基于变化更新的中东欧国家森林覆盖制图

doi:10.11707/j.1001-7488.LYKX20220097

摘要/Abstract

摘要：

目的: 分析中东欧国家森林资源覆盖状况，了解中东欧各国林业的现实情形和独特优势，为中国与中东欧国家加强林业合作提供基础数据。方法: 基于Google Earth Engine（GEE）平台获取中东欧国家2020年生长季无云影像，利用随机森林算法对影像进行分类，应用连续变化检测和分类算法得到森林变化信息并更新初步分类结果，最终获得中东欧国家土地覆盖分类结果。利用欧盟统计局土地利用/覆盖区域框架调查（LUCAS）数据、目视解译数据以及联合国粮农组织统计数据（FAOSTAT）对本研究分类结果进行验证和评估，结合谷歌高分辨率影像对制图结果进行分析。结果: 本研究分类结果利用LUCAS数据和目视解译数据的森林类型用户精度分别为0.930和0.911，生产者精度分别为0.860和0.956，总体精度分别为0.810和0.881，整体上优于GlobeLand30 2020产品（森林类型用户精度分别为0.920和0.900，生产者精度分别为0.690和0.840，总体精度分别为0.700和0.832）。根据研究结果计算出中东欧国家总体森林覆盖率为39.6%，相比GlobeLand30 2020结果（34.4%），与FAOSTAT的森林覆盖率（40.0%）更接近。目视上看，本制图结果在细节刻画上比GlobeLand30 2020产品更丰富，更能准确反映森林分布特征。截至2020年，中东欧国家森林覆盖呈北部、东南地区和西南地区丰富密集，中部分布广泛且相对较少的特点，爱沙尼亚、拉脱维亚、斯洛文尼亚、黑山等国家森林分布茂密，波兰、匈牙利等国家森林分布较为稀少。结论: 针对中东欧国家森林覆盖应用需求，本研究提出基于产品变化更新的森林覆盖制图方法，生产了中东欧国家2020年森林覆盖产品。本研究方法可为大区域森林覆盖制图提供新的借鉴和参考，结果有助于宏观了解中东欧国家森林覆盖状况。

关键词: 中东欧国家, Google Earth Engine, 森林资源, 遥感监测

Abstract:

Objective: The Central and Eastern European Countries (CEEC) which are the Belt and Road initiative are important forces in Europe and link Asia and Europe. Forest resources monitoring helps understand the forest development in CEEC, and is the basis for forestry cooperation between China and CEEC. Method: Taking CEEC as the study area, the cloudless images of CEEC in the growing season in 2020 are obtained based on the Google Earth Engine platform. The images are classified by random forest algorithm. The forest change information is obtained by the continue change detection and classification (CCDC) algorithm, and updated classification results by this information. Finally, the land cover results of CEEC are obtained. The results are evaluated and verified through the land use/cover area frame survey (LUCAS) data, visual interpretation data, and Food and Agricultural Organization of the United Nations statistical data (FAOSTAT). Finally, the mapping results are analyzed through Google high-resolution images. Result: Results show that our map obtained the users’ accuracy for forest type is 0.930 and 0.911 which validated by the LUCAS and visual interpretation data, the producer’s accuracy for forest type is 0.860 and 0.956, respectively, overall accuracy is 0.810 and 0.881, respectively. The accuracy of our map is better than the GlobeLand30 2020 products, in which the user’s accuracy for forest type is 0.920 and 0.900, while the producer’s accuracy for forest type is 0.690 and 0.840, respectively, and the overall accuracy is 0.700 and 0.832. According to the research results, the overall forest coverage rate of CEEC is 39.6%, this result is closer to the FAOSTAT result (40.0%) than the GlobeLand30 2020 result (34.4%). The final mapping results are richer in detail than the GlobeLand30 2020 product and can more accurately reflect the distribution characteristics of forests. In 2020, the forest cover of CEEC is rich in the north, southeast, and southwest, and less in the south. Estonia, Latvia, Slovenia, Montenegro, and other countries have rich forest distribution, while Poland, Hungary, and other countries have relatively rare forest distribution. Conclusion: This study has developed a forest cover mapping method based on change detection and update according to the application requirements of forest cover in CEEC and obtain a 2020 LC map. The product has high precision and provides a new reference for regional forest cover mapping research. This result is helpful in understanding the forest cover situation in CEEC.

