Inspection, detection, and sorting are of great importance in the fruit industry. Manual inspection is expensive, slow, and prone to error and inconsistency. Because of the great potential benefits to the fruit industry, improved machine vision systems are needed in order to extend the quality control task to include detection of more subtle phenomena such as bruises and degree of ripeness. The proposed work will provide essential design information for post harvest apple quality evaluation using a novel optical technique. The design information will be used in a Phase II to build a dedicated prototype imaging inspection system for sorting apples based on surface and internal defects as well as positive attributes of quality. OBJECTIVES: The primary objective is to investigate the feasibility of using Spatial-Frequency-Domain Imaging (SFDI) for the assessment of the post harvest quality of apples. This optical imaging technique uses multi-spatial frequencies for quantitatively determining the absorption and scattering properties of media in the near-infrared region between 600 nm and 1100 nm. The absorption and reduced scattering coefficients will then be correlated with the post harvest qualities of apples. APPROACH: The approach will be divided into three main tasks including 1) correlation between firmness and spatially resolved SFDI optical properties, 2) correlation between sugar content and spatially resolved SFDI optical properties and 3) correlation between bruising and spatially resolved SFDI optical properties. The latter will also include an investigation of the feasibility of SFDI for performing depth sectioned imaging of bruised tissues. The SFDI instrument at the Beckman Laser Institute will be used to acquire image data. A filter wheel containing band pass filters will be used to provide spectral selectivity in combination with the broadband light source that is currently in place. Spectral coverage from 500 to 1100 nm will be obtained using a combination of liquid crystal tunable filter with the rotating filter wheel. The system at BLI is optimized to work over the region from 500-720 nm. It is configured to allow serial combination with a rotating filter wheel which can be used to sweep through the 740 -1100 nm range. Images will be obtained for each spectral window as the spatial frequency of the illumination is swept. In all cases for which SFDI data is acquired we will also acquire a digital color image in order to provide a frame of reference and so that we can qualitatively illustrate the sensitivities of the SFDI method compared to simple color imaging
