Artificial neural network-based correction of log-scaling measurements collected with optical dendrometer

Authors

DOI:

https://doi.org/10.5380/rf.v56i1.102502

Keywords:

Forest Mensuration, Tapering, Multilayer Perceptrons, Resilient Propagation, Machine Learning

Abstract

The objective of this study was to correct standing-tree volume measurements using artificial neural networks (ANNs). Data were obtained from two forest inventories conducted in Tectona grandis (13.8 ha) and Schizolobium parahyba (46 ha) plantations in Pará, Brazil. Tree volume was measured with the Criterion RD 1000 using the Smalian method, with diameters collected at 0.1 m, 0.7 m, 1.3 m, and every 2 m up to the first live branch. Commercial and total heights, DBH measured with the Criterion, operator-to-tree distance, and species were used as input variables, whereas DBH measured with a tape was used as the output variable. A total of 84 Multilayer Perceptron networks were tested, varying the number of hidden-layer neurons and activation functions. The best-performing networks were selected based on the correlation coefficient and the percent root-mean-square error. Overall, 59% of the networks converged before reaching the maximum number of iterations, and the hyperbolic tangent function converged in 74% of the tests. The ANNs effectively corrected non-sampling errors present in the Criterion measurements. Networks with a single hidden layer performed best, and the logistic activation function yielded superior fits. Evidence of overfitting was identified in one of the top-ranked networks, while underfitting occurred in models with fewer iterations.

Author Biographies

Ingrid Raphaela Cromwell Pereira, Federal Rural University of the Amazon

Undergraduate student in Forestry Engineering at the Federal Rural University of the Amazon (UFRA) and scholarship recipient in the Institutional Program for Scientific Initiation Scholarships (PIBIC) 2020-2021 at the Laboratory for Measurement and Management of Forest Resources (LABFOR), where I work on projects that evaluate the growth and development of forests.

Quinny Soares Rocha, Universidade Estadual da Região Tocantina do Maranhão

Currently, a professor of economics, administration, and forest policy at UEMASUL. Post-doctoral fellow at the Federal Rural University of the Amazon in planted forest management and forest economics. PhD in Forest Science with a thesis on the economic viability of sustainable forest management in the Amazon rainforest. Master's degree in Forest Sciences with research on the economic viability of eucalyptus management subjected to different spacings and thinning weights aiming at multiple products. Specialization in Environmental Management, Licensing, and Auditing from the University of Northern Paraná, addressing the influence of Brazilian legislation on the increase of Conservation Units. Forestry engineer from the Federal University of Minas Gerais with work on the economic viability of eucalyptus plantations. Worked on Degraded Area Recovery Projects, in family farming cooperatives, and as an environmental technician in Agrarian Reform settlements.

Marina Mell Campos Bastos, Federal Rural University of the Amazon

Student in Forestry Engineering (8th semester) at the Federal Rural University of the Amazon (UFRA). Completed an internship at the Forest Resource Measurement and Management Laboratory (LabFor) for 2 years, carrying out a scientific initiation project on dendrometric bands in the first year and a scientific initiation project on decomposition classes in fallen trees in the second year.

Raylon Pereira Maciel, Federal Rural University of the Amazon

Raylon Pereira Maciel holds a degree in Animal Science (2008) from the Federal University of Tocantins, a Master's degree (2010) and a Doctorate (2014) in Tropical Animal Science from the Federal University of Tocantins, where he conducted research on the use of agro-industrial by-products for ruminant feed and the utilization of dairy bulls for meat production. He completed a sandwich doctorate at the Department of Animal Science - University of Florida (2013). He was a PNPD scholarship recipient in the Postgraduate Program in Tropical Animal Science (2014). Currently, he is an Adjunct Professor IV at the Federal Rural University of the Amazon, in the area of ​​Forage and Ruminant Production, and the local coordinator of the Integrated Postgraduate Program in Animal Science in the Tropics (UFNT/UFRA), UFRA/Parauapebas-PA Campus.