Key words: Central and Eastern European Countries (CEEC), Google Earth Engine (GEE), forest resource, remote sensing monitoring

中图分类号:

S771.8

王春玲,史锴源,庞勇,蒙诗栎. 基于变化更新的中东欧国家森林覆盖制图[J]. 林业科学, 2024, 60(5): 116-126.

Chunling Wang,Kaiyuan Shi,Yong Pang,Shili Meng. Forest Cover Mapping of Central and Eastern European Countries Based on Change Detection and Update[J]. Scientia Silvae Sinicae, 2024, 60(5): 116-126.

图/表 13

图1

表1

图2

表2

图3

图4

图5

表3

图6

图7

表4

图8

表5

参考文献 0

	陈　军, 陈　晋, 廖安平, 等. 全球30 m地表覆盖遥感制图的总体技术. 测绘学报, 2014, 43 (6): 551- 557.
	Chen J, Chen J, Liao A P, et al. Concepts and key techniques for 30 m global land cover mapping. Acta Geodaetica et Cartographica Sinica, 2014, 43 (6): 551- 557.
	陈　军, 廖安平, 陈　晋, 等. 全球30 m地表覆盖遥感数据产品——GlobeLand30. 地理信息世界, 2017, 24 (1): 1- 8. doi: 10.3969/j.issn.1672-1586.2017.01.002
	Chen J, Liao A P, Chen J, et al. 30-meter global land cover data product: GlobeLand30. Geomatics World, 2017, 24 (1): 1- 8. doi: 10.3969/j.issn.1672-1586.2017.01.002
	庞　勇, 蒙诗栎, 史锴源, 等. 中国天然林保护工程区森林覆盖遥感监测. 生态学报, 2021, 41 (13): 5080- 5092.
	Pang Y, Meng S L, Shi K Y, et al. Forest coverage monitoring in the natural forest protection project area of China. Acta Ecologica Sinica, 2021, 41 (13): 5080- 5092.
	王燕琴, 陈　洁, 顾亚丽. 中东欧地区林业发展现状及“16+1”合作前景分析. 林业资源管理, 2017, (1): 153- 159.
	Wang Y Q, Chen J, Gu Y L. Forestry development in central and eastern European countries and analysis on future“16+1”forestry cooperation. Forest Resources Management, 2017, (1): 153- 159.
	Ballin M, Barcaroli G, Masselli M, et al. Redesign sample for land use/cover area frame survey (LUCAS) 2018. Eurostat:Statistical Working Papers, 2018, 10, 132365.
	Cánovas-García F, Alonso-Sarría F, Gomariz-Castillo F, et al. Modification of the random forest algorithm to avoid statistical dependence problems when classifying remote sensing imagery. Computers & Geosciences, 2017, 103, 1- 11.
	Chen C, Liaw A, Breimenn L. 2004. Using random forest to learn imbalanced data. Statistics Department of University of California, Berkeley.
	Chen J, Ban Y F, Li S N. Open access to earth land-cover map. Nature, 2014, 514 (7523): 434.
	Chen J, Chen J, Liao A P, et al. Global land cover mapping at 30 m resolution: a POK-based operational approach. ISPRS Journal of Photogrammetry and Remote Sensing, 2015, 103, 7- 27. doi: 10.1016/j.isprsjprs.2014.09.002
	Chen J, Lu M, Chen X, et al. A spectral gradient difference based approach for land cover change detection. ISPRS Journal of Photogrammetry and Remote Sensing, 2013, 85, 1- 12. doi: 10.1016/j.isprsjprs.2013.07.009
	Cohen W B, Yang Z, Healey S P, et al. A LandTrendr multispectral ensemble for forest disturbance detection. Remote Sensing of Environment, 2018, 205, 131- 140. doi: 10.1016/j.rse.2017.11.015
	Foody G M. Thematic map comparison: evaluating the statistical significance of differences in classification accuracy. Photogrammetric Engineering & Remote Sensing, 2004, 70 (5): 627- 633.
	Foody G M. Explaining the unsuitability of the kappa coefficient in the assessment and comparison of the accuracy of thematic maps obtained by image classification. Remote Sensing of Environment, 2020, 239, 111630. doi: 10.1016/j.rse.2019.111630
