Spectral and image analysis of hyperspectral data for internal and external quality assessment of peach fruit

• Published in: Spectrochimica acta. Part A, Molecular and biomolecular spectroscopy Vol. 272; p. 121016
• Main Authors: Xuan, Guantao, Gao, Chong, Shao, Yuanyuan
• Format: Journal Article
• Language: English
• Published: England Elsevier B.V 05.05.2022
• Subjects: Fruit; Holistic non-invasive measurement; Hyperspectral imaging; Image Processing, Computer-Assisted; Least-Squares Analysis; Linear Models; Prunus persica; Quality attributes evaluation; Visualization; ‘Feicheng’ peach; Visualization; Holistic non-invasive measurement; ‘Feicheng’ peach; Quality attributes evaluation; Hyperspectral imaging
• ISSN: 1386-1425; 1873-3557
• DOI: 10.1016/j.saa.2022.121016

[Display omitted] •Employ the hyperspectral imaging to evaluate the internal and external quality.•Establish the prediction models of SSC and firmness using the normalized spectra.•Create the pixel-wise and object-wise visualization maps for SSC and firmness.•Obtain the pixel diameter by calculating the minimum bounding rectangle.•Develop the imaging-based model to estimate the weight of the fruit. Hyperspectral imaging was attempted to evaluate the internal and external quality of ‘Feicheng’ peach by providing the spectral and spatial data simultaneously. Mask-image was created from hyperspectral image at 810 nm and used to segment the fruit region where the average spectrum, after area normalization, was obtained for soluble solids content (SSC) and firmness evaluation. Pixel size and area were used for diameter and weight estimation. Then effective wavelengths were selected by competitive adaptive reweighted sampling (CARS) and random frog (RF), and employed to develop multiple linear regression (MLR) models. The more effective prediction performances emerged from CARS-MLR model withRV2 = 0.841, RMSEV = 0.546, RPD = 2.51 for SSC andRV2 = 0.826, RMSEV = 1.008, RPD = 2.401 for firmness, followed by creating pixel-wise and object-wise visualization maps for quantifying SSC and firmness. Furthermore, peach diameter was estimated by calculating the minimum bounding rectangle with an average percentage error of 1.01 %, and the MLR model forweightpredictionachieveda good performance ofRV2 = 0.957, RMSEV = 9.203, and RPD = 4.819. The overall results showed that hyperspectral imaging could be used as an effective and non-destructive tool for evaluating the internal and external quality attributes of ‘Feicheng’ peach, and provided a holistic approach to develop online grading systems for quality tiers identification.