Loading...
Classification of sweet corn based on storage time after harvest using near infrared spectroscopy
Author(s)
Date Issued
March 21, 2017
Type
Conference Paper
Abstract
The freshness of sweet corns is important for production of canned sweet corn. The quality of sweet corns changes rapidly after harvest. Sweet corns should be processed through a production line as fast as possible after harvest. Therefore, some methods of classification of sweet corns based on storage time after harvest are needed. In this study, near infrared (NIR) spectroscopy operating in reflectance mode (1000-2500 nm) and interactance mode (588-1091) were investigated as methods of classification. Sweet corns both with and without husk were tested. Samples (n=120) were scanned with a NIR spectrophotometer every 6 h after harvest. They were then classified into two groups (0 and 1) with a 24-h after harvest cut-off time between the two groups (<24 h and ≥24 h). Classification models were established and validated with a calibration set (n=80), and then the accuracies of the models were evaluated with a prediction set (n=40), using a partial least squares discriminant analysis (PLSDA). It was found that second derivative spectral pretreatment gave the best results for NIR operating in reflectance mode. Regarding prediction accuracy, showed the best accuracies for both unhusked and husked sweet corns (100%), while it was found that mean center and second derivative spectral pretreatment gave good results for NIR operating in interactance mode. The predictive accuracies for unhusked and husked sweet corns obtained 90 and 97.5%, respectively. All of the results demonstrated that NIR spectroscopy has a real potential for non-destructive classification of sweet corns based on storage time after harvest.
Citation
Acta Horticulturae, 1152, 47-54, 2017