在汽車工業領域中，車輛安全一直是被非常重視的議題，近年來隨著科技的發展，越來越多科技產品應用於汽車，特別是主動安全系統，防止事故的發生，有效提升車輛安全並為未來自動駕駛做準備。而這些主動安全系統，影像處理被廣泛的應用，例如車道辨識、路標辨識、行人偵測、車輛偵測、雨滴感應等，這些功能的開發為駕駛人帶來不只便利，更為駕駛人提升行車安全。
在駕駛車輛的過程中，駕駛人的視角受限於車輛的設計，無法自由的移動視野，因此常常會有重要的資訊被遮蔽，當遇到緊急狀況時，只能依靠前方車輛的動態，做為駕駛人反應的依據。因此，本研究利用現今廣泛被架設於車外後視鏡上的攝影機，以與其他辨識功能共用影像為前提，將自車周圍環境的資訊，轉換至二維鳥瞰圖，清楚明瞭的畫面以供駕駛人快速的判斷當下的情況，資訊包含與周圍車輛的相對位置、相對速度等資訊。本研究針對攝影機可以掌握而駕駛人無法看到的目標物，藉由不同的演算法進行辨識和追蹤，一層一層的累積信心指數，使得車輛辨識以及追蹤系統更為完善。縝密的考量運算時間、辨識準確度，使實驗成果能在150公尺內，日間以每畫格0.79秒的運算時間，達到97.64%的車輛辨識正確率；在夜間以每畫格1.28秒的運算時間，達到91.74%的車輛辨識正確率。

The vehicle safety has been always taken as an important issue in vehicle industry. With the development of technology, there are more and more high-tech devices installed on vehicles, especially for active safety systems and autonomous driving. When it comes to driving, due to the limit of cabin, the driver cannot change the view freely to see widely. There would be important information covered by other objects. Therefore, the driver can react only judging by the behavior of the frontal vehicle when facing emergency. The safety assistances nowadays cannot solve this problem because they detect first surrounding objects. In order to provide further environmental information, this research utilizes six cameras which are installed on rearview mirrors outside the vehicle. Including two frontal cameras, two telescope cameras and two rear cameras.
Comparing with other kinds of sensor, cameras are easily accessible and sharing the same scene with other functions. The system can establish aerial view on instrument panel. While marking where the obstacles are, it provides message like relative distance, relative speed and the behavior of street vehicles. Not only it is possible to alarm the driver if the speed is plummeted without relying frontal car but also provides two steps route suggestion to overtake. In spite of the fact that these functions are based on currently commercialized system like lane detection, its processing time, accuracy, detecting distance have been greatly improved in this research. However, the most difficult part is detecting vehicles which are easily covered by other objects. This research distinguishes these features and accomplishes function mentioned above successfully in 150 meters by several kinds of algorithm in nearly real time. It takes 0.79 second to process day time frame and has 97.64% accuracy in vehicle detection. While it takes 1.28 seconds to process night time frame and has 91.74% accuracy in vehicle detection.
Various aspects are considered such as processing time when combining several functions, feasibility of device installation for general users, and interaction with the driver in high speed cruise. For the purpose of improving driving safety and paving the way for autonomous driving, the system is kept making more perfect in this research.

[1] P. Hough. “Method and Means for Recognizing Complex Patterns.” U.S. Patent No. 3069654, 1962.
[2] R. Duda and P. Hart. “Use of The Hough Transformation to Detect Lines and Curves in Pictures.” Communications of the ACM, Vol.15, No. 1, pp. 11-15, 1972.
[3] A. Coates, B. Carpenter and C. Case. “Text Detection and Character Recognition in Scene Images with Unsupervised Feature Learning.” International Conference on Document Analysis and Recognition, pp. 440-445, 2011.
