引用本文:
【打印本页】   【HTML】   【下载PDF全文】   查看/发表评论  【EndNote】   【RefMan】   【BibTex】
←前一篇|后一篇→ 过刊浏览    高级检索
本文已被:浏览 230次   下载 72 本文二维码信息
码上扫一扫!
分享到: 微信 更多
一种基于U-Net的高分影像土地利用/覆盖变化检测方法
李聪毅1,2, 孔祥兵2, 杨娜1, 王逸男2, 杨刚凤1,2
1.河南理工大学 测绘与国土信息工程学院, 河南 焦作 454003;2.黄河水利科学研究院 水利部黄土高原水土保持重点实验室, 河南 郑州 450003
摘要:
[目的] 介绍一种基于U-Net的高分影像的土地利用/覆盖变化检测方法,为该模型在遥感影像变化检测方面的应用提供理论支持。[方法] 采用U型神经网络对河南省禹州市两期高分一号影像和WHU building dataset建筑物变化检测数据集中的变化图斑进行自动检测试验,并与FCN和SegNet两种模型进行比较。[结果] 在两个数据集的验证样本中,U型神经网络模型的F1值分别为0.699,0.66和0.673,均优于其他两种模型,并且漏检率较低,更加接近变化参考图。[结论] 采用U型神经网络对高分辨率遥感影像进行土地利用/覆盖变化检测是可行的,且能有较高的检测精度。
关键词:  变化检测  高分辨率遥感  U型神经网络  深度学习
DOI:10.13961/j.cnki.stbctb.2021.04.019
分类号:P237
基金项目:国家自然科学基金项目“基于同质区分析的高光谱影像混合像元稀疏分解研究”(61501200);国家重点研发计划项目(2017YFC0504501);河南省水利科技攻关计划项目(GG201942;GG201829);黄科院研究开发项目(HKY-YFXM-2020-02)
A U-Net Based Land Use/Cover Change Detection Method with High Resolution Image
Li Congyi1,2, Kong Xiangbing2, Yang Na1, Wang Yinan2, Yang Gangfeng1,2
1.School of Surveying and Land Information Engineering, Henan Polytechnic University, Jiaozuo, He'nan 454150, China;2.Key Laboratory of Soil and Water Conservation on the Loess Plateau of Ministry of Water Resources, Yellow River Institute of Hydraulic Research, Zhengzhou, He'nan 450000, China
Abstract:
[Objective] The U-Net based land use/cover change detection method with high resolution image was introduced to provide theoretical support for the application of the model in remote sensing image change detection. [Methods] The U-type neural network was used to detect the change spots in Gaofen-1 image of Yuzhou City, He’nan Province and WHU building data, and compared with FCN and SegNet. [Results] The experimental results showed that the F1 score of U-type neural network model were 0.699,0.66 and 0.673 respectively, which were better than the other two methods, and the missing rate was lower, which was closer to the change reference diagram. [Conclusion] It is feasible to use U-type neural network for change detection in high-resolution remote sensing images, and it could have high detection accuracy.
Key words:  change detection  high resolution remote sensing  U-type neural network  deep learning