Remote sensing parameters optimization for accurate land cover classification

doi:10.11947/j.AGCS.2024.20230327

Abstract

Abstract:

Sustainable natural resources management requires considerable accurate land cover information given the evident climate change impacts and human disturbances on wetlands. It is characterized by the convergence of numerous materials and energies, resulting in fragmented landscapes and frequent land cover changes. To address the challenges posed by the complexity of landforms, diversity of land cover types, and non-linearity of remote sensing image features in traditional remote sensing image classification methods, this paper proposes a feature parameter selection method based on the Gini index of random forests, with a 10% threshold decision. The aim is to identify the optimal combination of remote sensing feature parameters. Firstly, spectral features, texture features, thermal features, elevation features, and principal component features are selected to form a stack of remote sensing images. Then, multiple decision trees are set up to cross-validate the contributions of the features, and the feature ranking is determined based on the normalized mean importance of the features. Finally, a threshold is set to select the remote sensing feature parameters that meet the requirements, and the process is iterated. Experiments are conducted using Sentinel-2 remote sensing images covering the Yancheng Nature Reserve in Jiangsu province. The results show that the remote sensing feature parameters selected by this method have good representativeness. Compared with CART, SVM, KNN, and RF methods that only use band information, the proposed method produces clearer boundaries and more accurate category attributes in the classification results, with an overall accuracy of 96.20% and a Kappa coefficient of 0.955 6. This research can provide technical support for regional spatial planning and sustainable development.

Key words: land cover classification, random forest, feature optimizing, feature recursive elimination, Gini index

CLC Number:

P237

Chao CHEN, Jintao LIANG, Gang YANG, Weiwei SUN, Shaojun GONG, Jianqiang WANG. Remote sensing parameters optimization for accurate land cover classification[J]. Acta Geodaetica et Cartographica Sinica, 2024, 53(7): 1401-1416.

