基於車燈資訊夜間車輛偵測方法

在本篇論文中，我們提出了應用於一般交通監視器的夜間車輛偵測方法。隨著近年科技發展的進步，智慧監控系統的應用也更加廣泛。過去的交通監控系統常透過架設感測器的方式來達到監控目的，但架設與維護往往需要耗費大量金錢與勞力；相對地，利用大量架設於道路的交通監視器進行車輛分析，是較有效率且具彈性的方法。然而夜間環境下的可見資訊稀少，如何透過一般交通監視器畫面，在夜間環境下精確地偵測出車輛，成為智慧運輸系統領域中的一道難題。由於在夜間環境中，車燈是車輛唯一的穩定可見資訊，我們提出基於車燈資訊偵測車輛的方法，並將車燈種類分為頭燈及尾燈，透過CenSurE(Center Surround Extremas)偵測器與色彩空間分析，分別對兩類車燈進行偵測。為了進一步確保偵測出的燈源正確性，我們經由在日間環境下建立車流遮罩模型，取得夜間可偵測出車輛車燈的範圍。在偵測出可能的車燈位置後，利用統計出的車燈間距資訊進行匹配，並計算車燈配對區域的方向梯度直方圖(Histogram of Oriented Gradient, HOG)特徵，結合支撐向量機(Support Vector Machine, SVM)分類器進行驗證，以確認該對車燈是否為真實車燈。由於在偵測方法中可能僅偵測出一個車燈位置，我們以間距資訊提出假設的車燈區域，以提高車輛的偵測率。

關鍵字

車燈偵測；車輛偵測；支撐向量機；智慧監控系統

並列摘要

Vision-based vehicle detection plays an important role in Intelligent Transportation System (ITS) due to its low cost, flexibility and easy to maintain. However, it is difficult to detect vehicle via camera at nighttime because of the lack of stable appearance feature on vehicles. In this paper, we proposed a vehicle detection method using general traffic cameras at nighttime. We detect vehicles by their paired lights. We use Center Surround Extremas (CenSurE) detector to detect individual lights and perform color space analysis to extract taillight. In order to filter out street/building lights, we establish the road area in the daytime first. After detecting the lights inside road area, we model the distance between two lights of a vehicle, and find the potential vehicle light pairs based on the model. Then we extract Histogram of Oriented Gradient (HOG) feature from each paired light region and verify the paired lights by Support Vector Machine (SVM) for vehicle detection. We also propose a hypothesis to find the vehicles that only one of its lights can be detected by CenSurE detector.

並列關鍵字

Vehicle Light Detection ； Vehicle Detection ； Support Vector Machine ； Intelligent Transportation System

參考文獻

[28] 沈宗慶，“基於視訊之夜間車輛偵測”，中原大學資訊工程所碩士論文，Jan. 2017。

[30] 施養孝，“用於分析都市交通之自動化道路模型建置技術”，中原大學資訊工程所碩士論文，Jul. 2017。

[3] C. L. Huang and W. C. Liao, “A Vision-Based Vehicle Identification System,” Proc. IEEE International Conf. on Pattern Recognition (ICPR), vol. 4, pp. 364–367, Aug. 2004.

[4] B. Coifman, D. Beymer, P. McLauchlan and J. Malik, “A real-time computer vision system for vehicle tracking and traffic surveillance,” Transportation Research Part C: Emerging Technologies, vol. 6, no. 4, pp. 271–288, 1998.

[5] A. Ottlik and H. H. Nagel, “Initialization of Model-Based Vehicle Tracking in Video Sequences of Inner-City Intersections,” International Journal of Computer Vision, vol. 80, no. 2, pp. 211–225, Nov. 2008.

國際替代計量

基於車燈資訊夜間車輛偵測方法

全文下載

主題瀏覽