Progress 09/15/02 to 09/14/05

Outputs
Rapid and noninvasive measurements of spectral scattering and absorption coefficients are critical to the quantitative understanding of light interaction with turbid food products and the development of effective optical techniques for inspecting, sorting and grading fruits and other agricultural products. This research was aimed to develop an imaging spectroscopy technique for rapid measurement of the optical properties of apples and other turbid food products to predict their quality attributes such as firmness and soluble solids content. An imaging spectroscopy system was assembled for acquiring spatially resolved reflectance from turbid samples over the spectral range between 500 nm and 1000 nm. The system was validated with liquid simulation samples with known scattering and absorption coefficients. Monte Carlo simulations were performed to model light propagation and absorption in simulation samples. Hyperspectral scattering images were collected from apples and peaches as well as selected fruit and vegetable juices and milk samples with different fat levels over the spectral region of 500-1000 nm. A nonlinear search algorithm based on diffusion theory model was developed for determining scattering and absorption coefficients from the hyperspectral scattering images of these products. The major accomplishments of the research include: 1) An imaging spectroscopy (or hyperspectral imaging) method was developed for rapid, noninvasive measurement of spectral absorption and scattering properties of turbid food products over the spectral range between 500 nm and 1000 nm. The method is capable of measuring the scattering coefficient with the average error of 10% or less and the absorption coefficient with the error less than 20%. 2) New data on the spectral scattering and absorption coefficients for a variety of fruit and vegetable juices and milk have been collected and added to the literature. An excellent correlation (r=0.99) was found between the scattering and absorption coefficient at 650 nm and the fat level of milk. The technique could be useful for fast, accurate measurement of the fat content of milk. 3) The imaging spectroscopy technique showed promising for rapid measurement of multiple quality attributes of fruits. A neural network model was developed on relating hyperspectral scattering features to the firmness and soluble solids content of apple fruit. Good correlations were obtained for predicting the firmness and soluble solids content of Golden Delicious apples with values for the correlation coefficient equal to 0.85 and 0.89, respectively. An empirical model based on a Lorentzian distribution function was used for describing the spectral scattering profiles from peaches over the spectral range between 500 and 1040 nm. Lorentzian function parameters were highly correlated with peach fruit firmness with the correlation coefficient of 0.90. Research is continuing in quantifying the fruit size/shape effect for better determination of the scattering and absorption coefficients from the hyperspectral scattering images of apples. Data on the optical properties of apples will be made available in the near future.

Impacts
The imaging spectroscopy or hyperspectral imaging technique developed in this research provides a means for rapid, convenient measurement of the optical properties of apples and other agricultural and food products. The data on the optical properties of fruits and other food products will be of great value to researchers and instrumentation engineers for better understanding of light interaction with biological materials and the development of more effective optical systems for quality evaluation of agricultural products. The technique is promising for assessing multiple quality attributes of apples and other fruits, which will benefit the US fruit industry and the consumer for having better quality, more consistent fresh products.

Publications

• Lu, R. 2003. Imaging spectroscopy for assessing internal quality of apple fruit. ASAE Paper #036012, 10pp.
• Lu, R. 2004. Hyperspectral scattering for assessing internal quality of apple fruit. Postharvest Biology and Technology. (In review)
• Lu, R. and Y. Peng. 2004. Hyperspectral scattering for assessing peach fruit firmness. ASAE Paper #043007, 10pp.
• Lu, R. and Y. Peng. 2005. Hyperspectral scattering for assessing peach fruit firmness. Biosystems Engineering. (In print)
• Qin, J. and R. Lu. 2005. Determination of the optical properties of turbid materials by hyperspectral diffuse reflectance. ASAE Paper No. 053068, 12pp. 2005 (Proceeding)
• Qin, J. and R. Lu. 2005. Hyperspectral diffuse reflectance for determination of the optical properties of milk and fruit juice. In Optical Sensors and Sensing Systems for Natural Resources and Food Safety and Quality, edited by Y. R. Chen et al., Proceedings of SPIE Vol. 5996, 10pp. SPIE (The International Society for Optical Engineering), Bellingham, WA. (Proceeding)

