摘要： 提出一种基于图像处理的盲道识别算法。根据采集到的盲道图像，综合利用阈值分割、彩色连续性空间分割、纹理分割，结合直线检测算法，分析各种盲道特征以区分盲道与普通人行道。实验结果表明，该算法能有效分割多数盲道图像，检测出图片中的盲道区域和盲道行进方向，可以自适应选择速度最快且有效的分割方法，适用于实时导盲系统。

关键词: 图像分割, 基于纹理分割, 盲道识别

Abstract: This paper proposes a blind way recognition algorithm based on image processing, which contains thresholding, color image segmentation, texture-based segmentation and line detection technique. It can study the characters of the blind ways then segment them from the pavements. Experimental results show that this algorithm is effective for most of the images of the blind way. It can partition the area of the blind ways and tell the turning direction of the blind ways. As this algorithm can adaptively select the fastest method to run, it can be implemented in the real-time blind direction system.

Key words: image segmentation, texture-based segmentation, blind way recognition

中图分类号: 

• TN911.73