Arquitecturas YOLO y RF DETR para el diagnóstico temprano de enfermedades en Solanum tuberosum L
Revista Científica CEDIA. Revista de investigación en tecnologías de información y comunicación aplicadas.  ilustración de una flor en una maceta electrónica.
PDF (ES)
PDF (English)

Palabras clave

agricultura de precisión
aprendizaje profundo
localización de objetos
visión por computadora
aumento de datos

Resumen

Se compararon, bajo un protocolo común, tres detectores de objetos para el diagnóstico poscosecha en tubérculos: dos arquitecturas de la familia YOLO y un detector basado en transformadores (RF‑DETR). Se empleó un conjunto de datos público de seis clases (cinco patologías y una sana). Las lesiones se anotaron con bounding boxes en Roboflow y se aplicó aumento de datos geométrico y fotométrico. Tras la conversión a COCO para RF‑DETR, se mantuvieron idénticas particiones (train/val/test = 792/56/57), misma resolución (640 px) y el mismo protocolo de entrenamiento (60 épocas con early stopping). La evaluación se realizó exclusivamente en test con precision, recall, mAP@0.50 y mAP@0.50:0.95. En test, YOLOv12 alcanzó P = 78,72 %, R = 96,86 %, mAP@0.50 = 79,36 %, mAP@0.50:0.95 = 73,26 %; YOLOv26 obtuvo P = 79,34 %, R = 96,64 %, mAP@0.50 = 78,48 %, mAP@0.50:0.95 = 73,55 % (según los registros experimentales). RF‑DETR logró P = 70,86 %, R = 88,39 %, mAP@0.50 = 85,61 % y mAP@0.50:0.95 = 82,64 %. Para el tamizado de alta sensibilidad (recall alto), la arquitectura YOLO es más conveniente, mientras que RF‑DETR ofrece localización estricta (mAP superior en 0.50–0.95) y simplifica el posprocesamiento al prescindir de NMS. Un pipeline en dos etapas (YOLO → RF‑DETR) combina velocidad y precisión espacial. Como trabajo futuro, se sugiere calibración por clase y fusión con NIR/HSI para lesiones subepidérmicas.

https://doi.org/10.61854/rccedia.v1n1.005
PDF (ES)
PDF (English)

Citas

Aleinzi, M. (2025). Enhancing precision agriculture with YOLOv8: A deep learning approach to potato disease identification. International Journal of Advanced Computer Science and Applications, 16(4), 1156–1166. https://doi.org/10.14569/IJACSA.2025.01604110

Badgujar, C. M., Poulose, A., & Gan, H. (2024). Agricultural object detection with You Look Only Once (YOLO) algorithm: A bibliometric and systematic literature review. Computers and Electronics in Agriculture, 223. https://doi.org/10.1016/j.compag.2024.109090

Blok, P. M., Wang, H., Suh, H. K., Wang, P., Burridge, J., & Guo, W. (2025). PointRAFT: 3D deep learning for high-throughput prediction of potato tuber weight from partial point clouds [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2512.24193

Chaudhary, S., Lal, M., Sagar, S., Sharma, S., & Kumar, M. (2023). Black scurf of potato: Insights into biology, diagnosis, detection, host-pathogen interaction, and management strategies. Tropical Plant Pathology, 49(2), 169–192. https://doi.org/10.1007/s40858-023-00622-4

Dahiya, N., Prakash, D., Kundu, S., Kuttan, S. R., Suwalka, I., Ayadi, M., Dubale, M., & Hashmi, A. (2025). Optimised RFO tuned RF-DETR model for precision urine microscopy for renal and systemic disease diagnosis. Scientific Reports, 15(1), 25842. https://doi.org/10.1038/s41598-025-11725-0

Divyanth, L. G., Khanal, S. R., Paudel, A., Mattupalli, C., & Karkee, M. (2024). Efficient detection of eyes on potato tubers using deep-learning for robotic high-throughput sampling. Frontiers in Plant Science, 15, 1512632. https://doi.org/10.3389/fpls.2024.1512632

Ephytia. (2025, July 24). Potato—Rhizoctonia solani (stem canker & black scurf). https://ephytia.inra.fr/en/C/20899/Potato-Rhizoctonia-solani-Stem-canker-Black-scurf

Faria, F. T. J., Moin, M. Bin, & Wase, A. Al. (2023). Potato Diseases Datasets [Conjunto de datos]. Kaggle. https://doi.org/10.34740/KAGGLE/DSV/7203264

