基于改进全卷积神经网络模型的土地覆盖分类方法研究

doi:10.12082/dqxxkx.2023.220435

地球信息科学学报 ›› 2023, Vol. 25 ›› Issue (3): 495-509.doi: 10.12082/dqxxkx.2023.220435

• 地球信息科学理论与方法 • 上一篇下一篇

基于改进全卷积神经网络模型的土地覆盖分类方法研究

衡雪彪¹(), 许捍卫¹^,^*(), 唐璐¹, 汤恒¹, 许怡蕾²

1.河海大学水文水资源学院，南京 210024
2.南京师范大学地理科学学院，南京 210024

收稿日期:2022-06-23 修回日期:2022-08-02 出版日期:2023-03-25 发布日期:2023-04-19
通讯作者: * 许捍卫（1969— ），男，博士，副教授，主要从事地理大数据应用与开发。E-mail: xuhanwei@hhu.edu.cn
作者简介:衡雪彪（1999— ），男，河南焦作人，硕士生，从事深度学习遥感信息提取研究。E-mail: hxb719617378@163.com
基金资助:
国家自然科学基金项目(411771478)

Research on Land Cover Classification Method based on Improved Fully Convolutional Neural Network Model

HENG Xuebiao¹(), XU Hanwei¹^,^*(), TANG Lu¹, TANG Heng¹, XU Yilei²

1. School of Hydrology and Water Resources, Hohai University, Nanjing 210024, China
2. School of Geographical Sciences, Nanjing Normal University, Nanjing 210024, China

Received:2022-06-23 Revised:2022-08-02 Online:2023-03-25 Published:2023-04-19
Contact: XU Hanwei
Supported by:
National Natural Science Foundation of China(411771478)

摘要/Abstract

摘要：

遥感卫星数据是地球表面信息的重要来源，但利用传统的遥感分类方法进行土地覆盖分类局限性大、过程繁琐、解译精度依赖专家经验，而深度学习方法可以自适应地提取地物更多深层次的特征信息，适用于高分辨率遥感影像的土地覆盖分类。文中对高分辨率影像中水体、交通运输、建筑、耕地、草地、林地、裸土等进行高精度分类，结合遥感多地物分类的特点，以DeepLabV3+模型为基础，作出了以下改进：① 骨干网络的改进，使用ResNeSt代替ResNet作为骨干网络；② 空洞空间金字塔池化模块的改进，首先在并联的每个分支的前一层增加一个空洞率相对较小的空洞卷积，其次在分支后层加入串联的空洞率逐渐减小的空洞卷积层。使用土地覆盖样本库和自制样本库进行模型训练、测试。结果表明，改进模型在2个数据集的精度和时间效率均明显优于原始DeepLabV3+模型：土地覆盖样本库总体精度达到88.08%，自制样本库总体精度达到85.22%，较原始DeepLabV3+模型分别提升了1.35%和3.4%，时间效率每epoch减少0.39 h。改进模型能够为数据量以每日TB级增加的高分影像提供更加快速精确的土地覆盖分类结果。

关键词: 土地覆盖, 全卷积神经网络, 深度学习, 遥感影像分类, DeepLabV3+, ResNeSt, 高分辨率影像, 语义分割

Abstract:

Remote sensing satellite data are essential source of earth surface information. However, traditional remote sensing classification methods usually have limitations and include cumbersome processes, and the accuracy of interpretation depends on the expert experience. Deep learning methods can adaptively extract more detailed feature information from field objects and are suitable for land cover classification of high-resolution remote sensing images. Based on the DeepLabV3+ model, this paper makes the following improvements: (1) Improvement of the backbone network. We use ResNeSt instead of ResNet as the backbone network; (2) The improvement of the hole space pyramid pooling module. First, a hole convolution with a relatively small hole rate is added to the previous layer of the parallel branch, and then a series of hole convolution layers with a gradually decreasing hole rate are added to the back layer of the branch. We use the land cover sample database and the self-made sample database respectively for model training and classify water bodies, transportation, buildings, cultivated land, grasslands, forests, bare soil, etc. from high-resolution images. Our results show that the accuracy and time efficiency of the improved model using the two databases are significantly higher than those of the original DeepLabV3+ model. The overall accuracy using the land cover sample database and self-made sample database reach 88.08% and 85.22%, respectively, which are 1.35% and 3.4% higher than that using the original DeepLabV3+ model, respectively. Also the time cost decreases by 0.39h per epoch. The improved model can provide faster and more accurate land cover classification results for high-resolution imageries that increases in terabytes of data per day.

