呼包鄂榆城市群遥感土地覆被产品评估分析与集成学习

doi:10.12118/j.issn.1000-6060.2023.718

干旱区地理 ›› 2024, Vol. 47 ›› Issue (8): 1367-1379.doi: 10.12118/j.issn.1000-6060.2023.718

呼包鄂榆城市群遥感土地覆被产品评估分析与集成学习

夏子龙¹^,²^,³(), 杜培军¹^,²^,³(), 郭山川¹^,²^,³, 方宏¹^,²^,³, 张伟¹^,²^,³, 潘小全¹^,²^,³

1.南京大学地理与海洋科学学院，江苏南京 210023
2.自然资源部国土卫星遥感应用重点实验室，江苏南京 210023
3.江苏省地理信息资源开发与利用协同创新中心，江苏南京 210023

收稿日期:2024-02-16 修回日期:2024-03-17 出版日期:2024-08-25 发布日期:2024-09-02
通讯作者: 杜培军（1975-），男，教授，主要从事城市遥感、遥感信息智能处理与地学分析等方面的研究. E-mail: peijun@nju.edu.cn
作者简介:夏子龙（1996-），男，博士研究生，主要从事土地覆被遥感分析与应用等方面的研究. E-mail: zilongxia@smail.nju.edu.cn
基金资助:
国家自然科学基金项目(42330106);国家重点研发计划课题(2022YFC3800802)

Evaluation and ensemble learning of remote sensing land-cover products in the Hohhot-Baotou-Ordos-Yulin urban agglomeration

XIA Zilong¹^,²^,³(), DU Peijun¹^,²^,³(), GUO Shanchuan¹^,²^,³, FANG Hong¹^,²^,³, ZHANG Wei¹^,²^,³, PAN Xiaoquan¹^,²^,³

1. School of Geography and Ocean Science, Nanjing University, Nanjing 210023, Jiangsu, China
2. Key Laboratory for Land Satellite Remote Sensing Applications of Ministry of Natural Resources, Nanjing 210023, Jiangsu, China
3. Jiangsu Center for Collaborative Innovation in Geographical Information Resource Development and Application, Nanjing 210023, Jiangsu, China

Received:2024-02-16 Revised:2024-03-17 Published:2024-08-25 Online:2024-09-02

摘要/Abstract

摘要：

近年来国内外生产了多个大范围、中高分辨率的土地覆被产品，但这些产品在旱区城市群的可靠性还缺乏验证。评估不同产品在旱区城市群的空间一致性和分类精度，探究如何集成现有数据生产精度更高的产品，有助于推动和完善该区域相关研究。以呼包鄂榆城市群为研究区，基于现有5种主要土地覆被产品[GlobeLand30、GLC_FCS30、CLCD、FROM_GLC30和ESA WorldCover （ESAWC）]，从主要地类的构成相似性、空间一致性和分类精度3个方面分析这些产品在呼包鄂榆城市群土地覆被的空间差异和可靠性，提出一种多产品集成学习方法，融合5种土地覆被产品形成新土地覆被产品。结果表明：（1） 5种产品对呼包鄂榆城市群土地覆被的总体构成和空间分布描述具有较高的相似性，不同产品中分类完全一致和高度一致区域占呼包鄂榆城市群总面积的68.79%，其中草地识别的一致性最高。（2） 5种产品的总体精度在67.83%~76.96%之间，其中ESAWC产品的总体精度最高（76.96%），其次是GLC_FCS30（75.34%），所有产品在林地、草地、水体和裸地分类中精度较高，在灌木、湿地等占比较小的地类中精度较低。（3）提出的集成学习方案可以充分整合已有产品，实现多种土地覆被产品的集成应用，新产品分类精度达到83.36%，相较原始产品分类精度提高6.40%~15.53%。研究结果可以为旱区城市群专题研究选择和生产土地覆被产品提供有价值的参考。

关键词: 土地覆被产品, 空间一致性分析, 精度评价, 数据集成, 呼包鄂榆城市群

Abstract:

