Visión por computadora para detectar enfermedades fúngicas en cacao: revisión sistemática de arquitecturas de deep learning
DOI:
https://doi.org/10.55873/kt5n2q37Palabras clave:
detección de enfermedades del cacao, aprendizaje profundo, yolo, vision transformer, visión por computadoraResumen
Esta revisión sistemática de la literatura analiza 43 estudios científicos sobre la aplicación de técnicas de aprendizaje profundo (DL) para la detección temprana de enfermedades del cacao (Theobroma cacao). Mediante la metodología PRISMA y el marco PICOC, se examinaron modelos y arquitecturas aplicados a tareas de clasificación, detección y segmentación. Los resultados muestran que los modelos ligeros de la familia YOLO alcanzan precisiones de hasta el 96 % en la identificación de enfermedades del cacao, con capacidad para realizar inferencias en tiempo real. Asimismo, las arquitecturas Vision Transformer (ViT) y U-Net registran precisiones entre el 97 % y 99 % en entornos controlados y conjuntos de datos agrícolas relacionados, evidenciando su potencial para aplicaciones basadas en transferencia de aprendizaje. No obstante, persisten limitaciones relacionadas con la generalización entre dominios, la explicabilidad de los modelos y la disponibilidad de datos obtenidos en condiciones reales de campo. Se concluye que futuras investigaciones deben ampliar los conjuntos de datos públicos, incorporar inteligencia artificial explicable (XAI) y aprendizaje autosupervisado (SSL), así como optimizar modelos para dispositivos con recursos limitados, favoreciendo soluciones sostenibles y escalables para el cultivo inteligente del cacao.
Descargas
Referencias
Al-Obeidat, N., & Li, Z. (2025). A multimodal UAV-based pipeline for precision agriculture: Aerial stress detection with YOLO and high-fidelity disease classification using DeiT. Procedia Computer Science, 921–926. https://doi.org/10.1016/j.procs.2025.03.118
Bacus, J. A., & Linsangan, N. B. (2022). Detection and identification with analysis of Carica papaya leaf using Android. Journal of Advances in Information Technology, 13(2), 162–166. https://doi.org/10.12720/jait.13.2.162-166
Bhowmik, A. C., Ahad, M. T., Emon, Y. R., Ahmed, F., Song, B., & Li, Y. (2024). A customised vision transformer for accurate detection and classification of Java plum leaf disease. Smart Agricultural Technology, 8, 100500. https://doi.org/10.1016/j.atech.2024.100500
Borromeo, K. O., Bonotan, M. R., Llagas, A. P., Juan, J. M. O., Sanchez, P. D. C., & Berro, N. K. A. (2025). Classification of cacao pods ripeness using convolutional neural networks. Procedia Computer Science, 257, 1197–1204. https://doi.org/10.1016/j.procs.2025.03.160
Hafizal, M. T., et al. (2022). Implementation of expert systems in potassium deficiency in cocoa plants using forward chaining method. Procedia Computer Science, 136–143. https://doi.org/10.1016/j.procs.2022.12.120
Huang, X., et al. (2025). EConv-ViT: A strongly generalized apple leaf disease classification model based on the fusion of ConvNeXt and Transformer. Information Processing in Agriculture. https://doi.org/10.1016/j.inpa.2025.03.001
Ibtehaz, N., & Rahman, M. S. (2020). MultiResUNet: Rethinking the U-Net architecture for multimodal biomedical image segmentation. Neural Networks, 121, 74–87. https://doi.org/10.1016/j.neunet.2019.08.025
Jackulin, C., & Murugavalli, S. (2022). A comprehensive review on detection of plant disease using machine learning and deep learning approaches. Measurement: Sensors, 24, 100441. https://doi.org/10.1016/j.measen.2022.100441
Kazi, S. S., Palkar, B., & Mishra, D. (2025). VCNet: Optimized deep learning framework with deep feature extraction and genetic algorithm for multiclass rice crop disease detection. MethodsX, 15, 103551. https://doi.org/10.1016/j.mex.2025.103551
Kumi, S., Kelly, D., Woodstuff, J., Lomotey, R. K., Orji, R., & Deters, R. (2022). Cocoa Companion: Deep learning-based smartphone application for cocoa disease detection. Procedia Computer Science, 203, 87–94. https://doi.org/10.1016/j.procs.2022.07.013
