2025年04月05日 星期六

广西师范大学学报（自然科学版） ›› 2025, Vol. 43 ›› Issue (2): 121-132.doi: 10.16088/j.issn.1001-6600.2024032002

• 智能信息处理 • 上一篇    下一篇

基于层级尺度交互的U-Net遥感影像建筑物提取方法

余快1,2, 宋宝贵1,2, 邵攀1,2*, 余翱1,2   

  1. 1.水电工程智能视觉监测湖北省重点实验室(三峡大学), 湖北 宜昌 443002;
    2.三峡大学 计算机与信息学院, 湖北 宜昌 443002
  • 收稿日期:2024-03-20 修回日期:2024-06-20 出版日期:2025-03-05 发布日期:2025-04-02
  • 通讯作者: 邵攀(1985—), 男, 河南滑县人, 三峡大学副教授, 博士。E-mail: panshao@whu.edu.cn
  • 基金资助:
    国家自然科学基金青年项目(41901341); 湖北省自然科学基金(2024AFB867)

Hierarchical-scale Interaction-based U-Net for Remote Sensing Image Building Extraction

YU Kuai1,2, SONG Baogui1,2, SHAO Pan1,2*, YU Ao1,2   

  1. 1. Hubei Key Laboratory of Intelligent Vision Based Monitoring for Hydroelectric Engineering (China Three Gorges University), Yichang Hubei 443002, China;
    2. College of Computer and Information, China Three Gorges University, Yichang Hubei 443002, China
  • Received:2024-03-20 Revised:2024-06-20 Online:2025-03-05 Published:2025-04-02

PDF (PC)

0

摘要： 针对U-Net及其改进网络在跳跃链接中因忽略多层级特征间相互作用而导致对特征的表征能力不足问题,本文提出一种基于层级尺度交互的U-Net遥感影像建筑物提取方法。首先,在U-Net网络跳跃连接中设计层级尺度交互模块,实现多层级特征的交互增强,提升对特征的表征能力;然后,通过改进空洞空间金字塔池化模块,提出一种多尺度特征提取模块,并将其应用到最高层级特征,来提升网络提取多尺度特征的能力;最后,将自校准卷积引入到解码过程,促进浅层与深层特征更好地融合。在公开建筑物提取数据集WHU和Inria上,将本文方法与6种遥感影像建筑物提取方法进行对比,实验结果表明,本文方法的IoU分别为91.26%和79.23%,均优于对比方法。

关键词: 遥感影像, 建筑物提取, U-Net, 层级尺度交互, 多尺度, 注意力机制

Abstract: Aiming at the problem that U-Net and its improved network have insufficient feature characterisation ability due to ignoring the interactions between multi-level features in jump links, a building extraction method based on hierarchical scale interactions is proposed for U-Net remote sensing images. Firstly, a hierarchical scale interaction module is designed in the jump link of U-Net network to achieve the interaction enhancement of multilevel features and improve the characterisation ability of features. Then a multi-scale feature extraction module is proposed by improving the null-space pyramid pooling module and applying it to the highest level features to enhance the ability of the network to extract multi-scale features. Finally, self-calibrating convolution is introduced into the decoding process to promote better fusion of shallow and deep features. The method of this paper is compared with six remote sensing image building extraction methods on two publicly available building extraction datasets, WHU and Inria. The experimental results show that the IoU of the proposed method is 91.26% and 79.23%, respectively, which are better than the comparison methods.

Key words: remote sensing imagery, building extraction, U-Net, hierarchical-scale interaction, multi-scale, attention mechanism

