摘要
Hough变换算法应用于车道线识别,存在复杂度高、运算量大、运行时间长等不足,导致车道线识别系统实时性不高。为此本文提出一种基于参数空间分块即将参数空间划分为均匀的累积区域的改进的霍夫变换算法。首先进行目标区域划分、图像灰度化和滤波去噪,然后进行二值化,最后分别用Hough变换和改进的Hough变换实现车道线识别。实验结果表明,本文算法具有更强的抗噪性能,提升了运算速度,减少运行时间,提高车道线识别的实时性,识别精度高。
When Hough transform is applied to lane recognition,it has some disadvantages,such as high complexity,large amount of computation,long running time and so on,which leads to low real-time performance of lane recognition system.Therefore,this paper proposes an improved Hough transform algorithm based on parameter space partition,which divides the parameter space into uniform accumulation regions.Firstly,the target region is divided,the image is grayed and denoised,and then binaries.Finally,the lane recognition is realized by Hough transform and improved Hough transform respectively.The experimental results show that the algorithm has stronger anti-noise performance,improves the operation speed,reduces the running time,improves the real-time performance of lane recognition,and has high recognition accuracy.
作者
辛敏
罗山
Xin Min;Luo Shan(School of Intelligent Manufacturing, Panzhihua University, Panzhihua, Sichuan 617000, China)
出处
《山西电子技术》
2021年第6期40-42,46,共4页
Shanxi Electronic Technology
基金
攀枝花学院大学生创新创业训练计划项目(S202011360056)。
