Heliocot dataset: raw and processed field RGB images, annotations, and derived outputs

This dataset accompanies the Heliocot workflow and contains raw and processed top-down field RGB images of cotton, YOLO-format instance segmentation annotations exported from CVAT, plot-level metadata, and derived outputs (pixel-based projected leaf area time series, grid evaluation results, and analysis-ready tables). The dataset is intended for reproducible evaluation of segmentation-to-PLA processing and quantification of diurnal canopy orientation dynamics. The associated workflow repository and software releases are archived on Zenodo (see “Related/Identifiers”). “This dataset is provided as a multi-part 7z archive split into ~1 GB chunks (.001, .002, .003).” “To extract, download all parts into the same folder and open the .001 file with 7-Zip.” Keywords: cotton field phenotyping RGB imaging instance segmentation YOLO projected leaf area canopy orientation diurnal dynamics Heliocot