中图分类号:  TP751

[1] 杨俊, 任嘉义, 于文博. 局地气候分区视角下城市气候与人居环境研究进展[J]. 生态学报, 2024, 44(11): 4489-4506. DOI: 10.20103/j.stxb.202310222297.
[2] 慈慧, 郭朋辉, 秦勇, 等. 基于多时相遥感影像的滨海湿地监测方法研究[J]. 广西师范大学学报(自然科学版), 2013, 31(3): 144-151. DOI: 10.16088/j.issn.1001-6600.2013.03.025.
[3] LIASIS G, STAVROU S. Building extraction in satellite images using active contours and colour features[J]. International Journal of Remote Sensing, 2016, 37(5): 1127-1153. DOI: 10.1080/01431161.2016.1148283.
[4] 施仲添, 沈正伟, 杨四海. 基于前景感知的遥感影像建筑物提取方法[J]. 测绘通报, 2023(2): 134-138. DOI: 10.13474/j.cnki.11-2246.2023.0053.
[5] 王明常, 朱春宇, 陈学业, 等. 基于FPN Res-Unet的高分辨率遥感影像建筑物变化检测[J]. 吉林大学学报(地球科学版), 2021, 51(1): 296-306. DOI: 10.13278/j.cnki.jjuese.20190321.
[6] 王俊, 秦其明, 叶昕, 等. 高分辨率光学遥感图像建筑物提取研究进展[J]. 遥感技术与应用, 2016, 31(4): 653-662. DOI: 10.11873/j.issn.1004-0323.2016.4.0653.
[7] LI Q Y, MOU L C, SUN Y, et al. A review of building extraction from remote sensing imagery: geometrical structures and semantic attributes[J]. IEEE Transactions on Geoscience and Remote Sensing, 2024, 62: 4702315. DOI: 10.1109/TGRS.2024.3369723.
[8] FANG F, XU R, LI S W, et al. Semisupervised building instance extraction from high-resolution remote sensing imagery[J]. IEEE Transactions on Geoscience and Remote Sensing, 2023, 61: 5619212. DOI: 10.1109/TGRS.2023.3309918.
[9] CHEN J, HE P E, ZHU J R, et al. Memory-contrastive unsupervised domain adaptation for building extraction of high-resolution remote sensing imagery[J]. IEEE Transactions on Geoscience and Remote Sensing, 2023, 61: 5605615. DOI: 10.1109/TGRS.2023.3262576.
[10] 宋冠武, 陈知明, 李建军. 基于ResNet-50的级联注意力遥感图像分类[J]. 广西师范大学学报(自然科学版), 2023, 41(6): 80-91. DOI: 10.16088/j.issn.1001-6600.2023031702.
[11] JIN X Y, DAVIS C H. Automated building extraction from high-resolution satellite imagery in urban areas using structural, contextual, and spectral information[J]. EURASIP Journal on Applied Signal Processing, 2005, 2005: 2196-2206. DOI: 10.1155/ASP.2005.2196.
[12] 张浩, 赵云胜, 陈冠宇, 等. 基于支持向量机的遥感图像建筑物识别与分类方法研究[J]. 地质科技情报, 2016, 35(6): 194-199.
[13] 马为駽. 基于形态学的高分辨率遥感影像建筑物自动提取[D]. 武汉: 武汉大学, 2022. DOI: 10.27379/d.cnki. gwhdu.2019.001850.
[14] 顾炼, 许诗起, 竺乐庆. 基于FlowS-Unet的遥感图像建筑物变化检测[J]. 自动化学报, 2020, 46(6): 1291-1300. DOI: 10.16383/j.aas.c180122.
[15] 唐晴, 徐胜华, 高贤君, 等. 融合矢量数据的高分遥感影像建筑物轮廓优化方法[J]. 测绘科学, 2023, 48(12): 143-152. DOI: 10.16251/j.cnki.1009-2307.2023.12.015.
[16] LONG J, SHELHAMER E, DARRELL T. Fully convolutional networks for semantic segmentation[C] // 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2015: 3431-3440. DOI: 10.1109/CVPR.2015.7298965.
[17] RONNEBERGER O, FISCHER P, BROX T. U-Net:convolutional networks for biomedical image segmentation[C] // Medical Image Computing and Computer-Assisted Intervention: MICCAI 2015. Cham: Springer International Publishing AG Switzerland, 2015: 234-241. DOI: 10.1007/978-3-319-24574-4_28.
[18] BADRINARAYANAN V, KENDALL A, CIPOLLA R.SegNet: a deep convolutional encoder-decoder architecture for image segmentation[J]. IEEE transactions on pattern analysis and machine intelligence, 2017, 39(12): 2481-2495. DOI: 10.1109/TPAMI.2016.2644615.
