Study on fruit quality inspection based on its surface color in produce logistics

Abstract

A novel non-invasive and nondestructive fruit quality inspection method for produce logistics is proposed in this paper based on fruits' surface color. In this method, an image of fruits is firstly taken, which is in the RGB color model. The image is then transferred from the RGB color model to the HSI color model, and is segmented based on hue value to separate the fruits and its background. After that, the simplified histograms of hue H and saturation S of fruits' surface color are calculated, which are used as the input of a designed back propagation (BP) network. The output of the BP network is the quality description of the inspected fruits. After training, the quality of fruits is inspected by the BP network according to the simplified histograms of H and S of fruits' surface color. Experiments are conducted for the quality inspection of bananas with satisfied results, which show the feasibility and reliability of the proposed quick fruit quality inspection method. © 2010 IEEE.