Progress 10/01/03 to 09/30/04

Outputs
A quantitative understanding of light scattering and absorption is critical to developing an effective optical technique for nondestructive quality measurement of fruits and other agricultural products. An imaging spectroscopy technique was developed for measuring the optical properties and quality attributes of apples and other fruits. Experiments were performed on obtaining hyperspectral scattering images from phantom samples with known optical properties in the spectral region between 450 nm and 1000 nm in order to validate the system performance for measuring the optical properties of fruit. Monte Carlo simulations were performed to model light propagation and absorption in phantom samples. Results showed that Monte Carlo simulations were able to adequately describe light scattering and propagation in phantom samples. In order to accurately measure optical properties, light source must meet certain requirements in terms of beam size, profile and incident angle. As a result of this finding, modification to the light source was made and further experiments were performed on additional phantom samples. Monte Carlo simulations showed that this modified light source should lead to more accurate determination of the optical properties of apples. With the development of the imaging spectroscopy system and measurement procedures, we provide a new method of measuring the optical properties of apples and other fruits. This will allow us to establish a database on the optical properties of fruits and other agricultural products and have a better understanding of light interaction with fruit tissue. Hyperspectral scattering profiles were obtained from 'Golden Delicious' and 'Red Delicious' apples by using the imaging spectroscopy system. Mean and standard deviation (std) spectra were extracted from individual hyperspectral scattering images. Principal component (PC) analyses were applied to the mean and std spectra. PC scores were input into a backpropagation neural network (NN). The NN predicted fruit firmness and soluble solids content with the correlation (r) of 0.85 and 0.89, respectively, for 'Golden Delicious'; and r=0.72 and 0.78, respectively, for 'Red Delicious'. Experiments were also performed on peach fruit. A Lorentzian function with two parameters was proposed to describe scattering profiles at individual wavelengths between 500 nm and 1040 nm. PC analyses were applied to spectra of Lorentzian parameters. The backpropagation NN was then used to relate PC scores of Lorentzian parameters to peach firmness. A good correlation (r=0.90) was obtained between hyperspectral scattering profiles and fruit firmness. The imaging spectroscopy technique provides a new means for measuring quality attributes of fruits.

Impacts
Imaging and spectroscopy are widely used for quality evaluation of fruits and other agricultural products. Yet, our understanding of light interaction with biological materials is still limited. This research proposed an imaging spectroscopy technique for measuring the optical properties of apples and other agricultural products. Research demonstrated that the technique, coupled with the Monte Carlo simulation method, will allow us to measure the optical properties of fruits. Monte Carlo simulations enable us to understand how light propagates and scatters in fruit tissue. The imaging spectroscopy technique provides a new way of measuring the optical properties of fruits and other agricultural products. This research will establish a much needed database on the optical properties of fruits, which are critical for designing effective optical systems for quality evaluation of agricultural products. Our research also demonstrated that imaging spectroscopy is useful for determining important quality attributes (firmness, soluble solids, etc.) of apples and other fruits. The technique is relatively easy to implement and thus provides a new means for quality evaluation of fruits and other agricultural products.

Publications

• Lu, R. 2003. Imaging spectroscopy for assessing internal quality of apple fruit. ASAE Paper #036012. ASAE, St. Joseph, MI. 10pp.
• Lu, R. 2004. Hyperspectral scattering for assessing internal quality of apple fruit. Postharvest Biology and Technology. (In review)
• Lu, R. and Y. Peng. 2004. Hyperspectral scattering for assessing peach fruit firmness. ASAE Paper #043007. ASAE, St. Joseph, MI. 10pp.

Progress 10/01/02 to 09/30/03

Outputs
Progress was made during the first year on developing an imaging spectroscopy technique for determining the optical properties of apple fruit. Specifically, we have assembled an imaging spectroscopy system needed for acquiring light scattering from apple fruit over the visible and short wave near-infrared region. A specially designed broadband light source was built, which is essential for acquiring hyperspectral scattering from apple fruit. System calibration is an important step for the imaging spectroscopy system. Preliminary calibrations were performed on the imaging spectroscopy system for wavelength and spatial distance using spectral calibration lamps and laser light sources. A spectral calibration equation was developed for the system. Key personnel including one postdoctoral research associate and one graduate student were hired by the end of the first year.

Impacts
This project will lead to the development of a new method and technique for measuring the optical properties of apple fruit and other agricultural products. Computer simulations will enable us to gain a greater understanding of light propagation and scattering in fruit tissue and its relation with fruit quality attributes. The method and technique developed in this project will provide a new means for rapid, nondestructive assessment of fruit and other agricultural products to ensure and/or enhance their quality and consistency.

Publications

• No publications reported this period