摘要
针对背景复杂的大场景遥感影像,本文提出了一种识别机场区域并检测跑道的算法。该算法利用改进最大类间方差法分割机场区域和背景图像,并用最小外接矩算法提取疑似机场区域,再用基于fine-tuning机制的残差网络筛选出非机场区域;对识别为机场的区域用概率Hough变换算法预估机场跑道直线方向,并对该区域边缘图像沿直线方向进行边缘编组,通过对边缘组筛选和组合检测出机场跑道。实验结果表明,该算法能有效判断出大场景遥感影像中是否存在机场,进而检测出机场跑道。
This paper presents an algorithm to identify the airport area and detect the runway for the remote sensing images with complex background.This algorithm is based on the high gray level and integrity characteristics of the airport area.It uses the improved approximated to split the airport area and the background image,and uses the minimum external moment algorithm to get the approximated airport area.Then use the residual network based on the fine-tuning mechanism to filter out the non-airport area;and use the probabilistic Hough transform algorithm to predict the straight direction of the airport runway for the identified airport area.and the edge image of the region is grouped along the linear direction.The airport runway is finally determined by screening and combining the edge group.Experimental results show that the algorithm can effectively determine the existence of an airport in remote sensing image and detect the runway.
作者
杨玉莹
杜鹏
姜麟
Yang Yuying;Du Peng;Jiang Lin(Kunming University of Science and Technology,Kunming 650500,China)
出处
《城市勘测》
2020年第3期59-65,共7页
Urban Geotechnical Investigation & Surveying
关键词
遥感影像
残差网络
机场目标检测
边缘编组
边缘跟踪
remote sensing image
linear target detection
airport target detection
edge grouping
edge tracking