High-resolution RGB image dataset of blueberry plants with polygon-based annotations for canopy segmentation, area estimation, and size classification

This dataset contains 2,043 high-resolution RGB images of blueberry (Vaccinium spp.) plants collected at the Rohu Experimental Site of the Estonian University of Life Sciences, Tartu, Estonia. Images were acquired from a near top-view perspective under natural sunlight on four dates in August and September 2025. The dataset includes polygon-based canopy annotations for each image, from which canopy area is computed. Canopy size-class labels are derived from these area measurements. It is intended for applications in canopy segmentation, area estimation, size classification, plant phenotyping, and precision agriculture.