利用倒傳遞網路搭配透視轉換不變性之廣義霍夫轉換做路標的偵測與定位

Road Sign Detection and Locating Using PTIGHT Based on BP Neural Network

顏家俊(Chia-Chun Yen)

指導教授 ： 駱榮欽

國立臺北科技大學/電資學院/電腦與通訊研究所/碩士(2009年)

摘要

本篇論文，我們提出一種能自動偵測與辨識路標之系統，此系統可用以輔助路徑規劃與定位，此系統使用類神經分類器搭配型態學初步偵測出路標可能區，再配合透視轉換不變性之廣義霍夫轉換精確地偵測出路標。在實驗結果中，深度的誤差約為28cm，角度為11度；此外，由於是在自然環境下做偵測，因此也針對特別的形狀做處理;例如路標影像的歪斜。

關鍵字

立體視覺 ； 類神經網路 ； 路標偵測 ； 透視 ； 廣義霍夫轉換

並列摘要

In the paper, we propose a new automatic road sign detection and locating system to assist path planning and location the road sign. In the system, Back-Propagation Neural Networks has been first adopted to search the road sign of interest, and then the morphology is employed to smooth the RSOI. In the real environment road signs are usually affected by affine distortion. Therefore, we use perspective-transformation-invariant generalized Hough transformation to detect and locate the road signs accurately. Several experimental results are also included to demonstrate the practicable of the system.

並列關鍵字

Stereo Vision ； Neural Networks ； Road Sign Detection ； Perspective ； Generalized Hough Transform

參考文獻

[1]Computer Vision, A Modern Approach.

[2]Feature Extraction and Image Processing-Second edition.

[4]C. Y. Fang, S. W. Chen, and C. S. Fuh, "Road-sign detection and tracking," IEEE Trans. Vehicular Technology, vol. 52, no. 5, pp. 1329-1341, 2003.

[6]X. Chen, J. Yang, J Zhang, and A. Waibel, "Automatic detection and recognition of signs from natural scenes," IEEE Transactions on Image Processing, vol. 13, no. 1, 2004.

[7]塗韋程，基於多個到傳遞網路分類器搭配視差圖做戶外自動車導航之研究，國立台北科技大學機電整合所，2008。