Skip to main content
ARS Home » Midwest Area » East Lansing, Michigan » Sugarbeet and Bean Research » Research » Publications at this Location » Publication #328127

Research Project: Nondestructive Quality Assessment and Grading of Fruits and Vegetables

Location: Sugarbeet and Bean Research

Title: Vis/SWNIR spectroscopy and hyperspectral scattering for determining bulk density and particle size of wheat flour

item ZHU, QIBING - Jiangnan University
item XING, YONGCHUN - Jiangnan University
item Lu, Renfu
item HUANG, MIN - Jiangnan University
item NG, PERRY - Michigan State University

Submitted to: Journal of Near Infrared Spectroscopy
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/31/2017
Publication Date: 4/17/2017
Citation: Zhu, Q., Xing, Y., Lu, R., Huang, M., Ng, P. 2017. Vis/SWNIR spectroscopy and hyperspectral scattering for determining bulk density and particle size of wheat flour. Journal of Near Infrared Spectroscopy. 25(2):116-126.

Interpretive Summary: Diffuse reflectance spectroscopy is now being widely used for measuring the compositions and properties of food products. It provides approximate measurements of light absorption in food samples, which are related to the chemical compositions and properties. However, the technique cannot separate absorption from scattering, two basic phenomena when light interacts with the turbid food, the latter of which is known to be related to the physical properties of food (i.e., density, particle size and distribution, cellular structures, etc.). Several innovative techniques have been developed recently for better quantification of the absorption and scattering properties of biological materials and food products. Among them is a hyperspectral imaging-based spatially-resolved technique (also called hyperspectral scattering), developed by a USDA/ARS lab at East Lansing, Michigan, for rapid, noninvasive measurement of food and agricultural products. Compared to point measurements by conventional reflectance spectroscopy, hyperspectral scattering acquires spatially-resolved reflectance profiles over the visible and near-infrared spectral region of 400-1,000 nm for a food sample under the illumination of a small light beam. In this research, reflectance spectroscopy and hyperspectral scattering were used for determining the two important fundamental properties of food powder, i.e., bulk density and particle size. A total of 474 wheat flour samples with different particle sizes and bulk densities were measured. Prediction and classification models were developed for each instrumentation to predict the bulk density and classify the flour samples into five particle size classes. Results showed that hyperspectral scattering performed much better than reflectance spectroscopy for predicting the bulk density, with the prediction error of 30.20 mg/ml versus 57.17 mg/ml (or the correlation coefficient of 0.93 versus 0.74). Hyperspectral scattering achieved 98.2% accuracy for the particle size classification versus 96.8% by reflectance spectroscopy. This research demonstrated that hyperspectral scattering is more suitable than reflectance spectroscopy for assessing the bulk density and particle size of food power like wheat flour.

Technical Abstract: Particle size and bulk density are two important fundamental properties of food powder that directly affect processing and final product quality. The objective of this study was to evaluate and compare two optical sensing methods, i.e., visible/shortwave near-infrared (Vis/SWNIR) spectroscopy and hyperspectral scattering, for bulk density determination and particle size classification of wheat flour. Hyperspectral scattering images over the spectral region of 500-1,000 nm and Vis/SWNIR reflectance spectra covering the spectral region of 400-1,000 nm were acquired for 474 wheat flour samples. Partial least squares regression and discriminant analysis models for Vis/SWNIR spectra and mean spectra extracted from the hyperspectral scattering profiles were developed for determining the bulk density and classifying the particle size, of wheat flour samples. Hyperspectral scattering gave excellent prediction results for bulk density, with the R (correlation coefficient of prediction) value of 0.93 and the RMSEP (root mean squares error of prediction) value of 30.20 mg/ml, compared with the R of 0.74 and RMSEP of 57.13 mg/ml obtained by Vis/SWNIR spectroscopy. Moreover, hyperspectral scattering resulted in 98.2% classification accuracy for particle size, versus 96.8% by Vis/SWNIR spectroscopy. This research suggested that hyperspectral scattering is more suitable than Vis/SWNIR spectroscopy for bulk density determination and particle size classification of food powder.