基于点集优化和干扰点模糊化的车道线识别

Lane Detection Based on Points Set Optimization and Disturbance Set Fuzzying

提出了一种抗干扰车道线识别方法,通过对车道线点集的优化,能够去除道路上行驶的其他车辆造成的干扰。优化过程主要包括点集元素的预挑选、干扰点的去除以及有效点的补充。在干扰点去除的过程中,针对车辆边界很难精确定位的问题,对干扰点集进行了模糊化处理,引入隶属度函数对点集中元素的干扰程度进行描述。多工况试验证明,车道线识别方法能够稳定地对车道线进行识别,准确地提取车道线参数,并且算法体现出了对车辆干扰的良好抵抗能力。

An anti-disturbance lane detection method combined with vehicle detection was proposed. By the optimization of lane points set, the disturbance caused by other vehicles driving on the road can be eliminated. The optimization process included the pre-selection of points set members, the elimination of disturbance points and the complement of valid lane points. During the disturbance elimination, in order to solve the problem that the vehicle borders are hard to be located accurately, the fuzzy disturbance set was introduced and the membership degree was used to describe the disturbance degree of points set elements. Experiments under various conditions show that the lane detection method can recognize lanes steadily, obtain lane parameters accurately, and the algorithm owns resistance ability to vehicles disturbance.