摘要
为解决无人机航拍图像中输电线检测识别率低和误检率高的问题,分析航拍图像输电线的线性特征和区域分布特点,提出一种在复杂背景环境下输电线检测的改进随机霍夫变换方法。使用黑塞(Hessian)矩阵对输电线进行预处理,采用边界搜索和像素条分块对输电线区域和非输电线区域进行分割,降低输电线的误检率和漏检率,增强输电线目标检测识别率。实验结果表明,该方法比标准霍夫(Hough)变换方法和随机霍夫变换方法具有更快的处理速度和更优异的检测效果。
To deal with low recognition rate and high false detection rate for transmission line detection in unmanned aerial vehicle(UAV)aerial images,characteristics of the linear feature and regional distribution for aerial images were analyzed.The method of detecting transmission lines in complex background environment was proposed,an improved random Hough transform(RHT)method for detecting transmission lines in complex background environment was proposed.To reduce the transmission line of the false detection rate and missed rate,the transmission line region and the non-transmission line region were segmented using the boundary search and the pixel strip,and the Hessian matrix was also employed in the image preprocessing.Results of the simulation indicate that the proposed improved RHT method has better efficiency and optimization performance than standard Hough transform(SHT)and RHT.
作者
陈建
沈潇军
姚一杨
刘雄
琚小明
CHEN Jian;SHEN Xiao-jun;YAO Yi-yang;LIU Xiong;JU Xiao-ming(State Grid Zhejiang Electric Power Company Information and Communication Branch,Hangzhou 310007,China;School Computer Science and Software Engineering,East China Normal University,Shanghai 200062,China)
出处
《计算机工程与设计》
北大核心
2018年第4期1155-1160,共6页
Computer Engineering and Design
关键词
输电线检测
航拍图像
随机霍夫变换
黑塞矩阵
区域分割
transmission line detection
aerial image
random Hough transform
Hessian matrix
region segment
