Aplicação de modelos YOLO para detecção e contagem automática de pragas agrícolas: uma análise bibliométrica

Abstract

Agricultural pests are among the main factors responsible for reducing crop productivity, making the development of technologies capable of improving pest monitoring and control increasingly necessary. In this context, the YOLO (You Only Look Once) family of models has emerged as one of the leading Artificial Intelligence approaches for real-time object detection, showing promising applications in agriculture. Given the growing volume of scientific publications on this topic, this study aimed to analyze the scientific landscape of the application of YOLO models to agricultural pest detection and counting through a bibliometric analysis of the literature published between 2021 and 2026. The research was conducted using the Scopus database, following systematic criteria for document search, selection, and analysis. Data were processed using the Bibliometrix package through the Biblioshiny interface, enabling the analysis of scientific production, impact, collaboration, and thematic structure indicators. The results revealed a continuous increase in scientific publications throughout the analyzed period, with research concentrated in a limited number of journals, authors, institutions, and countries, as well as the consolidation of themes related to computer vision, deep learning, object detection, and precision agriculture. The thematic analyses also identified specialized research lines and emerging topics, highlighting opportunities for future investigations. It is concluded that the scientific production on YOLO models applied to agricultural pest detection and counting has experienced continuous growth and diversification, reflecting the increasing interest of the scientific community and providing valuable support for future research and the development of more efficient solutions for precision agriculture.