基于机器学习算法的樟子松立木材积预测

孙铭辰; 姜立春

doi:10.12302/j.issn.1000-2006.202104014

基于机器学习算法的樟子松立木材积预测

孙铭辰, 姜立春

南京林业大学学报（自然科学版） ›› 2023, Vol. 47 ›› Issue (1) : 31-37.

PDF(1588 KB)

国家林草科技领军期刊
中国精品科技期刊
中国高校百佳科技期刊
江苏省新闻出版政府奖期刊奖
RCCSE林学权威期刊（A+）
CSCD核心期刊
Scopus数据库收录期刊
中文核心期刊
SCD核心期刊

作者加群：102861116

微信公众号：南京林业大学学报

高级检索

PDF(1588 KB)

南京林业大学学报（自然科学版） ›› 2023, Vol. 47 ›› Issue (1) : 31-37. DOI: 10.12302/j.issn.1000-2006.202104014

专题报道Ⅰ:智慧林业之森林参数遥感估测

基于机器学习算法的樟子松立木材积预测

孙铭辰 ,
姜立春 ^*

作者信息 +

Standing volume prediction of Pinus sylvestris var. mongolica based on machine learning algorithm

SUN Mingchen ,
JIANG Lichun ^*

Author information +

文章历史 +

摘要

【目的】 通过非线性和多种机器学习算法构建并对比不同的立木材积模型,为樟子松(Pinus sylvestris var. mongolica)立木材积的精准预测提供理论依据。【方法】 以大兴安岭图强林业局184株樟子松伐倒木数据为基础,建立非线性二元材积模型(NLR),并通过十折交叉检验和袋外数据(OOB)误差检验的方法得到3种最优机器学习算法,包括:反向神经网络(BP)、ε-支持向量回归(ε-SVR)和随机森林(RF)。对比分析不同模型间的差异,得到最优立木材积模型。【结果】 机器学习算法在立木材积的拟合和预测中均优于传统二元材积模型,具体拟合结果排序为RF>BP>ε-SVR> NLR。其中RF的决定系数(R²)比传统模型的提高了2.00%,均方根误差(RMSE)、相对均方根误差(RMSE%)、平均绝对误差(MAE)分别降低了22.90%、22.93%、36.34%,且与真实值相比平均相对误差(MRB)的绝对值更低,证明了RF在立木材积预测中的优越性。【结论】 机器学习算法作为一种新兴的建模方法可以有效地提高立木材积的预测精度,为森林资源的精准调查和经营管理提供新的解决方案。

Abstract

【Objective】 Using various, nonlinear machine learning algorithms, different volume models were constructed and compared to provide a theoretical basis for the accurate prediction of the volume of Pinus sylvestris var. mongolica.【Method】 A total of 184 felled Pinus sylvestris var. mongolica trees in the Tuqiang Forestry Bureau of the Greater Khingan Mountains were used to establish a nonlinear binary volume model (NLR). Three optimal machine learning algorithms were obtained using the K-fold cross test and OOB error test, including back propagation neural network (BP), ε-support vector regression (ε-SVR), and random forest (RF). An optimal volume model was obtained by comparing and analyzing the differences between the different models. 【Result】 The results showed that the machine learning algorithm was superior to the traditional binary volume model in the fitting and prediction of standing volume, and the specific order was RF > BP > ε-SVR > NLR. Compared with the traditional model, the R² of RF increased by 2.00%; the RMSE, RMSE% and MAE decreased by 22.95%, 22.93% and 36.34%, respectively; and the absolute value of MRB was lower than the real value, which proved the superiority of RF in volume prediction. 【Conclusion】 Machine learning algorithms can effectively improve the accuracy at which standing volume can be predicted, providing a new solution for the accurate investigation and management of forest resources.

导出引用

孙铭辰, 姜立春. 基于机器学习算法的樟子松立木材积预测[J]. 南京林业大学学报（自然科学版）. 2023, 47(1): 31-37 https://doi.org/10.12302/j.issn.1000-2006.202104014

SUN Mingchen, JIANG Lichun. Standing volume prediction of Pinus sylvestris var. mongolica based on machine learning algorithm[J]. JOURNAL OF NANJING FORESTRY UNIVERSITY. 2023, 47(1): 31-37 https://doi.org/10.12302/j.issn.1000-2006.202104014

中图分类号： S791.253

参考文献

列表( 原文顺序 | 文献年度倒序 | 文中引用次数倒序 ) 可视化分析

[1]