Key words: land cover, fully convolutional neural network, deep learning, remote sensing image classification, DeepLabV3+, ResNeSt, high-resolution images, semantic segmentation

衡雪彪, 许捍卫, 唐璐, 汤恒, 许怡蕾. 基于改进全卷积神经网络模型的土地覆盖分类方法研究[J]. 地球信息科学学报, 2023, 25(3): 495-509.DOI:10.12082/dqxxkx.2023.220435

HENG Xuebiao, XU Hanwei, TANG Lu, TANG Heng, XU Yilei. Research on Land Cover Classification Method based on Improved Fully Convolutional Neural Network Model[J]. Journal of Geo-information Science, 2023, 25(3): 495-509.DOI:10.12082/dqxxkx.2023.220435

图/表 18

图1

图2

图3

图4

图5

图6

表1

图7

图8

表2

表3

图9

图10

图11

图12

表4

图13

图14

参考文献 28

[1]	刘纪远, 张增祥, 张树文, 等. 中国土地利用变化遥感研究的回顾与展望——基于陈述彭学术思想的引领[J]. 地球信息科学学报, 2020, 22(4):680-687. doi: 10.12082/dqxxkx.2020.200052
	[Liu J Y, Zhang Z X, Zhang S W, et al. Innovation and development of remote sensing-based land use change studies based on Shupeng chen's academic thoughts[J]. Journal of Geo-Information Science, 2020, 22(4):680-687.] DOI:CNKI:SUN:DQXX.0.2020-04-006 doi: CNKI:SUN:DQXX.0.2020-04-006
[2]	Di Gregorio A. Land cover classification system: classification concepts and user manual: LCCS[M]. Software version 2. Rome: Food and Agriculture Organization of the United Nations, 2005
[3]	Jin S M, Yang L M, Zhu Z, et al. A land cover change detection and classification protocol for updating Alaska NLCD 2001 to 2011[J]. Remote Sensing of Environment, 2017, 195:44-55. DOI:10.1016/j.rse.2017.04.021 doi: 10.1016/j.rse.2017.04.021
[4]	de Oliveira M L, dos Santos C A C, de Oliveira G, et al. Remote sensing-based assessment of land degradation and drought impacts over terrestrial ecosystems in Northeastern Brazil[J]. Science of the Total Environment, 2022, 835:155490. DOI:10.1016/j.scitotenv.2022.155490 doi: 10.1016/j.scitotenv.2022.155490
[5]	Stone R. Earth-observation summit endorses global data sharing[J]. Science, 2010, 330(6006):902. DOI:10.1126/science.330.6006.902 doi: 10.1126/science.330.6006.902
[6]	Zhu Q Q, Zhong Y F, Zhao B, et al. The bag-of-visual-words scene classifier combining local and global features for high spatial resolution imagery[C]// 2015 12th International Conference on Fuzzy Systems and Knowledge Discovery (FSKD). IEEE, 717-721. DOI:10.1109/FSKD.2015.7382030 doi: 10.1109/FSKD.2015.7382030
[7]	Tuxen K, Schile L, Stralberg D, et al. Mapping changes in tidal wetland vegetation composition and pattern across a salinity gradient using high spatial resolution imagery[J]. Wetlands Ecology and Management, 2011, 19(2):141-157. DOI:10.1007/s11273-010-9207-x doi: 10.1007/s11273-010-9207-x
[8]	周星宇, 张继贤, 高绵新, 等. 高分辨率遥感影像下沿海地区地表覆盖信息的提取[J]. 测绘通报, 2017(2):19-24.
	[Zhou X Y, Zhang J X, Gao M X, et al. Land cover information extraction based on high-resolution remote sensing image in coastal areas[J]. Bulletin of Surveying and Mapping, 2017(2):19-24.] DOI:10.13474/j.cnki.11-2246.2017.0041 doi: 10.13474/j.cnki.11-2246.2017.0041
[9]	Wu Q, Zhong R F, Zhao W J, et al. A comparison of pixel-based decision tree and object-based Support Vector Machine methods for land-cover classification based on aerial images and airborne lidar data[J]. International Journal of Remote Sensing, 2017, 38(23):7176-7195. DOI:10.1080/01431161.2017.1371864 doi: 10.1080/01431161.2017.1371864
[10]	Zhu Q Q, Guo X, Deng W H, et al. Land-use/land-cover change detection based on a Siamese global learning framework for high spatial resolution remote sensing imagery[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2022, 184:63-78. DOI:10.1016/j.isprsjprs.2021.12.005 doi: 10.1016/j.isprsjprs.2021.12.005
[11]	刘浩, 骆剑承, 黄波, 等. 基于特征压缩激活Unet网络的建筑物提取[J]. 地球信息科学学报, 2019, 21(11):1779-1789. doi: 10.12082/dqxxkx.2019.190285
	[Liu H, Luo J C, Huang B, et al. Building extraction based on SE-unet[J]. Journal of Geo-Information Science, 2019, 21(11):1779-1789.] DOI:CNKI:SUN:DQXX.0.2019-11-012 doi: CNKI:SUN:DQXX.0.2019-11-012
