基于传统边缘算子的车道线识别研究

Research on lane line recognition based on traditional edge operators

车道线识别检测技术是汽车实现自动驾驶的关键技术之一,利用图像低级特征提取技术实现边缘检测是车道线识别的最佳方法之一。为解决传统车道线检测易误检、漏检、效率低等问题,利用结构化道路条件下车道线图像,在去噪、滤波等预处理后对获取的图像进行ROI区域提取。基于传统边缘检测算子边缘提取方法,发现Roberts等算子在考虑环境因素下检测结果存在细节边缘提取不足的问题。通过比较发现,利用霍夫变换改进的Canny算子边缘检测方法不仅能够有效提高检测结果的精确度,而且能够同时提升运行速度。该方法能够有效避免传统边缘检测方法检测结果误差大的问题,具有较强的鲁棒性和抗干扰性。

Lane line recognition and detection technology is one of the key technologies for achieving autonomous driving of automobiles.Using low-level image feature extraction technology to achieve edge detection is one of the best methods for lane line recognition.Since the traditional lane line detection is prone to false detection and missed detection,and its detection efficiency is low,after the preprocessing of denoising and filtering,the obtained lane line images under the conditions of structured road are subjected to ROI(region of interest)extraction.On the basis of the traditional edge detection operator edge extraction method,it is found that Roberts and other operators have insufficient detail edge extraction in the detection results in case of considering environmental factors.By comparison,it is found that the Canny operator edge detection method improved by Hough transform can not only effectively improve the accuracy of the detection results,but also increase the running speed.This method can effectively avoid the large detection error of traditional edge detection methods,and has strong robustness and anti-interference.