適用於ETC多車道環境下的執法演算法

Abstract

在電子收費（Electronic toll collection, ETC）系統中，利用車上機（On-Board Unit, OBU）與路邊節點(Roadside Unit, RSU)快速的資料交換，可降低收費處理時間，進而提升車流量。ETC系統由四個模組構成：自動車輛辨識模組、自動車輛分類模組、扣款模組、影像執法模組。其中影像執法模組將每張車牌影像透過自動化車牌辨識 (Automatic License Plate Recognition, ALPR) 產生牌照號碼，再利用牌照號碼與車輛扣款資料比對找出未繳費與交易未成功之車輛。然而大量的車牌辨識會造成ETC系統的瓶頸。此外牌照號碼與車輛扣款資料比對之運算，多車道(Multi-Lane Free Flow, MLFF)比單車道(Single-Lane Free Flow, SLFF)來的複雜。在本篇論文中，我們提出一個適用於ETC多車道環境下之模組，將車牌影像資料和扣款資訊匹配關係轉換成雙分圖(Bipartite Graph)。針對雙分圖提出演算法以找出未繳費與交易未成功之車輛。模擬的結果顯示出，我們提出的演算法可大幅的降低影像辨識的次數並提高執法系統的成功率。

There are many benefits of electronic toll collection (ETC) system such as reducing toll paying time, increasing the capacity of toll station, decreasing fuel consumption, enhancing the convenience of traveler and so on. For a multi-lane free flow ETC system, how to find out the non-payment vehicles without recognizing all license plate images is an important research problem. In this thesis, we formulate the non-payment vehicle searching problem into a bipartite graph and propose an algorithm without recognizing all license plate images for solving it. Simulation results show that our algorithm can reduce the number of ALPR (Automatic License Plate Recognition) and increase success rate of enforcement.