基于模糊化SOBEL自适应算子的直线车道处理

Lane line processing based on the blur SOBEL adaptive operator

针对汽车安全辅助驾驶系统领域的车道线检验问题,提出了一种改进的基于模糊化SOBEL自适应算子直线车道线检验.首先在车道线边缘检验时,采用了模糊化SOBEL自适应算子进行边缘检验;为了更好地选取阈值,采用了一种基于体育项目评分标准的阈值选取方法,保证所保留的车道线信息的正确性,并给出了一种增修补车道线信息的数学模型;利用梯度函数的分布特性过滤掉噪声,最大限度地保留车道线分布信息,利用Hough变换拟合出最终的车道线.经MATLAB仿真,并在嵌入式系统上实现.实验表明,该方法简单,能保证系统的实时性.

For the lane line detection problems in the field of vehicle safety assistant driving system, the paper puts forward an improved straight line detection based on blur SOI3EL adaptive operator.First of all, on the edge of the lane line detection, the paper adopted blur SOBEL adaptive operator to detect edge of the lane line.In order to select a better threshold,the paper employed a method of threshold selection based on sports scoring criteria to guarantee the correctness of the reserved lane line information.The pa- per also gives the mathematical model to modify the lane line information. In order to maximally retain lane line distribution information features, gradient function was used to filter out the noise distribution. Finally Hough transform was adopted to present the lane line. The experimental results show that this method is simple, can guarantee the real-time performance of the system by MATLAB simulation and embedded system implementation.