结合遥感林龄因子的亚热带森林蓄积量估算方法

doi:10.11707/j.1001-7488.LYKX20210712

Abstract

Abstract:

Objective: The XGBoost algorithm was applied to establish a remote sensing factor-stock volume model containing forest age, to evaluate the effectiveness of combining the remote sensing estimated forest age factor with the remote sensing factor to improve the accuracy of forest volume estimation, and to provide a new idea and method to achieve efficient, fast and accurate forest volume estimation on a large scale. Method: Taking Jiangle County, Fujian Province as a case, firstly, based on the time-series Landsat images from 1987—2016, combined with the measured stock volume data of subcompartment of forest resource inventory and planning, the LandTrendr forest disturbance and restoration monitoring algorithm was used to monitor the annual stand turnover disturbance and estimate the forest age in the disturbance area; Second, based on the GF-1 image spectral, texture, and topography features, the recursive feature elimination random forest algorithm (RFE-RF) to estimate the forest age in the non-disturbed area; Finally, the GF-1 image spectral and texture factors were combined with the forest age factor by the extreme gradient boosting algorithm (XGBoost) to estimate the forest stock of the study area. The accuracy of forest stock estimation with and without the forest age factor was compared to further verify the importance of remote sensing forest age factor to improve the accuracy of forest stock estimation. Result: The error of forest age obtained by using LandTrendr algorithm in the forest disturbance area was only 1-2 years, and the accuracy of forest age estimation was significantly better than that of the traditional estimation of forest age using remote sensing factors (error of 4-12 years). When only conventional remote sensing factors were used to estimate the volume, the model R² of XGBoost was 0.59 and the average RMSE was 30.72 m³·hm^?2,the rRMSE was 16.46%; after adding the forest age factor, the model R² increased to 0.73, the average RMSE decreased to 23.73 m³·hm^?2, the rRMSE was 13.26%, and the average overall accuracy of the stock volume estimation improved by about 10.4% to 84.4%. Conclusion: The accuracy of the XGBoost algorithm combined with the forest age parameter for estimating the stock volume is close to the requirements of the relevant regulations of forest resources survey, which can provide important technical support for the rapid survey and assessment of forest resources on a large scale.

Key words: forest stock volume, forest age, time series remote sensing, RFE-RF, XGBoost

CLC Number:

Xiaocheng Zhou, Tingting Huang, Yuan Li, Xiangxi Xiao, Hongru Zhu, Yunzhi Chen, Zhiqing Feng. A Method for Estimating Subtropical Forest Stock by Combining Remotely Sensed Forest Age Factors[J]. Scientia Silvae Sinicae, 2023, 59(4): 88-99.

Figures/Tables 14

Fig.1

Fig.2

Fig.3

Table 1

The vegetation index of GF-1 image"

