基于多信息融合优化的鲁棒性车道检测算法

Robust lane detection algorithm based on multiple information fusion and optimizations

为了提高复杂环境下车道线检测的鲁棒性,提出一种基于多特征信息融合优化的鲁棒性车道线检测算法.首先构建了基于二次曲线空间道路模型图像中左右车道线数学模型;然后融合像素梯度值、梯度方向、像素灰度以及车道线结构等多特征信息,构造后验概率函数;最后采用基于免疫克隆策略的改进粒子群优化算法优化车道线模型参数,实现车道线提取.对实际道路图像的实验结果表明,引入多特征信息后,在道路中存在阴影、车辆和道路标记等干扰因素,以及车道线模糊、对比度较低的情况下,该算法也能快速准确地提取车道线,具有很强的鲁棒性.

To improve the robustness of lane detection under complex conditions,a robust lane detection approach based on multiple information fusion optimizations is proposed.First based on the spatial quadratic road model,the left and right lane model expression in image plane is constructed.Then combined with the gradient value,gradient direction,gray information and road structure information,the expression of the posterior probability is derived.Finally the particle swarm optimization combined with immune clone strategy is used for calculating the model parameters.The results of the real road image experimentation show that after involving the multiple features information,the proposed method can robustly and rapidly detect the lane markings even if there are some interference factors in the road such as shadow,vehicle and land mark etc.,as well as baded lane boundaries and relatively weak local contrast.