Geoffrey Manurung, D., Ryan Pinasthika, M., Obila Vasya, M. A., Dwi Setya Putri, R. A., Parasian Tampubolon, A., Fadhil Prayata, R., Khoirin Nisa, S., & Yudistira, N. (2024). Deteksi dan klasifikasi hama potato beetle pada tanaman kentang menggunakan YOLOV8. Jurnal Teknologi Informasi dan Ilmu Komputer, 11(4), 723–734. https://doi.org/10.25126/jtiik.1148092

Guillemette, A. M., Casanova, G. H., Hamilton, J. P., Pokorná, E., Dobrev, P. I., Motyka, V., Rashotte, A. M., & Leisner, C. P. (2024). The physiological and molecular responses of potato tuberization to projected future elevated temperatures. Plant Physiology, 197(1), 664. https://doi.org/10.1093/PLPHYS/KIAE664

He, Y., Peng, Y., Wei, C., Zheng, Y., Yang, C., & Zou, T. (2024). Automatic disease detection from strawberry leaf based on improved YOLOv8. Plants, 13(18). 2256 https://doi.org/10.3390/PLANTS13182556

Jayanthi, J., & Kumar, K. A. (2024). Transformative impact of AI-driven computer vision in agriculture. En Artificial intelligence techniques in smart agriculture (pp. 129–150). Springer https://doi.org/10.1007/978-981-97-5878-4_9

Jocher, G., & Qiu, J. (2026). Ultralytics YOLO26 [Software]. GitHub. https://github.com/ultralytics/ultralytics

Keithellakpam, L. B., Karunakaran, C., Singh, C. B., Jayas, D. S., & Danielski, R. (2026). A comprehensive review on pre- and post-harvest perspectives of potato quality and non-destructive assessment approaches. Applied Sciences, 16(1). 190 https://doi.org/10.3390/APP16010190

Liu, C. (2024). Advancing strawberry disease detection in agriculture: A transfer learning approach with YOLOv5 algorithm. International Journal of Advanced Computer Science and Applications, 15(3), 1013–1022. https://dx.doi.org/10.14569/IJACSA.2024.01503101

Ma, P., Li, C., Rahaman, M. M., Yao, Y., Zhang, J., Zou, S., Zhao, X., & Grzegorzek, M. (2022). A stateof-the-art survey of object detection techniques in microorganism image analysis: From classical methods to deep learning approaches. Artificial Intelligence Review, 56(2), 1627–1698. https://doi.org/10.1007/S10462-022-10209-1

Muñoz Espinoza, M., Moreno Castillo, W. E., y Palacios Ruiz, F. P. (2026). Detección de enfermedades en papa mediante deep learning usando YOLOv12. Ciencias de la Ingeniería y Aplicadas, 10(1), 119–129. https://doi.org/10.61236/CIYA.V10I1.1239

Moreno Mayhuire, J. S., & Herrera Quispe, J. A. (2023). Modelos de aprendizaje automático para la clasificación de enfermedades de la Solanum tuberosum: Una revisión sistemática de la literatura. Revista de Investigación de Sistemas e Informática, 16(2), 129–137. https://doi.org/10.15381/risi.v16i2.25980

Nagel, M., Dulloo, M. E., Bissessur, P., Gavrilenko, T., Bamberg, J., Ellis, D., & Giovannini, P. (2022). Global strategy for the conservation of potato. Leibniz Institute of Plant Genetics and Crop Plant Research. https://doi.org/10.5447/IPK/2022/29

Öztürk, L., & Şin, B. (2026). Real-time transformer-based object detection for advanced plant disease recognition: The grapevine fanleaf virus example. Applied Fruit Science, 68(1). https://doi.org/10.1007/s10341-026-01777-5

Palacio Betancur, S., & Bolaños Martínez, F. (2025). Tomato (Solanum lycopersicum L.) leaf disease detection using computer vision. Revista Facultad Nacional de Agronomía, 78(3), 11203–11212. https://doi.org/10.15446/rfnam.v78n3.116493

Parlak, C. (2026). Evaluation of cutting-edge object detection architectures on multi-object and single-object datasets. Black Sea Journal of Engineering and Science, (9)1. 287-294. https://doi.org/10.34248/bsengineering.1736319

