基于改进的DeepLabV3+网络的Sentinel-1影像水体提取

doi:10.13474/j.cnki.11-2246.2025.0311

测绘通报 ›› 2025, Vol. 0 ›› Issue (3): 66-70.doi: 10.13474/j.cnki.11-2246.2025.0311

• 学术研究 • 上一篇

基于改进的DeepLabV3+网络的Sentinel-1影像水体提取

赵兴旺^1,2,3, 赵妍^1,2,3, 刘超^1,2,3, 刘春阳^1,2,3

1. 安徽理工大学空间信息与测绘工程学院, 安徽淮南 232001;
2. 矿山采动灾害空天地协同监测与预警安徽普通高校重点实验室, 安徽淮南 232001;
3. 矿区环境与灾害协同监测煤炭行业工程研究中心, 安徽淮南 232001

收稿日期:2024-06-25 发布日期:2025-04-03
作者简介:赵兴旺(1982—),男,博士,教授,主要从事GNSS卫星遥感、遥感图像处理研究工作。E-mail:xwzhao2008@126.com
基金资助:
安徽省自然科学基金(2208085MD101;2108085QD171)

Water extraction from Sentinel-1 images based on improved DeepLabV3+ network

ZHAO Xingwang^1,2,3, ZHAO Yan^1,2,3, LIU Chao^1,2,3, LIU Chunyang^1,2,3

1. School of Geomatics, Anhui University of Science and Technology, Huainan 232001, China;
2. Key Laboratory of Aviation-Aerospace-Ground Cooperative Monitoring and Early Warning of Coal Mining-induced Disasters of Anhui Higher Education Institutes, Huainan 232001, China;
3. Coal Industry Engineering Research Center of Mining Area Environmental and Disaster Cooperative Monitoring, Huainan 232001, China

Received:2024-06-25 Published:2025-04-03

摘要/Abstract

摘要： 为了提高雷达影像提取水体的精度,本文以2023年Sentinel-1系列影像为数据源,在DeepLabV3+网络模型的基础上优化主干网络,并融合SE通道注意力机制,提出了一种改进的深度学习网络模型SEDeepLabV3+,针对改进的模型进行了消融试验,并以7月31日北京市昌平区水体提取为例,对该模型进行了验证。试验结果表明,使用改进后的SEDeepLabV3+方法提取水体时,平均交并比与像素准确率能够达到88.55%和93.49%,与DeepLabV3+、HRNet、U-Net相比,平均交并比分别提高了2.26%、2.31%和5.08%,平均像素准确率分别提高了0.76%、0.80%和3.07%,改进后的SEDeepLabV3+不仅具有更轻量级的网络结构,而且能够有效地提高水体提取精度和效率。

关键词: DeepLabV3+, 水体提取, SE通道注意力机制, Sentinel-1影像, 语义分割

Abstract: In order to improve the accuracy of water extraction from radar images, this paper uses Sentinel-1 series images from 2023 as the data source, optimizes the backbone network on the basis of the DeepLabV3+ network model, integrates the SE channel attention mechanism, and proposes an improved deep learning network model SEDeepLabV3+. The ablation experiment is carried out for the improved model, and the model is verified by the water body extraction in Changping district of Beijing on July 31. The experimental results show that when the improved SEDeepLabV3+ method is used to extract water body, the mean intersection over union and pixel accuracy can reach 88.55% and 93.49%. Compared with DeepLabV3+, HRNet and U-Net, the average intersection ratio is increased by 2.26%,2.31% and 5.08%,and the average pixel accuracy is increased by 0.76%,0.80% and 3.07%,respectively. The improved SEDeepLabV3+ not only has a lighter network structure,but also can effectively improve the accuracy and efficiency of water extraction.

Key words: DeepLabV3+, water extraction, SE channel attention mechanism, Sentinel-1 images, semantic segmentation

中图分类号:

P237

赵兴旺, 赵妍, 刘超, 刘春阳. 基于改进的DeepLabV3+网络的Sentinel-1影像水体提取[J]. 测绘通报, 2025, 0(3): 66-70.

ZHAO Xingwang, ZHAO Yan, LIU Chao, LIU Chunyang. Water extraction from Sentinel-1 images based on improved DeepLabV3+ network[J]. Bulletin of Surveying and Mapping, 2025, 0(3): 66-70.

