基于光流法与背景建模法融合的车道线识别算法研究

Based on Fusing Optical Flow Method and Background Modeling for Lane Line Recognition

为了准确高效地识别无人驾驶车辆行驶过程中的车道线信息,采用光流法与背景建模法相融合的车道线识别算法,针对车辆行驶中的连续视频,对比连续视频帧中车辆前方背景的相对运动,运用光流法检测出背景中特征点的移动方向和距离,再结合背景建模法将背景滤除,混合高斯模型去噪后进行ROI特征区域的提取进一步减少计算量,最后进行车道线的提取与拟合,达到车道线识别与提取的目的。

A lane line recognition algorithm based on fusing the light flow with the background modeling method is presented in order to accurately and efficiently identify the lane line information for driverless vehicle.For continuous video in vehicle driving,the relative motion of the vehicle's front background in the continuous video frame is compared.The moving direction and distance of the feature points in the background are detected by optical flow method,and then the background is filtered by combining with background modeling method.After noise removal,the ROI feature area is extracted to further reduce the computational load.Finally,the lane line is extracted and fitted to identify the lane line information.