張堂賢Cheng, Tang-Hsien臺灣大學:土木工程學研究所周辰儒Chou, Chen-JuChen-JuChou2010-06-302018-07-092010-06-302018-07-092008U0001-3006200811031800http://ntur.lib.ntu.edu.tw//handle/246246/187668本研究的目的為利用倒車輔助攝影機(Rear-end camera)建立先進安全車輛(Advance Safety Vehicle, ASV)的防被追撞警示系統(Rear-end collision avoidance and warning system)，係利用暨有影像處理系統為基礎，取得行車環境資訊，將資料分析及處理後，進一步發展防被追撞警示邏輯與系統。著車用電子產業發展迅速，駕駛人倒車輔助系統幾乎已為新車的基本配備，其中以超音波感測器和倒車攝影機最為普及。然而，這些設備目前只於倒車時開啟，因此本研究應用倒車攝影機於一般駕駛情況，提供駕駛者行車環境資訊，除降低車輛購置成本外，亦提升行車安全性。般防追撞警示系統主要再提醒駕駛者與前車的情況，降低因為駕駛者精神未集中而發生車禍的情況。針對前車駕駛人，本研究利用影像處理技術，偵測後方來車的縱向資訊，包括相對距離、相對速度以及相對加速度等，推估本車與後車的相對關係，並利用α-β-γ濾波器做出動態預測，並利用駕駛人感知反應時間，建立動態門檻值，適時對駕駛者發出警示。研究利用Borland C++ Builder程式開發軟體、工業電腦及車用影音設備，配合倒車攝影機建立防被追撞之行車預警模式。期望目標為建立一個可應用於實際道路狀況之安全警示系統，降低車輛追撞的發生率。The objective of this study is to construct the rear-end collision avoidance and warning system of advance safety vehicle (ASV) by rear-end parking camera. Based on image processing technology to get and analyze the driving environmental data, we developed the rear-end collision warning logic and system.ince the automobile electronics industry developed faster, the parking assist system, like the ultrasonic sensors and rear-end camera, almost become the basic equipment in vehicles. However, the parking assist equipments are only turned on when in parking mode. Therefore, to provide the following vehicle information for subjective one, this study tends to apply the rear-end camera in vehicle when moving forward. It not only reduces the equipment cost of the vehicle but also increases the safety for drivers.he main idea of warning system is to prevent accidents which caused by inattentive of drivers. This study used the image processing to detect the longitudinal data of the following vehicle, including the relative distance, velocity, and acceleration. We developed the dynamic thresholds by the relative data and the drivers’ perceptive reaction time to issue the warning signal for drivers. The α-β-γ filter was applied in this section to get the smoother relative data.his study built up a rear-end collision avoidance and warning system by rear-end camera. We constructed the warning program by Borland C++ Builder. The warning hardware of the system was used the industry personal computer and on-board video equipment. The experiment result was successful for the off-line video test. In other words, this study proves that the developed rear-end warning system could appropriately issue the warning signal for drivers.誌謝 i要 iiibstract vable of Contents viiist of Figures ixist of Tables xihapter One – Introduction 1.1 Background & Motivation 1.2 Objectives 4.3 Frameworks 5.4 Literature Reviews 7.5 Thesis Organization 10hapter Two – Fundamentals 12.1 Introduction 12.2 Fundamentals of System 12.2.1 Image Coordinate 13.2.2 Vehicle fixed coordinate 13.2.3 Definition of Parameters 14.3 Image Calibration 16.3.1 Calibration Experiment 16.3.2 Longitudinal Measurements Experiment 18.4 Data Collection 23.4.1 Longitudinal Distance Measurement 25hapter Three –Rear-end Collision Avoidance and Warning Module 29.1 Introduction 29.2 System Framework 30.3 Detection of Rear-end Objections 32.3.1 Collision Components 32.3.2 Warning Module Components 42hapter Four – Experiments and Results Analysis 46.1 Introduction 46.2 System Configurations 46.2.1 Hardware 46.2.2 Software 49.2.3 Constraints 50.3 Experiment Design 50.4 Result & Analysis 55.5 Summary 58hapter Five – Conclusions and Future Work 60.1 Conclusions 60.2 Future Work 61eference 631837959 bytesapplication/pdfen-US先進安全車輛倒車輔助攝影機被追撞α-β-γ濾波器駕駛人感知反應時間Advance Safety Vehicle(ASV)rear-end camerarear-end collision for leading vehicleα-β-γ filterdrivers’ perceptive reaction timethesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/187668/1/ntu-97-R95521507-1.pdf