEEE Computer Society, 1994.

3. M. Rossi and A. Bozzoli, “Tracking and Counting Moving People,” In ICIP94, pp. 212-216, 1994.

4. Q. Cai and J.K Aggarwal, “Tracking Human Motion Using Multiple Cameras,” In International Conference on Pattern Recognition, 1996.

5. C. Smith, C. Richards, S. Brandt, and N. Papanikolopoulos, “Visual Tracking for Intelligent Vehicle-Highway Systems,” IEEE Transactions on Vehicular Technology, pp. 732-743, 1996.

6. B. Heisele and C. Whler, “Motion-Based Recognition of Pedestrians,” International Conference on Pattern Recogni-tion, pp. 1325-1330, 1998.