Title: 多重感知整合式輔助安全車載資通訊系統
An Assistant Safety Telematics System with Integrated Multiple Sensors
Authors: 沈子貴
Shen, Tzu-Kuei
Lin, Chin-Teng
Keywords: 車載資通訊;影像處理;反透視模型;障礙物偵測;Telematics;Image Processing;Inverse Perspective Mapping;Obstacle Detection
Issue Date: 2010
Abstract: 近年來隨著都市人口和車輛數目不斷的增加,都市的交通問題越來越 嚴重,過多的車輛造成了交通壅塞的情況,交通意外事故的發生也更加的頻繁,這些事故造成了人民生命與財產的損失,讓每個人平均花在醫療上的費用大幅的增加,造成了龐大的社會成本與負擔,並降低國家的整體經濟競爭能力。因此,各個先進國家投入了相當多的人力在研究智慧型運輸系統(Intelligent Transportation Systems, ITS),ITS 的主要目的是利用先進科技於車輛及道路設施上,協助駕駛對車輛之控制,以減少事故,增進行車安全並達到提高用路效率與節能減碳的目標。本論文為一完整多重感知整合式輔助安全車載資通訊系統,其中包含車輛安全輔助系統與路口安全監控兩部份。車輛安全輔助系統利用單顆魚眼鏡頭攝影機對車輛進行周邊障礙物偵測與倒車輔助駕駛。周邊障礙物偵測系統是透過反映射模型將影像轉換至世界座標平面上,再根據此影像維度中具立體資訊的障礙物相對移動特性,進而定位出車輛周遭障礙物的位置與距離。倒車輔助駕駛系統是 以電腦視覺為基礎,自動估測倒車時的相對移動量,於畫面上估測出未來的行進軌跡。路口監控的部份,主要為一事件監控系統,開發一影像嵌入式系統平台用於收集路口相關資訊,並由路側單元接收與彙整,再經由DSRC與車上機進行資訊交流,整合車上與附近路口資訊給駕駛者參考,建立起完整的輔助安全車載資通訊系統。
In recent years, the urban population and vehicles increase continuingly. The problems of city traffic are more and more serious, such as too manyvehicles cause traffic jams and accidents frequently. These accidents do not only make people lost their life and fortune, but also waste lots of medical resources. Upper disastrous influences make enormous social costs, besides they also debase whole national economic competitiveness. For dealing with these problems, Intelligent Transportation System (ITS) becomes an important policy in each country. The main objective of ITS is to develop high-end technology on the electrical equipment in vehicles and traffic applications. Drivers can reduce the probability of traffic accidents and improve self-driving safety via controlling high-end assist driving technology and then achieve the goals of increasingly efficiency in road freight and energy saving and carbon reduction. This dissertation presents a whole integrated multi-sensors telematics safety system that can be divided into safety vehicle assistant system and intersection video surveillance system two parts. The first part contains obstacle detectionsystem and parking assistant system. The obstacle detection system one is transferred image coordinate into world coordinate by fisheye lens inverse perspective mapping modal (FLIPM) and follows the property of moving obstacle to position candidates’ location. The parking assistant system is based on computer vision algorithm via motion vector, and estimates the curve in the path of vehicle. The other segment is intelligent intersection surveillance system. Our concern is to consider a whole intersection events monitor system. It collects the traffic data from local intersection by embedded platform and then arranges these data for road side unit (RSU) to communicate with on board unit (OBU) in vehicles via DSRC protocol to set up an assistant safety telematics system.
Appears in Collections:Thesis

Files in This Item:

  1. 281001.pdf