LA-UNet网络模型在城市绿地遥感分类中的应用

doi:10.13287/j.1001-9332.202404.025

应用生态学报 ›› 2024, Vol. 35 ›› Issue (4): 1101-1111.doi: 10.13287/j.1001-9332.202404.025

LA-UNet网络模型在城市绿地遥感分类中的应用

徐亮亮^1,2,3, 马开森⁴, 王霞⁵, 李东胜⁵, 孙华^1,2,3*

¹中南林业科技大学林业遥感信息工程研究中心, 长沙 410004;
²林业遥感大数据与生态安全湖南省重点实验室, 长沙 410004;
³南方森林资源经营与监测国家林业与草原局重点实验室, 长沙 410004;
⁴湖南科技大学地理空间信息技术国家地方联合工程实验室, 湖南湘潭 411201;
⁵河北省林业和草原调查规划设计院, 石家庄 050011

收稿日期:2023-11-08 接受日期:2024-02-23 出版日期:2024-04-18 发布日期:2024-10-18
通讯作者: * E-mail: sunhua@csuft.edu.cn
作者简介:徐亮亮, 男, 1998年生, 硕士研究生。主要从事深度学习遥感分类研究。E-mail: 1104574017@qq.com
基金资助:
“十四五”国家重点研发计划项目(2023YFD2201703)和湖南省科技创新计划项目(2023RC1065)

Application of LA-UNet network model in remote sensing classification of urban green space

XU Liang-liang^1,2,3, MA Kaisen⁴, WANG Xia⁵, LI Dongsheng⁵, SUN Hua^1,2,3*

¹Research Center of Forestry Remote Sensing & Information Engineering, Central South University of Forestry and Technology, Changsha 410004, China;
²Key Laboratory of Forestry Remote Sensing Based Big Data & Ecological Security for Hunan Province, Changsha 410004, China;
³Key Laboratory of State Forestry & Grassland Administration on Forest Resources Management and Monitoring in Southern Area, Changsha 410004, China;
⁴National-Local Joint Engineering Laboratory of Geo-Spatial Information Technology, Hunan University of Science and Technology, Xiangtan 411201, Hunan, China;
⁵Hebei Forestry & Grassland Survey, Planning and Design Institute, Shijiazhuang 050011, China

Received:2023-11-08 Accepted:2024-02-23 Online:2024-04-18 Published:2024-10-18

摘要/Abstract

摘要： 城市绿地的精准识别和监测在城市规划和生态管理方面具有重要意义。城市绿地背景复杂,使用传统遥感分类技术容易出现错分粘连的问题。本研究以长沙市雨花区为研究区,高分二号(GF-2)遥感影像为数据源,提出一种基于LA-UNet模型的城市绿地遥感分类方法,该方法以UNet模型为基础,引入DWTCA通道注意力机制模块提升网络对绿地信息的关注度,并使用CARAFE模块对提取特征进行上采样,实现对城市复杂背景下乔木、灌草等多种地类的精准分类。结果表明: 利用标准假彩色遥感影像时,LA-UNet模型的城市绿地分类效果最优,总体准确率和平均交并比分别为96.3%和90.9%,比UNet模型分别提高2.8%和6.1%。在波茨坦公开数据集中,LA-UNet模型的总体准确率和平均交并比也同样优于UNet模型,分别提高0.9%和1.8%,表明LA-UNet模型具有较好的鲁棒性和通用性。本研究提出的LA-UNet模型能有效缓解城市绿地错分粘连问题,在城市绿地遥感分类中具有显著优势。改进后的LA-UNet模型比UNet模型具有更小的参数体积,能有效提升城市绿地分类精度。本研究将为城市绿地的精准分类及其空间分布规律研究提供方法参考。

