Identification of sub-compartment forest type based on multi-source data and three-tier models

doi:10.12302/j.issn.1000-2006.202109037

JOURNAL OF NANJING FORESTRY UNIVERSITY ›› 2022, Vol. 46 ›› Issue (1): 69-80.doi: 10.12302/j.issn.1000-2006.202109037

Special Issue: 第二届中国林草计算机大会论文精选

Previous Articles Next Articles

Identification of sub-compartment forest type based on multi-source data and three-tier models

HUANG Jian(), WU Dasheng, FANG Luming

College of Mathematics and Computer Science,Zhejiang Agriculture and Forestry University, Key Laboratory of National Forestry and Grassland Administration on Forestry Sensing Technology and Intelligent Equipment,Hangzhou 311300, China

Received:2021-09-19 Accepted:2021-11-02 Online:2022-01-30 Published:2022-02-09

Abstract

Abstract:

【Objective】 Currently, forest type identification research is mainly focused on small forest areas and forest farms. To explore the forest type identification method across a large range, this study used Sentinel-2 optical remote sensing imagery, forest resource survey data, a digital elevation model (DEM), and Sentinel-1 radar remote sensing image data to establish tree species identification model.【Method】 The research area was defined as Chun’an County, where seven forest types were modeled separately, including Phyllosstachys edulis forest, Camellia sinensis forest, Carya cathayensis forest, Cunninghamia lanceolata forest, Pinus massoniana forest, broad-leaved mixed forest, and other hard broad-leaved species. Models were divided into three layers. In the first layer, the RF algorithm was used to establish the identification model of forested land and non-forested land. In the second layer, forest structure was identified from forested land. The RF, XGBoost and LightGBM methods were used to build various models and analyze the experimental results. In the third layer, the forest structures were further divided into forest type.【Result】 The overall accuracy of the first-layer model based on the RF algorithm dividing samples into forested and non-forested land samples was 98.08%. In the second layer (i.e., forest structure recognition model), the performance of the three models under various feature combinations were compared. It was found that the LightGBM model had the highest overall accuracy of 81.43%. In the third layer, the performance indicators for seven forest type models were compared; based on the combination of all features and radar factors, the overall accuracy of the LightGBM model was 84.51%, after feature selection by the recursive feature elimination algorithm, the optimal accuracy was 83.21%.【Conclusion】 The green, red, near-infrared and red-edge bands from optical remote sensing imagery, and terrain factors from a DEM are effective in identifying the forest type. However, independent variables extracted from Sentinel-1 radar do not provide significant help to identify forest type.

Key words: forest type identification, optical remote sensing, radar remote sensing, digital elevation model, multi-layer models

CLC Number:

S771.8

HUANG Jian, WU Dasheng, FANG Luming. Identification of sub-compartment forest type based on multi-source data and three-tier models[J]. JOURNAL OF NANJING FORESTRY UNIVERSITY, 2022, 46(1): 69-80.