[19] CHEN L C, PAPANDREOU G, KOKKINOS I, et al.DeepLab: semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected CRFs[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018, 40(4): 834-848. DOI: 10.1109/TPAMI.2017.2699184.
[20] RAN S H, GAO X J, YANG Y W, et al. Building multi-feature fusion refined network for building extraction from high-resolution remote sensing images[J]. Remote Sensing, 2021, 13(14): 2794. DOI: 10.3390/rs13142794.
[21] 王一琛, 刘慧, 王海涛, 等. 面向遥感图像的建筑物轻量化语义分割方法[J]. 计算机工程与设计, 2022, 43(9): 2646-2653. DOI: 10.16208/j.issn1000-7024.2022.09.032.
[22] 林禹, 赵泉华, 沈昭宇, 等. 改进SegNet与迁移学习的遥感建筑物分割方法[J]. 测绘科学, 2022, 47(6): 78-89. DOI: 10.16251/j.cnki.1009-2307.2022.06.011.
[23] 王振庆, 周艺, 王世新, 等. IEU-Net高分辨率遥感影像房屋建筑物提取[J]. 遥感学报, 2021, 25(11): 2245-2254. DOI: 10.11834/jrs.20210042.
[24] KANG J, FERNANDEZ-BELTRAN R, SUN X, et al. Deep learning-based building footprint extraction with missing annotations[J]. IEEE Geoscience and Remote Sensing Letters, 2022, 19: 3002805. DOI: 10.1109/LGRS.2021.3072589.
[25] DIXIT M, CHAURASIA K, MISHRA K V. Dilated-ResUnet: a novel deep learning architecture for building extraction from medium resolution multi-spectral satellite imagery[J]. Expert Systems with Applications, 2021, 184: 115530. DOI: 10.1016/j.eswa.2021.115530.
[26] QIU W Y, GU L J, GAO F, et al. Building extraction from very high-resolution remote sensing images using refine-unet[J]. IEEE Geoscience and Remote Sensing Letters, 2023, 20: 6002905. DOI: 10.1109/LGRS.2023.3243609.
[27] CAI J H, CHEN Y M. MHA-Net: multipath hybrid attention network for building footprint extraction from high-resolution remote sensing imagery[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2021,14: 5807-5817. DOI: 10.1109/JSTARS.2021.3084805.
[28] WANG Z Q, ZHOU Y, WANG F T, et al. Fine-grained building extraction with multispectral remote sensing imagery using the deep model[J]. IEEE Transactions on Geoscience and Remote Sensing, 2023, 61: 4706013. DOI: 10.1109/TGRS.2023.3327370.
[29] GUO H N, SU X, WU C, et al. Decoupling semantic and edge representations for building footprint extraction from remote sensing images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2023, 61: 5613116. DOI: 10.1109/TGRS.2023.3287298.
[30] LIU J J, HOU Q B, CHENG M M, et al. Improving convolutional networks with self-calibrated convolutions[C] // 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2020: 10093-10102. DOI: 10.1109/CVPR42600.2020.01011.
[31] LI X T, YOU A S, ZHU Z, et al. Semantic flow for fast and accurate scene parsing[C] // Computer Vision-ECCV 2020: LNCS Volume 12346. Cham: Springer Nature Switzerland AG, 2020: 775-793. DOI: 10.1007/978-3-030-58452-8_45.
[32] ZHAO H S, SHI J P, QI X J, et al. Pyramid scene parsing network[C] // 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2017: 6230-6239. DOI: 10.1109/CVPR.2017. 660.
[33] 朱节中, 陈永, 柯福阳, 等. 基于Siam-UNet++的高分辨率遥感影像建筑物变化检测[J]. 计算机应用研究, 2021, 38(11): 3460-3465. DOI: 10.19734/j.issn.1001-3695.2021.01.0070.