关键词: 遥感分类, 城市绿地, 高分二号, LA-UNet, 注意力机制

Abstract: The accurate identification and monitoring of urban green space is of great significance in urban planning and ecological management. In view of the complex background of urban green space, the traditional remote sensing classification technology is prone to the problem of misalignment and adhesion. Taking Yuhua District of Changsha City as the research area and Gaofen-2 (GF-2) remote sensing image as the data source, we proposed a remote sensing classification method for urban green space based on the LA-UNet model, which was based on the UNet model. We introduced the DWTCA channel attention mechanism module to improve the attention of the network to green space information, and used the CARAFE module to up sample the extracted features to achieve accurate classification of trees, shrubs and other land types in the complex background of the city. The results showed that the LA-UNet model had the best classification effect of urban green space when using standard false color remote sensing images. The overall accuracy and mean intersection over union were 96.3% and 90.9%, which were 2.8% and 6.1% higher than the UNet model, respectively. In the Potsdam public dataset, the overall accuracy and mean intersection over union of the LA-UNet model were also better than those of the UNet model, which increased by 0.9% and 1.8%, respectively, indicating that the LA-UNet model had good robustness and versatility. In summary, the proposed LA-UNet model could effectively alleviate the problems of misalignment and adhesion of urban green space, with advantages in the remote sensing classification of urban green space. The improved LA-UNet model had a smaller parameter volume than the UNet model, which could effectively improve the classification accuracy of urban green space. This study would provide a methodological reference for the accurate classification and understanding the spatial distribution of urban green space.

Key words: remote sensing classification, urban green space, GF-2, LA-UNet, attention mechanism

徐亮亮, 马开森, 王霞, 李东胜, 孙华. LA-UNet网络模型在城市绿地遥感分类中的应用[J]. 应用生态学报, 2024, 35(4): 1101-1111.

XU Liang-liang, MA Kaisen, WANG Xia, LI Dongsheng, SUN Hua. Application of LA-UNet network model in remote sensing classification of urban green space[J]. Chinese Journal of Applied Ecology, 2024, 35(4): 1101-1111.