[4] K. Bhosale, J. Jadhav, S. Kalyankar, and P. R. R. Bhambare. “Number Plate Recognition System for Toll Collection.” International Journal of Emerging Technology and Advanced Engineering, Vol. 4, No. 4, pp. 729-732, 2014.
[5] J. A. Khan and M. A. Shah. “Car Number Plate Recognition (CNPR) System Using Multiple Template Matching.” International Conference on Automation and Computing, pp. 290-295, 2016.
[6] N. Dalal and B. Triggs. "Histograms of Oriented Gradients for Human Detection." IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Vol. 1, No. 20-25, pp. 886-893, 2005.
[7] P. Qi, S. Su and X. Hu. "Modeling Outer Products of Features for Image Classification." International Conference on Advanced Computational Intelligence, pp. 334-338, 2013.
[8] J. Kim, J. Baek and E. Kim. “A Novel On-Road Vehicle Detection Method Using pi HOG.” IEEE Transactions on Intelligent Transportation Systems, Vol. 16, No. 6, pp. 3414-3429, 2015.
[9] W. Pei, Y. Zhu, X. Zuo and M. Lu. “A Multi-Scale Vehicle Tracking Algorithm Based on Structured Output SVM.” IEEE International Congress on Image and Signal Processing, pp. 1-6, 2015.
[10] S. S. Teoh. “Performance Evaluation of HOG and Gabor Features for Vision-Based Vehicle Detection.” IEEE International Conference on Control System Computing and Engineering, pp. 66-71, 2015.
[11] J. F. Canny (17 November 2016), Available: https://www2.eecs.berkeley.edu/Faculty/Homepages/canny.html
[12] J. Flusser, S. Farokhi and C. Hoschi. “Recognition of Images Degraded by Gaussian Blur.” IEEE Transactions on Image Processing, Vol. 25, No. 2, pp. 790-806, 2016.
[13] Canny Edge Detection (17 November 2016), Available: http://140.115.11.235/~chen/course/vision/ch6/ch6.htm
[14] A. Mallareddy, R. C. Babu and K. D. Rani. "Vehicle Tracking Using Canny Edge Detector." International Journal of Research, Vol. 1, No. 7, 2014.
[15] F. Nelwamondo and G. M. Hocquet. “Fusion of Phase Congruency and Harris Algorithm for Extraction of Iris Corner Points.” International Conference on Artificial Intelligence, pp. 315-320, 2015.
[16] Harris and Shi-Tomasi Key Points Detection (17 November 2016), Available: http://fanli7.net/a/bianchengyuyan/C__/20120731/196260.html
[17] J. Shi and C. Tomasi. “Good Features to Track.” IEEE Computer Vision and Pattern Recognition, pp. 593-600, 1994.
[18] F. Lei, W. Ren and L. Lv. “Image Matching Algorithm Based on Improved Harris.” International Conference on Computational Intelligence and Communication Networks, pp. 378-381, 2015.
[19] B. Coifman, D. Beymer, P. McLauchlan and J. Malik. “A Real-Time Computer Vision System for Vehicle Tracking and Traffic Surveillance.” Transportation Research Part C, pp. 271-288, 1998.
[20] T. A. C. Bragatto, G. I. S. Ruas and V. A. P. Benso. “A New Approach to Multiple Vehicle Tracking in Intersections Using Harris Corners and Adaptive Background Subtraction.” IEEE Intelligent Vehicles Symposium, pp. 548-553, 2008.
[21] D. Seo and K. Jo. “Inverse Perspective Mapping Based Road Curvature Estimation.” IEEE/SICE International Symposium on System Integration, pp. 480-483, 2014.
[22] D. Comaniciu, V. Ramesh and P. Meer. “Real-Time Tracking of Non-Rigid Objects Using Mean Shift.” IEEE Conference on Computer Vision and Pattern Recognition, pp. 142-149, 2000.