参考文献

[1] 苏龙飞,李振轩,高飞,等.遥感影像水体提取研究综述[J].国土资源遥感,2021,33(1): 9-19.
[2] 都金康,黄永胜,冯学智,等.SPOT卫星影像的水体提取方法及分类研究[J].遥感学报,2001,5(3): 214-219.
[3] 刘宇晨,高永年.Sentinel时序影像的长江流域地表水体提取[J].遥感学报,2022,26(2): 358-372.
[4] RANA V K,SURYANARAYANA T M V.Evaluation of SAR speckle filter technique for inundation mapping[J].Remote Sensing Applications: Society and Environment,2019,16: 100271.
[5] 范荣双,陈洋,徐启恒,等.基于深度学习的高分辨率遥感影像建筑物提取方法[J].测绘学报,2019,48(1): 34-41.
[6] WU Peng,FU Junjie,YI Xiaomei,et al.Research on water extraction from high resolution remote sensing images based on deep learning[J].Frontiers in Remote Sensing,2023,4: 1283615.
[7] LÜ Suna,MENG Lingsheng,EDWING D,et al.High-performance segmentation for flood mapping of HISEA-1 SAR remote sensing images[J].Remote Sensing,2022,14(21): 5504.
[8] HARIKA A,SIVANPILLAI R,SAJITH VARIYAR V V,et al.Extracting water bodies in rgb images using DEEPLABV³⁺ algorithm[J].The International Archives of the Photogrammetry,Remote Sensing and Spatial Information Sciences,2022,XLVI-M-2: 97-101.
[9] SU Houcheng,PENG Yuhao,XU Chao,et al.Using improved DeepLabv3+ network integrated with normalized difference water index to extract water bodies in Sentinel-2A urban remote sensing images[J].Journal of Applied Remote Sensing,2021,15(1): 1682312.
[10] LUO Yuanjiang,FENG Ao,LI Hongxiang,et al.New deep learning method for efficient extraction of small water from remote sensing images[J].PLoS One,2022,17(8): e0272317.
[11] 张晗涛,胡荣明,姜友谊,等.改进DeeplabV³⁺模型的河流水体提取[J].遥感信息,2023,38(3): 146-152.
[12] 薛丰昌,吕鑫,陈笑娟.基于改进DeeplabV³⁺的城市内涝监测方法[J].测绘科学,2023,48(10): 216-224.
[13] 张秀再,张昊,杨昌军.基于DeeplabV³⁺网络的轻量化语义分割算法[J].科学技术与工程,2024,24(24): 10382-10393.
[14] 李聪妤,刘家奇,刘欣鑫,等.适应复杂区域的时序SAR影像洪水监测与分析[J].遥感学报,2024,28(2): 346-358.

[1]	李升海, 张俊, 唐海林. 基于改进SVM算法的Sentinel-2A MSI遥感影像水体提取[J]. 测绘通报, 2025, 0(2): 53-57,76.
[2]	李振轩, 黄敏儿, 高飞, 陶庭叶, 吴兆福, 朱勇超. 基于U-Net、U-Net++和Attention-U-Net网络的遥感影像水体提取[J]. 测绘通报, 2024, 0(8): 26-30.
[3]	马姗姗, 杨嘉葳, 王继燕, 熊俊楠. 基于光谱特征的城市水体提取[J]. 测绘通报, 2024, 0(6): 71-76.
[4]	张伟, 张朝龙, 王本林, 蔡安宁. 基于深度学习的多尺度无人机遥感图像道路提取[J]. 测绘通报, 2024, 0(6): 77-81.
[5]	王鸿雁, 车向红, 徐辛超, 徐胜华, 李洪胜. 利用DeepLabv3+模型提取分析街景图像绿视率——以北京三环内为例[J]. 测绘通报, 2024, 0(3): 88-94.
[6]	黄应华, 董振川, 李昊, 陈壮, 刘长睿, 张献州. 城市竣工测绘典型要素语义分割PointNet++深度学习模型适用性分析[J]. 测绘通报, 2024, 0(2): 85-89.
[7]	黄金彩, 李诗逸, 石岩. 利用行车记录仪视频提取路面车道线[J]. 测绘通报, 2024, 0(12): 1-5.
[8]	李锋, 薛梅, 詹勇, 杨元. 基于分割一切模型SAM的实景三维场景语义分割[J]. 测绘通报, 2024, 0(12): 101-105.
[9]	张宝文, 赵展. 利用多地物指数自动提取水体方法[J]. 测绘通报, 2024, 0(11): 21-26.
[10]	王星, 倪欢. 遥感图像农田识别的跨类别小样本分割方法[J]. 测绘通报, 2024, 0(10): 77-83.
[11]	孙常建, 尚永福, 王石岩, 窦小楠. 基于语义分割网络的冬小麦遥感分类及变化分析[J]. 测绘通报, 2024, 0(10): 151-156.
[12]	邹鑫, 王磊, 李靖宇, 滕超群, 黄金中, 李忠, 李世保. 融合Deep-ResUnet和PS-InSAR的城市道路网形变灾害监测方法——以合肥市为例[J]. 测绘通报, 2023, 0(8): 67-71.
[13]	马海超, 王崇倡, 何昭宁, 信晟, 王蕾. 面向GF-7影像的绿波改正归一化水体指数构建[J]. 测绘通报, 2023, 0(5): 38-43.
[14]	姚萌萌, 李晓明, 王伟玺, 谢林甫, 黄俊杰, 黄鸿盛, 汤圣君. 几何特征与神经网络联合优化的室内三维点云语义分割方法[J]. 测绘通报, 2023, 0(5): 56-61.
[15]	蔡军, 温日强, 江伟, 严娇, 卢丽娟. 传统村落航拍图像中民族建筑识别的DeepLabV3+改进算法[J]. 测绘通报, 2023, 0(4): 49-53.

基于改进的DeepLabV3+网络的Sentinel-1影像水体提取

Water extraction from Sentinel-1 images based on improved DeepLabV3+ network

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价