Dataset: Detecting leaf-cutting ants through computer vision in vineyards

Between 20% and 40% of global agricultural production is lost annually due to pests, generating significant economic impacts—estimated at 70 billion USD from invasive insects alone—according to the Food and Agriculture Organization of the United Nations (FAO). In addition, the expected rise in average global temperature is likely to increase the destructive potential of leaf-cutting ants [1], reinforcing the need for sustainable pest management strategies in modern agriculture. In this context, the present dataset aims to support the identification and geolocation of leaf-cutting ants and their foraging trails. The provided images and ground truth data were generated through the use of agricultural machinery equipped with cameras, combined with image preprocessing techniques to detect ants and their trails in real-world conditions. Data collection was conducted in a vineyard located in Uruguay, providing a valuable real-world case study to advance the Technology Readiness Level (TRL) of AI-based solutions aimed at supporting Smart Agriculture practices.