Detection of Wampee Damage based on Hyperspectral Imaging Technology

Abstract

Wampee is one of the characteristic fruits in southern China, and its brittle and thin skin can easily be damaged. In this study, principal components analysis (PCA) and minimum noise fraction (MNF) analysis were carried out on the two wampee varieties by hyperspectral imaging technology, and 680nm was determined to be the optimal characteristic wavelength. The accurate recognition rate obtained from PCA algorithm for wampee samples of two varieties was about 83.75%, and that obtained from MNF algorithm for two variety samples was 85%. It was indicated that the wampee damaged can be identified more accurately and effectively by MNF based on hyperspectral imaging technology