[12]	李森, 彭玲, 胡媛, 等. 基于FD-RCF的高分辨率遥感影像耕地边缘检测[J]. 中国科学院大学学报, 2020, 37(4):483-489. doi: 10.7523/j.issn.2095-6134.2020.04.007
	[Li S, Peng L, Hu Y, et al. FD-RCF-based boundary delineation of agricultural fields in high resolution remote sensing images[J]. Journal of University of Chinese Academy of Sciences, 2020, 37(4):483-489.] DOI:10.7523/j.issn.2095-6134.2020.04.007 doi: 10.7523/j.issn.2095-6134.2020.04.007
[13]	Liu H, Luo J, Huang B, et al. DE-Net: Deep encoding network for building extraction from high-resolution remote sensing imagery[J]. Remote Sensing, 2019, 11(20):2380. DOI:10.3390/rs11202380 doi: 10.3390/rs11202380
[14]	郭颖, 李增元, 陈尔学, 等. 一种改进的高空间分辨率遥感影像森林类型深度学习精细分类方法:双支FCN^-8s[J]. 林业科学, 2020, 56(3):48-60.
	[Guo Y, Li Z Y, Chen E X, et al. A deep learning method for forest fine classification based on high resolution remote sensing images: Two-branch FCN-8s[J]. Scientia Silvae Sinicae, 2020, 56(3):48-60.] DOI:CNKI:SUN:LYKE.0.2020-03-006 doi: CNKI:SUN:LYKE.0.2020-03-006
[15]	Zheng Z, Zhong Y F, Ma A L, et al. FPGA: fast patch-free global learning framework for fully end-to-end hyperspectral image classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2020, 58(8):5612-5626. DOI:10.1109/TGRS.2020.2967821 doi: 10.1109/TGRS.2020.2967821
[16]	Sherrah J. Fully convolutional networks for dense semantic labelling of high-resolution aerial imagery[EB/OL]. 2016: arXiv: 1606.02585. DOI:10.48550/arXiv.1606.02585 doi: 10.48550/arXiv.1606.02585
[17]	Persello C, Stein A. Deep fully convolutional networks for the detection of informal settlements in VHR images[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14(12):2325-2329. DOI:10.1109/LGRS.2017.2763738 doi: 10.1109/LGRS.2017.2763738
[18]	Zhang C, Sargent I, Pan X, et al. An object-based convolutional neural network (OCNN) for urban land use classification[J]. Remote Sensing of Environment, 2018, 216:57-70. DOI:10.1016/j.rse.2018.06.034 doi: 10.1016/j.rse.2018.06.034
[19]	Long J, Shelhamer E, Darrell T. Fully convolutional networks for semantic segmentation[C]// 2015 IEEE Conference on Computer Vision and Pattern Recognition. IEEE,: 3431-3440. DOI:10.1109/CVPR.2015.7298965 doi: 10.1109/CVPR.2015.7298965
[20]	Ronneberger O, Fischer P, Brox T. U-Net: Convolutional Networks for Biomedical Image Segmentation[C]// International Conference on Medical Image Computing and Computer-Assisted Intervention. Cham: Springer, 2015:234-241. DOI:10.1007/978-3-319-24574-4_28 doi: 10.1007/978-3-319-24574-4_28
[21]	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 doi: 10.1109/TPAMI.2016.2644615 pmid: 28060704
[22]	Zhu Q Q, Yang Y, Sun X L, et al. CDANet: contextual detail-aware network for high-spatial-resolution remote-sensing imagery shadow detection[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60:1-15. DOI:10.1109/TGRS.2022.3143886 doi: 10.1109/TGRS.2022.3143886
[23]	Zhu Q, Deng W, Zheng Z, et al. A spectral-spatial-dependent global learning framework for insufficient and imbalanced hyperspectral image classification[J]. IEEE Transactions on Cybernetics, 2021, 25. DOI:10.1109/TCYB.2021.3070577 doi: 10.1109/TCYB.2021.3070577
[24]	Chen L, Papandreou G, Kokkinos I, et al. Semantic image segmentation with deep convolutional nets and fully connected CRFs[J]. Computer Science, 2014(4):357-361.DOI:10.1080/17476938708814211 doi: 10.1080/17476938708814211
[25]	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 doi: 10.1109/TPAMI.2017.2699184
[26]	Chen L C, Papandreou G, Schroff F, et al. Rethinking Atrous Convolution for Semantic Image Segmentation[J]. 2017. DOI:10.1007/978-3-030-01234-2 doi: 10.1007/978-3-030-01234-2
[27]	Chen L C, Zhu Y K, Papandreou G, et al. Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation[C]// European Conference on Computer Vision. Cham: Springer, 2018:833-851. 10.1007/978-3-030-01234-2_49 doi: 10.1007/978-3-030-01234-2_49
[28]	Zhang H, Wu C, Zhang Z, et al. Resnest: Split-attention networks[J]. 2020. DOI:10.48550/arXiv.2004.08955 doi: 10.48550/arXiv.2004.08955