植被指数 Vegetation indices (VIS)	公式 Formulation	植被指数 Vegetation indices (VIS)	公式 Formulation
差值植被指数 Differential vegetation index (DVI)	${\rm{ DVI}} = {\rm{NIR}} - R$	归一化差异绿度指数 Normalized difference green vegetation index (NDGI)	${\rm{NDGI }}= { {\left( {G - R} \right)} \mathord{\left/ {\vphantom { {\left( {G - R} \right)} {\left( {G + R} \right)} } } \right. } {\left( {G + R} \right)} }$
环境植被指数 Environmental vegetation index (EVI)	${\rm{EVI } }= \left[ \left(2.5 \times { {\rm{NIR} } - R}\right) \right] / $ $ {\left( { {\rm{NIR} } + 6 \times R - 7.5 \times B + 1} \right)}$	重归一化植被指数 Renormalized difference vegetation index (RDVI)	${\rm{RDVI}} = {\rm{NIR}} - {R \mathord{\left/ {\vphantom {R {\sqrt {NIR + R} } } } \right. } {\sqrt {{\rm{NIR}} + R} } }$
归一化植被指数 Normalized difference vegetation index (NDVI)	${\rm{ NDVI}} = { {\left( {{\rm{NIR}} - R} \right)} \mathord{\left/ {\vphantom { {\left( {NIR - R} \right)} {\left( {NIR + R} \right)} } } \right. } {\left( {{\rm{NIR}} + R} \right)} }$	红边比值植被指数 Red edge ratio vegetation index (RGRI)	${\rm{RGRI}} = R/G$
红绿植被指数 Green-red normalized difference vegetation index (GRNDVI)	${\rm{GRNDVI }}= { {\left[ {{\rm{NIR}} - \left( {R + G} \right)} \right]} \mathord{\left/ {\vphantom { {\left( {NIR - \left( {R + G} \right)} \right)} {\left( {NIR + R + G} \right)} } } \right. } {\left( {{\rm{NIR}} + R + G} \right)} }$	比值植被指数 Ratio vegetation index (RVI)	${\rm{RVI}} = {\rm{NIR}}/R$
绿化率植被指数 Green red vegetation index (GRVI)	${\rm{GRVI}} = {\rm{NIR}}/G$	转换型植被指数 Transformed normalized difference vegetation index (TNDVI)	${\rm{ TNDVI}} = \sqrt {\left( {0.5 + { {\left( {{\rm{NIR}} - R} \right)} \mathord{\left/ {\vphantom { {\left( {NIR - R} \right)} {\left( {NIR + R} \right)} } } \right. } {\left( {{\rm{NIR}} + R} \right)} } } \right)}$
改进植被指数 Modified simple ratio index (MSR)	${\rm{MSR} } = { {\left[ { { { {\rm{NIR} } } \mathord{\left/ {\vphantom { {NIR} {\left( {R - 1} \right)} } } \right. } {\left( {R - 1} \right)} } } \right]} \mathord{\left/ {\vphantom { {\left( { { {NIR} \mathord{\left/ {\vphantom { {NIR} {\left( {R - 1} \right)} } } \right. } {\left( {R - 1} \right)} } } \right)} {\left( {\sqrt { { { {\rm{NIR} } } \mathord{\left/ {\vphantom { {NIR} R} } \right. } R} } + 1} \right)} } } \right. } {\left( {\sqrt { { {{\rm{NIR}}} \mathord{\left/ {\vphantom { {NIR} R} } \right. } R} } + 1} \right)} }$	GF-1波段GF-1 band	/

