影像監控系統交通參數擷取技術之研究

A Study on Traffic Parameters Extraction from Monitoring System via Image Processing

鄭其政(Chi-Cheng Cheng)

指導教授 ： 張堂賢

國立臺灣大學/工學院/土木工程學研究所/碩士(2006年)

https://doi.org/10.6342/NTU.2006.01605

摘要

交通監視系統Traffic Monitoring System (TMS)，主要做為收集、處理、分析交通量參數。透過綜合、管理訊息、電腦進行資料庫處理，並期許能及時將交通狀況傳送至交控中心，經整理後可以將訊息發送給用路人，如此可以提供用路人做最好的路線選擇，減少用路人旅行時間，並輔助交通工程師做幾何設計之用，並有效地管理控制交通。
以電腦影像視覺(Computer vision)為基礎之交通監視系統有著許多現有偵測技術所沒有之優點，透過影像處理，交通監視系統可以得知：車種、車軌跡、車行方向、路佔率、車隊長度、靜態或慢車之偵測、車輛軌跡之追蹤等，比現有的偵測技術得到更多的交通訊息，更可以對交通事故做紀錄。由於影像有大量資訊，可利用影像處理(Image processing)技術將相關訊息擷取出來，有許多研究以固定式攝影機影像（CCTV），加以處理擷取交通參數，有相當的成效，然而隨著網路數位化的發展，影像監控系統多以網際網路傳送數位資訊，然一般的研究並未考慮網路傳輸之影像，且對於可轉變角度之攝影機亦少有研究。本研究藉由可轉向型攝影機之影像，以及網路傳送影像之特性，輔以後端應用程式，期望由既有之影像監控系統之架構下，即時(real-time)且線上(on-line)系統擷取交通參數，而無須變更任何硬體設備。本研究考慮即時線上處理之需求，考慮以劃設虛擬偵測視窗之部分影像處理方式，能快速擷取出車流量參數，以及推估平均車速與平均密度。

關鍵字

交通監視系統 ； 影像處理 ； 網際網路 ； 轉向型攝影機 ； 虛擬偵測視窗 ； 即時線上系統

並列摘要

In general, CCTV for traffic control is used to monitor circumstances by the control officers. However, this research proposes a dual function of online traffic parameters extraction from the CCTV as a vehicle detector (VD) through image processing. The online parameters’ trajectories including traffic volume, average speed, and occupancy in terms of time are illustrated. About the image processing, background remove, moving objects identification, partial image analyzing, and parameters extraction are presented. Basically, initialization should be executed for measuring traffic parameters in each angle change of CCTV. Set virtual windows to detect vehicle is more efficient in computation. Real-time and on-line extracting traffic parameters over the image sequences via internet can be implemented with these virtual windows. Also, the measures should be adjusted in accordance with the CCTV angle. In fact, the quality of image should be fine enough for précising data acquisition.
Implementing a platform for traffic monitor system via internet should solve the difficulty of low frames rate. Partial window detection cost few computations for real-time and on-line system. In additional, virtual window extracts traffic parameters neglecting about the direction of vehicle flow. Some windows set on the same road extract referable parameters by means of statistical computing. The experimental result is tested by the level of statistical confidence.

並列關鍵字

vehicle detector (VD) ； CCTV ； image processing ； traffic parameter ； partial window

參考文獻

[32]李宗昇、謝正雄，「視訊影像處理在車流感測上的應用」，光學工程，第52期，頁11-16，84年12月。

[1]M. Fathy, M.Y. Siyal” An image detection technique based on morphological edge detection and background differencing for real-time traffic analysis.”Parrern Recognition Letters 16(1995) 1321-1330.

[2]Benjamin Coifman, David Beymer, Philip McLauchlan, Jitendra Malik, “A real-time computer vision system for vehicle tracking and traffic surveillance.” Transportation Research Part C 6 (1998) 271-288

[3]D. KOLLER, K. DANIILIDIS AND H.-H. NAGEL,” Model-Based Object Tracking in Monocular Image Sequences of Road Traffic Scenes.” International Journal of Computer Vision, 10:3, 257-281 (1993)

[4]M. Fathy and M.Y. Siyal, “Real-time image processing approach to measure traffic queue parameters”,IEE Proc.-Vis. Image Signal Process., Vol. 142, No.5, October 1995.