参考文献

[1] 闫明, 陈利顶, 孙然好. 城市生态品质的内涵及其核心指标体系构建. 应用生态学报, 2023, 34(6): 1459-1466
[2] Neyns R, Canters F. Mapping of urban vegetation with high-resolution remote sensing: A review. Remote Sen-sing, 2022, 14: 1031
[3] Shahtahmassebi AR, Li C, Fan Y, et al. Remote sen-sing of urban green spaces: A review. Urban Forestry & Urban Greening, 2021, 57: 126946
[4] Qian Y, Zhou W, Nytch CJ, et al. A new index to differentiate tree and grass based on high resolution image and object-based methods. Urban Forestry & Urban Greening, 2020, 53: 126661
[5] Men G, He G, Wang G. Concatenated residual attention UNet for semantic segmentation of urban green space. Forests, 2021, 12: 1441
[6] 辛动军, 袁梦, 陈建安, 等. 基于FDCT与ELM的遥感影像湿地类型分类: 以黄家湖国家湿地公园为例. 中南林业科技大学学报, 2018, 38(6): 30-35
[7] 赵庆展, 江萍, 王学文, 等. 基于无人机高光谱遥感影像的防护林树种分类. 农业机械学报, 2021, 52(11): 190-199
[8] 李哲, 张沁雨, 邱新彩, 等. 基于高分二号遥感影像树种分类的时相及方法选择. 应用生态学报, 2019, 30(12): 4059-4070
[9] 王佳玥, 蔡志文, 王文静, 等. 协同多源国产高分影像和面向对象方法的南方农作物遥感识别. 中国农业科学, 2023, 56(13): 2474-2490
[10] 薛明, 韦波, 李景文, 等. 优化BP神经网络的多特征融合遥感影像分类方法. 测绘科学, 2021, 46(11): 47-55
[11] 冯权泷, 牛博文, 朱德海, 等. 土地利用/覆被深度学习遥感分类研究综述. 农业机械学报, 2022, 53(3): 1-17
[12] Han W, Zhang XH, Wang Y, et al. A survey of machine learning and deep learning in remote sensing of geological environment: Challenges, advances, and opportunities. ISPRS Journal of Photogrammetry and Remote Sensing, 2023, 202: 87-113
[13] Xu ZY, Zhou Y, Wang SX, et al. A novel intelligent classification method for urban green space based on high-resolution remote sensing images. Remote Sensing, 2020, 12: 3845
[14] Chen SQ, Chen M, Zhao BY, et al. Urban tree canopy mapping based on double-branch convolutional neural network and multi-temporal high spatial resolution satellite imagery. Remote Sensing, 2023, 15: 765
[15] 陈婕, 刘纪平, 徐胜华. 增强边缘信息的全卷积神经网络遥感影像建筑物变化检测. 测绘通报, 2023(6): 61-67
[16] Zunair H, Hamza AB. Sharp U-Net: Depthwise convolutional network for biomedical image segmentation. Computers in Biology and Medicine, 2021, 136: 104699
[17] 王晶, 高帅, 郭亮, 等. 基于多尺度特征融合的U-Net网络高分影像不透水面提取研究. 遥感技术与应用, 2022, 37(4): 811-819
[18] Zhang XL, Wang ZS, Zhang JH, et al. MSANet: An improved semantic segmentation method using multi-scale attention for remote sensing images. Remote Sen-sing Letters, 2022, 13: 1249-1259
[19] Wang ZG, Wang QN, Zhang YJ, et al. Enhanced seismic data segmentation using an assembled scSE-Res-UNet deep neural network. Geoenergy Science and Engineering, 2023, 231: 212347
[20] Wang JQ, Chen K, Xu R, et al. Carafe: Content-aware reassembly of features. Proceedings of the IEEE/CVF International Conference on Computer Vision, Seoul, Korea, 2019: 3007-3016
[21] Cao YY, Zhao ZX, Huang Y, et al. Case instance segmentation of small farmland based on Mask R-CNN of feature pyramid network with double attention mechanism in high resolution satellite images. Computers and Electronics in Agriculture, 2023, 212: 108073
[22] Bartoldson B, Morcos A, Barbu A, et al. The generalization-stability tradeoff in neural network pruning. Advances in Neural Information Processing Systems, 2020, 33: 20852-20864
[23] LeCun Y, Boser B, Denker JS, et al. Back propagation applied to handwritten zip code recognition. Neural Computation, 1989, 1: 541-551
[24] 中华人民共和国国家质量监督检验检疫总局. GB/T 21010—2017: 土地利用现状分类. 北京: 中国标准出版社, 2017
[25] Yang QR, Ku T, Hu KY. Efficient attention pyramid network for semantic segmentation. IEEE Access, 2021, 9: 18867-18875
[26] 张亚茹, 孔雅婷, 刘彬. 多维注意力特征聚合立体匹配算法. 自动化学报, 2022, 48(7): 1805-1815
[27] 李龙, 张重阳. 基于改进YOLOv3的车辆尾灯检测方法. 计算机与现代化, 2021(7): 89-94
[28] 孙志同, 朱珊娜, 高郑杰, 等. 基于波段增强的DeepLabv3+多光谱影像葡萄种植区识别. 农业工程学报, 2022, 38(7): 229-236
[29] 徐知宇, 周艺, 王世新, 等. 面向GF-2遥感影像的U-Net城市绿地分类. 中国图象图形学报, 2021, 26(3): 700-713
[30] 何直蒙, 丁海勇, 安炳琪. 高分辨率遥感影像建筑物提取的空洞卷积E-Unet算法. 测绘学报, 2022, 51(3): 457-467
[31] 申传庆, 王凯, 王文杰. 基于ResNet-UNet的地表覆盖自动分类技术研究. 地理空间信息, 2023, 21(6): 21-23
[32] 袁德宝, 王子林, 李雪莹, 等. 基于改进U-Net的遥感影像城镇绿地提取. 遥感信息, 2023, 38(1): 33-39
[33] Zhang XY, Zhou XY, Lin MX, et al. Shufflenet: An extremely efficient convolutional neural network for mobile devices. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, New York, United States, 2018: 6848-6856
[34] Chen YP, Kalantidis Y, Li JS, et al. A²-nets: Double attention networks. 32^nd Conference on Neural Information Processing Systems (NeurIPS 2018), Montréal, Canada, 2018, http://arxiv.org/pdf/1810.11579.pdf
[35] He K, Zhang X, Ren S, et al. Deep residual learning for image recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, New York, United States, 2016: 770-778