In recent years, numerous large-scale, medium to high-resolution land-cover products have been produced both domestically and internationally. However, the reliability of these products in arid region urban agglomerations has not been adequately validated. Assessing the spatial consistency and classification accuracy of different products in arid urban agglomerations, and exploring ways to utilize existing data to produce more precise integrated products, can advance and refine relevant research in this region. Based on the five major existing land-cover products [GlobeLand30, GLC_FCS30, CLCD, FROM_GLC30, and ESA WorldCover (ESAWC)], this study analyzed the spatial differences and reliability of these products within the Hohhot-Baotou-Ordos-Yulin (HBOY) urban agglomeration, north China in terms of the similarity of composition type, spatial consistency, and classification accuracy. An effective multi-source product ensemble learning method was proposed, involving the fusion of five land-cover products to generate a novel, integrated land-cover product. The results show that: (1) The overall composition and spatial distribution of land-cover in the HBOY urban agglomeration is similar among the five products. The areas that are completely consistent and highly consistent among different products account for 68.79% of the total area, with grassland having the highest consistency. (2) The overall accuracy of the five products ranged from 67.83% to 76.96%, with the ESAWC product exhibiting the highest overall accuracy (76.96%), followed by GLC_FCS30 (75.34%). All the products demonstrated significant spatial accuracy in depicting ground reality in forests, grasslands, water bodies, and bare ground, with relatively lower accuracy observed in scrubland and wetlands. (3) The proposed product integration scheme harnesses the strengths of existing products and prior knowledge, enabling the integrated application of multiple land-cover products. The new product achieves an overall accuracy of 83.36%, representing an improvement of 6.40% to 15.53% compared to the accuracy of the original products. The results of the study provide valuable references for the selection and production of land-cover products for thematic studies in arid region urban agglomerations.

Key words: land-cover products, spatial consistency analysis, accuracy assessment, integrated application, HBOY urban agglomeration

夏子龙, 杜培军, 郭山川, 方宏, 张伟, 潘小全. 呼包鄂榆城市群遥感土地覆被产品评估分析与集成学习[J]. 干旱区地理, 2024, 47(8): 1367-1379.

XIA Zilong, DU Peijun, GUO Shanchuan, FANG Hong, ZHANG Wei, PAN Xiaoquan. Evaluation and ensemble learning of remote sensing land-cover products in the Hohhot-Baotou-Ordos-Yulin urban agglomeration[J]. Arid Land Geography, 2024, 47(8): 1367-1379.

