[1]卢德彬,等.基于GEE和双层随机森林模型的土地利用分类——以黔中城市群为实证[J].山地学报,2025,(5):749-762.[doi:10.16089/j.cnki.1008-2786.000927]
　LU Debin,SHI Zhang,SUO Chao,et al.Land-Use Classification Based on GEE and a Two-Layer Random-Forest Model: A Case Study of the Central Guizhou Urban Agglomeration[J].Mountain Research,2025,(5):749-762.[doi:10.16089/j.cnki.1008-2786.000927]
点击复制

基于GEE和双层随机森林模型的土地利用分类——以黔中城市群为实证()
分享到：

《山地学报》[ISSN:1008-2186/CN:51-1516]

卷:
期数:
2025年第5期
页码:
749-762
栏目:
山区发展
出版日期:
2025-12-30

文章信息/Info

Title:
Land-Use Classification Based on GEE and a Two-Layer Random-Forest Model: A Case Study of the Central Guizhou Urban Agglomeration
文章编号:
1008-2786-(2025)5-749-14
作者:
卢德彬1;  2; 石 彰1; 索 超2; 韩 畅1; 耿敬杰1;  3
(1. 铜仁学院 a. 旅游与地理系; b.贵州省高等学校山地国土空间智能监测与政策仿真工程研究中心,贵州 铜仁 554300; 2.贵州大学 公共管理学院,贵阳 550025; 3.四川大学 法学院,成都 610207)
Author(s):
LU Debin1; 2;  SHI Zhang1; SUO Chao2;  HAN Chang1; GENG Jingjie1; 3
(1.a. Department of Tourism & Geography; b. Guizhou Provincial Engineering Research Center for Intelligent Monitoring & Policy Simulation of Mountain Territorial Space, Tongren University, Tongren 554300, Guizhou, China; 2. School of Public Administration, Guizhou University, Guiyang 550025, China; 3. School of Law, Sichuan University, Chengdu 610207, China)
关键词:
Sentinel 1/2影像;  随机森林模型;  土地利用分类;  黔中城市群
Keywords:
Sentinel-1/2 imagery;  random-forest model;  land-use classification;  central Guizhou urban agglomeration
分类号:
F301.2
DOI:
10.16089/j.cnki.1008-2786.000927
文献标志码:
A
摘要:
在山地多云雾环境下,地表辐射信号畸变严重,地物识别困难,亟需高精度土地利用/覆被变化(LUCC)数据支撑城市群国土空间精细化治理。然而,现有适宜样本、影像与分类方法所导致的本地化高精度制图仍然不足。本文以黔中城市群为研究区,基于Google Earth Engine平台获取Sentinel-1/2影像,经去云、阴影和雪预处理,提取光谱、指数与SAR波谱三类特征变量; 借鉴Stacking思想,构建双层随机森林框架:第一层输出地类概率,第二层以高概率区重采样生成新特征并二次分类。结果表明:(1)双层模型总体精度达0.941,Kappa系数为0.929,比单层随机森林分别提升11.75%与14.55%;(2)耦合样本迁移法快速生产2018—2024年度尺度LUCC序列数据,平均总体精度0.92,Kappa系数为0.91,采用多月份集成方式,有效利用了稀疏影像,显著提升年图空间连续性与完整性,降低噪声影响;(3)样本点概率纯化策略显著提高样本质量与分类可靠性。本研究为多云雾山地城市群提供了一种高效、稳健的LUCC制图新框架,可为基础数据生产与国土空间治理提供科学参考。
Abstract:
In cloudy and foggy mountain environments, severe distortion of surface radiation signals makes land-cover identification difficult, and high-accuracy land-use/land-cover change(LUCC)data are urgently needed to support precise territorial-space governance of urban agglomerations. However, appropriate samples, imagery and classification methods for local high-precision mapping were still insufficient.
This study took the urban agglomeration in central Guizhou as the research area. Sentinel-1/2 images were acquired via the Google Earth Engine platform and preprocessed to remove clouds, shadows, and snow. Three types of feature variables—spectral, index, and SAR spectral—were extracted. Drawing on the Stacking concept, a double-layer random forest framework was constructed: the first layer outputs land class probabilities, while the second layer resamples high-probability areas to generate new features and performs secondary classification.
(1)Results show that the two-layer model achieved an overall accuracy of 0.941 and a Kappa coefficient of 0.929, increasing by 11.75% and 14.55% compared with the single-layer random forest.
(2)Coupled with a sample-transfer scheme, monthly-to-annual LUCC sequences from 2018 to 2024 were produced rapidly; the average overall accuracy was 0.92 and the Kappa coefficient was 0.91. Multi-month integration effectively utilized sparse images, significantly improving the spatial continuity and integrity of annual maps and reducing noise impacts.
(3)The above probability purification framework notably improved sample quality and classification reliability.
This study provides an efficient and robust LUCC-mapping framework for cloudy and foggy mountain urban agglomerations, and offers scientific reference for basic-data production and territorial-space governance.