Figures/Tables 15

Table 1

Table 2

Fig.1

Table 3

Fig.2

Table 4

Table 5

Fig.3

Fig.4

Fig.5

Fig.6

Table 6

Fig.7

Fig.8

Fig.9

References 29

[1]	樊雪, 刘清旺, 谭炳香. 基于机载PHI高光谱数据的森林优势树种分类研究[J]. 国土资源遥感, 2017, 29(2):110-116.
	FAN X, LIU Q W, TAN B X. Classification of forest species using airborne PHI hyperspectral data[J]. Remote Sens Land Resour, 2017, 29(2):110-116.DOI: 10.6046/gtzyyg.2017.02.16. doi: 10.6046/gtzyyg.2017.02.16
[2]	王佳, 张隆裕, 吕春东, 等. 基于地面激光雷达点云数据的树种识别方法[J]. 农业机械学报, 2018, 49(11):180-188.
	WANG J, ZHANG L Y, LYU C D, et al. Tree species identification methods based on point cloud data using ground-based LiDAR[J]. Trans Chin Soc Agric Mach, 2018, 49(11):180-188.DOI: 10.6041/j.issn.1000-1298.2018.11.021. doi: 10.6041/j.issn.1000-1298.2018.11.021
[3]	李煜, 李崇贵, 刘思涵, 等. 应用哨兵2A多时相遥感影像对树种的识别[J]. 东北林业大学学报, 2021, 49(3):44-47,51.
	LI Y, LI C G, LIU S H, et al. Tree species recognition with sentinel-2A multitemporal remote sensing Image[J]. J Northeast For Univ, 2021, 49(3):44-47,51.DOI: 10.13759/j.cnki.dlxb.2021.03.008. doi: 10.13759/j.cnki.dlxb.2021.03.008
[4]	张悦楠, 房磊, 乔泽宇, 等. 亚热带常绿林型遥感识别及尺度效应[J]. 生态学杂志, 2020, 39(5):1636-1650.
	ZHANG Y N, FANG L, QIAO Z Y, et al. Remote sensing-based identification of forest types and the scale effect in subtropical evergreen forests[J]. Chin J Ecol, 2020, 39(5):1636-1650.DOI: 10.13292/j.1000-4890.202005.016. doi: 10.13292/j.1000-4890.202005.016
[5]	蔡林菲, 吴达胜, 方陆明, 等. 基于XGBoost的高分二号影像树种识别[J]. 林业资源管理, 2019(5):44-51.
	CAI L F, WU D S, FANG L M, et al. Tree species identification using XGBoost based on GF-2 Image[J]. For Resour Manag, 2019(5):44-51.DOI: 10.13466/j.cnki.lyzygl.2019.05.009. doi: 10.13466/j.cnki.lyzygl.2019.05.009
[6]	刘丽娟, 庞勇, 范文义, 等. 机载LiDAR和高光谱融合实现温带天然林树种识别[J]. 遥感学报, 2013, 17(3):679-695.
	LIU L J, PANG Y, FAN W Y, et al. Fused airborne LiDAR and hyperspectral data for tree species identification in a natural temperate forest[J]. J Remote Sens, 2013, 17(3):679-695.
[7]	赵颖慧, 张大力, 甄贞. 基于非参数分类算法和多源遥感数据的单木树种分类[J]. 南京林业大学学报(自然科学版), 2019, 43(5):103-112.
	ZHAO Y H, ZHANG D L, ZHEN Z. Individual tree species classification based on nonparametric classification algorithms and multi-source remote sensing data[J]. J Nanjing For Univ (Nat Sci Ed), 2019, 43(5):103-112.DOI: 10.3969/j.issn.1000-2006.201810041. doi: 10.3969/j.issn.1000-2006.201810041
[8]	王瑞瑞, 李文静, 石伟, 等. 基于多源遥感数据的输电线走廊树种分类[J]. 农业机械学报, 2021, 52(3):226-233.
	WANG R R, LI W J, SHI W, et al. Tree species classification of power line corridor based on multi-source remote sensing Data[J]. Trans Chin Soc Agric Mach, 2021, 52(3):226-233.
[9]	徐逸, 甄佳宁, 蒋侠朋, 等. 无人机遥感与XGBoost的红树林物种分类[J]. 遥感学报, 2021, 25(3):737-752.
	XU Y, ZHEN J N, JIANG X P, et al. Mangrove species classification with UAV-based remote sensing data and XGBoost[J]. Natl Remote Sens Bull, 2021, 25(3):737-752.
[10]	皋厦, 申鑫, 代劲松, 等. 结合LiDAR单木分割和高光谱特征提取的城市森林树种分类[J]. 遥感技术与应用, 2018, 33(6):1073-1083.
