Evaluation of melon drying using hyperspectral imaging technique in the near infrared region

The evaluation of the melon drying process under different pretreatments using hyperspectral images in near infrared (HSI-NIR) is presented as a quick and non-destructive means. Hyperspectral images of melon slices (3.0 cm × 3.0 cm x 0.3 cm) were obtained for different drying times. Principal Component Analysis applied to the images of cross sections of the dried samples without pre-treatment indicated that the water evaporation occurred unidirectionally. PLS models were generated for each drying pretreatment condition according to the moisture content based on their respective reference values. Prediction errors (RMSEP) obtained for these models were 3.79; 3.69; 3.24 and 2.98% moisture content under the following conditions, respectively: no pretreatment, ultrasound, vacuum, and the combination of vacuum and ultrasound. Chemical maps reveal that the use of vacuum and the combination of vacuum and ultrasound result in uniform drying, corroborating the results obtained by the drying kinetic curve, indicating that these pretreatments facilitate water diffusion in the samples. •Vacuum and ultrasound pre-treatments reduced melon drying time.•PCA in melon-slice HIS-NIR images suggests that water evaporation is unidirectional.•PLS models predicted moisture content accurately, regardless of pre-treatment.•Chemical maps revealed that combining vacuum + ultrasound resulted in uniform drying.