以監視器影像執行闖紅燈車牌之辨識

摘要在本篇論文中，我們提出一個以監視器影像執行闖紅燈車牌知辨識方法，主要目的是用監視器影像拍下闖紅燈車輛之車牌，有助於警察進行取締違規車輛，以降低發生車禍事故。本文主要分成四個部分，第一部分是介紹如何辨識出紅燈區域，以圖像辨識和紅色變是來尋找紅燈區域，而紅色辨識運用了YCbCr色彩空間與RGB紅色模型，圖像辨識運用了圓度值範圍與面積範圍值,第二部要辨識紅燈時車輛超越白色停止線，運用HSV色彩空間來辨識白線停止線位置，HSV色彩空間較不易受到光線影響,且HSV顏色空間彼此是獨立的,第三部要辨識出闖紅燈之車牌，車牌辨識系統流程包含車牌影像前處理、車牌定位與字元的萃取、車牌字元辨識等三大部分,最後一部分利用matlab模擬研究步驟並呈現實驗結果。本文的研究貢獻如下: (1)更方便有效:現在馬路上都以埋設管線來進行感測闖紅燈之車輛，而我們提出以監視器影像執行闖紅燈車牌之辨識方法。 (2)低成本:利用馬路上現有監視器就能執行闖紅燈車牌辨識，有效節省成本

關鍵字

闖紅燈車牌、車牌字元辨識、紅燈區域

並列摘要

Abstract In this thesis, we will propose a scheme for detecting red light running by license plate recognition based on video surveillance. The main purpose is to detect the license plate of the vehicles passing through the red lights on the road by video surveillance to assist the police to ban irregular vehicle for reducing possibility of the traffic accidents. There are fourth main parts in this thesis. The first part introduces red light detection by using color and graphic detection. YCbCr color space and RGB color space are used to do the color detection. Roundness values range and area size are used to do graphic detection. The second part introduces red light stop line detection by using HSV color space. The reason of using HSV is stated below.HSV color space is not easily affected by light. In addition, HSV color space is independent of each other. In the third part we detect the red light running by license plate recognition. License plate detection system can be divided into three phases which are (1) pre-processing (2)vehicle plate locating and characters extracting and (3) the vehicle plate characters recognition. In the last part, simulation procedures are executed by matlab to present the results. In this thesis, the contributions of the research are as follows: (1) More convenient and efficient scheme: In the worldwide trend, government of countries apply buried pipeline for red light running detection. In this research, we detect red light running by license plate recognition based on video surveillance. It is more convenient and efficient than the former. (2) Low cost: While detecting red light running by license plate recognition based on video surveillance, there is no need to use buried pipeline. The cost of our scheme is much lower than the scheme by using buried pipeline.