MAO

, QIN

L J

, HAO

M Y

, et al. An improved approach to estimate above-ground volume and biomass of desert shrub communities based on UAV RGB images[J]. Ecol Indic, 2021, 125: 107494. DOI:10.1016/j.ecolind.2021.107494.

https://doi.org/10.1016/j.ecolind.2021.107494

https://linkinghub.elsevier.com/retrieve/pii/S1470160X2100159X

本文引用 [3]

[2]	MUUKKONEN P. Generalized allometric volume and biomass equations for some tree species in Europe[J]. Eur J Forest Res, 2007, 126(2): 157-166. DOI:10.1007/s10342-007-0168-4. https://doi.org/10.1007/s10342-007-0168-4 http://link.springer.com/10.1007/s10342-007-0168-4 本文引用 [2]

[3]

廖祖辉. 福建桉树人工林材积表和蓄积量表编制的研究[J]. 福建林业科技, 2005, 32(2):17-20.

LIAO

Z H

. Studies of the volume table and stocking table compilations of Eucalyptus plantations[J]. J Fujian For Sci Technol, 2005, 32(2): 17-20. DOI:10.13428/j.cnki.fjlk.2005.02.005.

https://doi.org/10.13428/j.cnki.fjlk.2005.02.005

本文引用 [1]

[4]

吕勇, 刘辉, 王才喜. 杉木林分蓄积量不同测定方法的比较[J]. 中南林学院学报, 2001, 21(4): 50-53.

, LÜ

, WANG

C X

. Comparison of the different measurement methods of stand volume[J]. J Central South For Univ, 2001, 21(4): 50-53. DOI:10.3969/j.issn.1673-923X.2001.04.006.

https://doi.org/10.3969/j.issn.1673-923X.2001.04.006

http://link.springer.com/10.1007/s11771-014-1914-7

本文引用 [1]

[5]

WIANT

H V JR

, WOOD

G B

, WILLIAMS

. Comparison of three modern methods for estmating volume of sample trees using one or two diameter measurements[J]. For Ecol Manag, 1996, 83(1/2): 13-16. DOI:10.1016/0378-1127(96)03708-5.

https://doi.org/10.1016/0378-1127(96)03708-5

https://linkinghub.elsevier.com/retrieve/pii/0378112796037085

本文引用 [1]

[6]	OZCELIK R, WIANT H, BROOKS J. Accuracy using xylometry of log volume estimates for two tree species in Turkey[J]. Scan J Forest Res, 2008, 23(3): 272-277. DOI: 10.1080/02827580801995323. https://doi.org/10.1080/02827580801995323 本文引用 [1]

[7]	曾伟生. 杉木相容性立木材积表系列模型研建[J]. 林业科学研究, 2014, 27(1): 6-10. ZENG W S. Establishment of compatible tree volume equation systems of Chinese fir[J]. For Res, 2014, 27(1): 6-10. DOI:10.13275/j.cnki.lykxyj.2014.01.002. https://doi.org/10.13275/j.cnki.lykxyj.2014.01.002 本文引用 [1]

[8]	许晴, 李晓莎, 许中旗, 等. 塞罕坝地区樟子松立木材积表研究[J]. 林业资源管理, 2017(1): 57-62. XU Q, LI X S, XU Z Q, et al. Study on the volume table of scotch pine in Saihanba aera[J]. For Resour Manag, 2017(1): 57-62. DOI:10.13466/j.cnki.lyzygl.2017.01.011. https://doi.org/10.13466/j.cnki.lyzygl.2017.01.011 本文引用 [1]

[9]	雷相东. 机器学习算法在森林生长收获预估中的应用[J]. 北京林业大学学报, 2019, 41(12): 23-36. LEI X D. Applications of machine learning algorithms in forest growth and yield prediction[J]. J Beijing For Univ, 2019, 41(12): 23-36. DOI:10.12171/j.1000-1522.20190356. https://doi.org/10.12171/j.1000-1522.20190356 本文引用 [2]

[10]

CUTLER

D R

, EDWARDS

T C J

, BEARD

K H

, et al. Random forests for classification in ecology[J]. Ecology, 2007, 88(11): 2783-2792. DOI:10.1890/07-0539.1.

