本研究開發一種適用於隧道環境下利用影像之方位推估法。傳統的方位推估法實作方式係以尺度和方位感測器來實現，然而累積誤差現象是不可避免的。本論文藉由不同的連續影像截取，從影像中分析取得方位和位移來實現方位推估運算。由於影像隨機的特性，方位推估累積誤差的現象低於一般的傳統方位推估法。

演算法以隧道燈光做為影像特徵。這些特徵則以多項式形式用來描述路徑曲線，且藉由不同之連續影像中燈光位置差異計算出車輛之位移量。從實驗顯示出高精確度位移量計算之結果，但對於角位移之精確度仍然需要進一步改善。以連續影像實現方位推估法所求得之結果証實無需傳統之感測器，此方法開啟新的方式利用影像處理來完成單純隧道情況之導航。

A vision-based dead-reckoning method was developed for vehicle navigation in tunnel environments. Traditionally, dead-reckoning method is implemented by distance and heading sensors and error accumulations become inevitable with sensor biases and noises. Using differences of consecutive images, displacement vectors between images can be derived. And dead-reckoning can be achieved. The randomness characteristics of the vision derived displacement errors result in less severe error accumulation in the dead-reckoning operation.

The algorithm utilizes the lightings in tunnels as feature points. These lights are used to derive curves described by polynomials and the relative displacements of lights in different images are used to calculate the vehicle displacements between images. The experimental results show very good accuracy in the longitudinal displacement calculation but there are still rooms for the angular displacement accuracy improvement. Nevertheless, acceptable dead-reckoning results can be obtained by processing sequential images within the tunnel without the need of any traditional dead-reckoning sensors, and this method opens new possibilities of vehicle navigation in tunnel situations purely based on vision-processing.

[1] Y. Wanga, E. K. Teoh and D. Shen, “Lane detection and tracking using B-Snake,” in International Conf. on Information Intell. and Syst. , October 2003.
[2] D. J. Kang, J. W. Choi and I.S. Kweon, “Finding and tracking road lanes using line-snakes,” Proceedings of Conference on Intelligent Vehicle, 1996.
[3] C. Rasmussen, “Texture-Based Vanishing Point Voting for Road Shape Estimation,” British Machine Vision Conference, 2004.
[4] Y. Wang, D. Shen and E. K. Teoh, “Lane Detection Using Catmull-Rom Spline,” IEEE Intelligent Vehicles Symposium, Oct. 1998.
[5] A. Minagawa and N. Tagawa, “Vanishing point and vanishing line estimation with line clustering,” IEICE Trans Info & Syst, 2000.
[6] C.F. McLean and D. Koyyuri, “Vanishing point detection by line clustering,” IEEE, Pattern Anal. & Mach. Intell., 1995.
[7] D. Liebowitz and A. Zisserman, “Metric rectification for perspective images of planes,” IEEE International Conference on Computer Vision & Pattern Recognition, Santa Barbara, CA, June, 1998.
[8] J. Semple and G. Kneebone, “Algebraic Projective Geometry,” Oxford University Press, 1979.
[9] R. Mohr and B. Triggs, “Projective geometry for image analysis,” Technical report, International Society for Photogrammetry and Remote Sensing, Vienna Congress, July 1996.
[10] P. Gamba, A. Mecocci and U. Salvatore, “Vanishing point. detection by a voting scheme,” 1996.
[11] C. Brauer-Burchardt, and K.Voss, “Robust Vanishing, Point Determination in Noisy Images,” Friedrich Schiller. University Jena, Jena, Alemania, 2000.
[12] R. Cipolla, T. W. Drummond and D. Robertson, “Camera Calibration from Vanishing Points in Images of Architectural Scenes,” British Machine Vision Conf. , T. Pridmore and D. Elliman, Nottingham, UK, 1999.
[13] R. Y. Tsai, “A versatile camera calibration technique for high-accuracy 3d machine vision metrology using off-the-shelf TV cameras and lenses,” IEEE J.Robotics Automat. , 1987.
[14] C. Chen, C. Yu and Y. Hung, “New calibration-free approach for augmented reality based on parameterized cuboid structure,” of the IEEE ICCV99, 1999.
[15] X. Cao and M. Shah, “Camera calibration and light source estimation from images with shadows,” IEEE CVPR, 2005.
[16] F. Devernay and O. Faugeras, “Automatic calibration and removal of distortion from scenes of structured environments,” In SPIE, San Diego, CA, July 1995.
[17] J. Canny, “A computational approach to edge detection,” IEEE Trans PAMI, 1986.
[18] D. G. Lowe, “Distinctive Image Features from Scale-Invariant Keypoints,” International Journal of Computer Vision, 2004.