Figures/Tables 18

Fig.1

Tab.1

Fig.3

Tab.2

Fig.5

Fig.6

Fig.7

Tab.3

Fig.9

Fig.10

Fig.11

Fig.12

Tab.4

Fig.14

References 64

[1]	侯婉, 侯西勇. 考虑湿地精细分类的全球海岸带土地利用/覆盖遥感分类系统[J]. 热带地理, 2018, 38(6):866-873.
	HOU Wan, HOU Xiyong. Remote sensing classification system of land use and land cover for global coastal zone considering fine classification of wetlands[J]. Tropical Geography, 2018, 38(6):866-873.
[2]	冯如意, 王力哲, 曾铁勇. 高光谱遥感图像亚像元信息提取方法综述[J]. 测绘学报, 2023, 52(7):1187-1201. DOI: 10.11947/j.AGCS.2023.20220491.
	FENG Ruyi, WANG Lizhe, ZENG Tieyong. Review of hyperspectral remote sensing image subpixel information extraction[J]. Acta Geodaetica et Cartographica Sinica, 2023, 52(7):1187-1201. DOI: 10.11947/j.AGCS.2023.20220491.
[3]	梁锦涛, 陈超, 孙伟伟, 等. 基于Landsat长时间序列卫星影像和GEE云平台的杭州湾土地利用/覆被变化时空格局演变分析 [J]. 遥感学报, 2023, 27(6):1480-1495. DOI: 10.11834/jrs.20232614.
	LIANG Jintao, CHEN Chao, SUN Weiwei, et al. Spatio-temporal land use/cover change dynamics in Hangzhou Bay, China using long-term Landsat time series and GEE platform[J]. National Remote Sensing Bulletin, 2023, 27(6):1480-1495. DOI: 10.11834/jrs.20232614.
[4]	侯婉, 侯西勇. 全球海岸带多源土地利用/覆盖遥感分类产品一致性分析[J]. 地球信息科学学报, 2019, 21(7):1061-1073.
	HOU Wan, HOU Xiyong. Consistency of the multiple remote sensing-based land use and land cover classification products in the global coastal zones[J]. Journal of Geo-Information Science, 2019, 21(7):1061-1073.
[5]	GENG Jiachen, SHEN Shi, CHENG Changxiu, et al. A hybrid spatiotemporal convolution-based cellular automata model (ST-CA) for land-use/cover change simulation [J]. International Journal of Applied Earth Observation and Geoinformation, 2022, 110:102789. DOI: 10.1016/j.jag.2022.102789.
[6]	骆剑承, 胡晓东, 吴田军, 等. 高分遥感驱动的精准土地利用与土地覆盖变化信息智能计算模型与方法研究[J]. 遥感学报, 2021, 25(7):1351-1373.
	LUO Jiancheng, HU Xiaodong, WU Tianjun, et al. Research on intelligent calculation model and method of precision land use/cover change information driven by high-resolution remote sensing[J]. National Remote Sensing Bulletin, 2021, 25(7):1351-1373.
[7]	陈超, 陈慧欣, 陈东, 等. 舟山群岛海岸线遥感信息提取及时空演变分析[J]. 国土资源遥感, 2021, 33(2):141-152.
	CHEN Chao, CHEN Huixin, CHEN Dong, et al. Coastline extraction and spatial-temporal variations using remote sensing technology in Zhoushan Islands[J]. Remote Sensing for Land & Resources, 2021, 33(2):141-152.
[8]	孙根云, 付航, 张爱竹, 等. 高光谱影像奇异谱分析特征提取方法：综述与评价 [J]. 测绘学报, 2023, 52(7):1148-1163. DOI: 10.11947/j.AGCS.2023.20220542.
	SUN Genyun, FU Hang, ZHANG Aizhu, et al. Singular spectrum analysis method for hyperspectral imagery feature extraction: a review and evaluation [J]. Acta Geodaetica et Cartographica Sinica, 2023, 52(7):1148-1163. DOI: 10.11947/j.AGCS.2023.20220542.
[9]	彭江涛, 孙伟伟, 魏天慧, 等. 高分五号高光谱影像的关联对齐域适应与分类[J]. 遥感学报, 2020, 24(4):417-426.
	PENG Jiangtao, SUN Weiwei, WEI Tianhui, et al. A modified correlation alignment algorithm for the domain adaptation of GF-5 hyperspectral image[J]. Journal of Remote Sensing, 2020, 24(4):417-426.
[10]	侯西勇, 邸向红, 侯婉, 等. 中国海岸带土地利用遥感制图及精度评价[J]. 地球信息科学学报, 2018, 20(10):1478-1488.
	HOU Xiyong, DI Xianghong, HOU Wan, et al. Accuracy evaluation of land use mapping using remote sensing techniques in coastal zone of China[J]. Journal of Geo-Information Science, 2018, 20(10):1478-1488.
[11]	LOVELAND T R, REED B C, BROWN J F, et al. Development of a global land cover characteristics database and IGBP DISCover from 1 km AVHRR data[J]. International Journal of Remote Sensing, 2000, 21(6/7):1303-1330.
[12]	HANSEN M C, REED B. A comparison of the IGBP DISCover and University of Maryland 1 km global land cover products[J]. International Journal of Remote Sensing, 2000, 21(6/7):1365-1373.
[13]	FRIEDL M A, MCLVER D K, HODGES J C F, et al. Global land cover mapping from MODIS: algorithms and early results[J]. Remote Sensing of Environment, 2002, 83(1/2):287-302.