参考文献/References:

[1] 方创琳, 张国友, 薛德升. 中国城市群高质量发展与科技协同创新共同体建设[J]. 地理学报, 2021, 76(12): 2898-2908. [FANG Chuanglin, ZHANG Guoyou, XUE Desheng. High-quality development of urban agglomerations in China and construction of science and technology collaborative innovation community [J]. Acta Geographica Sinica, 2021, 76(12): 2898-2908] DOI: 10.11821/dlxb202112003
[2] CAO Haojie, LI Yu, WENG Min, et al. Deciphering the evolving trajectories of China's megaregions from 1992 to 2020: A novel morphological approach based on global land cover products [J]. Applied Geography, 2024, 164: 103205. DOI: 10.1016/j.apgeog.2024.103205
[3] FOLEY J A, DEFRIES R, ASNER G P, et al. Global consequences of land use [J]. Science, 2005, 309(5734): 570-574. DOI: 10.1126/science.111177
[4] NEWBOLD T, HUDSON L N, HILL S L L, et al. Global effects of land use on local terrestrial biodiversity [J]. Nature, 2015, 520(7545): 45-50. DOI: 10.1038/nature14324
[5] LU Debin, XU Jianhua, YUE Wenze, et al. Response of PM2.5 pollution to land use in China [J]. Journal of Cleaner Production, 2020, 244: 118741. DOI: 10.1016/j.jclepro.2019.118741
[6] TURNER B L, MEYER W B, SKOLE D L. Global land-use/land-cover change: Towards an integrated study [J]. Ambio, 1994, 23(1): 91-95.
[7] WINKLER K, FUCHS R, ROUNSEVELL M, et al. Global land use changes are four times greater than previously estimated [J]. Nature Communications, 2021, 12(1): 2501. DOI: 10.1038/s41467-021-22702-2
[8] 杨钟贤, 苏春江. 平原与山区土地利用/覆被变化对比—以双流县和米易县为例[J]. 山地学报, 2009, 27(5): 585-592. [YANG Zhongxian, SU Chunjiang. A comparative study between plain and mountain on LUCC: Case study of Shuangliu County and Miyi County, Sichuan [J]. Mountain Research, 2009, 27(5): 585-592]
[9] 梁顺波, 孙林, 杜永明, 等. 基于 GEE 平台的每年夏冬两期土地覆盖分类方法研究[J]. 北京师范大学学报(自然科学版), 2024, 60(6): 831-842. [LIANG Shunbo, SUN Lin, DU Yongming, et al. Land cover classification in summer and winter based on the Google Earth Engine platform [J]. Journal of Beijing Normal University(Natural Science), 2024, 60(6): 831-842] DOI: 10.12202/j.0476-0301.2023250
[10] 毛丽君, 李海涛, 薛晓明, 等. 基于 GEE 平台的国家公园土地覆盖变化遥感检测方法构建[J]. 南京林业大学学报(自然科学版), 2022, 46(2): 213-220. [MAO Lijun, LI Haitao, XUE Xiaoming, et al. Developing remote sensing based methods for land cover change detection in national parks from GEE platform: A case study from the Qianjiangyuan National Park pilot area [J]. Journal of Nanjing Forestry University(Natural Sciences Edition), 2022, 46(2): 213-220] DOI: 10.12302/j.issn.1000-2006.202101013
[11] 李培林, 刘小平, 黄应淮, 等. 基于 GEE 平台的广州市主城区不透水面时间序列提取[J]. 地球信息科学学报, 2020, 22(3): 638-648. [LI Peilin, LIU Xiaoping, HUANG Yinghuai, et al. Mapping impervious surface dynamics of Guangzhou Downtown based on Google Earth Engine[J]. Journal of Geo-information Science, 2020, 22(3): 638-648] DOI: 10.12082/dqxxkx.2020.190047
[12] GONG Peng, LIU Han, ZHANG Meinan, et al. Stable classification with limited sample: Transferring a 30-m resolution sample set collected in 2015 to mapping 10-m resolution global land cover in 2017 [J]. Science Bulletin, 2019, 64(6): 370-373. DOI: 10.1016/j.scib.2019.03.002
[13] ZANAGA D, VAN DE KERCHOVE R, DE KEERSMAECKER W, et al. ESA WorldCover 10 m 2020 v100 [DB]. Zenodo. https://doi.org/10.5281/zenodo.5571936
[14] KARRA K, KONTGIS C, STATMAN-WEIL Z, et al. Global land use/land cover with Sentinel 2 and deep learning [C]//Proceedings of 2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS. Brussels, Belgium: IEEE, 2021: 4704-4707. DOI: 10.1109/IGARSS47720.2021.9553499
[15] BROWN C F, BRUMBY S P, GUZDER-WILLIAMS B, et al. Dynamic world, near real-time global 10 m land use land cover mapping [J]. Scientific Data, 2022, 9(1): 251. DOI: 10.1038/s41597-022-01307-4
[16] WANG Jun, YANG Xiaomei, WANG Zhihua, et al. Consistency analysis and accuracy assessment of three global ten-meter land cover products in rocky desertification region: A case study of southwest China [J]. ISPRS International Journal of Geo-Information, 2022, 11(3): 202. DOI: 10.3390/ijgi11030202
[17] VENTER Z S, BARTON D N, CHAKRABORTY T, et al. Global 10 m land use land cover datasets: A comparison of dynamic world, world cover and esri land cover [J]. Remote Sensing, 2022, 14(16): 4101. DOI: 10.3390/rs14164101
[18] 李恒凯, 王利娟, 肖松松. 基于多源数据的南方丘陵山地土地利用随机森林分类[J]. 农业工程学报, 2021, 37(7): 244-251. [LI Hengkai, WANG Lijuan, XIAO Songsong. Random forest classification of land use in hilly and mountaineous areas of southern China using multi-source remote sensing data [J]. Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE), 2021, 37(7): 244-251] DOI: 10.11975/j.issn.1002-6819.2021.07.030
[19] INGLADA J, VINCENT A, ARIAS M, et al. Operational high resolution land cover map production at the country scale using satellite image time series [J]. Remote Sensing, 2017, 9(1): 95. DOI: 10.3390/rs9010095
[20] GRIFFITHS P, NENDEL C, HOSTERT P. Intra-annual reflectance composites from Sentinel-2 and Landsat for national-scale crop and land cover mapping [J]. Remote Sensing of Environment, 2019, 220: 135-151. DOI: 10.1016/j.rse.2018.10.031
