Night curve recognition algorithm based on K-means clustering and improved Hough transform

2019 3rd International Workshop on Renewable Energy and Development

能源学;生态环境科学

Lin, Zhiqian^1 ; Yongze, Zhang^2

Wuhan University of Technology, Wuhan, Hubei

430070, China^1

University of Electronic Science and Technology of China, Chengdu, Sichuan

611700, China^2

关键词: Line recognition;    Local optimal;    Real time performance;    Recognition algorithm;    The region of interest (ROI);   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/267/4/042074/pdf
DOI  :  10.1088/1755-1315/267/4/042074

学科分类：环境科学（综合）

来源: IOP

【 摘 要 】

In order to solve the problem of poor robustness and real-time performance of existing lane line recognition algorithms in night curve recognition, a fast night curve recognition algorithm based on K-means clustering and improved Hough transform is proposed and implemented. The algorithm firstly intercepts the lower two thirds of the image as the region of interest ( ROI ), then binarizes the ROI image by using the block local optimal threshold method, and then obtains the starting point of the lane line by using the improved Hough transform. Finally, based on the starting point, the lane line in the image is calibrated by using the K-means clustering algorithm. The experimental results show that the method can detect the lane lines of roads turning at night, and is suitable for other roads at day and night, with good real-time performance and robustness.