Lina Bufalino, Federal Rural University of the Amazon

Possui graduação em Engenharia Florestal pela Universidade Federal de Lavras (UFLA) (2003-2008) e mestrado (2008-2010) e doutorado (2010-2014) em Ciência e Tecnologia da Madeira pela UFLA. Atua principalmente nas áreas de tecnologia de produtos florestais madeireiros e não-madeireiros, bioenergia, bioprodutos, fibras e celulose. Foi professora Adjunta I no curso de Engenharia Florestal da Universidade do Estado do Amapá (UEAP) por três anos (2014-2017) e atualmente é professora Adjunta do curso de Engenharia Florestal da Universidade Federal Rural da Amazônia (UFRA). É professora permanente de três programas de pós-graduação: Ciências Florestais (PPGCF) da UFRA; Rede BIONORTE da Universidade Federal do Amazonas; e Ciências Ambientais da Universidade Federal do Amapá. É editora de área do periódico Cerne da UFRA (2015-2017). É membro do Biodiversity Research Consortium Brazil-Norway. Tem mais de 100 artigos publicados em periódicos nacionais e internacionais. Teve experiência no desenvolvimento de pesquisas em rede pela Rede Brasileira de Compósitos e Nanocompósitos (RELIGAR) da UFLA. Teve as seguintes experiências em gestão universitária: chefe de divisão de pesquisa (2014) - UEAP; chefe de divisão de pós-graduação (2016-2017) - UEAP; membro do comitê de pós-graduação e de iniciação científica (2014-2017); comissão de revisão de regimento acadêmico (2017) - UEAP; núcleo docente estruturante (2015-2016) UEAP; e coordenadora do PPGCF-UFRA (2017-2022)

Rodrigo Geroni Mendes Nascimento, Federal Rural University of the Amazon

He is a professor at the Institute of Agricultural Sciences of the Federal Rural University of the Amazon (UFRA). With a degree in Forestry Engineering, and a master's, doctorate, and post-doctorate in forest measurement and management, he dedicates himself primarily to growth and production modeling in tropical forests, as well as research on forest dynamics in the Amazon. He teaches undergraduate and graduate courses and supervises studies in measurement, management, and economics of biological assets. His work seeks to integrate the management of forest ecosystems and socioeconomic needs, offering support for forest practices and public policies. He collaborates with national and international institutions and reviews articles in specialized journals.

References

AMORIM, D. C. R. Quantificação de biomassa e altura das árvores em uma floresta de mata atlântica em Belo Horizonte - MG, comparação entre levantamento Laser Scanner e inventário florestal por área fixa. Revista Gestão & Sustentabilidade Ambiental, Palhoça, v. 9, n. 3, p. 711, 2020.

ARAÚJO, E. F.; FERNANDES, M. M.; SILVA, J. P. M.; KUNZ, S. H.; FERNANDES, M. R. de M. Estimativa da altura de espécies florestais em regeneração natural utilizando redes neurais artificiais. Revista de Ciências Ambientais, Canoas, v. 14, n. 3, p. 27–37, 2020.

BIAZATTI, S. C.; MÔRA, R.; SCCOTI, M. S. V.; BRITO JÚNIOR, J. F. de; QUEIROZ, N. dos S.; CURTO, R. de A. Criterion dendrometer as a non-destructive method for dendrometric estimations of native species in western Amazon. Revista Ibero-Americana de Ciências Ambientais, Aracaju, v. 11, n. 6, p. 59–70, 2020.

BUENO, G. F.; COSTA, E. A.; CRISTINA, A. N.; SOARES, A. A. V.; MIRANDA, R. O. V. de; SCHONS, C. T. Efeito do número de neurônios na camada oculta para relações hipsométricas de eucalipto usando redes neurais artificiais. BIOFIX Scientific Journal, Curitiba, v. 5, n. 2, p. 222–230, 2020.

CHAMMA, W. D. S.; BATISTELLA, D.; CRISIGIOVANNI, E. L.; VICTORINO, H. da S.; LIMA, V. A. Aprendizado de máquina aplicado em imagens de satélite para classificação de telhados. Brazilian Journal of Development, São José dos Pinhais, v. 7, n. 7, p. 72558–72576, 2021.

CURTO, R. D. A.; LAURO, A. C.; TONINI, H.; KOHLER, S. V.; ARAÚJO, E. J. G. de; BIAZATTI, S. C. Cubagem de árvores em pé com dendrômetro óptico em sistema de integração lavoura-pecuária-floresta. Pesquisa Florestal Brasileira, Colombo, v. 39, n. 1, p. 1–11, 2019.

DANTAS, D.; PINTO, L. O. R.; GONÇALVES, A. F. A.; TERRA, M. de C. N. S.; CALEGÁRIO, N. Predição volumétrica por meio da Krigagem pontual reduz o esforço de amostragem em inventários florestais pré-corte. Caderno de Ciências Agrárias, Montes Claros, v. 12, p. 1–9, 2020.

DONGARE, A. D.; KHARDE, R. R.; KACHARE, A. D. Introduction to artificial neural network. International Journal of Engineering and Innovative Technology, Huwei, v. 2, n. 1, p. 189–194, 2012.