	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. Remote Sensing, 2020, 12 (21): 3479. doi: 10.3390/rs12213479
	Griffiths P, van der Linden S, Kuemmerle T, et al. A pixel-based Landsat compositing algorithm for large area land cover mapping. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2013, 6 (5): 2088- 2101. doi: 10.1109/JSTARS.2012.2228167
	Hansen M C, Potapov P V, Moore R, et al. High-resolution global maps of 21st-century forest cover change. Science, 2013, 342 (6160): 850- 853. doi: 10.1126/science.1244693
	Hermosilla T, Wulder M A, White J C, et al. Mass data processing of time series Landsat imagery: pixels to data products for forest monitoring. International Journal of Digital Earth, 2016, 9 (11): 1035- 1054. doi: 10.1080/17538947.2016.1187673
	Hu Y F, Dong Y. An automatic approach for land-change detection and land updates based on integrated NDVI timing analysis and the CVAPS method with GEE support. ISPRS Journal of Photogrammetry and Remote Sensing, 2018, 146, 347- 359. doi: 10.1016/j.isprsjprs.2018.10.008
	Jin S, Yang L, Zhu Z, et al. A land cover change detection and classification protocol for updating Alaska NLCD 2001 to 2011. Remote Sensing of Environment, 2017, 195, 44- 55. doi: 10.1016/j.rse.2017.04.021
	Karydas C G, Gitas I Z, Kuntz S, et al. Use of LUCAS LC point database for validating country-scale land cover maps. Remote Sensing, 2015, 7 (5): 5012- 5041. doi: 10.3390/rs70505012
	Kelley L C, Pitcher L, Bacon C. Using google earth engine to map complex shade-grown coffee landscapes in northern Nicaragua. Remote Sensing, 2018, 10 (6): 952. doi: 10.3390/rs10060952
	Kennedy R E, Yang Z, Cohen W B. Detecting trends in forest disturbance and recovery using yearly Landsat time series: 1. LandTrendr-temporal segmentation algorithms. Remote Sensing of Environment, 2010, 114 (12): 2897- 2910. doi: 10.1016/j.rse.2010.07.008
	Li X J, Chen W T, Cheng X W, et al. A comparison of machine learning algorithms for mapping of complex surface-mined and agricultural landscapes using Ziyuan-3 stereo satellite imagery. Remote Sensing, 2016, 8 (6): 514. doi: 10.3390/rs8060514
	Mahdianpari M, Salehi B, Mohammadimanesh F, et al. Random forest wetland classification using ALOS-2 L-band, RADARSAT-2 C-band, and TerraSAR-X imagery. ISPRS Journal of Photogrammetry and Remote Sensing, 2017, 130, 13- 31. doi: 10.1016/j.isprsjprs.2017.05.010
	Malinowski R, Lewiński S, Rybicki M, et al. Automated production of a land cover/use map of Europe based on Sentinel-2 imagery. Remote Sensing, 2020, 12 (21): 3523. doi: 10.3390/rs12213523
	Maxwell A E, Strager M P, Warner T A, et al. Large-area, high spatial resolution land cover mapping using random forests, GEOBIA, and NAIP orthophotography: findings and recommendations. Remote Sensing, 2019, 11 (12): 1409. doi: 10.3390/rs11121409
	Maxwell A E, Warner T A, Fang F. Implementation of machine-learning classification in remote sensing: an applied review. International Journal of Remote Sensing, 2018, 39 (9): 2784- 2817. doi: 10.1080/01431161.2018.1433343
