Work

Nondestructive Assessment of Internal Bruise Damage and Postharvest Quality in Peaches Using Smartphone Near‐Contact Light Scattering

Year: 2026
Type: article
Abstract: Internal bruising and early browning in peaches reduce postharvest marketability and can remain hidden before surface symptoms become obvious. This study evaluated smartphone near‐contact light scattering as a controlled screening method for internal peach damage referenced by physicochemical measurements. “Jinxiu” peaches were subjected to controlled impact treatments and accelerated deterioration at 25°C and 85% relative humidity. Near‐contact images were acquired under fixed geometry with smartphones, whereas destructive measurements of firmness, color difference, browning area ratio, and damage depth defined four operational quality grades. A YOLOv10‐NWD front end localized fruit and halo regions, and structured optical‐response descriptors were mapped to terminal quality outputs through constrained latent‐state inference. Under controlled‐input fruit‐level evaluation, classification accuracy was 93.5 % ± 1.4 % across five independent training runs. In full‐pipeline fruit‐level evaluation, accuracy was 92.5% for internal Device A (37/40; 95% Wilson CI, 80.1%–97.4%), 92.5% for internal Device B (37/40; 95% Wilson CI, 80.1%–97.4%), and 88.0% for the independent external‐device feasibility cohort assessed with Device D (44/50; 95% Wilson CI, 76.2%–94.4%), which consisted of fruit of the same cultivar and source evaluated under the same controlled deterioration protocol. Controlled‐input regression yielded R 2 values of 0.92 for firmness, 0.89 for color difference, 0.85 for browning area ratio, and 0.94 for damage depth. The results support smartphone near‐contact scattering as a secondary postharvest screening approach under the tested conditions, whereas broader commercial use still requires paired multidevice, multicultivar, multiregion, multiseason, and cold‐chain validation.
Language: English
Field: Chemistry
Open Access status: hybrid
Estimated APC paid: $2,700
OpenAlex ID: W7214471069
FWCI:
Cites: 31
Cited by:
Related to: 9