[14]	LI Wei, CIAIS P, MACBEAN N, et al. Major forest changes and land cover transitions based on plant functional types derived from the ESA CCI land cover product[J]. International Journal of Applied Earth Observation and Geoinformation, 2016, 47:30-39.
[15]	ZANAGA D, VAN DE KERCHOVE R, DAEMS D, et al. ESA WorldCover 10 m 2021 v200 [DB/OL]. [2023-08-01]. https://research.wur.nl/en/datasets/esa-worldcover-10-m-2021-v200.
[16]	宫鹏, 张伟, 俞乐, 等. 全球地表覆盖制图研究新范式[J]. 遥感学报, 2016, 20(5):1002-1016.
	GONG Peng, ZHANG Wei, YU Le, et al. New research paradigm for global land cover mapping[J]. Journal of Remote Sensing, 2016, 20(5):1002-1016.
[17]	CHEN Xuehong, CAO Xin, LIAO Anping, et al. Global mapping of artificial surfaces at 30-m resolution[J]. Science China: Earth Sciences, 2016, 59(12):2295-2306.
[18]	MYINT S W, GIRI C P, WANG Le, et al. Identifying mangrove species and their surrounding land use and land cover classes using an object-oriented approach with a lacunarity spatial measure[J]. GIScience & Remote Sensing, 2008, 45(2):188-208.
[19]	孙根云, 王鑫, 安娜, 等. 基于多源遥感的大尺度高分辨率不透水面深度学习提取方法 [J]. 测绘学报, 2023, 52(2):272-282. DOI: 10.11947/j.AGCS.2023.20210546.
	SUN Genyun, WANG Xin, AN Na, et al. A method for large-scale and high-resolution impervious surface extraction based on multi-source remote sensing and deep learning [J]. Acta Geodaetica et Cartographica Sinica, 2023, 52(2):272-282. DOI: 10.11947/j.AGCS.2023.20210546.
[20]	SMEECKAERT J, MALLET C, DAVID N, et al. Large-scale classification of water areas using airborne topographic lidar data[J]. Remote Sensing of Environment, 2013, 138:134-148.
[21]	ZHANG Fang, YANG Xiaojun. Improving land cover classification in an urbanized coastal area by random forests: the role of variable selection[J]. Remote Sensing of Environment, 2020, 251:112105.
[22]	MA Lei, LIU Yu, ZHANG Xueliang, et al. Deep learning in remote sensing applications: a meta-analysis and review[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2019, 152:166-177.
[23]	ZHU Xiaoxiang, TUIA D, MOU Lichao, et al. Deep learning in remote sensing: a comprehensive review and list of resources[J]. IEEE Geoscience and Remote Sensing Magazine, 2017, 5(4):8-36.
[24]	YUAN Qiangqiang, SHEN Huanfeng, LI Tongwen, et al. Deep learning in environmental remote sensing: achievements and challenges[J]. Remote Sensing of Environment, 2020, 241:111716.
[25]	杨彬, 毛银, 陈晋, 等. 深度学习的遥感变化检测综述：文献计量与分析 [J]. 遥感学报, 2023, 27(9):1988-2005. DOI: 10.11834/jrs.20222156.
	YANG Bin, MAO Yin, CHEN Jin, et al. Review of remote sensing change detection in deep learning: bibliometric and analysis[J]. National Remote Sensing Bulletin, 2023, 27(9):1988-2005. DOI: 10.11834/jrs.20222156.
[26]	SAHA S, SAHA A, HEMBRAM T K, et al. Prediction of spatial landslide susceptibility applying the novel ensembles of CNN, GLM and random forest in the Indian Himalayan region[J]. Stochastic Environmental Research and Risk Assessment, 2022, 36(10):3597-3616.
[27]	HUAN Juan, LI Mingbao, XU Xiangen, et al. Multi-step prediction of dissolved oxygen in rivers based on random forest missing value imputation and attention mechanism coupled with recurrent neural network[J]. Water Supply, 2022, 22(5):5480-5493.
[28]	SHEN Yulin, MERCATORIS B, CAO Zhen, et al. Improving wheat yield prediction accuracy using LSTM-RF framework based on UAV thermal infrared and multispectral imagery[J]. Agriculture, 2022, 12(6):892.
[29]	钟燕飞, 王心宇, 胡鑫, 等. 高光谱高空间分辨率遥感观测、处理与应用 [J]. 测绘学报, 2023, 52(7):1212-1226. DOI: 10.11947/j.AGCS.2023.20220715.
	ZHONG Yanfei, WANG Xinyu, HU Xin, et al. Hyperspectral with high-spatial resolution remote sensing from observation, processing to applications [J]. Acta Geodaetica et Cartographica Sinica, 2023, 52(7):1212-1226. DOI: 10.11947/j.AGCS.2023.20220715.
