VIC02

The dataset employed in this study comprises high-resolution images of ripe Macauba fruits (Acrocomia aculeata) collected in Araponga, Minas Gerais, Brazil, on February 25, 2026. The fruits were obtained from naturally occurring palm trees in the region, thereby ensuring a diverse and representative sample set. After collection, the fruits were carefully cleaned and maintained in a shaded environment for 48 hours in order to preserve their physiological and external characteristics prior to image acquisition.

Subsequently, high-resolution photographs were captured to document biometric and visual attributes relevant to detection and classification tasks. The dataset was systematically organized to encompass natural variability in fruit size, shape, and coloration, thereby improving the model’s capacity for generalization and robustness under practical field conditions. This dataset provides a valuable foundation for the advancement of automated agro-industrial processes, particularly in fruit sorting, quality assessment, and waste reduction. Overall, it highlights the applicability of artificial intelligence techniques to precision agriculture and post-harvest management systems.