[21] GARNOT V S F, LANDRIEU L, CHEHATA N. Multi-modal temporal attention models for crop mapping from satellite time series [J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2022, 187: 294-305. DOI: 10.1016/j.isprsjprs.2022.03.012
[22] TANG Yunwei, QIU Fang, WANG Bangjin, et al. A deep relearning method based on the recurrent neural network for land cover classification [J]. GIScience & Remote Sensing, 2022, 59(1): 1344-1366. DOI: 10.1080/15481603.2022.2115589
[23] CHE Xianghong, ZHANG Hankui, LI Zhongbin, et al. Linearly interpolating missing values in time series helps little for land cover classification using recurrent or attention networks[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2024, 212: 73-95. DOI: 10.1016/j.isprsjprs.2024.04.021
[24] MAXWELL A E, WARNER T A, FANG Fang. Implementation of machine-learning classification in remote sensing: An applied review [J]. International Journal of Remote Sensing, 2018, 39(9): 2784-2817. DOI: 10.1080/01431161.2018.1433343
[25] MAO Wanliu, LU Debin, HOU Li, et al. Comparison of machine-learning methods for urban land-use mapping in Hangzhou City, China [J]. Remote Sensing, 2020, 12(17): 2817. DOI.10.3390/rs12172817
[26] HOSSEINI M M, ZOEJ M J V, DEHKORDI A T, et al. Cropping intensity mapping in Sentinel-2 and Landsat-8/9 remote sensing data using temporal transfer of a stacked ensemble machine learning model within Google Earth Engine [J]. Geocarto International, 2024, 39(1): 2387786. DOI: 10.1080/10106049.2024.2387786
[27] 潘娇, 李超, 彭文忆, 等. 基于随机森林和支持向量机的云南省土地利用分类[J]. 科学技术与工程, 2024, 24(17): 7043-7051. [PAN Jiao, LI Chao, PENG Wenyi, et al. Land use classification in Yunnan Province based on random forest and support vector machines [J]. Science Technology and Engineering, 2024, 24(17): 7043-7051] DOI: 10.12404/j.issn.1671-1815.2306050
[28] DAS B, RATHORE P, ROY D, et al. Comparison of bagging, boosting and stacking algorithms for surface soil moisture mapping using optical-thermal-microwave remote sensing synergies [J]. Catena, 2022, 217: 106485. DOI: 10.1016/j.catena.2022.106485
[29] RAHMATI A, ZOEJ M J V, DEHKORDI A T. Early identification of crop types using Sentinel-2 satellite images and an incremental multi-feature ensemble method(Case study: Shahriar, Iran)[J]. Advances in Space Research, 2022, 70(4): 907-922. DOI: 10.1016/j.asr.2022.05.038
[30] ABDALI E, ZOEJ M J V, DEHKORDI A T, et al. A parallel-cascaded ensemble of machine learning models for crop type classification in Google Earth Engine using multi-temporal sentinel-1/2 and landsat-8/9 remote sensing data [J]. Remote Sensing, 2023, 16(1): 127. DOI: 10.3390/rs16010127
[31] ZHANG Chengkang, ZHANG Hongyan, TIAN Sijing. Phenology-assisted supervised paddy rice mapping with the Landsat imagery on Google Earth Engine: Experiments in Heilongjiang Province of China from 1990 to 2020 [J]. Computers and Electronics in Agriculture, 2023, 212: 108105. DOI: 10.1016/j.compag.2023.108105
[32] GITELSON A A, KAUFMAN Y J, MERZLYAK M N. Use of a green channel in remote sensing of global vegetation from EOS-MODIS [J]. Remote Sensing of Environment, 1996, 58(3): 289-298.
[33] MATURANO J G, SANTANA J D M, AGUILAR-GARCIA A L, et al. Remote sensing of illegal dumps through supervised classification of satellite images: Application in Oaxaca, Mexico [J]. Geographical Research Letters, 2024, 50(2): 157-177. DOI: 10.18172/cig.6273
[34] CHEN Junyi, CHEN Suozhong, YANG Chao, et al. A comparative study of impervious surface extraction using Sentinel-2 imagery [J]. European Journal of Remote Sensing, 2020, 53(1): 274-292.
[35] WANG Yongji, QI Qingwen, LIU Ying. Unsupervised segmentation evaluation using area-weighted variance and Jeffries-Matusita distance for remote sensing images [J]. Remote Sensing, 2018, 10(8): 1193. DOI: 10.3390/rs10081193
[36] HUANG Huabing, WANG Jie, LIU Caixia, et al. The migration of training samples towards dynamic global land cover mapping [J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2020, 161: 27-36. DOI: 10.1016/j.isprsjprs.2020.01.010
[37] WOLPERT D H. Stacked generalization [J]. Neural Networks, 1992, 5(2): 241-259. DOI: 10.1016/S0893-6080(05)80023-1
[38] MOHARRAMI M, ATTARCHI S, GLOAGUEN R, et al. Integration of Sentinel-1 and Sentinel-2 data for ground truth sample migration for multi-temporal land cover mapping [J]. Remote Sensing, 2024, 16(9): 1566. DOI: 10.3390/rs16091566
[39] XU Panpan, TSENDBAZAR N E, HEROLD M, et al. Comparative validation of recent 10 m-resolution global land cover maps [J]. Remote Sensing of Environment, 2024, 311: 114316. DOI: 10.1016/j.rse.2024.114316
[40] WANG Jun, YANG Xiaomei, WANG Zhihua, et al. Consistency analysis and accuracy assessment of three global ten-meter land cover products in rocky desertification region: A case study of southwest China [J]. ISPRS International Journal of Geo-Information, 2022, 11(3): 202. DOI: 10.3390/ijgi11030202