[34] JI S P, WEI S Q, LU M. Fully convolutional networks for multisource building extraction from an open aerial and satellite imagery data set[J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(1): 574-586. DOI: 10.1109/TGRS.2018.2858817.
[35] MAGGIORI E, TARABALKA Y, CHARPIAT G, et al. Can semantic labeling methods generalize to any city? the inria aerial image labeling benchmark[C] // 2017 IEEE International Geoscience and Remote Sensing Symposium (IGARSS). Piscataway, NJ: IEEE, 2017: 3226-3229. DOI: 10.1109/IGARSS.2017.8127684.
[36] PASZKE A, CHAURASIA A, KIM S, et al. ENet: a deep neural network architecture for realtime semantic segmentation[EB/OL]. (2016-06-07)[2024-03-20]. https://arxiv.org/abs/1606.02147. DOI: 10.48550/arXiv.1606.02147.
[37] ZHANG H, ZHENG X C, ZHENG N S, et al. A multiscale and multipath network with boundary enhancement for building footprint extraction from remotely sensed imagery[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2022,15: 8856-8869. DOI: 10.1109/JSTARS.2022.3214485.
[38] CHEN J, JIANG Y X, LUO L B, et al. ASF-Net: adaptive screening feature network for building footprint extraction from remote-sensing images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60: 4706413. DOI: 10.1109/TGRS.2022.3165204.
[1] 郭翔羽, 石天怡, 陈燕楠, 南新元, 蔡鑫. 基于YOLO-CDBW模型的列车接触网异物检测研究[J]. 广西师范大学学报（自然科学版）, 2025, 43(2): 56-69.
[2] 苏春海, 夏海英. 抗噪声双约束网络的面部表情识别[J]. 广西师范大学学报（自然科学版）, 2025, 43(2): 70-82.
[3] 刘玉娜, 马双宝. 基于改进YOLOv8n的轻量化织物疵点检测算法[J]. 广西师范大学学报（自然科学版）, 2025, 43(2): 83-94.
[4] 戴林华, 黎远松, 石睿, 何忠良, 李雷. HSED-YOLO:一种轻量化的带钢表面缺陷检测模型[J]. 广西师范大学学报（自然科学版）, 2025, 43(2): 95-106.
[5] 卢家辉, 陈庆锋, 王文广, 余谦, 何乃旭, 韩宗钊. 基于多尺度注意力的器官图像分割方法[J]. 广西师范大学学报（自然科学版）, 2024, 42(6): 138-148.
[6] 杜帅文, 靳婷. 基于用户行为特征的深度混合推荐算法[J]. 广西师范大学学报（自然科学版）, 2024, 42(5): 91-100.
[7] 田晟, 胡啸. 基于Transformer模型的车辆轨迹预测[J]. 广西师范大学学报（自然科学版）, 2024, 42(3): 47-58.
[8] 王天雨, 袁嘉伟, 齐芮, 李洋. 多类型知识增强的微博立场检测模型[J]. 广西师范大学学报（自然科学版）, 2024, 42(1): 79-90.
[9] 肖宇庭, 吕晓琪, 谷宇, 刘传强. 基于拆分残差网络的糖尿病视网膜病变分类[J]. 广西师范大学学报（自然科学版）, 2024, 42(1): 91-101.
[10] 席凌飞, 伊力哈木·亚尔买买提, 刘雅洁. 基于改进YOLOv5的铝型材表面缺陷检测方法[J]. 广西师范大学学报（自然科学版）, 2024, 42(1): 111-119.
[11] 宋冠武, 陈知明, 李建军. 基于ResNet-50的级联注意力遥感图像分类[J]. 广西师范大学学报（自然科学版）, 2023, 41(6): 80-91.
[12] 郭嘉梁, 靳婷. 基于语义增强的多模态情感分析[J]. 广西师范大学学报（自然科学版）, 2023, 41(5): 14-25.
[13] 吴正清, 曹晖, 刘宝锴. 基于注意力卷积神经网络的中文虚假评论检测[J]. 广西师范大学学报（自然科学版）, 2023, 41(5): 26-36.
[14] 唐侯清, 辛斌斌, 朱虹谕, 乙加伟, 张冬冬, 武新章, 双丰. 基于多尺度注意力倒残差网络的轴承故障诊断[J]. 广西师范大学学报（自然科学版）, 2023, 41(4): 109-122.
[15] 黄叶祺, 王明伟, 闫瑞, 雷涛. 基于改进的YOLOv5金刚石线表面质量检测[J]. 广西师范大学学报（自然科学版）, 2023, 41(4): 123-134.
Viewed
Full text
0
HTML PDF
Just accepted Online first Issue Just accepted Online first Issue
0 0 0 0 0 0


Abstract
2
Just accepted Online first Issue
0 0 2
  From Others
  Times 2
  Rate 100%

Cited
Web of Science  Crossref   ScienceDirect  Search for Citations in Google Scholar >>
 
This page requires you have already subscribed to WoS.
  Shared   
  Discussed   
No Suggested Reading articles found!
版权所有 © 广西师范大学学报（自然科学版）编辑部
地址：广西桂林市三里店育才路15号　邮编：541004
电话：0773-5857325　E-mail: gxsdzkb@mailbox.gxnu.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发