Semantic Segmentation of Coffee Regions From Satellite Images Using Deep Learning

João Marcelo Ribeiro; Vitor Augusto Negrão de Assis; Lais Bueno Palos; Jorge Otávio de Souza Nogueira; Jhonathas Emiliano Candido; André Ricardo Backes

doi:10.22456/2175-2745.150916

Authors

João Marcelo Ribeiro Instituto Federal de Educação, Ciência e Tecnologia do Sul de Minas (IFSULDEMINAS)
Vitor Augusto Negrão de Assis Instituto Federal de Educação, Ciência e Tecnologia do Sul de Minas (IFSULDEMINAS)
Lais Bueno Palos Instituto Federal de Educação, Ciência e Tecnologia do Sul de Minas (IFSULDEMINAS)
Jorge Otávio de Souza Nogueira Instituto Federal de Educação, Ciência e Tecnologia do Sul de Minas (IFSULDEMINAS)
Jhonathas Emiliano Candido Instituto Federal de Educação, Ciência e Tecnologia do Sul de Minas (IFSULDEMINAS)
André Ricardo Backes Universidade Federal de São Carlos (UFSCar) https://orcid.org/0000-0002-7486-4253

DOI:

https://doi.org/10.22456/2175-2745.150916

Keywords:

coffee cultivation, remote sensing, convolutional neural networks (CNNs), semantic segmentation

Abstract

Coffee is one of the main agricultural products exported by Brazil, with Minas Gerais standing out as the largest producer and responsible for a significant portion of the country's production. Due to its importance, this work investigates various Convolutional Neural Networks (CNNs) for mapping coffee plantation areas in the municipality of Muzambinho, MG. Using satellite images, we evaluated four semantic segmentation network architectures: FCN, MA-NET, SegNet, and DeepLabV3. Experiments using a dataset with 100 satellite images demonstrated that the FCN and MA-NET networks achieve better performance in identifying crops than SegNet and DeepLabV3. The latter, however, presented generalization challenges in the test set, indicating sensitivity to variations in lighting and shading. The study highlights the potential of artificial intelligence combined with remote sensing to improve agricultural management, and the need for more robust and diversified datasets to optimize the generalization and robustness of the models in real conditions.

Downloads

Download data is not yet available.

References

[1] STATISTA. Coffee Production Worldwide in 2020, by Leading Country. 2024. Acesso em: 15 abr. 2024. Disponível em: ⟨https://www.statista.com/statistics/277137/world-coffee-production-by-leading-countries/⟩.

[2] Organização Internacional do Café (OIC). Relatório sobre o mercado do café (janeiro/2024). 2024.

[3] ALVES, D. P.; LINDNER, E. O mercado do café no Brasil e a importância de Minas Gerais. Revista Brasileira de Economia Agrícola, v. 15, n. 3, p. 456–478, 2020.

[4] COOXUPÉ. Produção de Café Aumenta 13 Vezes em Quase 50 Anos no Sul de Minas. 2016. Acesso em: 15 abr. 2024. Disponível em: ⟨https://www.cooxupe.com.br/noticias/producao-de-cafe-aumenta-13-vezes-em-quase-50-anos-no-sul-de-minas/⟩.

[5] PAULA, R. Z. A. d. Região e regionalização: Um estudo da formação regional da Zona da Mata de Minas Gerais. Revista de História Econômica & Economia Regional Aplicada, v. 1, n. 1, p. 66–80, 2006.

[6] GONÇALVES, H. T. et al. Caracterização da cultura cafeeira em relação à altimetria e declividade no município de Muzambinho-MG. In: IFSULDEMINAS. Jornada Científica e Tecnológica e Simpósio da Pós-Graduação do IFSULDEMINAS. Muzambinho, MG, Brazil, 2020.

[7] MASKELL, G. M. et al. Integration of sentinel optical and radar data for mapping smallholder coffee production systems in Vietnam. Remote Sensing of Environment, Elsevier, v. 266, p. 112709, 2021.