[30]	REZAEE M, MAHDIANPARI M, ZHANG Yun, et al. Deep convolutional neural network for complex wetland classification using optical remote sensing imagery[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2018, 11(9):3030-3039.
[31]	FANG Bei, LI Ying, ZHANG Haokui, et al. Collaborative learning of lightweight convolutional neural network and deep clustering for hyperspectral image semi-supervised classification with limited training samples[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2020, 161:164-178.
[32]	JIA Sen, JIANG Shuguo, LIN Zhijie, et al. A survey: deep learning for hyperspectral image classification with few labeled samples[J]. Neurocomputing, 2021, 448:179-204.
[33]	ZHAO Chuanpeng, JIA Mingming, WANG Zongming, et al. Toward a better understanding of coastal salt marsh mapping: a case from China using dual-temporal images[J]. Remote Sensing of Environment, 2023, 295:113664.
[34]	ZHANG Xiao, LIU Liangyun, ZHAO Tingting, et al. GWL_FCS30: a global 30 m wetland map with a fine classification system using multi-sourced and time-series remote sensing imagery in 2020[J]. Earth System Science Data, 2023, 15(1):265-293.
[35]	GALL J, LEMPITSKY V. Class-specific Hough forests for object detection[M]//Decision forests for computer vision and medical image analysis. London: Springer London, 2013: 143-157.
[36]	ISHWARAN H, KOGALUR U B, GORODESKI E Z, et al. High-dimensional variable selection for survival data[J]. Journal of the American Statistical Association, 2010, 105(489):205-217.
[37]	MEINSHAUSEN N. Quantile regression forests[J]. Journal of Machine Learning Research, 2006, 7(6):983-999.
[38]	CHAN J C W, PAELINCKX D. Evaluation of random forest and Adaboost tree-based ensemble classification and spectral band selection for ecotope mapping using airborne hyperspectral imagery[J]. Remote Sensing of Environment, 2008, 112(6):2999-3011.
[39]	PRINZIE A, VAN DEN POEL D. Random forests for multiclass classification: random multinomial logit[J]. Expert Systems with Applications, 2008, 34(3):1721-1732.
[40]	BADER-EL-DEN M, TEITEI E, PERRY T. Biased random forest for dealing with the class imbalance problem[J]. IEEE Transactions on Neural Networks and Learning Systems, 2019, 30(7):2163-2172.
[41]	GUO Xiaojiao, ZHANG Chengcai, LUO Weiran, et al. Urban impervious surface extraction based on multi-features and random forest[J]. IEEE Access, 2020, 8:226609-226623.
[42]	霍轩琳, 牛振国, 张波, 等. 高寒湿地分类的遥感特征优选研究 [J]. 遥感学报, 2023, 27(4):1045-1060. DOI: 10.11834/jrs.20222080.
	HUO Xuanlin, NIU Zhenguo, ZHANG Bo, et al. Remote sensing feature selection for alpine wetland classification [J]. National Remote Sensing Bulletin, 2023, 27(4):1045-1060. DOI: 10.11834/jrs.20222080.
[43]	MILLARD K, RICHARDSON M. On the importance of training data sample selection in random forest image classification: a case study in peatland ecosystem mapping[J]. Remote Sensing, 2015, 7(7):8489-8515.
[44]	ZHOU Xinzhi, WEN Haijia, ZHANG Yalan, et al. Landslide susceptibility mapping using hybrid random forest with GeoDetector and RFE for factor optimization[J]. Geoscience Frontiers, 2021, 12(5):101211.
[45]	MENZE B H, KELM B M, MASUCH R, et al. A comparison of random forest and its Gini importance with standard chemometric methods for the feature selection and classification of spectral data[J]. BMC Bioinformatics, 2009, 10:213.
[46]	GUYON I, ELISSEEFF A. An introduction to variable and feature selection[J]. Journal of Machine Learning Research, 2003, 3:1157-1182.
[47]	李景霞, 付碧宏. 江苏盐城滨海湿地系统格局变化及其对丹顶鹤生境的影响[J]. 遥感学报, 2021, 25(12):2507-2519.
	LI Jingxia, FU Bihong. Pattern change of the coastal wetland system and its dynamic impact on the habitat of red-crowned cranes in Yancheng, Jiangsu province[J]. National Remote Sensing Bulletin, 2021, 25(12):2507-2519.
