Durian (Durio zibethinus) ripeness detection using thermal imaging with multivariate analysis

• Thermal imaging was used to predict quality attributes of durians. Classification accuracy is higher than 93 % for each classification model. Feature extraction based on thermal images were selected durian classification. The detection ripeness using an essential study geared towards improving the current analytical methods which rely heavily routine analysis and human labour skills. investigated in this order evaluate relationship physicochemical properties image parameters. durians acquired at three different ripening stages (unripe, ripe, overripe) soluble solids content, pH, firmness, moisture colour changes determined. Partial least squares (PLS) regression develop quantitative prediction models with R 2 values greater 0.94 all Principal component (PCA) showed successful clustering ability levels Linear discriminant (LDA), k-nearest neighbour (kNN), support vector machine (SVM) applied establishment optimal modelling algorithms. SVM classifier gave overall best performance discrimination a 97 %. feasibility coupled multivariate demonstrated huge potential non-destructive evaluation levels.