Articles
Blueberry bruise detection relative to fruit firmness
Article number
1440_3
Pages
23 – 28
Language
English
Abstract
Firm blueberries have a longer postharvest life and a higher consumer liking.
Machine harvesting for fresh market is greatly desired to reduce labor, but internal fruit bruising and subsequent loss of firmness are common outcomes.
Detection and analysis of internal bruising can be slow and challenging.
Additionally, determination of a minimal firmness value that would reduce internal bruising would be helpful in breeding blueberries better suited to machine harvest.
In this experiment, blueberries from 53 cultivars of soft to firm types were subjected to bruising by dropping fruit from a 40 cm height onto a stainless steel surface in the 2021-2023 harvest seasons.
Fruit were held for 18 h at 21-23°C, cut in half, and digitally photographed.
Digital photos were then analyzed by a deep learning model (Blueberry Bruising Assessment) and predictions of bruised area as a ratio of bruise to segmented fruit area were generated through box annotated images, berry/bruise annotation, and bruise segmentation.
Corresponding measurements of fruit mechanical hardness on separate blueberry samples best correlated with Young’s modulus 20% burst strain (YM20_BrSt, MPa/%) from a 2-mm flat probe and texture analyzer (TA.XTPlus). Correlation of YM20_BrSt to bruising ratio was moderately high (R2=0.60) and was more effective at distinguishing softer from firmer blueberry cultivars.
Generally, softer blueberries were more likely to exhibit high bruising ratios.
This program offers a simple method to evaluate large sample sizes of blueberry fruit for bruising.
Overall, it appears that blueberries with a YM20_BrSt value greater than 6 will be more resistant to bruising.
Machine harvesting for fresh market is greatly desired to reduce labor, but internal fruit bruising and subsequent loss of firmness are common outcomes.
Detection and analysis of internal bruising can be slow and challenging.
Additionally, determination of a minimal firmness value that would reduce internal bruising would be helpful in breeding blueberries better suited to machine harvest.
In this experiment, blueberries from 53 cultivars of soft to firm types were subjected to bruising by dropping fruit from a 40 cm height onto a stainless steel surface in the 2021-2023 harvest seasons.
Fruit were held for 18 h at 21-23°C, cut in half, and digitally photographed.
Digital photos were then analyzed by a deep learning model (Blueberry Bruising Assessment) and predictions of bruised area as a ratio of bruise to segmented fruit area were generated through box annotated images, berry/bruise annotation, and bruise segmentation.
Corresponding measurements of fruit mechanical hardness on separate blueberry samples best correlated with Young’s modulus 20% burst strain (YM20_BrSt, MPa/%) from a 2-mm flat probe and texture analyzer (TA.XTPlus). Correlation of YM20_BrSt to bruising ratio was moderately high (R2=0.60) and was more effective at distinguishing softer from firmer blueberry cultivars.
Generally, softer blueberries were more likely to exhibit high bruising ratios.
This program offers a simple method to evaluate large sample sizes of blueberry fruit for bruising.
Overall, it appears that blueberries with a YM20_BrSt value greater than 6 will be more resistant to bruising.
Publication
Authors
M. Iorizzo, P. Perkins-Veazie, C. Tan, C. Li, H. Oh, R. Xu, M. Mainland
Keywords
V. corymbosum, V. asheii, bruise ratio, Young’s Modulus, mechanical harvest
Online Articles (74)