LA-UNet网络模型在城市绿地遥感分类中的应用

Application of LA-UNet network model in remote sensing classification of urban green space

PDF (PC)

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

[1]	邱瑶, 罗涛, 王琼, 蒋思雨. 夏热冬暖地区绿量构成对城市热环境的影响——以福州居住区为例 [J]. 应用生态学报, 2023, 34(7): 1932-1940.
[2]	唐雨倩, 吴晓奕, 宗桦. 成都青羊区背街小巷行道树孢粉致敏风险分析 [J]. 应用生态学报, 2022, 33(6): 1615-1621.
[3]	李海防, 俞洁蕾, 邵西宁, 周春玲. 半湿润地区城市绿地灌木的截留集水功能及其影响因素 [J]. 应用生态学报, 2022, 33(5): 1363-1369.
[4]	李肖肖, 唐丽玉, 彭巍, 陈建新, 麻霞. 基于背包式激光雷达测量系统的城市绿地树木三维绿量估算方法 [J]. 应用生态学报, 2022, 33(10): 2777-2784.
[5]	韩继刚, 李刚, 张维维, 刘文, 刘舒, 马想, 张浪, 朱永官. 城市绿地土壤健康质量问题与对策 [J]. 应用生态学报, 2022, 33(1): 268-276.
[6]	陈樟昊, 黄甘霖. 城市绿地供需的差异与联系研究进展 [J]. 应用生态学报, 2020, 31(11): 3925-3934.
[7]	王启为, 宫兆宁, 关鸿亮, 张磊, 井然, 汪星. 异质背景下黄河三角洲潮沟的遥感提取方法 [J]. 应用生态学报, 2019, 30(9): 3097-3107.
[8]	李哲, 张沁雨, 邱新彩, 彭道黎. 基于高分二号遥感影像树种分类的时相及方法选择 [J]. 应用生态学报, 2019, 30(12): 4059-4070.
[9]	苟睿坤, 陈佳琦, 段高辉, 杨瑞, 卜元坤, 赵君, 赵鹏祥. 基于GF-2的油松人工林地上生物量反演 [J]. 应用生态学报, 2019, 30(12): 4031-4040.
[10]	李琪, 陈文波, 郑蕉, 谢涛, 卢陶捷. 南昌市中心城区绿地景观对PM_2.5的影响 [J]. 应用生态学报, 2019, 30(11): 3855-3862.
[11]	孙守家, 雷帅, 仇兰芬, 李春友, 舒健骅. 北京城市绿地与周边道路空气CO₂浓度和δ¹³C值的差异及影响因素 [J]. 应用生态学报, 2019, 30(11): 3844-3854.
[12]	曹雅琴, 陈樟昊, 黄甘霖, 陈力原, 姜亚琼, 张征恺, 屠星月, 华野毓. 城市绿地格局与居民社会经济特征关系研究进展 [J]. 应用生态学报, 2019, 30(10): 3303-3315.
[13]	刘凤凤, 闫伟姣, 孔繁花, 尹海伟, 班玉龙, 徐文彬. 基于气温实地调查的城市绿地降温效应研究现状与未来展望 [J]. 应用生态学报, 2017, 28(4): 1387-1396.
[14]	孙守家1,2**,孟平1,2,张劲松1,2,舒健骅3,郑宁1,2 . 城市绿地系统内外CO₂浓度、δ¹³C、δ¹⁸O差异和来源及影响因素 [J]. 应用生态学报, 2015, 26(10): 3000-3010.
[15]	姚正阳,刘建军**. 西安市4种城市绿化灌木单株生物量估算模型 [J]. 应用生态学报, 2014, 25(1): 111-116.