Table 1

Fig.4

Table 2

Fig.5

Fig.6

Fig.7

Table 3

Fig.8

Fig.9

Table 4

Fig.10

References 0

	曹　霖, 彭道黎, 王雪军, 等. 应用Sentinel-2A卫星光谱与纹理信息的森林蓄积量估算. 东北林业大学学报, 2018, 46 (9): 54- 58. doi: 10.3969/j.issn.1000-5382.2018.09.012
	Cao L, Peng D L, Wang X J, et al. Estimation of forest stock volume with spectral and textural information from the Sentinel-2A. Journal of Northeast Forestry University, 2018, 46 (9): 54- 58. doi: 10.3969/j.issn.1000-5382.2018.09.012
	霍常富, 王　朋, 陈龙池, 等. 杉木人工林蓄积量和生态系统碳数量成熟龄的关系. 中南林业科技大学学报, 2018, 38 (9): 94- 99. doi: 10.14067/j.cnki.1673-923x.2018.09.015
	Huo C F, Wang P, Chen L C, et al. Quantitative maturity age for stand volume and ecosystem carbon storage of Cunninghamia lanceolata plantation . Journal of Central South University of Forestry & Technology, 2018, 38 (9): 94- 99. doi: 10.14067/j.cnki.1673-923x.2018.09.015
	李洛晞, 沈润平, 李鑫慧, 等. 基于MODIS时间序列森林扰动监测指数比较研究. 遥感技术与应用, 2016, 31 (6): 1083- 1090.
	Li L X, Shen R P, Li X H, et al. Comparison of forest disturbance indices based on MODIS time-series data. Remote Sensing Technology and Application, 2016, 31 (6): 1083- 1090.
	刘笑笑, 王　亮, 徐胜华, 等. 一种后向迭代的森林生物量遥感特征选择方法. 测绘科学, 2017, 42 (5): 100- 105. doi: 10.16251/j.cnki.1009-2307.2017.05.017
	Liu X X, Wang L, Xu S H, et al. A remote sensing feature selection method of forest biomass estimation based on RF-RFE. Science of Surveying and Mapping, 2017, 42 (5): 100- 105. doi: 10.16251/j.cnki.1009-2307.2017.05.017
	罗洪斌, 岳彩荣, 张国飞, 等. 基于哑变量和因子选择的森林蓄积量估算研究. 西北林学院学报, 2022, 37 (1): 205- 210. doi: 10.3969/j.issn.1001-7461.2022.01.30
	Luo H B, Yue C R, Zhang G F, et al. Forest volume estimation based on dummy variables and factor selection. Journal of Northwest Forestry University, 2022, 37 (1): 205- 210. doi: 10.3969/j.issn.1001-7461.2022.01.30
	罗凯健, 许晓东, 龙江平, 等. 结合Landsat 8与PALSAR-2影像的龙南县针叶林蓄积量遥感估算研究. 林业资源管理, 2021, (1): 69- 76.
	Luo K J, Xu X D, Long J P, et al. Research on estimation of coniferous forest volume in Longnan County based on Landsat 8 and PALSAR-2 images. Forest Resources Management, 2021, (1): 69- 76.
	孟宪宇. 2006. 测树学. 北京: 中国林业出版社.
	Meng X Y. 2006. Forest mensuration. Beijing: China Forestry Publishing House.［in Chinese］
	沈文娟, 李明诗. 基于长时间序列Landsat影像的南方人工林干扰与恢复制图分析. 生态学报, 2017, 37 (5): 1438- 1449.
	Shen W J, Li M S. Mapping disturbance and recovery of plantation forests in southern China using yearly Landsat time series observations. Acta Ecologica Sinica, 2017, 37 (5): 1438- 1449.
	苏　迪, 高心丹. 基于无人机航测数据的森林郁闭度和蓄积量估测. 林业工程学报, 2020, 5 (1): 156- 163.
	Su D, Gao X D. Estimation of forest canopy density and stock volume based on UAV aerial survey data. Journal of Forestry Engineering, 2020, 5 (1): 156- 163.
	孙忠秋, 高金萍, 吴发云, 等. 基于机载激光雷达点云和随机森林算法的森林蓄积量估算. 林业科学, 2021, 57 (8): 68- 81. doi: 10.11707/j.1001-7488.20210807
	Sun Z Q, Gao J P, Wu F Y, et al. Estimating forest stock volume via small-footprint LiDAR point cloud data and random forest algorithm. Scientia Silvae Sinicae, 2021, 57 (8): 68- 81. doi: 10.11707/j.1001-7488.20210807
	唐少飞, 田庆久, 徐凯健, 等. Sentinel-2卫星落叶松林龄信息反演. 遥感学报, 2020, 24 (12): 1511- 1524.
	Tang S F, Tian Q J, Xu K J, et al. Age information retrieval of Larix gmelinii forest using Sentinel-2 data . Journal of Remote Sensing, 2020, 24 (12): 1511- 1524.
	徐期瑚, 林丽平, 薛春泉, 等. 基于树高-年龄分级的广东枫香生长模型. 林业资源管理, 2018, (5): 47- 53. doi: 10.13466/j.cnki.lyzygl.2018.05.009
	Xu Q H, Lin L P, Xue C Q, et al. Height-age growth model for Liquidambar formosana in Guangdong using the classified height method . Forest Resources Management, 2018, (5): 47- 53. doi: 10.13466/j.cnki.lyzygl.2018.05.009
