A comprehensive dataset of Halyomorpha halys for automated detection and monitoring in orchards

The Halyomorpha halys (HH) (Brown marmorated stink bug (BMSB)) is a globally invasive pest responsible for severe economic losses in fruit production, yet open, large-scale monitoring datasets remain scarce. Here we present a multi-year, multi-modal dataset acquired over three consecutive growing seasons in an experimental pear orchard in Carpi (Modena), Italy, designed to support research in entomology, agriculture, and Machine Learning (ML). The release comprises (i) three years of Unmanned Aerial Vehicle (UAV)-acquired RGB imagery annotated with bounding boxes for HH; (ii) one season of stationary canopy-level imaging from five fixed cameras targeting insects in occluded regions; (iii) temporally aligned environmental time series of microclimatic variables; and (iv) complementary 2021 acquisitions including smartphone field images, laboratory images, and a synthetic subset generated by compositing insect silhouettes onto orchard backgrounds. The second and third years follow a standardized UAV protocol, while the first year combines heterogeneous consumer-grade modalities. All annotations were manually verified, and the public release includes only images containing at least one annotated insect. This multi-modal design supports applications including small-object detection, sensor fusion, behavioral analysis, and data-driven pest-management tools.