Leite, D., Brito, A., & Faccioli, G. (2024). Advancements and outlooks in utilizing convolutional neural networks for plant disease severity assessment: A comprehensive review. Smart Agricultural Technology. https://doi.org/10.1016/j.atech.2024.100573
Li, N., Zhang, A., Han, H., & Duan, Y. (2026). RCPU-Net: A multi-scale multi-object segmentation model for coal and gangue under uneven lighting based on improved U-Net. Digital Signal Processing, 168, 105484. https://doi.org/10.1016/j.dsp.2025.105484
Madadum, H., Nasir, F. E., & Haruehansapong, K. (2025). Optimizing watermelon leaf disease detection using SAM-based augmentation with YOLO for practical agricultural solutions. Smart Agricultural Technology, 12, 101326. https://doi.org/10.1016/j.atech.2025.101326
Mashamba, M. M., Telukdarie, A., Munien, I., Onkonkwo, U., & Vermeulen, A. (2024). Detection of bacterial spot disease on tomato leaves using a convolutional neural network (CNN). Procedia Computer Science, 602–609. https://doi.org/10.1016/j.procs.2024.05.145
Nkuna, B. L., Abutaleb, K., Chirima, J. G., Newete, S. W., van der Walt, A. J., & Nyamugama, A. (2025). Identification of maize leaf diseases using red, green, blue-based images with convolutional neural network (CNN) and residual network (ResNet50) models. Smart Agricultural Technology, 12, 101226. https://doi.org/10.1016/j.atech.2025.101226
Nyarko, B. N. E., Wu, W., Zhou, Z. J., & Mohd, M. J. (2024). Improved YOLOv5m model based on Swin Transformer, K-means++, and Efficient Intersection over Union (EIoU) loss function for cocoa tree (Theobroma cacao) disease detection. Journal of Plant Protection Research, 64(3), 265–274. https://doi.org/10.24425/jppr.2024.151253
Porfírio, R. P., Madeira, R. N., & Santos, P. A. (2025). Exploring explainable AI techniques for plant disease classification in digital agriculture. Procedia Computer Science, 175–182. https://doi.org/10.1016/j.procs.2025.07.170
Ruhad, F. M., Fahim, M., Hossain, M. S., Monir, M. F., Islam, A., & Amin, M. A. (2025). Beyond classification: Benchmarking object detection models for efficient tomato leaf disease identification on a real-world dataset. Smart Agricultural Technology, 12, 101336. https://doi.org/10.1016/j.atech.2025.101336
Sapkota, R., & Karkee, M. (2026). Object detection with multimodal large vision-language models: An in-depth review. Information Fusion, 126, 103575. https://doi.org/10.1016/j.inffus.2025.103575
Sykes, J. R., Denby, K. J., & Franks, D. W. (2024). Computer vision for plant pathology: A review with examples from cocoa agriculture. Applications in Plant Sciences, 12(2), e11559. https://doi.org/10.1002/aps3.11559
Tong, J., Zhang, L., Tian, J., Yu, Q., & Lang, C. (2025). ToT-Net: A generalized and real-time crop disease detection framework via task-level meta-learning and lightweight multi-scale transformer. Smart Agricultural Technology, 12, 101249. https://doi.org/10.1016/j.atech.2025.101249
Vallabhajosyula, S., Sistla, V., & Kolli, V. K. K. (2024). A novel hierarchical framework for plant leaf disease detection using residual vision transformer. Heliyon, 10(9), e29912. https://doi.org/10.1016/j.heliyon.2024.e29912
Yao, J., Li, Y., Xia, Z., Nie, P., Li, X., & Li, Z. (2025). WTAD-YOLO: A lightweight tomato leaf disease detection model based on YOLO11. Smart Agricultural Technology, 12, 101349. https://doi.org/10.1016/j.atech.2025.101349
Yulita, I. N., Amri, N. A., & Hidayat, A. (2023). Mobile application for tomato plant leaf disease detection using a dense convolutional network architecture. Computation, 11(2), 20. https://doi.org/10.3390/computation11020020
Zahra, A., et al. (2024). Current advances in imaging spectroscopy and its state-of-the-art applications. Expert Systems with Applications. https://doi.org/10.1016/j.eswa.2023.122172
Descargas
Publicado
Número
Sección
Licencia
Derechos de autor 2026 Yomer Ysidro-Santos, Andy Roel Vasquez-Bueno, Isait Camizán-Ojeda
Esta obra está bajo una licencia internacional Creative Commons Atribución 4.0.