[48]	江碧涛. 我国空间对地观测技术的发展与展望[J]. 测绘学报, 2022, 51(7):1153-1159. DOI: 10.11947/j.AGCS.2022.20220199.
	JIANG Bitao. The development and prospect of China's space earth observation technology[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(7):1153-1159. DOI: 10.11947/j.AGCS.2022.20220199.
[49]	程丽娜, 钟才荣, 李晓燕, 等. Sentinel-2密集时间序列数据和Google Earth Engine的潮间带湿地快速自动分类[J]. 遥感学报, 2022, 26(2):348-357.
	CHENG Lina, ZHONG Cairong, LI Xiaoyan, et al. Rapid and automatic classification of intertidal wetlands based on intensive time series Sentinel-2 images and Google Earth Engine[J]. National Remote Sensing Bulletin, 2022, 26(2):348-357.
[50]	孙伟伟, 杨刚, 陈超, 等. 中国地球观测遥感卫星发展现状及文献分析[J]. 遥感学报, 2020, 24(5):479-510.
	SUN Weiwei, YANG Gang, CHEN Chao, et al. Development status and literature analysis of China's earth observation remote sensing satellites[J]. Journal of Remote Sensing, 2020, 24(5):479-510.
[51]	陈慧欣, 陈超, 张自力, 等. 一种基于Google Earth Engine云平台的潮间带遥感信息提取方法[J]. 自然资源遥感, 2022, 34(4):60-67.
	CHEN Huixin, CHEN Chao, ZHANG Zili, et al. A remote sensing information extraction method for intertidal zones based on Google Earth Engine[J]. Remote Sensing for Natural Resources, 2022, 34(4):60-67.
[52]	HARALICK R M. Statistical and structural approaches to texture[J]. Proceedings of the IEEE, 1979, 67(5):786-804.
[53]	徐涵秋. Landsat 8热红外数据定标参数的变化及其对地表温度反演的影响[J]. 遥感学报, 2016, 20(2):229-235.
	XU Hanqiu. Change of Landsat 8 TIRS calibration parameters and its effect on land surface temperature retrieval[J]. Journal of Remote Sensing, 2016, 20(2):229-235.
[54]	唐新明, 李世金, 李涛, 等. 全球数字高程产品概述[J]. 遥感学报, 2021, 25(1):167-181.
	TANG Xinming, LI Shijin, LI Tao, et al. Review on global digital elevation products[J]. National Remote Sensing Bulletin, 2021, 25(1):167-181.
[55]	夏誉玲, 李小娟, 王涛. 基于数字高程模型的混合流向算法[J]. 测绘学报, 2018, 47(5):683-691. DOI: 10.11947/j.AGCS.2018.20170614.
	XIA Yuling, LI Xiaojuan, WANG Tao. A hybrid flow direction algorithm for water routing on DEMs[J]. Acta Geodaetica et Cartographica Sinica, 2018, 47(5):683-691. DOI: 10.11947/j.AGCS.2018.20170614.
[56]	ABDI H, WILLIAMS L J. Principal component analysis [J]. Wiley Interdisciplinary Reviews: Computational Statistics, 2010, 2(4):433-459.
[57]	BREIMAN L. Random forests[J]. Machine Language, 2001, 45(1):5-32.
[58]	KAVZOGLU T, BILUCAN F. Effects of auxiliary and ancillary data on LULC classification in a heterogeneous environment using optimized random forest algorithm[J]. Earth Science Informatics, 2023, 16(1):415-435.
[59]	PRASETIYOWATI M I, MAULIDEVI N U, SURENDRO K. Determining threshold value on information gain feature selection to increase speed and prediction accuracy of random forest[J]. Journal of Big Data, 2021, 8(1):84.
[60]	王庚泽, 靳海亮, 顾晓鹤, 等. 基于改进分离阈值特征优选的秋季作物遥感分类[J]. 农业机械学报, 2021, 52(2):199-210.
	WANG Gengze, JIN Hailiang, GU Xiaohe, et al. Remote sensing classification of autumn crops based on hybrid feature selection model combining with relief F and improved separability and thresholds[J]. Transactions of the Chinese Society for Agricultural Machinery, 2021, 52(2):199-210.
[61]	UZUN B, BALLI S. A novel method for intrusion detection in computer networks by identifying multivariate outliers and ReliefF feature selection[J]. Neural Computing and Applications, 2022, 34(20):17647-17662.
[62]	LI Zhuo, HE Wei, CHENG Mofan, et al. SinoLC-1: the first 1-meter resolution national-scale land-cover map of China created with the deep learning framework and open-access data [J]. Earth System Science Data Discussions, 2023, 15(11):1-38.
[63]	YANG Jie, HUANG Xin. 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.
[64]	ZHANG Xiao, LIU Liangyun, CHEN Xidong, 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.

