This paper presents a method of structured road lane detection for blind travel aid. Median filter is implemented to process image firstly, it then marks off the region of interesting in the initial image, Canny edge enhancement, threshold are used to segment the image, road lane is fitted by modified Hough transformation. Finally, according to the detected region judge whether racing course of blind has a deviation. The experimental results prove this algorithm is very robust, accurate and real-time.%实现对盲人辅助行走装置中结构化道路的检测和行走方向是否偏离道路的检测:首先对单目摄像头采集的道路图像进行中值滤波,抑制随机噪声；然后从整幅图像中划分感兴趣区域(ROI),通过Canny算子对ROI区域的图像进行初次边缘检测,接着采用基于灰度的阈值分割方法对图像进一步分割,选用改进的Hough变换拟合出完整、清晰的道路边缘线,最后根据检测出的道路区域判定行走方向是否偏离道路.使用Halcon图像处理软件进行仿真实验,结果表明上述算法具有较好的实时性、准确性和鲁棒性.
展开▼