一级大类	代码	二级小类	分类示例
水体	1	水体	河流、湖泊、水库、坑塘、沟渠、海洋
交通运输		道路	快速路、主干路、次干路、支路
	2	机场	飞行区、航站楼等
		火车站	火车站站台及候车厅
建筑	3	建筑物	住宅、工厂等人工建筑
耕地	4	普通耕地	普通耕地
耕地	4	农业大棚	耕地另一种形式，如温室
草地	5	自然草地	自然草地
草地	5	绿地绿化	包括公园绿地、生态景观绿地、防护绿地等人工草地
林地	6	自然林	乔木林、灌木林
林地	6	人工林	果园林、苗圃等
裸土	7	自然裸土	自然裸土、戈壁、沙漠
裸土	7	人为裸土	因建筑需求等原因破坏地表覆盖而造成裸土裸露的地表
其他		光伏	太阳能光伏发电板
	8	停车场	社会停车场用地、公共交通场站
	8	操场	篮球场、排球场、羽毛球场、网球场等
		其他无法确定地物	-

	DeepLabV3+	U-Net	SegNet
水体	94.23	93.84	93.91
交通运输	82.47	80.26	80.53
建筑	90.39	87.13	87.70
耕地	84.09	81.37	81.76
草地	82.81	81.04	81.73
林地	91.18	89.65	89.03
裸土	83.94	82.17	82.33
其他	81.44	78.47	79.21
OA	86.73	84.76	85.68
Kappa	84.47	82.67	83.54
F1-score	86.39	84.65	85.42

	D	D+R	D+R+A
水体	94.23	94.56	94.52
交通运输	82.47	86.07	86.53
建筑	90.39	91.25	91.70
耕地	84.09	84.39	84.26
草地	82.81	83.04	83.33
林地	91.18	92.68	93.03
裸土	83.94	86.17	86.10
其他	81.44	82.47	83.31
OA	86.73	87.76	88.08
Kappa	84.47	85.67	86.04
F1-score	86.39	87.65	87.92
平均每个epoch训练时间/h	1.89	1.48	1.50