图/表 12

图1

表1

表2

图2

图3

图4

图5

图6

图7

图8

表3

图9

参考文献 36

[1]	Reynolds J F, Dms S, Lambin E F, et al. Global desertification: Building a science for dryland development[J]. Science, 2007, 316(5826): 847-851. doi: 10.1126/science.1131634 pmid: 17495163
[2]	Ren Q, He C Y, Huang Q X, et al. Impacts of urban expansion on natural habitats in global drylands[J]. Nature Sustainability, 2022, 5(10): 869-878.
[3]	He C Y, Li J W, Zhang X L, et al. Will rapid urban expansion in the drylands of northern China continue: A scenario analysis based on the land use scenario dynamics-urban model and the shared socioeconomic pathways[J]. Journal of Cleaner Production, 2017, 165: 57-69.
[4]	Zhou S Y, We L, Lu Z G, et al. An analysis of multiple ecosystem services in a large-scale urbanized area of northern China based on the food-energy-water integrative framework[J]. Environmental Impact Assessment Review, 2023, 98: 106913, doi: 10.1016/j.eiar.2022.106913.
[5]	Song S X, He C Y, Liu Z F, et al. Evaluating the influences of urban expansion on multiple ecosystem services in drylands[J]. Landscape Ecology, 2022, 37(11): 2783-3802.
[6]	Song S X, Liu Z F, He C Y, et al. Evaluating the effects of urban expansion on natural habitat quality by coupling localized shared socioeconomic pathways and the land use scenario dynamics-urban model[J]. Ecological Indicators, 2020, 112: 106071, doi: 10.1016/j.ecolind.2020.106071.
[7]	Yu L, Du Z R, Dong R M, et al. FROM-GLC Plus: Toward near real-time and multi-resolution land cover mapping[J]. GIScience & Remote Sensing, 2022, 59(1): 1026-1047.
[8]	Li J W, Liu Z F, He C Y, et al. Are the drylands in northern China sustainable? A perspective from ecological footprint dynamics from 1990 to 2010[J]. Science of the Total Environment, 2016, 553: 223-231.
[9]	孙泽祥, 刘志锋, 何春阳, 等. 中国快速城市化干燥地区的生态系统服务权衡关系多尺度分析——以呼包鄂榆地区为例[J]. 生态学报, 2016, 36(15): 4881-4891.
	[Sun Zexiang, Liu Zhifeng, He Chunyang, et al. Multi-scale analysis of ecosystem service trade-offs in urbanizing drylands of China: A case study in the Hohhot-Baotou-Ordos-Yulin region[J]. Acta Ecologica Sinica, 2016, 36(15): 4881-4891.]
[10]	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[J]. 农业机械学报, 2022, 53(3): 1-17.
	[Feng Quanlong, Niu Bowen, Zhu Dehai, et al. Review for deep learning in land use and land cover remote sensing classification[J]. Transactions of the Chinese Society for Agricultural Machinery, 2022, 53(3): 1-17.]
[11]	McCallum I, Obersteiner M, Nilsson S, et al. A spatial comparison of four satellite derived 1 km global land cover datasets[J]. International Journal of Applied Earth Observation and Geoinformation, 2006, 8(4): 246-255.
[12]	Kaptué A, Roujean J L, De Jong S M. Comparison and relative quality assessment of the GLC2000, GLOBCOVER, MODIS and ECOCLIMAP land cover data sets at the African continental scale[J]. International Journal of Applied Earth Observation and Geoinformation, 2011, 13(2): 207-219.
[13]	胡云锋, 张千力, 戴昭鑫, 等. 多源遥感土地覆被产品在欧洲地区的一致性分析[J]. 地理研究, 2015, 34(10): 1839-1852. doi: 10.11821/dlyj201510003
	[Hu Yunfeng, Zhang Qianli, Dai Zhaoxin, et al. Agreement analysis of multi-sensor satellite remote sensing derived land cover products in the Europe continent[J]. Geographical Research, 2015, 34(10): 1839-1852.]
[14]	Pan N, Wang S, Liu Y X, et al. Quantifying responses of net primary productivity to agricultural expansion in drylands[J]. Land Degradation & Development, 2020, 32(5): 2050-2060.
[15]	Maestre F T, Benito B M, Berdugo M, et al. Biogeography of global drylands[J]. New Phytologist, 2021, 231: 540-558. doi: 10.1111/nph.17395 pmid: 33864276
[16]	田绍鸿, 张显峰. 采用随机森林法的天绘数据干旱区城市土地覆盖分类[J]. 国土资源遥感, 2016, 28(1): 43-49.
	[Tian Shaohong, Zhang Xianfeng. Random forest classification of land cover information of urban areas in arid regions based on TH-1 data[J]. Remote Sensing for Land and Resources, 2016, 28(1): 43-49.]
[17]	陈逸聪, 邵华, 李杨. 多源土地覆被产品在长三角地区的一致性分析与精度评价[J]. 农业工程学, 2021, 37(6): 142-150.
	[Chen Yicong, Shao Hua, Li Yang. Consistency analysis and accuracy assessment of multi-source land cover products in the Yangtze River Delta[J]. Transactions of the Chinese Society of Agricultural Engineering, 2021, 37(6): 142-150.]
[18]	陈逸聪, 邵华, 李杨, 等. 2015年长三角地区30 m土地覆被融合数据[J]. 中国科学数据, 2022, 7(1): 149-167.
	[Chen Yicong, Shao Hua, Li Yang, et al. A dataset of 30 m-resolution land cover fusion in Yangtze River Delta in 2015[J]. Science Data Bank, 2022, 7(1): 149-167.]
[19]	白燕, 冯敏. 全球尺度多源土地覆被数据融合与评价研究[J]. 地理学报, 2018, 73(11): 2223-2235. doi: 10.11821/dlxb201811013
	[Bai Yan, Feng Min. Data fusion and accuracy evaluation of multi-source global land cover datasets[J]. Acta Geographica Sinica, 2018, 73(11): 2223-2235.] doi: 10.11821/dlxb201811013
[20]	Huang A Q, Shen R P, Li Y Q, et al. A methodology to generate integrated land cover data for land surface model by improving Dempster-Shafer theory[J]. Remote Sensing, 2022, 14: 972, doi: 10.3390/rs14040972.
[21]	王莉. 基于IKONOS影像融合的土地覆盖分类及居民地信息提取研究[D]. 徐州: 中国矿业大学, 2010.
	[Wang Li. Study on land cover classification and residential areas extraction using IKONOS imagery based on data fusion[D]. Xuzhou: China University of Mining and Technology, 2010.]
[22]	宋世雄, 张金茜, 刘志锋, 等. 旱区城市扩展过程区位因素研究——以中国呼包鄂榆城市群为例[J]. 自然资源学报, 2021, 36(4): 1021-1035. doi: 10.31497/zrzyxb.20210416
	[Song Shixiong, Zhang Jinxi, Liu Zhifeng, et al. Study on the location factors of urban expansion in the drylands: A case study in the Hohhot-Baotou-Ordos-Yulin urban agglomeration, China[J]. Journal of Natural Resources, 2021, 36(4): 1021-1035.] doi: 10.31497/zrzyxb.20210416
[23]	魏乐, 周亮, 孙东琪, 等. 黄河流域城市群扩张的时空格局演化及情景模拟——以呼包鄂榆城市群为例[J]. 地理研究, 2022, 41(6): 1610-1622. doi: 10.11821/dlyj020210566
	[Wei Le, Zhou Liang, Sun Dongqi, et al. The evolution of spatio-temporal pattern and scenario simulation of urban agglomeration expansion in the Yellow River Basin: A case study in the Hohhot-Baotou-Ordos-Yulin urban agglomeration[J]. Geographical Research, 2022, 41(6): 1610-1622.]
[24]	萨日盖, 包玉海, 窦银银, 等. 近20 a内蒙古高原城乡开发建设对生态系统生产力的影响[J]. 干旱区地理, 2023, 46(6): 922-933.
	[Sa Rigai, Bao Yuhai, Dou Yinyin, et al. Impacts of urban and rural construction on ecosystem productivity in Inner Mongolia Plateau from 2000 to 2020[J]. Arid Land Geography, 2023, 46(6): 922-933.]
[25]	张梦圆, 荣丽华, 李伊彤, 等. 基于“三生”空间的农牧交错区城市土地利用转型及生态环境效应分析——以包头市为例[J]. 干旱区地理, 2023, 46(6): 958-967.
	[Zhang Mengyuan, Rong Lihua, Li Yitong, et al. Land use function transformation in the agro-pastoral ecotone based on ecological-production-living spaces and associated eco-environment effects: A case of Baotou City[J]. Arid Land Geography, 2023, 46(6): 958-967.]
[26]	Chen J, Chen J, Liao A P, et al. Global land cover mapping at 30 m resolution: A POK-based operational approach[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2014, 103: 7-27.
[27]	Zhang X, Liu L Y, Chen X D, et al. GLC_FCS30: Global land-cover product with fine classification system at 30 m using time-series Landsat imagery[J]. Earth System Science Data, 2021, 13(6): 2753-2776.
[28]	Yang J, Huang X. The 30 m annual land cover dataset and its dynamics in China from 1990 to 2019[J]. Earth System Science Data, 2021, 13(8): 3907-3925.
[29]	戴昭鑫, 胡云锋, 张千力. 多源卫星遥感土地覆被产品在南美洲的一致性分析[J]. 遥感信息, 2017, 32(2): 137-148.
	[Dai Zhaoxin, Hu Yunfeng, Zhang Qianli. Agreement analysis of multi source land cover products derived from remote sensing in South America[J]. Remote Sensing Information, 2017, 32(2): 137-148.]
[30]	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.
[31]	Gao Y, Liu L Y, Zhang X, et al. Consistency analysis and accuracy assessment of three global 30-m land-cover products over the European Union using the LUCAS dataset[J]. Remote Sensing, 2020, 12(21): 3479, doi: 10.3390/rs12213479.
[32]	Olofsson P, Giles M F, Herold M, et al. Good practices for estimating area and assessing accuracy of land change[J]. Remote Sensing of Environment, 2014, 148: 42-57.
[33]	Olofsson P, Arévalo P, Andres B, et al. Mitigating the effects of omission errors on area and area change estimates[J]. Remote Sensing of Environment, 2020, 236: 111492, doi: 10.1016/j.rse.2019.111492.
[34]	仝冉, 杨雅萍, 陈晓娜. 多源30 m分辨率土地覆被数据在蒙古高原的一致性分析和精度评价[J]. 地球信息科学学报, 2022, 24(12): 2420-2434. doi: 10.12082/dqxxkx.2022.220578
	[Tong Ran, Yang Yaping, Chen Xiaona. Consistent analysis and accuracy evaluation of multisource land cover datasets in 30 m spatial resolution over the Mongolian Plateau[J]. Journal of Geo-information Science, 2022, 24(12): 2420-2434.]
[35]	Kang J M, Yang X M, Wang Z H, et al. Comparison of three ten meter land cover products in a drought region: A case study in northwestern China[J]. Land, 2022, 11(3): 427, doi: 10.3390/land11030427.
[36]	热合曼·如克亚, 卡斯木·阿里木江, 阿布来提·哈力木拉提, 等. 基于InVEST模型的天山北坡城市群生境质量时空演化研究[J]. 生态与农村环境学报, 2022, 38(9): 1112-1121.
	[Rukeya Reheman, Alim Kasim, Halmurat Ablat, et al. Research on the temporal and spatial evolution of habitat quality in urban agglomeration on the northern slope of Tianshan Mountains based on InVEST model[J]. Journal of Ecology and Rural Environment, 2022, 38(9): 1112-1121.]

