摘要
目前无人机电力线巡检已成为热点,但采用人工解译的方法对航拍图像中电力线异物判别效率低下。因此,为实现无人机智能高效巡检,提出一种基于电力线中线的异物检测方法。首先,采用先验知识与最小二乘算法对电力线图像中电力线的中心线准确提取;其次,对获取的中线经过像元的变化设定阈值判断电力线上是否有显著性异物,通过二值图像判断像元偏离电力线中心线距离进而辨别是否有颜色相近的异物;最后,采用RCF边缘检测和区域生长法对提取异物范围。该方法以电力线中线为基础,判断电力线上是否有颜色相似的异物,从而达到复杂背景下异物检测的目的。通过实验证明了该方法能有效地检测出输电线上存在的异物,并且准确度高、轮廓提取完整。
Uav power line inspection has become a hot topic,but it is inefficient to distinguish foreign bodies in aerial images by manual interpretation. In order to realize intelligent and efficient inspection of uav,this paper proposes a foreign substances detection method based on the middle line of power line. Firstly,the center line of power line in power line image was extracted accurately by using prior knowledge and least square algorithm. Then the threshold value was set for the change of the center line passing through pixels to judge whether there were significant foreign substances on the power line,and the distance of pixels deviating from the center line of the power line was determined by binary image to judge whether there were foreign substances with similar colors. Finally,RCF edge detection and region growth method were used to extract the foreign substances range.Based on the middle line of power line,this method can judge whether there are foreign substances with similar color on the power line,so as to achieve the purpose of foreign substances detection in complex background. Experiments showed that this method can effectively detect foreign substances on transmission lines with high accuracy and complete contour extraction.
作者
于国军
邹梓龙
付小
彭佳琪
施陈敬
YU Guojun;ZOU Zilong;FU Xiao;PENG Jiaqi;SHI Chenjing(Faculty of Geomaticn,East Chino University of Technology,330013,Nanchang,PRC;Faculty of Earth Sciencen,East China University of Technology,330013,Nanchang,PRC;Guangdong Guodi Institute of Resource and Evironment,510000,Guangzhou)
出处
《江西科学》
2022年第2期223-228,共6页
Jiangxi Science
基金
国家自然科学基金项目(51708098、52168010)
江西省自然科学基金项目(20212BAB204003)。
