高速公路路面识别技术研究
Study on Highway Road Detection Based on Computer Vision
Abstract
根据高速公路路面的基本特征,通过分道线特征点投影位置的计算,提取出了比较完整的当前车道分道线特征点。建立道路边缘灰度模型,采用gAbOr变换多分辨率的特点,对图像按方向进行滤波处理,实现道路边缘线检测和拟合。实验结果表明,所讨论的方法在工程实践中有实际应用价值。 According to the basic characteristics of the road and through marginal extraction,the location of current lane line can be acquired.After calculating projection location of the lane feature points,feature points of a relatively complete current lane dividers can be distilled.Establishing the model of road edge grey scale and adopting the multi-resolution characteristics of Gabor transform,they can filter the image so that they can realize the detection and fitting of the road edge line.Experimental results show us that the method discussed in this text has a certain value in practice.