Desarrollo de aplicación Android móvil utilizando Visión Artificial y Deep Learning para la identificación de aguacates Hass con la plaga Monalonion, en la finca “Las Palmas”, ubicada en el municipio de San Agustín, Huila

Abstract

In the department of Huila, in recent years there has been an increase in Hass avocado crops, which is why there is also evidence of the presence of pests such as Monalonion, considered the main one according to farmers, the appearance of th losses between 50 and 100% (Eugenia & Zuluaga, 2020) . T is can generate economic he purpose of this project was to develop a mobile Android application using Artificial Vision and Deep Learning for the identification of the fruit with Monalonion . The data set consolidated by the author contains real images of avocados, the Deep Learning transfer learning technique and the EfficientDetLite3 model are used, it was evaluated in the Google Colab environment with the mAP (mean Average Precision) metric of 9 0 which i % was obtaineds very popular among different competitions and an average accuracy