Automatic sorting of Chinese jujube (Zizyphus jujuba Mill. cv. ‘hongxing’) using chlorophyll fluorescence and support vector machine

Fruit classification is important to improve quality during processing, storage and marketing. The aim of the study was to determine if a new system combining chlorophyll fluorescence (ChlF) and C-support vector machine (C-SVM) might assist the classification of jujube fruits based on postharvest quality, including ascorbic acid and total phenols contents and 2,2′-diphenyl-1-picrylhydrazyl (DPPH) radical-scavenging activity. Our results showed that the best classification accuracy of fruit quality was up to 93.33% using the RBF SVM classifier (C = 2, γ = 0.5), and the correct classification rates of 86.67% was achieved for the sigmoid (C = 2, γ = 0.5) SVM classifier as well as the polynomial (C = 2, γ = 0.5, d = 1) SVM classifier. The proposed SVM classifier achieved the best classification accuracy, showing that the SVM-ChlF system can provide a potential tool for automatically classifying the quality of not only jujube fruits, but also any other chlorophyll-containing fruits in packing lines.