A Preliminary Study on Convolutional Neural Network–Based Classification of Salvinia biloba Raddi Growth Stages

• Brenda Diniz UFMT https://orcid.org/0009-0009-8726-3455
• Wellynton Diniz FASTECH https://orcid.org/0000-0002-6268-1513
• Roselene M. Schneider UFMT https://orcid.org/0000-0002-3167-5997
• Leandro D. Battirola UFMT https://orcid.org/0000-0002-5920-5997
• Ricardo L. T. de Andrade UFMT https://orcid.org/0000-0002-2202-6813

DOI: https://doi.org/10.5753/wcama.2026.21112

Resumo

Salvinia biloba Raddi (Salviniaceae) is used in phytoremediation systems and requires monitoring to avoid pollutant release during senescence. This paper reports a preliminary computer vision approach to classify three growth stages (young, intermediate, and advanced) from images using CNNs. We evaluated DenseNet, ResNet, MobileNet, and a custom CNN with transfer learning and data augmentation on a dataset of 312 labeled images. All models achieved accuracy above 89%, and MobileNet obtained the best performance (95.74% accuracy; 0.95 macro F1-score). Errors occurred only between adjacent stages, suggesting an ordinal pattern. Results indicate that lightweight CNNs can support automated monitoring in phytoremediation contexts.

Referências

Ansari, A. A., Naeem, M., Gill, S. S., and AlZuaibr, F. M. (2020). Phytoremediation of contaminated waters: An eco-friendly technology based on aquatic macrophytes application. Egyptian Journal of Aquatic Research, 46(4):371–376.

Bai, Y. and Bai, X. (2024). Deep learning-based aquatic plant recognition technique and natural ecological aesthetics conservation. Crop Protection, 184:106765.

Boas, L. V., Rodrigues, L., e Silva, R. M., Lima, D., and Moreira, L. (2025). Ai-driven approach for digital agriculture: A case study on coffee leaf disease. In Anais do XVI Workshop de Computação Aplicada à Gestão do Meio Ambiente e Recursos Naturais, pages 1–10, Porto Alegre, RS, Brasil. SBC.

Chairat, S. and Gheewala, S. H. (2024). The conceptual quantitative assessment framework for nature-based solutions (nbs). Nature-Based Solutions, 6:100152.

Choudhury, M. I., Nilsson, J. E., Hylander, S., Hauber, M., Ehde, P. M., Weisner, S. E., and Liess, A. (2024). Enhancing nitrogen removal through macrophyte harvest and installation of woodchips-based floating beds in surface-flow constructed wetlands. Chemosphere, 359:142284.

Francisco, R., Pedroso, G., and Ventura, T. (2024). Aplicação de redes neurais convolucionais para classificação de imagens de estágios de maturação da banana prata catarina. In Anais do XV Workshop de Computação Aplicada à Gestão do Meio Ambiente e Recursos Naturais, pages 141–148, Porto Alegre, RS, Brasil. SBC.

Freitas, F., Solera, K., Ferreira, L. F., Sheng, L. Y., Moreno, M. I. C., Jacinto, M. J., Battirola, L. D., and de Andrade, R. L. T. (2025). Use of salvinia biloba raddi biomass in the remediation of solutions contaminated by nanoparticles and silver ions. Brazilian Journal of Biology, 85:e283123. Open Access.

Kröger, R., Holland, M., Moore, M., and Cooper, C. (2007). Plant senescence: A mechanism for nutrient release in temperate agricultural wetlands. Environmental Pollution, 146(1):114–119.

Levachou, Y. and Stonevičius, E. (2025). Near-infrared reflectance thresholding for macrophyte identification in temperate lakes using sentinel-2. Environmental Conservation, 52(4):239–244.

Machado, T., Rodrigues, L., Travençolo, B., Costa, C., Lima, D., and Moreira, L. (2025). Evaluation of convolutional neural networks for coffee leaf rust classification. In Anais do XVI Workshop de Computação Aplicada à Gestão do Meio Ambiente e Recursos Naturais, pages 11–20, Porto Alegre, RS, Brasil. SBC.

Mancuso, G., Bencresciuto, G. F., Lavrnić, S., and Toscano, A. (2021). Diffuse water pollution from agriculture: A review of nature-based solutions for nitrogen removal and recovery. Water, 13(14).

Palanikkumar, D., Anuradha, T., Ramalingam, J., and Sivaraju, S. (2025). A hybrid machine learning strategy for aquatic plant surveillance in sustainable aqua-ecosystems using iot attributes. Aquaculture, 609:742779.

Patil, M. L., Mehta, V., Sigtia, S., Pandey, G., and Kumari, S. (2025). Decoding water hyacinth growth stages with deep learning. In 2025 International Conference on Next Generation Communication Information Processing (INCIP), pages 692–696.

Pawaiya, A. and Suthar, S. (2025). Dual application of azolla- and wolffia-based floating constructed wetland in sewage treatment and phytomass production under the circular economy model. Cleaner Water, 4:100186.

Wang, J., Wang, W., Xiong, J., Li, L., Zhao, B., Sohail, I., and He, Z. (2021). A constructed wetland system with aquatic macrophytes for cleaning contaminated runoff/storm water from urban area in florida. Journal of Environmental Management, 280:111794.

Zevallos, W. T., Salvatierra, L. M., Loureiro, D. B., Morató, J., and Pérez, L. M. (2018). Evaluation of the autochthonous free-floating macrophyte salvinia biloba raddi for use in the phytoremediation of water contaminated with lead. Desalination and Water Treatment, 103:282–289.