	SVM	DeepLabV3+	改进的DeepLabV3+
水体	91.05	89.87	92.51
交通运输	80.79	82.18	84.77
建筑	83.05	85.63	89.18
耕地	75.44	81.39	83.55
草地	80.69	81.56	83.06
林地	86.96	82.14	87.80
裸土	77.82	81.38	85.80
其他	69.58	79.70	82.43
OA	79.02	81.82	85.22
Kappa	74.39	76.65	80.85
F1-score	80.25	83.58	87.11

基于改进全卷积神经网络模型的土地覆盖分类方法研究

Research on Land Cover Classification Method based on Improved Fully Convolutional Neural Network Model

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

图/表 18

参考文献 28

相关文章 15

Metrics

本文评价

推荐阅读 0

[1]	左溪冰, 刘智, 金飞, 林雨准, 王淑香, 刘潇, 李美霖. 面向高光谱影像小样本分类的全局-局部特征自适应融合方法[J]. 地球信息科学学报, 2023, 25(8): 1699-1716.
[2]	林娜, 何静, 王斌, 唐菲菲, 周俊宇, 郭江. 结合植被光谱特征与Sep-UNet的城市植被信息智能提取方法[J]. 地球信息科学学报, 2023, 25(8): 1717-1729.
[3]	张春菊, 刘文聪, 张雪英, 叶鹏, 汪陈, 朱少楠, 张达玉. 基于本体的金矿知识图谱构建方法[J]. 地球信息科学学报, 2023, 25(7): 1269-1281.
[4]	张彤, 刘仁宇, 王培晓, 高楚林, 刘杰, 王望舒. 感知物理先验的机器学习及其在地理空间智能中的研究前景[J]. 地球信息科学学报, 2023, 25(7): 1297-1311.
[5]	曹兴文, 吴孟泉, 郑雪婷, 郑宏伟, 李映祥, 张安安. 室内空间布局约束下的在线跟踪注册学习方法[J]. 地球信息科学学报, 2023, 25(7): 1418-1431.
[6]	柳林, 谢华芳, 岳瀚. 不同街景影像微环境提取方法对街面财产犯罪解释的差异分析[J]. 地球信息科学学报, 2023, 25(7): 1432-1447.
[7]	赵丹, 杜萍, 刘涛, 令振飞. 融合图自编码器与GRU的城市盗窃犯罪时空分布预测模型[J]. 地球信息科学学报, 2023, 25(7): 1448-1463.
[8]	金广垠, 沙恒宇, 张金雷, 黄金才. 基于路段-路口联合建模的对偶图卷积网络的行程时间估计方法[J]. 地球信息科学学报, 2023, 25(7): 1500-1513.
[9]	刘潇, 刘智, 林雨准, 王淑香, 左溪冰. 面向遥感影像场景分类的类中心知识蒸馏方法[J]. 地球信息科学学报, 2023, 25(5): 1050-1063.
[10]	侯慧太, 蓝朝桢, 徐青. 基于卫星影像全局和局部深度学习特征检索的无人机绝对定位方法[J]. 地球信息科学学报, 2023, 25(5): 1064-1074.
[11]	张金雷, 陈奕洁, Panchamy Krishnakumari, 金广垠, 王骋程, 杨立兴. 基于注意力机制的城市轨道交通网络级多步短时客流时空综合预测模型[J]. 地球信息科学学报, 2023, 25(4): 698-713.
[12]	李心雨, 闫浩文, 王卓, 王炳瑄. 街景图像与机器学习相结合的道路环境安全感知评价与影响因素分析[J]. 地球信息科学学报, 2023, 25(4): 852-865.
[13]	于方圆, 曹家玮, 李发源, 李思进. 顾及对象特征的地面式光伏电站提取及减碳效益评估[J]. 地球信息科学学报, 2023, 25(3): 529-545.
[14]	高建文, 管海燕, 彭代锋, 许正森, 康健, 季雅婷, 翟若雪. 基于局部-全局语义特征增强的遥感影像变化检测网络模型[J]. 地球信息科学学报, 2023, 25(3): 625-637.
[15]	江宝得, 许少芬, 王进, 王淼. 基于条件生成对抗网络的电子地图可视水印去除方法[J]. 地球信息科学学报, 2023, 25(2): 288-297.