https://doi.org/10.1890/07-0539.1

https://www.ncbi.nlm.nih.gov/pubmed/18051647

本文引用 [1] 摘要

Classification procedures are some of the most widely used statistical methods in ecology. Random forests (RF) is a new and powerful statistical classifier that is well established in other disciplines but is relatively unknown in ecology. Advantages of RF compared to other statistical classifiers include (1) very high classification accuracy; (2) a novel method of determining variable importance; (3) ability to model complex interactions among predictor variables; (4) flexibility to perform several types of statistical data analysis, including regression, classification, survival analysis, and unsupervised learning; and (5) an algorithm for imputing missing values. We compared the accuracies of RF and four other commonly used statistical classifiers using data on invasive plant species presence in Lava Beds National Monument, California, USA, rare lichen species presence in the Pacific Northwest, USA, and nest sites for cavity nesting birds in the Uinta Mountains, Utah, USA. We observed high classification accuracy in all applications as measured by cross-validation and, in the case of the lichen data, by independent test data, when comparing RF to other common classification methods. We also observed that the variables that RF identified as most important for classifying invasive plant species coincided with expectations based on the literature.

[11]	GUAN B T, GERTNER G. Modeling red pine tree survival with an artificial neural network[J]. For Sci, 1991, 37(5): 1429-1440. DOI:10.1093/forestscience/37.5.1429. https://doi.org/10.1093/forestscience/37.5.1429 本文引用 [2]

[12]	MARIA J D. Artificial neural networks as an alternative tool in pine bark volume estimation[J]. Comput Electron Agric, 2005, 48(3): 235-244. DOI:10.1016/j.compag.2005.04.002. https://doi.org/10.1016/j.compag.2005.04.002 https://linkinghub.elsevier.com/retrieve/pii/S0168169905000839 本文引用 [1]

[13]

MARIA

J D

, ÖZÇELIK

, YAVUZ

. Tree-bark volume prediction via machine learning: a case study based on black alder’s tree-bark production[J]. Comput Electron Agric, 2018, 151: 431-440. DOI:10.1016/j.compag.2018.06.039.

https://doi.org/10.1016/j.compag.2018.06.039

https://linkinghub.elsevier.com/retrieve/pii/S0168169917311766

本文引用 [1]

[14]	COLIN J G, JONGHO I. Forest biomass estimation from airborne LiDAR data using machine learning approaches[J]. Remote Sens Environ, 2012, 125: 80-91. DOI: 10.1016/j.rse.2012.07.006. https://doi.org/10.1016/j.rse.2012.07.006 https://linkinghub.elsevier.com/retrieve/pii/S0034425712002787 本文引用 [1]

[15]

MARIA

J D

, MILIOS

. Modelling total volume of dominant pine trees in reforestations via multivariate analysis and artificial neural network models[J]. Biosyst Eng, 2010, 105(3): 306-315. DOI: 10.1016/j.biosystemseng.2009.11.010.

https://doi.org/10.1016/j.biosystemseng.2009.11.010

https://linkinghub.elsevier.com/retrieve/pii/S1537511009003420

本文引用 [2]

[16]

ÖZCELIK

, MARIA

J D

, BROOKS

J R

, et al. Estimating tree bole volume using artificial neural network models for four species in Turkey[J]. J Environ Manag, 2010, 91(3): 742-753. DOI: 10.1016/j.jenvman.2009.10.002.

https://doi.org/10.1016/j.jenvman.2009.10.002

https://linkinghub.elsevier.com/retrieve/pii/S0301479709003430

本文引用 [1]

[17]

J W

, YAO

, CHOI

, et al. A comparative study of predicting DBH and stem volume of individual trees in a temperate forest using airborne waveform LiDAR[J]. IEEE Geosci Remote Sens Lett, 2015, 12(11): 2267-2271. DOI:10.1109/LGRS.2015.2466464.

https://doi.org/10.1109/LGRS.2015.2466464

http://ieeexplore.ieee.org/document/7229269/

本文引用 [3]

[18]	MUSHAR S H, AHMAD S S, KASMIN F, et al. Machine learning approach for estimating tree volume[J]. J Phys: Conf Ser, 2020, 1502(1): 012039. DOI:10.1088/1742-6596/1502/1/012039. https://doi.org/10.1088/1742-6596/1502/1/012039 https://doi.org/10.1088/1742-6596/1502/1/012039 本文引用 [2]

[19]

BHERING

L L

, CRUZ

C D

, DE

AZEVEDO P L

, et al. Application of neural networks to predict volume in eucalyptus[J]. Crop Breed Appl Biotechnol, 2015, 15(3): 125-131. DOI:10.1590/1984-70332015v15n3a23.

https://doi.org/10.1590/1984-70332015v15n3a23

http://www.scielo.br/scielo.php?script=sci_arttext&pid=S1984-70332015000300125&lng=en&tlng=en

本文引用 [3]

[20]

靳晓东, 姜立春. 基于树干不同形率的樟子松立木材积方程研建[J]. 北京林业大学学报, 2020, 42(3):78-86.