土地覆被类型	GLC_FCS30		GlobeLand30			CLCD			FROM_GLC30			ESAWC （30 m）			ESAWC （10 m）		集成产品
土地覆被类型	PA	UA	PA		UA	PA		UA	PA		UA	PA		UA	PA	UA	PA	UA
耕地	60.49	72.59	48.46		81.48	63.98		76.30	56.07		71.85	73.33		76.74	79.34	71.11	78.47	83.70
林地	74.36	58.00	59.18		58.00	72.00		72.00	62.69		84.00	62.00		86.11	85.29	58.00	82.93	68.00
灌木	6.25	2.00	28.30		30.00	100.00		34.00	40.00		8.00	80.56		82.48	60.00	24.00	100.00	14.00
草地	78.55	84.55	75.00		73.26	73.98		90.70	80.19		71.93	22.00		47.83	82.72	81.89	82.40	93.36
湿地	76.19	64.00	80.95		34.00	0.00		22.00	53.85		14.00	36.00		100.00	100.00	36.00	100.00	42.00
水体	89.36	84.00	79.63		86.00	82.46		94.00	81.63		80.00	94.00		79.66	82.76	96.00	74.63	100.00
不透水面	97.22	70.00	65.57		80.00	74.42		64.00	50.00		52.00	76.00		95.00	97.56	80.00	93.02	80.00
裸地	75.27	74.05	84.03		54.05	86.32		44.32	61.94		89.73	87.03		57.71	58.45	89.73	89.82	81.08
OA	75.34			67.83			74.57			69.54			75.94		76.96		83.36