	杨　军. 卫星遥感技术在森林资源调查中的运用分析. 林业勘查设计, 2020, 50 (2): 79- 81. doi: 10.3969/j.issn.1673-4505.2020.02.024
	Yang J. Analysis on the application of satellite remote sensing technology in forest resources investigation. Forest Investigation Design, 2020, 50 (2): 79- 81. doi: 10.3969/j.issn.1673-4505.2020.02.024
	杨　柳, 冯仲科, 岳德鹏, 等. 结合纹理因子和地形因子的森林蓄积量多光谱估算模型. 光谱学与光谱分析, 2017, 37 (7): 2140- 2145.
	Yang L, Feng Z K, Yue D P, et al. Forest stock volume estimation model using textural and topographic factors of Landsat8OLI. Spectroscopy and Spectral Analysis, 2017, 37 (7): 2140- 2145.
	张智锋, 张　燕, 汪新岩, 等. 多源遥感数据森林蓄积量估算研究现状、问题及对策. 陕西林业科技, 2020, (6): 68- 73. doi: 10.3969/j.issn.1001-2117.2020.06.016
	Zhang Z F, Zhang Y, Wang X Y, et al. Forest stock estimation based on multi-sourced remote sensing data: status, problems and countermeasures. Shaanxi Forest Science and Technology, 2020, (6): 68- 73. doi: 10.3969/j.issn.1001-2117.2020.06.016
	曾伟生, 孙乡楠, 王六如, 等. 基于机载激光雷达数据的森林蓄积量模型研建. 林业科学, 2021, 57 (2): 31- 38. doi: 10.11707/j.1001-7488.20210204
	Zeng W S, Sun X N, Wang L R, et al. Development of forest stand volume models based on airborne laser scanning data. Scientia Silvae Sinicae, 2021, 57 (2): 31- 38. doi: 10.11707/j.1001-7488.20210204
	钟　莉, 陈芸芝, 汪小钦. 基于Landsat时序数据的森林干扰监测. 林业科学, 2020, 56 (5): 80- 88. doi: 10.11707/j.1001-7488.20200509
	Zhong L, Chen Y Z, Wang X Q. Forest disturbance monitoring based on time series of landsat data. Scientia Silvae Sinicae, 2020, 56 (5): 80- 88. doi: 10.11707/j.1001-7488.20200509
	周小成, 何　艺, 黄洪宇, 等. 基于两期无人机影像的针叶林伐区蓄积量估算. 林业科学, 2019, 55 (11): 117- 125. doi: 10.11707/j.1001-7488.20191113
	Zhou X C, He Y, Huang H Y, et al. Estimation of forest stand volume on coniferous forest cutting area based on two periods unmanned aerial vehicle images. Scientia Silvae Sinicae, 2019, 55 (11): 117- 125. doi: 10.11707/j.1001-7488.20191113
	向安民, 刘凤伶, 于宝义, 等. 基于k-NN方法和GF遥感影像的森林蓄积量估算. 浙江农林大学学报, 2017, 34 (3): 406- 412. doi: 10.11833/j.issn.2095-0756.2017.03.004
	Xiang A M, Liu F L, Yu B Y, et al. Forest stock volume estimation based on the k-NN method and GF remote sensing data. Journal of Zhejiang A & F University, 2017, 34 (3): 406- 412. doi: 10.11833/j.issn.2095-0756.2017.03.004
	Bausch W C, Halvorson A D, Cipra J, et al. Quickbird satellite and ground-based multispectral data correlations with agronomic parameters of irrigated maize grown in small plots. Biosystems Engineering, 2008, 101 (3): 306- 315. doi: 10.1016/j.biosystemseng.2008.09.011
	Bouvier M, Durrieu S, Fournier R A, et al. Generalizing predictive models of forest inventory attributes using an area-based approach with airborne LiDAR data. Remote Sensing of Environment, 2015, 156, 322- 334. doi: 10.1016/j.rse.2014.10.004
	Chen T, Guestrin C. 2016. XGBoost: a scalable tree boosting system. arXiv: 1603.02754. https://arxiv.org/abs/1603.02754.
	Cong X, Bruce M, Justin M. Evaluation of modelling approaches in predicting forest volume and stand age for small-scale plantation forests in New Zealand with RapidEye and LiDAR. International Journal of Applied Earth Observations and Geoinformation, 2018, 73, 386- 396. doi: 10.1016/j.jag.2018.06.021
	Dye M, Mutanga O, Ismail R. Combining spectral and textural remote sensing variables using random forests: predicting the age of Pinus patula forests in KwaZulu-Natal, South Africa . Journal of Spatial Science, 2012, 57 (2): 193- 211. doi: 10.1080/14498596.2012.733620
