Experimental Study of Determining Technique for Table Grape Qualities using Visible Wavelength of Imaging and Spectroscopy

Abstract

Imaging and spectroscopy are non-destructive techniques for determining fruit qualities. The qualities of table grapes (Vitis vinifera) such as soluble solids content (SSC), pH, fi rmness and seedlessness are key parameters. This research was focused on comparison between imaging and spectroscopy in laboratory and fi eld. The results of Partial least squares regression (PLSR) showed that the best coeffi cient of determination (R2) for prediction (R2 pred) on SSC for laboratory was 0.8085, for fi eld was 0.8169, and for imaging was 0.7994. The best R2 pred on fi rmness for laboratory was 0.6925, for fi eld was 0.5737, and for imaging was 0.6216. The best R2 pred on pH for laboratory was 0.6820, for fi eld was 0.7101 and for imaging was 0.6494. Partial least squares discriminant analysis (PLS-DA) was analyzed the successful percentage of seedlessness classifi cation: 89.66%, 93.10% and 81.25% for spectroscopy in laboratory, fi eld and imaging, respectively. The results of SSC and seedlessness in fi eld are almost same effi cient as in laboratory. That means farmer can do spectroscopy on SSC and seedlessness anywhere and non-destructively. By the way, we can use both techniques as effi cient non-destructive techniques for determining these key parameters of table grape qualities.