[23] Mean Shift (17 November 2016), Available: http://www.cnblogs.com/liqizhou/archive/2012/05/12/2497220.html
[24] J. G. Allen, R. Y. D. Xu and J. S. Jin. “Object Tracking Using Camshift Algorithm and Multiple Quantized Feature Spaces.” Proceedings of the Pan-Sydney area workshop on Visual information processing, pp. 3-7, 2004.
[25] Meanshift and Camshift (17 November 2016), Available: http://docs.opencv.org/trunk/db/df8/tutorial_py_meanshift.html
[26] S. G. Kim, J. E. Kim and K. Yi. “Detection and Tracking of Overtaking Vehicle in Blind Spot Area at Night Time.” IEEE International Conference on Consumer Electronics, pp. 47-48, 2017.
[27] K. M. Jeong and B. C. Song. “Night Time Vehicle Detection Using Rear-Lamp Intensity.” IEEE International Conference on Consumer Electronics-Asia, pp. 1-3, 2016.
[28] X. Zhang and Z. Shi. “Study on Lane Boundary Detection in Night Scene.” IEEE Intelligent Vehicles Symposium, pp. 538-541, 2009.
[29] Y. C. Lin, C. C. Lin and L. T. Chen. “Adaptive IPM-Based Lane Filtering for Night Forward Vehicle Detection.” IEEE International Conference on Consumer Electronics, pp. 1568-1573, 2011.
[30] H. Guan, W. Xingang and W. Wenqi. “Real-Time Lane-Vehicle Detection and Tracking System.” Chinese Control and Decision Conference, pp. 4438-4443, 2016.
[31] H. Tan, Y. Zhou and Y. Zhu. “A Novel Curve Lane Detection Based on Improved River Flow and RANSAC.” International Conference on Intelligent Transportation Systems, pp. 133-138, 2014.
[32] C. R. Jung, C. R. Kelber. “Lane Following and Lane Departure Using a Linear-Parabolic Model.” Image and Vision Computing, Vol. 23, No. 13, pp. 1192-1202, 2005.
[33] C. R. Jung and C. R. Kelber. “An Improved Linear-Parabolic Model for Lane Following and Curve Detection.” Symposium on Computer Graphics and Image Processing, pp. 131-138, 2005.
[34] Q. Chen and H. Wang. “A Real-Time Lane Detection Algorithm Based On a Hyperbola-Pair Model.” IEEE Intelligent Vehicles Symposium, pp. 510-515, 2006.
[35] T. Akita and Y. Yamada. “Image Recognition of Vehicle Applying Fusion of Structured Heuristic Knowledge and Machine Learning.” International Journal of Automotive Engineering, Vol. 5, No. 3, pp. 101-108, 2014.
[36] Y. Chen and Q. Wu. “Moving Vehicle Detection Based on Optical Flow Estimation of Edge.” International Conference on Natural Computation, pp. 754-758, 2015.
[37] P. K. Bhaskar, S. P. Yong and L. T. Jung. “Enhanced and Effective Parallel Optical Flow Method for Vehicle Detection and Tracking.” International Symposium on Mathematical Sciences and Computing Research, pp. 138-143, 2015.
[38] Garmin GDR 530 (17 November 2016), Available: http://www.garmin.com.tw/products/ontheroad/gdr-c530/
[39] CANSONIC Z1 (8 June 2017), Available: https://cansonic.com/usa/product/ultraduo.html
[40] C. Rother. “A New Approach to Vanishing Point Detection in Architectural Environments.” Image and Vision Computing, Vol. 20, No. 9-10, pp. 647-655, 2002.
[41] H. Bay. “Surf: Speeded Up Robust Features.” European conference on computer vision, pp. 404-417, 2006.
[42] Template Matching (17 November 2016), Available: http://docs.opencv.org/2.4/doc/tutorials/imgproc/histograms/template_matching/template_matching.html
[43] Raindrop Terminal Speed (2 August 2017), Available: http://www5b.biglobe.ne.jp/~saturn/meteology/04.htm