臺灣博碩士論文加值系統

當車輛在高速公路上快速行駛時，鄰近車輛的動向是駕駛者不可忽略的一項重要資訊。駕駛技術不純熟或對車況不熟悉的駕駛者，常因為必須同時注意許多事而疏忽了與鄰近車輛保持適當的距離；即使是熟練的駕駛者，在長時間駕駛之後，往往也因為太疲憊而無法時時刻刻集中精神來注意鄰近車輛的動向，等到發現可能和其它車輛產生碰撞時，已經來不及採取因應措施，悲劇於是發生。因此，鄰近車輛的動向分析是駕駛安全輔助系統中不可或缺的一部分。本研究即以電腦視覺為基礎，針對在高速公路上駕駛時拍攝到的影像序列來做前方鄰近車輛的動向分析，透過適當的警示訊息來輔助駕駛者注意前方鄰近車輛的動向。
在車載型攝影機所拍攝的影像序列輸入之後，先藉由偵測道路標線來找出路面區域，然後在路面區域中做鄰近車輛特徵的擷取，一方面減少計算量以加快處理速度，同時也可減少路面上方可能出現的雜訊干擾。將擷取到的車輛特徵輸入STA類神經模組後所形成的注意力焦點（focus of attention）顯示出鄰近車輛存在的位置，本系統利用以分子系統（particle system）為基礎的level-set方法來分割並追蹤偵測到的鄰近車輛。最後由動向分析模組針對追蹤的結果來分析鄰近車輛的動向，一旦分析出鄰近車輛與本車有明顯的相對運動，即輸出訊息供駕駛者參考。最後的實驗結果顯示出本研究所提出的方法能快速且準確地分析前方鄰近車輛的動向，在未來若加上拍攝後方鄰近車輛的分析，將可提供更豐富的訊息給駕駛者，讓駕駛者能更輕鬆地開車，同時也提高駕駛的安全性。

When driving at high speed on a freeway, the motions of nearby vehicles is very important for a driver. An inexperienced driver has to pay attention to so many things that the motions of nearby vehicles are sometimes ignored. Even an experienced driver may occasionally ignore motion of the nearby vehicles after a long period of driving. Hence, motion analysis in a driver-assistance system plays a very important role for improving driving safety. The objective of this research is to propose a vision-based system for analyzing the motions of nearby vehicles on a freeway.
When the video sequences captured from a forward-looking camcorder mounted in a vehicle are input to vehicle feature extraction is performed in the area of the road surface, which has been located by detecting the lane markings. The feature extraction results are then fed into the STA neural module. Once the focuses of attention, which indicate the possible locations of vehicles, are formed, the segmentation and tracking module is activated. The tracking results of consecutive frames are analyzed using an accumulated method. The experimental results show that our method can quickly and correctly recognize the motion of nearby vehicles in front of our vehicle. If the motion analysis of nearby vehicles at the rear of our vehicle can be combined with our system, drivers can drive more easily and driving safety can be promoted.

Chapter 1 Introduction
1.1 Background
1.2 Problem Formulation and Objective
1.3 Organization of this Thesis
Chapter 2 System Configuration
2.1 Pre-attention Process
2.2 Nearby Vehicle Detection
2.3 Segmentation and Tracking of Focuses of Attention
2.4 Motion Analysis of Focuses of Attention
Chapter 3 Nearby Vehicle Detection Module
3.1 Extraction of Pre-attention Map
3.1.1 Feature Extraction of Lane Markings
3.1.2 Straight line fitting
3.2 Nearby Vehicle Detection
Chapter 4 Segmentation and Tracking
4.1 Classical Level Set Methods
4.2 A Level Set Method Using a Particle System
4.3 Implementation of the Particle System
4.3.1 Initialization of Particles
4.3.2 Energy Minimization
4.3.3 Tracking of Particles
Chapter 5 Motion Analysis
5.1 Vehicle Motions
5.2 Motion Analysis
5.2.1 Lane Change Analysis (LCA)
5.2.2 Relative Distance Change Analysis (RDCA)
Chapter 6 Experimental Results
6.1 Nearby Vehicle Detection
6.2 Segmentation and Tracking of Focuses of Attention
6.3 Motion Analysis
6.4 Discussion
Chapter 7 Conclusions and Future Work
7.1 Summary
7.2 Future Research
Appendix A Projection of Lane Markings
Appendix B Spatiotemporal Attention Neural Network
Bibliography

[Bet00] M. Betke, E. Haritaoglu, and L. S. Davis, "Real-Time Multiple Vehicle Detection and Tracking from a Moving Vehicle," Machine Vision and Applications, vol. 12, pp. 69-83, 2000.
[Bla95] C. Blanc and C. Schlick, "Extended field functions for soft objects," Implicit Surfaces’ 95, pp. 18-29, Grenoble, France, Apr. 1995.
[Bly02] P. Blythe, "Video-based Vehicle and Pedestrian Tracking and Motion Modeling," Road Transport Information and Control, pp. 19-21, Mar. 2002.
[Fan03] C. Y. Fang, S. W. Chen, and C. S. Fuh, "Automatic Change Detection of Driving Environments in a Vision-based Driver Assistance System," IEEE Transactions on Neural Networks, pp. 646-657, 2003.
[Gar96] W. F. Gardner and D. T. Lawton, "Interactive Model-based Vehicle Tracking," IEEE Transaction on Pattern Analysis and Machine Intelligence, vol. 18, pp. 1115-1121, Nov. 1996.
[Gon92] R. C. Gonzalez and R. E. Woods, Digital Image Processing, Addison Wisely, New York, 1992.
[Jeo01] S. G. Jeong, C. S. Kim, K. S. Yoon, J. N. Lee, J.I. Bae, and M. H. Lee, “Real-Time Lane Detection for Autonomous Navigation,” IEEE Intelligent Transportation Systems Conference Proceedings, pp. 508-513, Oakland (CA), USA, Aug. 2001.
[Kim01] J. B. Kim, C. W. Lee, K. M. Lee, T. S. Yun, and H. J. Kim, "Wavelet-Based Vehicle Tracking for Automatic Traffic Surveillance," IEEE TENCON'01, Vol. 1 pp. 313-316, Singapore, Aug. 2001.
[Tea99] M. K. Teal and A. A. W. Smith, "Learning Structural Change for Identification and Tracking of Vehicles Moving in Open World Scenes," in Proc. IEE 7th International Conference on Image Processing & its Applications (IPA'99), pp. 706-710, Manchester, UK, July 12-15, 1999.
[Set01] J. A. Sethian, Level Set Methods and Fast Marching Methods: Evolving Interfaces in Computational Geometry, Fluid Mechanics, Computer Vision, and Materials Science, Cambridge University Press, New York, 1999.
[Yah98] H. Yahia, J. Berroir, and G. Mazars, "A particle system formulation for tracking deformable structures," Oceans'98, Nice, France, Sep. 1998.