	Millard K, Richardson M. On the importance of training data sample selection in random forest image classification: a case study in peatland ecosystem mapping. Remote Sensing, 2015, 7 (7): 8489- 8515. doi: 10.3390/rs70708489
	Pflugmacher D, Rabe A, Peters, M, et al. Mapping pan-European land cover using Landsat spectral-temporal metrics and the European LUCAS survey. Remote Sensing of Environment, 2019, 221, 583- 595. doi: 10.1016/j.rse.2018.12.001
	Potapov P V, Turubanova S A, Tyukavina A, et al. Eastern Europe’s forest cover dynamics from 1985 to 2012 quantified from the full Landsat archive. Remote Sensing of Environment, 2015, 159, 28- 43. doi: 10.1016/j.rse.2014.11.027
	Soenen S A, Peddle D R, Coburn C A. SCS C: a modified sun-canopy-sensor topographic correction in forested terrain. IEEE Transactions on Geoscience and Remote Sensing, 2005, 43 (9): 2148- 2159. doi: 10.1109/TGRS.2005.852480
	Tang D M, Fan H, Yang K, et al. Mapping forest disturbance across the China-Laos border using annual Landsat time series. International Journal of Remote Sensing, 2019, 40 (8): 2895- 2915. doi: 10.1080/01431161.2018.1533662
	Teluguntla P, Thenkabail P S, Oliphant A, et al. A 30-m Landsat-derived cropland extent product of Australia and China using random forest machine learning algorithm on Google Earth Engine cloud computing platform. ISPRS Journal of Photogrammetry and Remote Sensing, 2018, 144, 325- 340. doi: 10.1016/j.isprsjprs.2018.07.017
	Verbesselt J, Hyndman R, Newnham G, et al. Detecting trend and seasonal changes in satellite image time series. Remote Sensing of Environment, 2010, 114 (1): 106- 115. doi: 10.1016/j.rse.2009.08.014
	White J C, Wulder M A, Hobart G W, et al. Pixel-based image compositing for large-area dense time series applications and science. Canadian Journal of Remote Sensing, 2014, 40 (3): 192- 212. doi: 10.1080/07038992.2014.945827
	Xia J S, Falco N, Benediktsson J A, et al. Hyperspectral image classification with rotation random forest via KPCA. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2017, 10 (4): 1601- 1609. doi: 10.1109/JSTARS.2016.2636877
	Zhang X, Liu L Y, Chen X D, et al. Fine land-cover mapping in China using Landsat DataCube and an operational SPECLib-based approach. Remote Sensing, 2019, 11 (9): 1056. doi: 10.3390/rs11091056
	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. Earth System Science Data, 2021, 13, 2753- 2776. doi: 10.5194/essd-13-2753-2021
	Zhu Z, Woodcock C E, Olofsson P. Continuous monitoring of forest disturbance using all available Landsat imagery. Remote Sensing of Environment, 2012, 122, 75- 91. doi: 10.1016/j.rse.2011.10.030
	Zhu Z, Woodcock C E. Continuous change detection and classification of land cover using all available Landsat data. Remote Sensing of Environment, 2014, 144, 152- 171. doi: 10.1016/j.rse.2014.01.011
	Zhu Z, Zhang J X, Yang Z Q, et al. Continuous monitoring of land disturbance based on Landsat time series. Remote Sensing of Environment, 2020, 238, 111116. doi: 10.1016/j.rse.2019.03.009