呼包鄂榆城市群遥感土地覆被产品评估分析与集成学习

Evaluation and ensemble learning of remote sensing land-cover products in the Hohhot-Baotou-Ordos-Yulin urban agglomeration

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 36

相关文章 1

Metrics

本文评价

推荐阅读 0

数据名称	数据源	研发单位	空间分辨率/m	年份	地类	分类方法
GLC_FCS30	Landsat TM/ETM+/OLI	中科院	30	2020	29	随机森林
GlobeLand30	Landsat TM/ETM+，HJ-1	国家基础信息中心	30	2020	10	POK法
CLCD	Landsat TM/ETM+/OLI	武汉大学	30	2019	9	随机森林
FROM_GLC30	Landsat TM/ETM+	清华大学	30	2017	10	随机森林
ESAWC	Sentinel-1/2	欧洲航天局	10	2020	11	深度学习

土地覆被类型	GLC_FCS30	GlobeLand30	CLCD	FROM_GLC30	ESAWC
耕地	10、11、20	10	1	1	40
林地	50~100	20	2	2	10
灌木	120、121、122	40	3	4	20
草地	130、150	30	4	3	30
湿地	180	50	9	5	90
水体	210	60	5	6	80
不透水面	190	80	8	8	50
裸地	200	90	7	9	60