利用車燈偵測分析夜間交通壅塞情況

Nighttime Traffic Congestion Analysis Using Vehicle Headlight Detection

陳華總(Hua-Tsung Chen) ； 蔡立武(Li-Wu Tsai) ； 古蕙媜(Hui-Zhen Gu) ； 李素瑛(Suh-Yin Lee) ； 林寶樹(Bao-Shuh P. Lin)

《前瞻科技與管理》 2卷1期 (2012/05) Pp. 67-79

摘要

近年來，越來越多的國家致力於智慧型運輸系統（ITS，Intelligent Transportation Systems）的開發，以期能透過即時交通訊息的傳輸與整合來提升交通運輸品質。由數位攝影機所構成的交通監控系統已成為當前分析即時交通情況的主流。目前多數研究著眼於日間交通情況的分析，然而夜間交通的監控亦相當重要。夜間交通監控影像因受亮度不足、能見度不佳與光線反射等嚴苛因素之影響，相較之下更具分析之難度。在此，本研究提出一夜間交通監控系統，利用車燈偵測來分析交通壅塞情況，並提供用路人即時的道路交通資訊，讓他們夠考量交通情況，規劃安排行車路線，避開壅塞路段。如此一來，將可以分散交通流量，並大幅提升交通效率。本研究以台灣高速公路夜間之監控影片為測試資料，實驗結果證明我們所提出之夜間車輛偵測與交通壅塞情況分析演算法具有高準確率與高效率。

關鍵字

智慧化交通運輸系統 ； 特徵擷取 ； 圖樣辨識 ； 夜間交通監控系統 ； 車輛偵測

並列摘要

For traffic safety, efficiency and service, the development of Intelligent Transportation System (ITS) is an emerging trend in recent years. Most of the existing works concentrate on the daytime traffic surveillance, but they are unable to handle the nighttime conditions well. However, the demand for nighttime traffic surveillance is no less than the daytime one because high traffic flows as well as incidents may happen during the night. Compared to daytime traffic surveillance, more tough obstacles such as poor visibility, low luminance and light reflection should be overcome under the nighttime condition. In this paper, we propose a nighttime traffic surveillance system capable of traffic congestion analysis based on vehicle headlight detection. The instant traffic information can be provided to the road users or the police in real time, which greatly improves the transportation efficiency. We conduct experiments on Taiwan freeway surveillance videos in nighttime conditions, and satisfactory results validate the effectiveness and efficiency of the proposed framework.

並列關鍵字

Intelligent Transportation Systems (ITS) ； feature extraction ； pattern recognition ； nighttime traffic surveillance ； vehicle detection

參考文獻

＜即時路況資訊＞，《交通部台灣區國道高速公路局》，2011年5月17日，（ 瀏覽日期：2011年5月17日）

Bertozzi, M.(2000).Vision-based Intelligent Vehicles: State of the Art and Perspectives.Robotics and Autonomous Systems.32,1-16.

Bruzzone, L.(1999).A Neural-Statistical Approach to Multitemporal and Multisource Remote-Sensing Image Classification.IEEE Transaction on Geoscience and Remote Sensing.37,1350-1359.

Cai, Y.(2006).Context Enhancement of Nighttime Surveillance by Image Fusion.Proceedings of International Conference on Pattern Recognition.(Proceedings of International Conference on Pattern Recognition).:

Chen, Y.L.(2009).Real-Time Vision- Based Multiple Vehicle Detection and Tracking for Nighttime Traffic Surveillance.Proceedings of IEEE International Conference on Systems, Man, and Cybernetics.(Proceedings of IEEE International Conference on Systems, Man, and Cybernetics).: