基于核线匹配的高速公路智能识别方法的研究  

Research on Intelligent Recognition Method of Expressway Based on Epipolar Line Matching

作　　者：任靖福[1] 王瀚波[1] 胡晓光[2] REN Jingfu;WANG Hanbo;HU Xiaoguang(Wuxi City College of Vocational Technology,Wuxi 214153,China;Jiangsu Academy of Transportation Sciences Co.,Ltd.,Nanjing 210017,China)

机构地区：[1]无锡城市职业技术学院,江苏无锡214153 [2]江苏省交通科学研究院有限公司,江苏南京210017

出　　处：《现代信息科技》2020年第3期68-72,共5页Modern Information Technology

基　　金：无锡城市职业技术学院课题(WXCY-2017-KZ-03)。

摘　　要：针对车辆在高速运动情况下,各种噪声、干扰、周围环境等影响图像匹配结果的速度、精度以及可靠性的难题,以核线影像数据匹配为研究重点,探讨了关于核线的匹配方法、核线影像数据中特征点提取和匹配的方法,提出了两种基于特征段相似度的核线匹配算法,通过对三组距离不同的高速公路立体影像、不同的同名核线位置下的影像匹配结果的研究,验证了基于相邻特征段相似度的核线匹配算法匹配的可靠性和优越性。In view of the problems of noise,interference and surrounding environment that affect the speed,accuracy and reliability of image matching results when the vehicle is moving at high speed,this paper focuses on the epipolar line image data matching,discusses the epipolar line matching method,the feature point extraction and matching method in the epipolar line image data,and proposes two epipolar line matching algorithms based on the similarity of feature segments.Through the research on the matching results of three groups of expressway stereo images with different distances and different locations of the same name epipolar line,the reliability and superiority of the epipolar line matching algorithm based on the similarity of adjacent feature segments are verified.

关 键 词：图像匹配 高速公路 核线匹配 特征点 特征段 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]