	GAO S, SHEN X, DAI J S, et al. Tree species classification in urban forests based on LiDAR point cloud segmentation and hyperspectral metrics extraction[J]. Remote Sens Technol Appl, 2018, 33(6):1073-1083.DOI: 10.11873/j.issn.1004-0323.2018.6.1073. doi: 10.11873/j.issn.1004-0323.2018.6.1073
[11]	PERSSON M, LINDBERG E, REESE H. Tree species classification with multi-temporal sentinel-2 data[J]. Remote Sens, 2018, 10(11):1794.DOI: 10.3390/rs10111794. doi: 10.3390/rs10111794
[12]	DALPONTE M, BRUZZONE L, GIANELLE D. Tree species classification in the Southern Alps based on the fusion of very high geometrical resolution multispectral/hyperspectral images and LiDAR data[J]. Remote Sens Environ, 2012, 123:258-270.DOI: 10.1016/j.rse.2012.03.013. doi: 10.1016/j.rse.2012.03.013
[13]	田颖, 陈卓奇, 惠凤鸣, 等. 欧空局哨兵卫星Sentinel-2A/B数据特征及应用前景分析[J]. 北京师范大学学报(自然科学版), 2019, 55(1):57-65.
	TIAN Y, CHEN Z Q, HUI F M, et al. ESA Sentinel-2A/B satellite:characteristics and applications[J]. J Beijing Norm Univ (Nat Sci), 2019, 55(1):57-65.DOI: 10.16360/j.cnki.jbnuns.2019.01.007. doi: 10.16360/j.cnki.jbnuns.2019.01.007
[14]	邱布布, 徐丽华, 张茂震, 等. 基于Landsat OLI和ETM+的杭州城市绿地地上生物量估算比较研究[J]. 西北林学院学报, 2018, 33(1):225-232.
	QIU B B, XU L H, ZHANG M Z, et al. Estimation of above-ground biomass of urban green land in Hangzhou based on landsat OLI and ETM+ data[J]. J Northwest For Univ, 2018, 33(1):225-232.DOI: 10.3969/j.issn.1001-7461.2018.01.37. doi: 10.3969/j.issn.1001-7461.2018.01.37
[15]	LIM J, KIM K M, KIM E H, et al. Machine learning for tree species classification using sentinel-2 spectral information,crown texture,and environmental variables[J]. Remote Sens, 2020, 12(12):2049.DOI: 10.3390/rs12122049. doi: 10.3390/rs12122049
[16]	傅锋, 王新杰, 汪锦, 等. 高分二号影像树种识别及龄组划分[J]. 国土资源遥感, 2019, 31(2):118-124.
	FU F, WANG X J, WANG J, et al. Tree species and age groups classification based on GF-2 image[J]. Remote Sens Land &Resour, 2019, 31(2):118-124.DOI: 10.6046/gtzyyg.2019.02.17. doi: 10.6046/gtzyyg.2019.02.17
[17]	郝泷, 陈永富, 刘华, 等. 基于纹理信息CART决策树的林芝县森林植被面向对象分类[J]. 遥感技术与应用, 2017, 32(2):386-394.
	HAO ( L /S), CHEN Y F, LIU H, et al. Object-oriented forest classification of Linzhi County based on CART decision tree with texture information[J]. Remote Sens Technol Appl, 2017, 32(2):386-394.DOI: 10.11873/j.issn.1004-0323.2017.2.0386. doi: 10.11873/j.issn.1004-0323.2017.2.0386
[18]	吕杰, 郝宁燕, 李崇贵, 等. 利用随机森林和纹理特征的森林类型识别[J]. 遥感信息, 2017, 32(6):109-114.
	LV J, HAO N Y, LI C G, et al. Identification of forest type based on random forest and texture Characteristics[J]. Remote Sens Inf, 2017, 32(6):109-114.
[19]	杜雨菲, 吴保国, 陈玉玲. 基于机器学习算法的广西桉树适宜性研究[J]. 浙江农林大学学报, 2020, 37(1):122-128.
	DU Y F, WU B G, CHEN Y L. Eucalyptus suitability in Guangxi based on machine learning algorithms[J]. J Zhejiang A&F Univ, 2020, 37(1):122-128.DOI: 10.11833/j.issn.2095-0756.2020.01.016. doi: 10.11833/j.issn.2095-0756.2020.01.016