[8] ARRIOLA-VALVERDE, S. et al. A comparative study of deep learning frameworks applied to coffee plant detection from close-range UAS-RGB imagery in Costa Rica. Remote Sensing, MDPI, v. 16, n. 24, p. 4617, 2024.

[9] WANG, X. et al. A coffee plant counting method based on dual-channel NMS and YOLOv9 leveraging UAV multispectral imaging. Remote Sensing, MDPI, v. 16, n. 20, p. 3810, 2024.

[10] ZHANG, C.; KOVACS, J. M. The application of small unmanned aerial systems for precision agriculture: A review. Precision Agriculture, Springer, v. 13, p. 693–712, 2012.

[11] KAMILARIS, A.; PRENAFETA-BOLDÚ, F. X. Deep learning in agriculture: A survey. Computers and Electronics in Agriculture, Elsevier, v. 147, p. 70–90, 2018.

[12] MA, L. et al. Deep learning in remote sensing applications: A meta-analysis and review. ISPRS Journal of Photogrammetry and Remote Sensing, Elsevier, v. 152, p. 166–177, 2019.

[13] McKinsey & Company. The state of AI in early 2024: Gen AI adoption spikes and starts to generate value. 2024. Acesso em: 26 out. 2024. Disponível em: ⟨https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai#/⟩.

[14] CHAGAS, C. S. et al. Atributos topográficos e dados do Landsat 7 no mapeamento digital de solos com uso de redes neurais. Pesquisa Agropecuária Brasileira, Brasília, v. 45, n. 5, p. 497–507, 2010.

[15] VIEIRA, C. A. O.; LACERDA, J. D.; BOTELHO, M. L. Sensoriamento Remoto e Sistemas de Informação Geográfica: Fundamentos e Aplicações. 2. ed. São Paulo, Brazil: Érica, 2009.

[16] KAWAKUBO, F. S.; MACHADO, N. G. Mapping coffee crops in southeastern Brazil using spectral mixture analysis and data mining classification. International Journal of Remote Sensing, Taylor & Francis, v. 37, n. 18, p. 4390–4405, 2016.

[17] PEREIRA, F. V.; SILVA, J. P.; OLIVEIRA, M. R. Estimating coffee crop parameters through multispectral imaging and machine learning algorithms. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, X-3-2024, p. 317–324, 2024.

[18] ALAMANDA, F.; RACHMAT, A.; WIBOWO, B. Visual segmentation and classification of coffee beans after roasting. Journal of Applied Informatics and Computing, v. 8, n. 2, p. 111–120, 2024.

[19] LU, S.; WANG, H.; LIU, X. Rice plantation mapping using fully convolutional networks on sentinel-1 sar and sentinel-2 optical data. Remote Sensing, MDPI, v. 12, n. 18, p. 2975, 2020.

[20] SUN, H.; ZHANG, Y.; LIN, W. Deep learning for crop type classification and segmentation in agriculture using remote sensing imagery. Sensors, MDPI, v. 20, n. 19, p. 5559, 2020.

[21] SANTOS, A.; SOUZA, B. A. de; GOMES, H. L. Coffee leaf disease detection using a deep convolutional neural network. In: IEEE. 2020 International Conference on Systems, Signals and Image Processing (IWSSIP). [S.l.], 2020. p. 188–192.

[22] BAZAME, R.; SOUZA, B. A. de; SANTOS, A. Deep learning for coffee cherry maturation classification using computer vision. Computers and Electronics in Agriculture, Elsevier, v. 188, p. 106338, 2021.

[23] DOSOVITSKIY, A. et al. An image is worth 16x16 words: Transformers for image recognition at scale. In: International Conference on Learning Representations (ICLR). [S.l.: s.n.], 2021.

[24] HE, X. et al. Multi-scale vision transformer for semantic segmentation in remote sensing imagery. IEEE Transactions on Geoscience and Remote Sensing, IEEE, v. 60, p. 1–13, 2022.

[25] AFONSO, M. et al. Vision transformers for crop classification in satellite imagery: a comprehensive study. Remote Sensing, MDPI, v. 15, n. 6, p. 1672, 2023.