	Eggers J J, Bauml R, Tzschoppe R, et al. Scalar costa scheme for information embedding. IEEE Transactions on Signal Processing, 2003, 51 (4): 1003- 1019. doi: 10.1109/TSP.2003.809366
	Gang C, Jean-Claude T, Sutee A, et al. Stand age estimation of rubber (Hevea brasiliensis) plantations using an integrated pixel- and object-based tree growth model and annual Landsat time series . ISPRS Journal of Photogrammetry and Remote Sensing, 2018, 144, 94- 104. doi: 10.1016/j.isprsjprs.2018.07.003
	Haralick R M, Shanmugam K, Dinstein I. 1973. Textural features for image classification. IEEE Transactions on Systems, Man, and Cybernetics, SMC-3(6): 610-621.
	Hu Y, Xu X L, Wu F Y, et al. Estimating forest stock volume in Hunan Province, China, by integrating in situ plot data, Sentinel-2 images, and linear and machine learning regression models. Remote Sensing, 2020, 12 (1): 186. doi: 10.3390/rs12010186
	Jensen J R, Qiu F, Ji M H. Predictive modelling of coniferous forest age using statistical and artificial neural network approaches applied to remote sensor data. International Journal of Remote Sensing, 1999, 20 (14): 2805- 2822. doi: 10.1080/014311699211804
	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.
	Kou W L, Xiao X M, Dong J W, et al. Mapping deciduous rubber plantation areas and stand ages with PALSAR and Landsat images. Remote Sensing, 2015, 7 (1): 1048. doi: 10.3390/rs70101048
	Li D, Ju W, Fan W, et al. Estimating the age of deciduous forests in northeast China with enhanced thematic mapper plus data acquired in different phenological seasons. Journal of Applied Remote Sensing, 2014, 8 (10): 21.
	Loboda T V, Chen D. Spatial distribution of young forests and carbon fluxes within recent disturbances in Russia. Global Change Biology, 2017, 23 (1): 138- 153. doi: 10.1111/gcb.13349
	Nilsson M. Estimation of tree heights and stand volume using an airborne LiDAR system. Remote Sensing of Environment, 1996, 56 (1): 1- 7. doi: 10.1016/0034-4257(95)00224-3
	Peng D L, Zhang H L, Liu L Y, et al. Estimating the aboveground biomass for planted forests based on stand age and environmental variables. Remote Sensing, 2019, 11 (19): 2270. doi: 10.3390/rs11192270
	Prokhorenkova L, Gusev G, Vorobev A, et al. 2017. CatBoost: unbiased boosting with categorical features. arXiv: 1706.09516. https://arxiv.org/abs/1706.09516.
	Robert E K, Yang Z Q, Gorelick N, et al. Implementation of the LandTrendr algorithm on google earth engine. Remote Sensing, 2018, 10 (5): 691. doi: 10.3390/rs10050691
	Roy D P, Kovalskyy V, Zhang H K, et al. Characterization of Landsat-7 to Landsat-8 reflective wavelength and normalized difference vegetation index continuity. Remote Sensing of Environment, 2016, 185 (1): 57- 70.
	Shogoro F, Kei-Ichi O, Shogo N, et al. Estimation of the stand ages of tropical secondary forests after shifting cultivation based on the combination of WorldView-2 and time-series Landsat images. ISPRS Journal of Photogrammetry and Remote Sensing, 2016, 119, 280- 293. doi: 10.1016/j.isprsjprs.2016.06.008
	Wunderle A L, Franklin S E, Guo X G. Regenerating boreal forest structure estimation using SPOT-5 pan-sharpened imagery. International Journal of Remote Sensing, 2007, 28 (19): 4351- 4364. doi: 10.1080/01431160701244849
	Zhang Q F, Pavlic G, Chen W J, et al. Deriving stand age distribution in boreal forests using SPOT VEGETATION and NOAA AVHRR imagery. Remote Sensing of Environment, 2004, 91 (3): 405- 418.

