A non-invasive method for assessing changes in strawberry hardness under the influence of pulsed electric field

This study presents a non-destructive method for assessing changes in strawberry hardness induced by pulsed electric field (PEF) treatment using optical coherence tomography (OCT). Strawberry samples were subjected to PEF at field strengths ranging from 0.3 to 1.5 kV/cm, and structural changes were evaluated using both mechanical hardness tests and OCT-based image analysis of subsurface tissue layers. Texture descriptors, including Haralick features and lacunarity measures, were extracted from OCT images to characterize microstructural heterogeneity. Due to high feature redundancy, principal component analysis (PCA) was applied, preserving over 99.7% of data variance in the textural feature space. A linear prediction model supported by RANSAC regression was developed to estimate hardness from PCA-transformed features, ensuring robustness to outliers. Model evaluation showed moderate predictive performance (mean R 2 ≈ 0.6), with variability across data splits indicating sensitivity to sample heterogeneity. SHAP analysis revealed that the first principal component dominates the prediction. An additional linear regression model of changes in strawberry hardness under the influence of PEF with 7 different intensities was also introduced. Statistical analysis confirmed a significant effect of PEF on hardness ( p = 0.002, η 2 =0.365). The results show that OCT combined with machine learning shows promising potential for exploring the non-invasive monitoring of PEF-induced structural changes associated with fruit tissue hardness.