[20]	孙杰杰, 沈爱华, 黄玉洁, 等. 浙江省大叶榉树生境地群落数量分类与排序[J]. 南京林业大学学报(自然科学版), 2019, 43(4):85-93.
	SUN J J, SHEN A H, HUANG Y J, et al. Quantitative classification and ordination of Zelkova schneideriana habitat in Zhejiang Province[J]. J Nanjing For Univ (Nat Sci Ed), 2019, 43(4):85-93.DOI: 10.3969/j.issn.1000-2006.201809027. doi: 10.3969/j.issn.1000-2006.201809027
[21]	刘博文, 戴永寿, 金久才, 等. 基于空间分布与统计特性的海面远景目标检测方法[J]. 海洋科学, 2018, 42(1):88-92.
	LIU B W, DAI Y S, JIN J C, et al. Marine farsighted target-detection method based on spatial distribution and statistical characteristics[J]. Mar Sci, 2018, 42(1):88-92.DOI: 10.11759/hykx20171011005. doi: 10.11759/hykx20171011005
[22]	林海军, 张绘芳, 高亚琪, 等. 基于马氏距离法的荒漠树种高光谱识别[J]. 光谱学与光谱分析, 2014, 34(12):3358-3362.
	LIN H J, ZHANG H F, GAO Y Q, et al. Mahalanobis distance based hyperspectral characteristic discrimination of leaves of different desert tree species[J]. Spectrosc Spectr Anal, 2014, 34(12):3358-3362.DOI: 10.3964/j.issn.1000-0593(2014)12-3358-05. doi: 10.3964/j.issn.1000-0593(2014)12-3358-05
[23]	陶江玥, 刘丽娟, 庞勇, 等. 基于机载激光雷达和高光谱数据的树种识别方法[J]. 浙江农林大学学报, 2018, 35(2):314-323.
	TAO J Y, LIU L J, PANG Y, et al. Automatic identification of tree species based on airborne LiDAR and hyperspectral data[J]. J Zhejiang A F Univ, 2018, 35(2):314-323.
[24]	陈继龙, 魏雪馨, 刘洋, 等. 基于多时相遥感观测的板栗林分布提取研究[J]. 遥感技术与应用, 2020, 35(5):1226-1236.
	CHEN J L, WEI X X, LIU Y, et al. Extraction of chestnut forest distribution based on multi-temporal remote sensing Observations[J]. Remote Sens Technol Appl, 2020, 35(5):1226-1236.
[25]	张沁雨, 李哲, 夏朝宗, 等. 高分六号遥感卫星新增波段下的树种分类精度分析[J]. 地球信息科学学报, 2019, 21(10):1619-1628. doi: 10.12082/dqxxkx.2019.190116
	ZHANG Q Y, LI Z, XIA C Z, et al. Tree species classification based on the new bands of GF-6 remote sensing satellite[J]. J Geo Inf Sci, 2019, 21(10):1619-1628.
[26]	BOLYN C, MICHEZ A, GAUCHER P, et al. Forest mapping and species composition using supervised per pixel classification of Sentinel-2 imagery[J]. Biotechnol Agron Société et Environnement, 2018, 22(3):172-187.
[27]	TRAN A T, NGUYEN K A, LIOU Y A, et al. Classification and observed seasonal phenology of broadleaf deciduous forests in a tropical region by using multitemporal sentinel-1A and landsat 8 data[J]. Forests, 2021, 12(2):235.DOI: 10.3390/f12020235. doi: 10.3390/f12020235
[28]	WAN H M, TANG Y W, JING L H, et al. Tree species classification of forest stands using multisource remote sensing data[J]. Remote Sens, 2021, 13(1):144.DOI: 10.3390/rs13010144. doi: 10.3390/rs13010144
[29]	胥为, 周云轩, 沈芳, 等. 基于Sentinel-1A雷达影像的崇明东滩芦苇盐沼植被识别提取[J]. 吉林大学学报(地球科学版), 2018, 48(4):1192-1200.
	XU W, ZHOU Y X, SHEN F, et al. Recognition and extraction of Phragmites australis salt marsh vegetation in Chongming tidal flat from sentinel-1A SAR Data[J]. J Jilin Univ (Earth Sci Ed), 2018, 48(4):1192-1200.DOI: 10.13278/j.cnki.jjuese.20170004. doi: 10.13278/j.cnki.jjuese.20170004

