Health diagnostics of fruit plants using Deep learning in mobile applications: RSL (#416)

Authors

Antonio-Miranda, Nieves

Lopez-Lavado, Leonardo

Huaman Torres, Ivan Gustavo

Abstract

In modern agriculture, identifying diseases or pests in fruit plants is one of the biggest challenges. It is true that there are various methods for making diagnoses, mostly with the help of specialists who traverse the fields checking for any signs of disease in the plants. Given the aforementioned as stated in article [1], many of these techniques are slow and very costly, and even with errors, since diseases are not detected until they have a significant degree of damage. In the present systematic literature review, recent studies that approach the diagnosis of plant health through deep learning in mobile applications were analyzed, with the intention of identifying emerging trends and assessing the efficiency of the employed technologies. The methodology employed consists of the comprehensive search and analysis of each scientific article from academic literature sources, for instance, Web of Science and Scopus. In the selection of these articles, keywords related to fruit plants, mobile applications, deep learning, health, and diagnosis were used, following the established inclusion and exclusion criteria, considering that they are RSL articles or conference papers and open access. Subsequently, these articles were analyzed, comparing different approaches using various deep learning models such as RestNet50, VGG16, and other pre-trained models that achieve a high level of accuracy in image capture through mobile devices. This RSL analyzed the use of innovative tools such as Deep learning that systematize knowledge in this area, optimizing crop production and reducing significant losses, which is crucial for future research.