分类产品 Land cover products	发布机构 Issued organization	数据来源 Data source	产品年份 Production year	空间分辨率 Spatial resolution/m
GlobeLand30	国家基础地理信息中心 National Geomatics Center of China	Landsat TM/ETM+/OLI HJ-1 A/B	2000，2010，2020	30
GLC_FCS30	中国科学院 Chinese Academy of Sciences	Landsat TM/ETM+	2015，2020	30
The Land Cover Map of Europe 2017	欧洲航天局 European Space Agency	Sential-2	2017	10

年份 Year	耕地 Cropland	森林 Forest	草地 Grass	灌木 Shrub	湿地 Wetland	水体 Water	人造地表 Impervious	裸地 Bare land	合计 Total
2000	132 587	90 865	26 799	12 272	2 601	4 058	12 608	640	282 430
2020	107 706	121 412	29 644	8 563	1 930	4 435	7 854	886	282 430

类别 Category	耕地 Cropland	森林 Forest	草地 Grass	灌木 Shrub	湿地 Wetland	水体 Water	人造地表 Impervious	裸地 Bare land	合计 Total	用户精度 User accuracy
耕地 Cropland	101 568	10 412	9 704	4 656	386	503	2 550	276	130 055	0.781
森林 Forest	2 149	103 879	2 789	1 934	217	398	654	77	112 097	0.930
草地 Grass	2 510	2 273	13 632	778	29	10	104	78	19 414	0.700
灌木 Shrub	329	1 828	468	541	14	17	46	22	3 265	0.702
湿地 Wetland	68	643	288	221	1 250	201	19	24	2 714	0.461
水体 Water	16	86	62	50	27	3 235	2	34	3 512	0.920
人造地表 Impervious	1 061	2 256	2 651	339	7	61	4 478	328	11 181	0.401
裸地 Bare land	5	35	50	44	0	10	1s	47	192	0.245
合计 Total	107 706	121 412	29 644	8 563	1 930	4 435	7854	886	282 430
生产精度 Producer accuracy	0.943	0.860	0.500	0.060	0.650	0.730	0.570	0.050		0.810

类别 Category	耕地 Cropland	森林 Forest	草地 Grass	灌木 Shrub	湿地 Wetland	水体 Water	人造地表 Impervious	合计 total	用户精度 User accuracy
耕地 Cropland	797	12	17	21	1	1	10	859	0.928
森林 Forest	29	607	15	13	0	0	2	666	0.911
草地 Grass	24	8	33	17	0	0	0	82	0.402
灌木 Shrub	9	6	2	10	0	0	0	27	0.370
湿地 Wetland	5	2	1	3	16	2	0	29	0.552
水体 Water	0	0	1	0	1	25	0	27	0.926
人造地表 Impervious	7	0	2	2	0	0	59	70	0.831
合计 Total	871	635	71	66	18	28	71	1 760
生产精度 Producer accuracy	0.915	0.956	0.465	0.152	0.889	0.893	0.831		0.880

国家 Country	本文研究结果 This paper（%）	GLobaLand30 2020（%）	FAOSTAT （%）
阿尔巴尼亚Albania	39.0	51.1	29.5
爱沙尼亚Estonia	56.5	45.5	59.7
保加利亚Bulgaria	40.1	34.2	41.1
波黑Bosnia and Herzegovina	64.5	50.0	54.2
波兰Poland	34.8	30.8	35.1
黑山Montenegro	56.4	66.6	71.6
捷克Czech	38.8	25.7	41.7
克罗地亚Croatia	47.9	37.2	50.0
拉脱维亚Latvia	56.5	43.3	51.0
立陶宛Lithuania	36.3	34.4	34.4
罗马尼亚Romania	34.5	20.2	34.3
马其顿Macedonia	42.7	62.1	43.7
塞尔维亚Serbia	36.0	36.1	35.7
斯洛伐克Slovakia	46.8	39.2	49.4
斯洛文尼亚Slovenia	65.6	58.2	68.0
匈牙利 Hungary	21.9	20.0	26.6
总计 Total	39.6	34.4	40.0