Metrics

Comments

www.nldxb.njfu.edu.cn

Edited ＆ Published by Editorial Department of Nanjing Forestry University(Natural Sciences Edition)
Address： No.159 Longpan Road,Nanjing 210037 Jiangsu,P.R.China
E-mail ：xuebao@njfu.edu.cn , xuebao@njfu.com.cn
Telephone +86-25-85428247,85427076
Distributed by China International Book Trading Corporation P.O. Box 399, Beijing ,P.R. China

树种结构 forest structure	样本信息数量 sample size	林型 forest type	样本信息数量 sample size
阔叶林 (含阔叶混交林) broad-leaved forest (including broad- leaved mixed forest)	16 414	山核桃林 Carya cathayensis forest	6 602
		阔叶混交林 broad-leaved mixed forest	4 644
		其他硬阔林 other hard broad- leaved forest	5 168
针叶林 coniferous forest	15 550	马尾松林 Pinus massoniana forest	5 965
针叶林 coniferous forest	15 550	杉木林 Cunninghamia lancelata forest	9 585
经济林 no-woods forest	18 601	毛竹林 Phyllosstachys edulis forest	9 830
经济林 no-woods forest	18 601	茶树林 Camellia sinensis forest	8 771

自变量因子类别 category of independent variable factors	自变量因子名称 name of independent variable factors	量符号或公式 variable or the formula	中文名称 Chinese name
遥感影像波段信息 band information of remote sensing image	B2	B₂	遥感影像像元值
	B3	B₃
	B4	B₄
	B5	B₅
	B6	B₆
	B7	B₇
	B8	B₈
	B8A	B₈_A
	B11	B₁₁
	B12	B₁₂
植被指数 vegetation index	RVI	B₅/B₄	比值植被指数
	EVI	2.5×(B₈-B₄)/(B₈+ 6×B₄-7.5×B₂+1)	增强型植被指数
	DVI	B₈/B₄	差值环境植被指数
	NDVI	(B₈-B₄)/(B₈+B₄)	归一化植被指数
光学波段组合因子 optical band combination factor	B8A_B4	(B_8A-B₄)/(B_8A+B₄)
	B8A_B7	(B_8A-B₇)/(B_8A+B₇)
	B7_B6	(B₇-B₆)/(B₇+B₆)
	B8A_B5	(B_8A-B₅)/(B_8A+B₅)
	B6_B5	(B₆-B₅)/(B₆+B₅)
	B8A_B6	(B_8A-B₆)/(B_8A+B₆)

过采样算法 oversampling algorithm	用户精度/% UA		总体精度/% OA
过采样算法 oversampling algorithm	林地 forest	非林地 non-forest	总体精度/% OA
SMOTE	99.06	93.10	98.70
Borderline-SMOTE	98.28	92.94	97.96
SVM-SMOTE	98.35	92.62	98.00
ADASYN	98.48	94.95	98.27

