Development of a Detection System Based on Yolov8 for the Automatic Detection of the Anthracnose in Papayas

León León R.A.; Acevedo Chávez J.I.; Bringas Alvarez E.S. · Memorias de la Conferencia Iberoamericana de Complejidad, Informatica y Cibernetica, CICIC (2025) · DOI: 10.54808/cicic2025.01.260

Resumen

The anthracnose represents a serious threat to papaya cultivation, causing significant economic losses by reducing the quality and commercial value of the fruits. This disease, which can affect up to 90% of production under favorable conditions, particularly impacts small farmers, who lack accessible methods for early detection. Traditional diagnostic methods are slow and costly, making effective prevention difficult. To improve this situation, a visual detection system based on deep neural networks was developed, using the YOLO v8 model. This model was trained in Python to efficiently recognize anthracnose in high-resolution images of papayas, captured under controlled lighting conditions. The YOLO v8 network was chosen for its speed and accuracy in uncontrolled environments, which is crucial for field application. The results show that this system can detect anthracnose with high precision and speed of 93-98%, surpassing manual methods and offering a practical solution for the phytosanitary control of papayas. This innovative approach has the potential to significantly reduce economic losses and improve the sustainability of crops. Copyright 2025. © por el International Institute of Informatics and Systemics.

Instituciones: Universidad Privada del Norte