Quantification of the Optical Properties of Two-Layer Turbid Materials Using a Hyperspectral Imaging-based Spatially-resolved Technique

Authors: CEN, HAIYAN - Michigan State University; Lu, Renfu

Submitted to: Applied Optics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/10/2009
Publication Date: 10/7/2009
Citation: Cen, H., Lu, R. 2009. Quantification of the Optical Properties of Two-Layer Turbid Materials Using a Hyperspectral Imaging-based Spatially-resolved Technique. Applied Optics. 48(29):5612-5623.

Interpretive Summary: Optical techniques are useful for assessing internal quality of fruit and other horticultural products. However, the presence of fruit skin can negatively affect the optical measurement of fruit flesh because the structure of fruit skin is distinctively different from that of fruit flesh. Hence, it is desirable to minimize or separate the effect of fruit skin in the optical measurement of fruit flesh, so that more accurate assessment of internal quality can be achieved. This research was aimed at developing a nondestructive method to determine the spectral absorption and scattering properties of two-layer materials with the characteristics of fruit. A hyperspectral imaging system, which provides both spatial and spectral information about an object, was used to acquire diffuse reflectance images from the model samples with known optical properties for the spectral region of 500-1000 nm. A mathematical model for describing light transfer in two-layer samples was validated using computer simulation and experimental data. A computer algorithm was developed to estimate the optical properties of each layer from the hyperspectral image data of the model samples. The algorithm gave good estimation of the optical properties of each layer, with the average error of less than 15%, when the computer simulation data were used. The error became greater (~23%) when the algorithm was used to estimate the model samples. This research showed that the proposed technique can be used for estimating the optical properties of two-layer materials. With further improvements in the algorithm and the imaging system, the technique can be used for measuring the optical properties of fruit skin and flesh. This could open a new avenue for better assessment of internal quality of fruit.

Technical Abstract: Recent research has shown that a hyperspectral imaging-based spatially-resolved technique is useful for determining the optical properties of homogeneous fruits and food products. To better characterize fruit properties and quality attributes, it is desirable to consider fruit to be composed of two homogeneous layers of skin and flesh. This research was aimed at developing a nondestructive method to determine the spectral absorption and scattering properties of two-layer turbid materials with the characteristics of fruit. An inverse algorithm along with the sensitivity coefficient analysis for a two-layer diffusion model was developed for the extraction of optical properties from the spatially-resolved diffuse reflectance data acquired using a hyperspectral imaging system. The diffusion model and the inverse algorithm were validated with Monte Carlo simulations and experimental measurements from solid model samples. The average errors of determining two and four optical parameters were 6.8% and 15.3%, respectively, for Monte Carlo reflectance data. The optical properties of the first layer of the model samples were determined with errors of less than 23.0% for the absorption coefficient and 18.4% for the reduced scattering coefficient. The two-layer diffusion model coupled with the hyperspectral imaging-based spatially-resolved technique has the potential to measure the optical properties of turbid materials like fruits and food products.