特征参数集	特征参数	卫星或数据源
光谱特征参数	Blue（蓝波段）、Green（绿波段）、Red（红波段）、VNIR（可见光-近红外波段）、SWIR1（短波红外波段1）、SWIR2（短波红外波段2）、NDVI（归一化差异植被指数）、EVI（增强植被指数）、RVI（比值植被指数）、SAVI（土壤调整植被指数）、NDBI（归一化差异建筑指数）、MNDWI（改进的归一化差异水体指数）	Sentinel-2
纹理特征参数	ASM（角二阶距）、CONT（对比度）、CORR（相关度）、VAR（方差）、IDM（逆差距）、SAVG（求和平均值）、SENT（总熵）、ENT（熵）、DENT（差分熵）	Sentinel-2
温热特征参数	LST（地表温度）、TIR1（热红外波段1）、TIR2（热红外波段2）	Landsat 8
高程特征参数	DEM（高程）、Slope（坡度）、Aspect（坡向）	SRTMV3
主成分特征参数	PC1（第一主成分分量）、PC2（第二主成分分量）、PC3（第三主成分分量）	Sentinel-2

土地利用类型	定义	样本的数量
建设用地	城乡居民点及工矿交通等用地	120
海洋	盐城东部的黄海区域	50
内陆水体	天然陆地水域、水利设施用地、养殖池塘	70
林地	树木、灌木和其他林地	80
耕地	可正常耕作的农田	150
裸地	地表为土壤或岩石，基本不被植被覆盖	120
滩涂	永久或季节性被浅水覆盖的区域	50
湿地	高潮位和低潮位之间的潮汐淹没区	100

土地利用类型	最高精度	最多特征	RF	CART	SVM	KNN
建设用地	1.000 0	1.000 0	0.985 9	1.000 0	1.000 0	0.956 5
海水	1.000 0	0.967 7	0.967 7	0.967 7	0.967 7	0.914 3
内陆水体	1.000 0	0.975 6	0.976 7	0.977 8	0.758 6	0.950 0
林地	0.958 3	0.938 8	0.913 0	0.938 8	0.938 8	0.938 8
耕地	0.955 6	0.966 3	0.943 8	0.945 1	0.955 6	0.913 0
裸地	0.927 6	0.869 6	0.840 6	0.759 5	0.864 9	0.783 8
滩涂	0.928 6	0.928 6	0.928 6	0.857 1	0.000 0	0.727 3
湿地	0.945 1	0.905 3	0.905 3	0.781 6	0.914 9	0.761 9

模型	总体精度/（%）	Kappa系数
本文模型	96.20	0.955 6
UNet	92.37	0.871 7
DeepLabv3+	93.50	0.889 8