Prasetyo, E. W., Amanah, H. Z., Farras, I., Pahlawan, M. F. R., & Masithoh, R. E. (2024). Detection of Fusarium spp. infection in potato (Solanum tuberosum L.) during postharvest storage through visible-near-infrared and shortwave-near-infrared reflectance spectroscopy. Open Agriculture, 9(1). https://doi.org/10.1515/opag-2022-0295

Qin, R., Wang, Y., Liu, X., & Yu, H. (2024). Advancing precision agriculture with deep learning enhanced SIS-YOLOv8 for Solanaceae crop monitoring. Frontiers in Plant Science, 15, 1485903. https://doi.org/10.3389/fpls.2024.1485903

Rahaman Sams, F., Kazi Supti, S., Binte Hamid, S., Junayed, R., Fahim A Bari, K. M., Junaeid Ali, M., Gani, R., Shams, K., Rifat Ahmmad Rashid, M., & Ul Islam, R. (2026). Real-time sunflower detection using semi-supervised and self-supervised deep learning for precision agriculture. Smart Agricultural Technology, 13, 101684. https://doi.org/10.1016/J.ATECH.2025.101684

Redmon, J. (2016). You only look once: Unified, real-time object detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.

Robinson, I., Robicheaux, P., Popov, M., Ramanan, D., & Peri, N. (2026). RF-DETR: Neural architecture search for real-time detection transformers [Preprint]. arXiv. https://arxiv.org/abs/2511.09554

Roboflow. (2025). Roboflow: Computer vision tools for developers and enterprises. https://roboflow.com/

Sapkota, R., Cheppally, R. H., Sharda, A., & Karkee, M. (2025). YOLO26: Key architectural enhancements and performance benchmarking for real-time object detection [Preprint]. arXiv. https://arxiv.org/pdf/2509.25164

Sharma, S., Tiwari, R. K., Sagar, V., & Maharana, C. (2024). Soil- and tuber-borne diseases of potato. En Approaches for potato crop improvement and stress management (pp. 179–231). https://doi.org/10.1007/978-981-97-1223-6_7

Solankey, S. S. (Ed.). (2025). Advances in research on potato production. Springer. https://doi.org/10.1007/978-3-031-82710-5

Thibaut, L. (2023, August 13). COCO evaluation metrics explained. Picsellia. https://www.picsellia.com/post/coco-evaluation-metrics-explained

Tian, Y., Ye, Q., & Doermann, D. (2025, February 20). YOLO12: Attention-centric real-time object detectors [Preprint]. arXiv. https://arxiv.org/abs/2502.12524

Upeniece, L., Bimšteine, G., & Stramkale, V. (2025). The most important potato diseases and microbiological agents for their control: A review. Research for Rural Development 2025: Annual 31st International Scientific Conference Proceedings (Vol. 1, pp. 17–23). Latvia University of Life Sciences and Technologies. https://doi.org/10.22616/RRD.31.2025.003

Yurdakul, M., Sazak, H., Kotan, M., & Ta¸sdemir, T. (2026). A review of YOLO family from YOLOv1 to YOLO26 [Preprint]. Preprints.org https://doi.org/10.20944/PREPRINTS202602.1844.V1

Zhang, F., Wang, W. X., Wang, C. S., Zhou, J., Pan, Y., & Sun, J. F. (2024). Study on hyperspectral detection of potato dry rot in gley stage based on convolutional neural network. Spectroscopy and Spectral Analysis, 44(2), 480-489. https://doi.org/10.3964/j.issn.1000-0593(2024)02-0480-10

Zhang, J., Xie, J., Zhang, F., Gao, J., Yang, C., Song, C., Rao, W., & Zhang, Y. (2024). Greenhouse tomato detection and pose classification algorithm based on improved YOLOv5. Computers and Electronics in Agriculture, 216, 108519. https://doi.org/10.1016/J.COMPAG.2023.108519

Zhao, Y., Lv, W., Xu, S., Wei, J., Wang, G., Dang, Q., Liu, Y., & Chen, J. (2024). DETRs beat YOLOs on real-time object detection. En Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 16965–16974). https://zhao-yian.github.io/RTDETR.

Zhou, H., Liu, Y., & Wang, J. (2025). Railway fastener defect detection using RFD-DETR: A lightweight real-time transformer-based approach. PLOS ONE, 20(11), e0331513. https://doi.org/10.1371/JOURNAL.PONE.0331513

Creative Commons License

Esta obra está bajo una licencia internacional Creative Commons Atribución 4.0.

Derechos de autor 2026 Manolo Muñoz Espinoza