相似文献/References:

[1]赵广东,陈 健,史作民*,等.川西亚高山暗针叶林土壤温度和水分动态及其与气象因子的关系[J].山地学报,2021,(2):183.[doi:10.16089/j.cnki.1008-2786.000586]
　ZHAO Guangdong,CHEN Jian,SHI Zuomin*,et al.Soil Temperature and Moisture Dynamics and Their Relationship with Meteorological Factors of Subalpine Dark Coniferous Forest in Western Sichuan,China[J].Mountain Research,2021,(5):183.[doi:10.16089/j.cnki.1008-2786.000586]
[2]管家琳a,黄炎和a,林金石a,等.基于信息量模型与随机森林模型的崩岗风险对比评估[J].山地学报,2021,(4):539.[doi:10.16089/j.cnki.1008-2786.000618]
　GUAN Jialina,HUANG Yanhea,LIN Jinshia,et al.Comparisons Between Benggang Risk Assessments Based on Information Model and Random Forest Model[J].Mountain Research,2021,(5):539.[doi:10.16089/j.cnki.1008-2786.000618]

备注/Memo

备注/Memo:
收稿日期(Received date): 2025- 05-12; 改回日期(Accepted data):2025- 09-22
基金项目(Foundation item): 国家自然科学基金(42361042); 贵州省基础研究(自然科学)项目(黔科合基础-ZK[2023]一般461); 铜仁学院博士人才启动基金(trxyDH2218)。[National Natural Science Foundation of China(42361042); Guizhou Provincial Basic Research Program(Natural Science)(ZK(2023)-461); Doctoral Research Start-up Fund of Tongren University(trxyDH2218)]
作者简介(Biography): 卢德彬(1987-),男,博士,教授,主要研究方向:土地利用变化监测与政策仿真。[LU Debin(1987-), male, Ph.D., professor, research on land use change monitoring and policy simulation] E-mail:sooluo@163.com
更新日期/Last Update: 2025-10-20