Development of an Artificial Vision System using Yolov8 for the Detection of the Fungus Sphaceloma Perseae in Avocado

León León R.A.; Aguirre Vasquez H.Y.; Cruz Galdos I. · Memorias de la Conferencia Iberoamericana de Complejidad, Informatica y Cibernetica, CICIC (2025) · DOI: 10.54808/cicic2025.01.357

Resumen

This study develops an automated system for detecting the Sphaceloma persea fungus in avocados based on quality criteria, utilizing the YOLOv8 model. A total of 117 images were manually collected and labeled, then segmented into training, validation, and testing sets. The YOLOv8 model was trained and fine-tuned using the Make sense platform, leveraging a CORE i5 processor with 8 GB of memory to optimize processing resources. Model parameters, such as batch size and learning rate, were adjusted to maximize classification accuracy. Regarding results, the system achieved an average accuracy of ()%. The implementation of YOLOv8 in the agri-food industry offers an effective solution to enhance efficiency and competitiveness among producers, ensuring uniformity and quality in the final product. Copyright 2025. © por el International Institute of Informatics and Systemics.

Instituciones: Universidad Privada del Norte