透過您的圖書館登入
IP:3.144.172.115
  • 學位論文

應用YOLO於交通安全與科技執法

Application of YOLO for Traffic Safety and Technology Enforcement

指導教授 : 鄭立德
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


科技執法是現今改善交通安全的重要方法之一,因交通事故件數與事故率逐年攀升,輕則財損,重則人亡,而事故發生原因皆起於交通違規,加上台灣高齡化與少子化社會問題遲遲無法得到改善,導致警察人力嚴重缺乏,科技執法在未來更是不可或缺的重要課題。本論文研究目標即為此問題發展出一套科技執法之技術,將YOLO(You Only Look Once)物件偵測技術應用於科技執法。當偵測之車輛符合違規條件時,則記錄違規車種數量,並統計一時段內發生違規之頻率。往後將結合車牌辨識,完成違規取締之目標。

關鍵字

科技執法 YOLO 物件偵測 車牌辨識

並列摘要


Technology enforcement is one of the important ways for improving traffic safety. Because of the trend of fewer children, the number of police-officers won’t be enough. Therefore, the traffic violations will often occur, and increase the car accidents. Most car accidents are caused by traffic violations, in severe cases, it will cause death. Based on above description, the study target of this paper is to develop a technology enforcement for this issue. We use YOLO(You Only Look Once) object detection technology for technology enforcement. When the detected vehicle meets the violation conditions, the number of violations will be recorded and counted for a period of time. In the future, it will be combined with license plate recognition to complete the violations of banning.

參考文獻


[1] 維基百科,<少子化>
[2] 維基百科,<各國警察數量列表>
[3] 內政部<警政統計查詢網>
[4] Medium, 深度學習-物件偵測:You Only Look Once(YOLO)
[5] Joseph Redmon and Ali Farhadi. YOLO v3: An incremental improvement. arXiv preprint arXiv:1804.02767, 2018

延伸閱讀