DREW, P. J.; MONSON, J. R. T.; KINGDOM, U. Artificial Neural Networks. Surgery, Philadelphia, v. 127, p. 3–11, 2000.

DUARTE, L. T.; SILVA, D. B. do N.; BRITO, J. A. de M. Análise de paradados do censo demográfico 2010: uma investigação de fatores associados a erros não amostrais do levantamento de dados. Revista Brasileira de Estudos de População, Rio de Janeiro, v. 33, n. 3, p. 679–701, 2016.

FRITSCH, S.; GUENTHER, F.; WRIGHT, M. N.; SULING, M.; MUELLER, S. M. Package “neuralnet”: Training of neural networks, 2019.

JHA, S. YANG, S.; BRANDEIS, T. J.; KUEGLER, O.; MARCANO-VEGA, H. Evaluation of regression methods and competition indices in characterizing height-diameter relationships for temperate and pantropical tree species. Frontiers in Forests and Global Change, Lausanne, v. 6, p. e1282297, 2023.

KROGH, A. What are artificial neural networks? Nature Biotechnology, Pittsburgh, v. 26, n. 2, p. 195–197, 2008.

LEAL, F. A.; LEAL, G. da S. A.; SILVA, T. C. da. Redes Neurais Artificiais e Modelos Alométricos Aplicados Para Estimativa de Volume e Altura Em Eucalyptus Urophylla S.T.Blacke. Advances in Forestry Science, Cuiabá, v. 7, n. 3, p. 1181–1188, 2020.

MELO, J. A. F. de; ROCHA, K. da S. Aplicação de redes neurais artificiais na classificação e zoneamento de focos de calor na reserva extrativista chico mendes - Acre, Brasil. Observatório de la Economía Latinoamericana, Curitiba, v. 21, n. 10, p. 16740–16763, 2023.

NASCIMENTO, R. L. X.; SOUZA, C. C. de; GRASSI, G. OLIVEIRA, M. A. das N. de. Caderno de caracterização do estado do Pará. Brasília: Codevasf, 2022, 146 p.

NICOLETTI, M. F.; MACHADO, S. do A.; CARVALHO, S. de P. C. e; FIGUEIREDO FILHO, A.; SANQUETTA, C. R. Descrição do perfil do fuste de árvores por meio das funções de afilamento. Scientia Agraria Paranaensis, Marechal Cândido Rondon, v. 15, n. 4, p. 365–372, 2016.

RAFIQ, M. Y.; BUGMANN, G.; EASTERBROOK, D. J. Neural network design for engineering applications. Computers and Structures, England, v. 79, n. 17, p. 1541–1552, 2001.

REIS, L. P.; SOUZA, A. L. de; REIS, P. C. M. dos R.; MAZZEI, L. BINOTI, D. H. B.; LEITE, H. G. Diameter structure in a community of shrub-tree caatinga, municipality of Floresta, state of Pernambuco, Brazil. Floresta, Curitiba, v. 48, n. 1, p. 133–142, 2018.

SIMON, G.; ALIFERIS, C. Artificial intelligence and machine learning in health care and medical sciences: best practices and pitfalls. Cham: Springer, 2024, 810 P.

SYDENHAM, P. H.; THORN, R. Handbook of measuring system design. 3rd. ed. Stillwater: Wiley, 2005, 1648 p.

WANG, S.-C. Interdisciplinary computing in Java programming language. Amsterdam: Kluwer Academic Publishers, 2003, 266 p.

WICKHAM, H.; CHANG, W.; HENRY, L.; PEDERSEN T. L.; TAKAHASHI, K. WILKE, C.; WOO, K.; YUTANI, H.; DUNNINGTON, D.; BRAND, T. van den. Package "ggplot2": create elegant data visualisations using the grammar of graphics, 2019.

WU, Y. C.; FENG, J. W. Development and application of artificial neural network. Wireless Personal Communications, Berlin, v. 102, n. 2, p. 1645–1656, 2018.

YANG, G. R.; WANG, X. J. Artificial neural networks for neuroscientists: a primer. Neuron, Maryland Heights, v. 107, n. 6, p. 1048–1070, 2020.

Published

2026-04-22

How to Cite

Pereira, I. R. C., Rocha, Q. S., Bastos, M. M. C., Maciel, R. P., Bufalino, L., & Nascimento, R. G. M. (2026). Artificial neural network-based correction of log-scaling measurements collected with optical dendrometer. Floresta, 56(1), e102502. https://doi.org/10.5380/rf.v56i1.102502

Issue

Section

Artigos