说明：  提出了基于2层粒子滤波数据融合的道路检测框架.通过道路检测算法的2个主要模块:特征 提取和道路参数估计,同时引入粒子滤波算法,并以支持向量机(SVM)识别道路模型、初始化粒子滤波器提高算法的环境适应性,以期找到基于机器视觉的道路检测算法在鲁棒性和实时性之间更好的平衡点.
(Since the normal monocular-vision based lane detection algorithms for outdoor applications sel- dom consider the balance between the robustness and the real-time requirements, this paper proposed a no- vel lane detection framework to balance these requirements by introducing the particle filter theory into fea- ture extraction and road parameters estimation models at the same time. )