摘要
为了从复杂无人机航拍图像中快速有效地检测与提取出电力线,基于电力线特征,提出一种复杂地物背景下的电力线识别算法。采用改进的Ratio算子并结合基于轮廓特征的背景去噪算法完成电力线的边缘检测,利用基于Hough变换的直线动态编组拟合与筛选算法进行边缘提取,最终识别出完整的电力线。实验结果表明,所提算法可有效排除噪声干扰,在多种航拍场景下的识别准确率达96.37%以上,识别精度与时效性比常见的同类算法均有明显的提升,具有较好的电力巡线应用价值。
In order to detect and extract power lines from complex unmanned aerial vehicle(UAV)aerial images quickly and effectively,in this paper,based on the characteristics of power lines,a power line recognition algorithm in background of complex ground objects is proposed.Firstly,the edge of the power line is detected by an improved Ratio operator combined with a background-denoising algorithm based on contour features,then an algorithm of line dynamic grouping-fitting and screening based on Hough transform is used for edge extraction,and finally the power lines are completely identified.The experimental results show that the proposed algorithm can effectively eliminate noise interference,and that its recognition accuracy is more than 96.37%in a variety of aerial scenes.Its recognition rate and timeliness are significantly improved in comparison with the common similar algorithms,which makes it has higher application value for the power line inspection.
作者
骆顾平
陈静
苏成悦
刘信宏
徐胜
王泰圣
LUO Guping;CHEN Jing;SU Chengyue;LIU Xinhong;XU Sheng;WANG Taisheng(School of Physics and Optoelectronic Engineering,Guangdong University of Technology,Guangzhou 510006,China)
出处
《应用科技》
CAS
2021年第5期23-28,34,共7页
Applied Science and Technology
基金
广东省科技计划项目(2017A020208063)
广州市科技计划项目(201804010384)。
关键词
电力线识别
复杂地物
边缘检测与提取
Ratio算子
轮廓特征
HOUGH变换
动态编组拟合
直线筛选
power line recognition
complex ground objects
edge detection and extraction
Ratio operator
contour features
Hough transform
dynamic grouping-fitting
lines screening