JIN

X D

, JIANG

L C

. Equation construction on standing tree volume of Pinus sylvestris var. mongolica based on different form quotients of trunk[J]. J Beijing For Univ, 2020, 42(3):78-86. DOI:10.12171/j.1000-1522.20190047.

https://doi.org/10.12171/j.1000-1522.20190047

本文引用 [1]

[21]	张明铁. 单株立木材积测定方法的研究[J]. 林业资源管理, 2004(1):24-26. ZHANG M T. Study on volume measurement of single trees[J]. For Resour Manag, 2004(1): 24-26. DOI:10.13466/j.cnki.lyzygl.2004.01.005. https://doi.org/10.13466/j.cnki.lyzygl.2004.01.005 本文引用 [1]

[22]	RUMELHART D E, HINTON G E, WILLIAMS R J. Learning internal representations by error propagation[J]. Read Cognit Sci, 1988, 323(6088):399-421. DOI:10.1016/B978-1-4832-1446-7.50035-2. https://doi.org/10.1016/B978-1-4832-1446-7.50035-2 本文引用 [1]

[23]

王轶夫, 孙玉军, 郭孝玉. 基于BP神经网络的马尾松立木生物量模型研究[J]. 北京林业大学学报, 2013, 35(2): 17-21.

WANG

Y F

, SUN

Y J

, GUO

X Y

. Single-tree biomass modeling of Pinus massoniana based on BP neural network[J]. J Beijing For Univ, 2013, 35(2): 17-21. DOI: 10.13332/j.1000-1522.2013.02.018.

https://doi.org/10.13332/j.1000-1522.2013.02.018

本文引用 [1]

[24]	徐晓明. SVM参数寻优及其在分类中的应用[D]. 大连: 大连海事大学, 2014. XU X M. SVM parameter optimization and its application in the classification[D]. Dalian: Dalian Maritime University, 2014. 本文引用 [1]

[25]	曹正凤. 随机森林算法优化研究[D]. 北京: 首都经济贸易大学, 2014. CAO Z F. Study on optimization of random forests algorithm[D]. Beijing: Capital University of Economics and Business, 2014. 本文引用 [1]

[26]	BREIMAN L. Bagging predictors[J]. Mach Learn, 1996, 24(2): 123-140. DOI:10.1007/bf00058655. https://doi.org/10.1007/bf00058655 本文引用 [1]

[27]	DING S, CHANG X H, WU Q H. A study on approximation performances of improved BP neural networks based on LM algorithms[J]. Appl Mech Mater, 2013, 411/412/413/414: 1935-1938. DOI:10.4028/www.scientific.net/amm.411-414.1935. https://doi.org/10.4028/www.scientific.net/amm.411-414.1935 本文引用 [1]

[28]

杜军岗, 魏汝祥, 刘宝平. 基于PSO优化LS-SVM的小样本非线性协整检验与建模研究[J]. 系统工程理论与实践, 2014, 34(9): 2322-2331.

https://doi.org/10.12011/1000-6788(2014)9-2322

摘要

针对小样本非线性时间序列，根据非线性协整的定义，利用基于粒子群优化最小二乘支持向量机的方法，对小样本非线性协整关系检验与非线性误差修正模型建模进行研究，设计了方法的逻辑流程. 对舰船维修费指数与物价指数进行实证研究，在协整关系类型判断的基础上，实现了小样本非线性协整关系的检验，建立了预测舰船维修费指数的非线性误差修正模型，并与线性向量自回归模型进行分析比较. 研究表明：基于粒子群优化最小二乘支持向量机的小样本非线性协整检验与建模方法，刻画了小样本系统的非线性协整关系，所建立的非线性误差修正模型具有较好的预测效果，能够有效地预测小样本非线性系统.

J G

, WEI

R X

, LIU

B P

. Nonlinear cointegration test and error correction modeling based on LS-SVM optimized by PSO in small sample[J]. Syst Eng-theory Pract, 2014, 34(9): 2322-2331. DOI: CNKI:SUN:XTLL.0.2014-09-015.

https://doi.org/CNKI:SUN:XTLL.0.2014-09-015.

本文引用 [1]

[29]	HAJAR A S, AZIZAH A N, HASZLINNA M N. Adoption of machine learning techniques in software effort estimation: an overview[J]. IOP Conf Ser: Mater Sci Eng, 2019, 551(1): 012074. DOI:10.1088/1757-899x/551/1/012074. 本文引用 [1]