[26] SILVA, L.; SANTOS, J. A.; PEREIRA, M. Transformer-based panoptic segmentation for coffee crop mapping in brazilian regions. IEEE Geoscience and Remote Sensing Letters, IEEE, v. 20, p. 1–5, 2023.

[27] LIU, X.; WANG, H.; LU, S. A transformer-based model for cotton plantation segmentation in high-resolution remote sensing imagery. Remote Sensing, MDPI, v. 13, n. 17, p. 3430, 2021.

[28] GARIA, L.; MUTHUSAMY, H. Dual-tree complex wavelet pooling and attention-based modified u-net architecture for automated breast thermogram segmentation and classification. Journal of Imaging Informatics in Medicine, Springer, p. 1–15, 2024.

[29] YAO, S. et al. Radar-camera fusion for object detection and semantic segmentation in autonomous driving: A comprehensive review. IEEE Transactions on Intelligent Vehicles, IEEE, v. 9, n. 1, p. 2094–2128, 2023.

[30] ARAKAKI, L. G. et al. Evaluation of u-net backbones for cloud segmentation in satellite images. In: VISIGRAPP (4: VISAPP). [S.l.: s.n.], 2023. p. 452–458.

[31] MCCARTHY, J. What is Artificial Intelligence? 2007. Computer Science Department, Stanford University.

[32] GOLDSCHMIDT, R. R. Uma Introdução à Inteligência Computacional: Fundamentos, Ferramentas e Aplicações. 1. ed. Rio de Janeiro, Brazil: IST-Rio, 2010.

[33] FLECK, L. et al. Redes neurais artificiais: Princípios básicos. Revista Eletrônica Científica Inovação e Tecnologia, Medianeira, v. 1, n. 13, p. 47–57, 2016.

[34] HAYKIN, S. Neural Networks: A Comprehensive Foundation. 2. ed. Upper Saddle River, NJ, USA: Prentice Hall, 2001.

[35] GU, J. et al. Recent advances in convolutional neural networks. 2015. Disponível em: ⟨http://arxiv.org/abs/1512.07108⟩. Acesso em: 03 jun. 2025.

[36] LECUN, Y. et al. Gradient-based learning applied to document recognition. Proceedings of the IEEE, v. 86, n. 11, p. 2278–2324, 1998.

[37] LONG, J.; SHELHAMER, E.; DARRELL, T. Fully convolutional networks for semantic segmentation. In: IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Boston, MA, USA: [s.n.], 2015. p. 3431–3440.

[38] FAN, T. et al. MA-Net: A multi-scale attention network for liver and tumor segmentation. IEEE Access, v. 8, p. 179656–179665, 2020.

[39] CHANG, J. et al. Multi-scale attention network for building extraction from high-resolution remote sensing images. Sensors, v. 24, n. 3, p. 1010, 2024.

[40] BADRINARAYANAN, V.; KENDALL, A.; CIPOLLA, R. SegNet: A deep convolutional encoder–decoder architecture for image segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence, v. 39, n. 12, p. 2481–2495, 2017.

[41] CHEN, L.-C. et al. DeepLab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected CRFs. IEEE Transactions on Pattern Analysis and Machine Intelligence, v. 40, n. 4, p. 834–848, 2017.

[42] Instituto Brasileiro de Geografia e Estatística (IBGE). Portal de Mapas. 2014. Acesso em: 08 out. 2024. Disponível em: ⟨https://portaldemapas.ibge.gov.br/portal.php#mapa207423⟩.

[43] ZHENG, Z. et al. Distance-IoU loss: Faster and better learning for bounding box regression. In: AAAI Conference on Artificial Intelligence (AAAI-20). [S.l.: s.n.], 2020. v. 34, p. 12993–13000.

[44] REZATOFIGHI, H. et al. Generalized intersection over union: A metric and a loss for bounding box regression. In: IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Long Beach, CA, USA: [s.n.], 2019. p. 658–666.

[45] JORDAN, J. Evaluating image segmentation models. 2018. Acesso em: 03 jun. 2025. Disponível em: ⟨https://www.jeremyjordan.me/evaluating-image-segmentation-models⟩.