基于改进Hough变换的非结构化道路检测  

Unstructured Road Detection Based on Improved Hough Transform

作　　者：黄闽江 刘康 陈勇 徐则中[1] HUANG Minjiang;LIU Kang;CHEN Yong;XU Zezhong(School of Computer Information Engineering,Changzhou Institute of Technology,Changzhou 213032)

机构地区：[1]常州工学院计算机信息工程学院,江苏常州213032

出　　处：《常州工学院学报》2022年第6期19-22,54,共5页Journal of Changzhou Institute of Technology

基　　金：常州工学院大学生创新创业训练计划项目(202111055010Z)。

摘　　要：针对非结构化路面存在着阴影和水迹等干扰情况,提出了一种鲁棒的非结构化道路检测方法。把道路RGB图像转换成HSV图像,对S分量进行逆透视变换,在物理空间检测路面。运用改进的Hough变换,通过函数拟合直接得到线性矩形框,表示路面可行驶区域。该方法消除了透视成像的影响,避免了边缘检测过程,提高了算法的抗干扰能力。Considering the interference of shadow and water stains on unstructured roads,a robust unstructured road detection method is proposed.The RGB image of the roadis converted into HSV image,and inverse perspective transformation is conducted on the S component to detect the road surface in physical space.By using the improved Hough transform,a linear rectangular box is directly obtained through function fitting,which represents the drivable area of the road surface.This method eliminates the influence of perspective imaging and avoids the edge detection process,thus improving the anti-interference ability of the algorithm.

关 键 词：非结构化道路 逆透视变换 HOUGH变换 道路检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]