误差 Error	个数 Number	百分比 Percentage（%）
<1年 <1 year	85	90.59
=1年 =1 year	112
=2年 =2 year	34
=3年 =3 year	8	9.41
>3年 >3 year	16	9.41

光谱因子 Spectral factors		相关性 Correlation	纹理因子 Texture factors	相关性 Correlation
	G	?0.32^**	GLCM_Homog_NIR	0.24^**
	R	?0.41^**	GLCM_Ang_G	?0.23^**
	TNDVI	0.43^**	GLCM_Entro_B	?0.35^**
	RGRI	0.35^**	GLDV_Entro_NIR	0.12^**
	MSR	0.29^**	GLCM_Corre_B	0.26^**
	GRVI	0.35^**
	EVI	0.41^**
	DVI	0.43^**

基于常规遥感因子蓄积量估算 Estimation of volume based on ordinary remote sensing factors		结合林龄因子蓄积量估算 Estimation of volume by combining forest age parameters
小班平均每公顷蓄积量 Average accumulation per ha in subcompartment/ (m³·hm^?2)	XGBoost估算蓄积量平均精度 XGBoost estimated volume average accuracy(%)	小班平均每公顷蓄积量 Average accumulation per ha in subcompartment/ (m³·hm^?2)	XGBoost估算蓄积量平均精度 XGBoost estimated volume average accuracy
75~150	71.0	75~150	83.0
150~195	76.5	150~195	87.2
>195	74.1	>195	81.3

[1]	Sisheng Luo,Bizhen Luo,Shujing Wei,Haiqing Hu,Xiaochuan Li,Zhenshi Wang,Yufei Zhou,Zhao Song,Yingxia Zhong. Characteristics of Soil Carbon Pool in Pinus massoniana Forest One Year after Moderate Forest Fires [J]. Scientia Silvae Sinicae, 2022, 58(9): 25-35.
[2]	Zhongqiu Sun,Jinping Gao,Fayun Wu,Xianlian Gao,Yang Hu,Jianxin Gao. Estimating Forest Stock Volume via Small-Footprint LiDAR Point Cloud Data and Random Forest Algorithm [J]. Scientia Silvae Sinicae, 2021, 57(8): 68-81.
[3]	Baoguo Yang, Hongyan Jia, Jian Hao, Yunxing Li, Shengjiang Pang, Shiling Liu, Pei Zhang, Changhai Niu, Daoxiong Cai. Growth Variation of Heartwood and Sapwood of Teak (Tectona grandis) Plantations at Different Ages [J]. Scientia Silvae Sinicae, 2020, 56(1): 65-73.
[4]	Liu Yanhui;Wang Yanhui;Yu Pengtao;Xiong Wei;Hao Jia;Zhang Xiaobei;Xu Lihong. Soil Organic Carbon Contents of Larix principis-rupprechtii Plantations in the Southern Part of Liupan Mountains [J]. Scientia Silvae Sinicae, 2012, 48(12): 1-9.
[5]	Li Guolei;Liu Yong;Li Ruisheng;Lü Ruiheng;Xu Yang. Change of Soil Quality Affected by Forest Age of Pinus tabulaeformis Plantations [J]. Scientia Silvae Sinicae, 2008, 44(9): 76-81.

A Method for Estimating Subtropical Forest Stock by Combining Remotely Sensed Forest Age Factors

RichHTML

PDF (PC)

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 14

References 0

Related Articles 5

Recommended Articles

Metrics

Comments