模型 model	方案 plan	用户精度/% UA			总体精度/% OA	模型训练时间/s model training time
模型 model	方案 plan	阔叶林 (含阔叶混交林) broad-leaved forest (including broad- leaved mixed forest)	针叶林 coniferous forest	经济林 no-woods forest	总体精度/% OA	模型训练时间/s model training time
RF	1	66.34	78.11	80.40	75.13	16.16
	2	66.84	79.16	80.70	75.73	15.15
	3	67.86	79.38	81.21	76.31	18.90
XGBoost	1	70.94	79.74	83.84	78.44	49.45
	2	71.75	80.22	84.02	78.92	52.89
	3	72.19	81.06	85.04	79.70	59.76
LightGBM	1	72.42	80.03	84.07	79.05	1.29
	2	72.80	80.58	85.27	79.79	1.35
	3	73.96	81.71	85.93	80.76	1.45

窗口大小 sliding window size	用户精度/% UA			总体精度/% OA
窗口大小 sliding window size	阔叶林 (含阔叶混交林) broad-leaved mixed forest	针叶林 coniferous forest	经济林 no-woods forest	总体精度/% OA
3×3	73.69	81.25	87.15	80.99
7×7	73.79	82.52	87.22	81.43
11×11	74.04	81.82	86.87	81.17
15×15	73.79	81.71	86.95	81.09
19×19	73.81	81.73	86.57	80.95
23×23	73.90	81.45	86.71	80.95
27×27	73.77	81.21	86.51	80.76
31×31	73.57	80.99	86.19	80.51
35×35	74.00	80.95	86.05	80.59
39×39	74.10	80.80	86.35	80.68
43×43	73.98	81.08	86.32	80.71
47×47	73.94	81.06	86.23	80.66
51×51	74.06	80.64	86.32	80.61

Identification of sub-compartment forest type based on multi-source data and three-tier models

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 15

References 29

Related Articles 3

Recommended Articles

Metrics

Comments

Author

Reader

Reviewer

林型 forest type	雷达遥感因子及特征选择方案 radar remote sensing factor and feature selection scheme
	方案A plan A		方案B plan B		方案C plan C
	用户精度/% UA	生产者精度/% PA	用户精度/% UA	生产者精度/% PA	用户精度/% UA	生产者精度/% PA
山核桃林 Carya cathayensis forest	91.87	92.93	92.32	93.24	92.32	93.05
阔叶混交林 broad-leaved mixed forest	63.36	59.96	62.93	59.79	63.22	58.97
其他硬阔林 other hard broad-leaved forest	60.43	63.00	59.97	63.34	57.97	62.27
杉木林 Cunninghamia lanceolata forest	85.93	81.79	86.28	81.82	86.42	81.47
毛竹林 Phyllosstachys edulis forest	94.86	94.32	94.46	94.19	95.03	93.83
茶树林 Camellia sinensis forest	92.27	92.27	92.76	93.03	92.15	93.31
马尾松林 Pinus massoniana forest	69.77	75.51	69.94	75.79	69.14	75.86
总体精度/% OA	83.08		83.21		82.91
Kappa系数	0.80		0.80		0.80

[1]	ZHAO Ting, BAI Hongying, DENG Chenhui, TA Zhijie. A quantitative divided method for the vegetation vertical belt based on NDVI and DEM: a case study of Taibai Mountain on the south slope [J]. JOURNAL OF NANJING FORESTRY UNIVERSITY, 2023, 47(5): 165-171.
[2]	CAO Lin, ZHU Xingzhou, DAI Jinsong, XU Ziqian, SHE Guanghui. Error analysis of DEM interpolation based on small-footprint airborne LiDAR in subtropical hilly forests [J]. JOURNAL OF NANJING FORESTRY UNIVERSITY, 2014, 38(04): 7-13.
[3]	XU Maosong, LI Kun, XIE Chou, ZHU Song, LUO Hongzhang, ZHANG Fengli, WANG Xuejun, XIA Zhongsheng, DANG Yongfeng. Synergistically monitoring the snowstorm damaged forest with polarimetric SAR and optical remote sensing data [J]. JOURNAL OF NANJING FORESTRY UNIVERSITY, 2014, 38(04): 1-6.