High-throughput phenotyping tools for blueberry count, weight, and size estimation based on modified YOLOv5s
The increasing popularity of blueberry has led to expanded blueberry production in many parts of the world. Berry size and average berry weight are key factors in determining the price and marketability of blueberries and therefore are important traits for breeders and researchers to evaluate. Manua...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Maximum Academic Press
2025-01-01
|
| Series: | Fruit Research |
| Subjects: | |
| Online Access: | https://www.maxapress.com/article/doi/10.48130/frures-0025-0006 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849698432651886592 |
|---|---|
| author | Xingjian Li Sushan Ru Zixuan He James D. Spiers Lirong Xiang |
| author_facet | Xingjian Li Sushan Ru Zixuan He James D. Spiers Lirong Xiang |
| author_sort | Xingjian Li |
| collection | DOAJ |
| description | The increasing popularity of blueberry has led to expanded blueberry production in many parts of the world. Berry size and average berry weight are key factors in determining the price and marketability of blueberries and therefore are important traits for breeders and researchers to evaluate. Manual measurement of berry size and average berry weight is labor-intensive and prone to human error. This study developed an automated algorithm and smartphone application for accurate blueberry count and size estimation. Two different computer vision pipelines based on traditional methods and deep neural networks were implemented to detect and segment individual blueberries from Red-Green-Blue (RGB) images. The first pipeline used traditional algorithms such as Hough Transform, Watershed, and filtering. The second pipeline deployed YOLOv5 models with additional modifications using the Ghost module and bi-Feature Pyramid Network (biFPN). A total of 198 images of blueberries, together with manually measured berry count and average berry weight, were used to train and test the model performance. The YOLOv5-based model miscounted four berries in 4,604 total berries across the 198 images. The mean average precision was 92.3%, averaged across an intersection-over-union threshold between 0.50−0.95. The model-derived average berry size was highly correlated with measured average berry weight (R2 > 0.93), which translated to a mean absolute error of around 0.14 g (8.3%). An Android application was also developed in this study to allow easier access to implemented models for berry size and weight phenotyping. |
| format | Article |
| id | doaj-art-c2c0706849f640d5aae8543c925c03c3 |
| institution | DOAJ |
| issn | 2769-4615 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Maximum Academic Press |
| record_format | Article |
| series | Fruit Research |
| spelling | doaj-art-c2c0706849f640d5aae8543c925c03c32025-08-20T03:18:54ZengMaximum Academic PressFruit Research2769-46152025-01-01511910.48130/frures-0025-0006frures-0025-0006High-throughput phenotyping tools for blueberry count, weight, and size estimation based on modified YOLOv5sXingjian Li0Sushan Ru1Zixuan He2James D. Spiers3Lirong Xiang4Department of Electrical and Computer Engineering, North Carolina State University, Raleigh, NC 27606, USADepartment of Horticulture, Auburn University, Auburn, AL 36849, USADepartment of Biological Systems Engineering, Washington State University, Pullman, WA 99164, USADepartment of Horticulture, Auburn University, Auburn, AL 36849, USANorth Carolina Plant Science Initiative, North Carolina State University, Raleigh, NC 27606, USAThe increasing popularity of blueberry has led to expanded blueberry production in many parts of the world. Berry size and average berry weight are key factors in determining the price and marketability of blueberries and therefore are important traits for breeders and researchers to evaluate. Manual measurement of berry size and average berry weight is labor-intensive and prone to human error. This study developed an automated algorithm and smartphone application for accurate blueberry count and size estimation. Two different computer vision pipelines based on traditional methods and deep neural networks were implemented to detect and segment individual blueberries from Red-Green-Blue (RGB) images. The first pipeline used traditional algorithms such as Hough Transform, Watershed, and filtering. The second pipeline deployed YOLOv5 models with additional modifications using the Ghost module and bi-Feature Pyramid Network (biFPN). A total of 198 images of blueberries, together with manually measured berry count and average berry weight, were used to train and test the model performance. The YOLOv5-based model miscounted four berries in 4,604 total berries across the 198 images. The mean average precision was 92.3%, averaged across an intersection-over-union threshold between 0.50−0.95. The model-derived average berry size was highly correlated with measured average berry weight (R2 > 0.93), which translated to a mean absolute error of around 0.14 g (8.3%). An Android application was also developed in this study to allow easier access to implemented models for berry size and weight phenotyping.https://www.maxapress.com/article/doi/10.48130/frures-0025-0006blueberriesphenotypingyolov5weight estimationcomputer vision |
| spellingShingle | Xingjian Li Sushan Ru Zixuan He James D. Spiers Lirong Xiang High-throughput phenotyping tools for blueberry count, weight, and size estimation based on modified YOLOv5s Fruit Research blueberries phenotyping yolov5 weight estimation computer vision |
| title | High-throughput phenotyping tools for blueberry count, weight, and size estimation based on modified YOLOv5s |
| title_full | High-throughput phenotyping tools for blueberry count, weight, and size estimation based on modified YOLOv5s |
| title_fullStr | High-throughput phenotyping tools for blueberry count, weight, and size estimation based on modified YOLOv5s |
| title_full_unstemmed | High-throughput phenotyping tools for blueberry count, weight, and size estimation based on modified YOLOv5s |
| title_short | High-throughput phenotyping tools for blueberry count, weight, and size estimation based on modified YOLOv5s |
| title_sort | high throughput phenotyping tools for blueberry count weight and size estimation based on modified yolov5s |
| topic | blueberries phenotyping yolov5 weight estimation computer vision |
| url | https://www.maxapress.com/article/doi/10.48130/frures-0025-0006 |
| work_keys_str_mv | AT xingjianli highthroughputphenotypingtoolsforblueberrycountweightandsizeestimationbasedonmodifiedyolov5s AT sushanru highthroughputphenotypingtoolsforblueberrycountweightandsizeestimationbasedonmodifiedyolov5s AT zixuanhe highthroughputphenotypingtoolsforblueberrycountweightandsizeestimationbasedonmodifiedyolov5s AT jamesdspiers highthroughputphenotypingtoolsforblueberrycountweightandsizeestimationbasedonmodifiedyolov5s AT lirongxiang highthroughputphenotypingtoolsforblueberrycountweightandsizeestimationbasedonmodifiedyolov5s |