Bioimpedance Spectroscopy Tool with Embedded Hybrid PCA-SVM Model for Post-Harvest Carrot Decay Detection

Carrot postharvest decay leads to major economic losses but is difficult to detect early before visible symptoms emerge. This study introduces a novel approach for non-destructive assessment of decay stages in carrots using bioimpedance spectroscopy integrated with principal component analysis (PCA) and multiclass support vector machine (SVM) modeling. Carrot samples undergoing varying degrees of controlled decay over 14 days were evaluated by measuring electrical bioimpedance from 5-15 kHz using a longitudinal tetrapolar electrode configuration. The complex multivariate bioimpedance data was reduced into principal component features via PCA and classified using a radial SVM model. This integrated spectroscopy and machine learning approach achieved 94% accuracy in categorizing normal, moderate and severe decay grades in the carrots. The high precision of decay level discrimination demonstrates the efficacy of this technique for sensitive early detection of postharvest damage prior to visual spoilage. This provides a promising tool to optimize crop storage conditions, reduce losses, and transform quality evaluation practices in the fresh produce industry through rapid, non-invasive quantification of initial deterioration.