A Novel Public Dataset for Strawberry (Fragaria × Ananassa) Ripeness Detection and Comparative Evaluation of YOLO ‐Based Models
- 1Computer Engineering Department Kırıkkale University Kırıkkale Türkiye
- 2Computer Engineering Department Hacı Bektaş Veli University Nevsehir Türkiye
- 3Computer Engineering Department Selçuk University Konya Türkiye
Abstract
The strawberry (Fragaria × ananassa), known worldwide for its economic value and nutritional richness, is a widely cultivated fruit. Determining the correct ripeness level during the harvest period is crucial for both preventing losses for producers and ensuring consumers receive a quality product. However, traditional methods, that is, visual assessments alone, can be subjective and have a high margin of error. Therefore, computer-assisted systems are needed. However, the scarcity of comprehensive datasets accessible to everyone in the literature makes it difficult to compare studies in this field. In this study, a new and publicly available strawberry ripeness dataset, consisting of 566 images and 1169 labeled objects across three ripeness classes (fully-ripe, semi-ripe, and unripe), prepared under variable light and environmental conditions in two different greenhouses in Turkey, is presented to the literature. Comparative tests conducted on the dataset using twelve YOLOv8, YOLOv9, and YOLO11 detection variants showed that the highest precision value was 90.94% in the YOLOv9c model, while the highest recall value was 83.74% in the YOLO11s model. In terms of the general performance criterion mAP@50, YOLOv8s was the best performing model with a success rate of 86.09% (mAP@50:95 = 75.52%). On this dataset, smaller and medium-sized models (approximately 10-12 million parameters, e.g., YOLOv8s and YOLO11s) achieved accuracy comparable to the largest variants at substantially lower computational cost and higher inference speed, establishing a reproducible reference point for smart-agriculture applications. These findings are specific to the present dataset and should be validated on larger, more diverse data before being generalized.
Related
- Genome-Wide Identification of the PP2A Gene Family in Strawberry (Fragaria × ananassa) and Pharmacological Evidence for the Involvement of PP2A in ALA-Mediated Salt Tolerance
- Foliar application of microbial consortium formulations suppresses Phomopsis leaf blight, promotes vegetative growth, and modulates biochemical defence in strawberry (Fragaria × ananassa Duch. cv. Camarosa) under protected cultivation
- Genetic resources of garden strawberry Fragaria × ananassa as a basis for breeding for quality and resistance: world collections and the Russian gene pool
- Comparative evaluation of sample preparation and enrichment methods for culture and molecular detection of Burkholderia pseudomallei in naturally contaminated soils from Thailand over one year.
- LMF-YOLO: a multi-scale inception model for edge-based strawberry disease detection
BibTeX
@article{yurdakul2026novel,
title = {{A Novel Public Dataset for Strawberry (Fragaria × Ananassa) Ripeness Detection and Comparative Evaluation of YOLO ‐Based Models}},
author = {Mustafa Yurdakul and Zeynep Sena Bastug and Ali Emre Gök and Şakir Taşdemir},
journal = {Food Science \& Nutrition},
year = {2026},
month = jul,
doi = {10.1002/fsn3.72112},
url = {https://doi.org/10.1002/fsn3.72112},
note = {CC BY},
}
RIS
TY - JOUR
TI - A Novel Public Dataset for Strawberry (Fragaria × Ananassa) Ripeness Detection and Comparative Evaluation of YOLO ‐Based Models
AU - Mustafa Yurdakul
AU - Zeynep Sena Bastug
AU - Ali Emre Gök
AU - Şakir Taşdemir
PY - 2026
DA - 2026/07/01
JO - Food Science & Nutrition
DO - 10.1002/fsn3.72112
UR - https://doi.org/10.1002/fsn3.72112
AB - The strawberry (Fragaria × ananassa), known worldwide for its economic value and nutritional richness, is a widely cultivated fruit. Determining the correct ripeness level during the harvest period is crucial for both preventing losses for producers and ensuring consumers receive a quality product. However, traditional methods, that is, visual assessments alone, can be subjective and have a high margin of error. Therefore, computer-assisted systems are needed. However, the scarcity of comprehensive datasets accessible to everyone in the literature makes it difficult to compare studies in this field. In this study, a new and publicly available strawberry ripeness dataset, consisting of 566 images and 1169 labeled objects across three ripeness classes (fully-ripe, semi-ripe, and unripe), prepared under variable light and environmental conditions in two different greenhouses in Turkey, is presented to the literature. Comparative tests conducted on the dataset using twelve YOLOv8, YOLOv9, and YOLO11 detection variants showed that the highest precision value was 90.94% in the YOLOv9c model, while the highest recall value was 83.74% in the YOLO11s model. In terms of the general performance criterion mAP@50, YOLOv8s was the best performing model with a success rate of 86.09% (mAP@50:95 = 75.52%). On this dataset, smaller and medium-sized models (approximately 10-12 million parameters, e.g., YOLOv8s and YOLO11s) achieved accuracy comparable to the largest variants at substantially lower computational cost and higher inference speed, establishing a reproducible reference point for smart-agriculture applications. These findings are specific to the present dataset and should be validated on larger, more diverse data before being generalized.
SN - 2048-7177
ER -