Authors
Jaime Adan Cuevas-Ramirez, Carlos Alberto Villaseñor-Padilla, Javier Gomez-Avila, Carlos Lopez-Franco, Jesus Hernandez-Barragan, Gehova Lopez-Gonzalez, Nancy Arana-Daniel
Published in
Data in brief. Volume 69. Pages 113253. Epub Sep 16, 2026.
Abstract
We present RaspberryJalisco, a field-collected annotated image dataset of raspberry (Rubus idaeus) fruits at distinct phenological stages, captured in a commercial orchard in Jocotepec, Jalisco, Mexico, during November-December 2025. The dataset comprises 11,884 tiled images and 174,763 bounding-box annotations distributed across two complementary labeling schemes: a fine-grained 5-class configuration (Bud, BrightRed C4, DarkRed C5, Green, Orange (red dot)) and a coarser 3-class configuration (Immature, BrightRed C4, Orange (red dot)) designed for yield-prediction and harvest-planning tasks. Images were acquired using a DSLR camera under natural midday illumination and manually annotated by three domain-expert annotators with LabelImg. In response to the risk of parent-image leakage after tiling, the revised release includes persistent parent-image identifiers, random and parent-image grouped split manifests, a leakage audit, corrected random-split baselines, and leakage-resistant grouped YOLOv8 benchmarks. All annotations are provided in YOLO normalized format, with dataset descriptors, scripts, source tables, figures, and YOLOv8 benchmark weights. The data can be reused for in-domain fruit detection, fruit counting, phenological analysis, model benchmarking, and method development, while cross-site or cross-season deployment should be validated externally.
PMID:
42861360
Bibliographic data and abstract were imported from PubMed on 12 Oct 2026.
Read full publication at:
Please sign in
to see all details.
Advertisement
Stats
- Recommendations n/a n/a positive of 0 vote(s)
- Views 9
- Comments 0