應用機器視覺於視障者穿越道路 輔助系統之研究

A Study of Road-Crossing Auxiliary System for Visually Impaired using Machine Vision

蔡宜哲(Yi-Che Tsai)

指導教授 ： 吳明川

國立臺北科技大學/機電學院/機電整合研究所/碩士(2011年)

摘要

視障者在穿越道路時，以盲人音響號誌做為判斷依據，但是道路口有許多聲音干擾，容易造成視障者的誤判，因此本研究利用機器視覺技術發展一套視障者穿越道路輔助系統，進行行人穿越道位置、穿越道路距離與行人穿越道路指示燈秒數的辨識，提供道路資訊來幫助視障者安全穿越道路。
本研究在辨識行人穿越道路指示燈上，運用HSV色彩空間提取秒數顏色，再利用影像重疊技術辨識其特徵，最後使用歐氏最小距離做圖樣比對，完成秒數辨識。行人穿越道辨識上，先運用行人穿越道路黑白相間的顏色特性，將穿越道標記在相同區域，連通標記提取出穿越道，最後計算直線線段量，並利用線段量辨識出位置與距離。

關鍵字

視障者 ； 行人穿越道 ； 行人穿越道路指示燈

並列摘要

When blind people want to cross the road, they need some signals to help them. However, the environment noise causes difficulties for the blind to achieve this task. Therefore, the purpose of this study is to develop a road-crossing auxiliary system for the blind to assist blind people in crossing the road by using machine visions. This system provides road information to help blind people safely cross the road and contains two detection systems: detection of pedestrian signals and detection of crosswalks.
The pedestrian signal detection process of this research: First, using the HSV color space features to extract the countdown time on the pedestrian signal. Next, enhance characteristics by overlapping. Finally, recognize the time character through labeling location area and then template matching by Euclidean distance.
The crosswalk detection process of this research: First, characterize the color of the crosswalk. Next, obtain labeling same area of the crosswalk. Finally, perform calculation of lines to detect the location and distance of crosswalk.

並列關鍵字

The Blind ； Crosswalks ； Pedestrian Signal

參考文獻

[8] 曹舜賢，視障者人行交通號誌辨識輔具之研發，碩士論文，國立台灣大學電機工程學系，台北，2007。

[1] Bruno Ando, “Electronic Sensory Systems for the Visually Impaired,” IEEE Instrumentation & Measurement Magazine, 2003.

[6] A. D. L. Escalera, E. Moreno, M. A. Salichs, and J. M. Armingol, “Road Traffic Sign Detection and Classification,” IEEE Transactions on Industrial Electronics, vol. 44, 1997.

[7] M. S. Uddin, and T. Shioyama, “Bipolarity- and Projective Invariant-Based Zebra-Crossing Detection for the Visually Impaired,” Proc. IEEE Computer Society Conf. Computer Vision and Pattern Recognition (CVPR), San Diego, California, USA, June 2005.

[9] Tadayoshi Shioyama, Haiyuan Wu, Naoki Nakamura and Suguru Kitawaki, “Measurement of the lenth of pedestrian crossings and detection of traffic lights form image data,” Measurement Science and Technology, 2002, pp1440–1447.