Skip to main content
ARS Home » Pacific West Area » Albany, California » Western Regional Research Center » Healthy Processed Foods Research » Research » Publications at this Location » Publication #413236
Research Project: New Technologies and Methodologies for Increasing Quality, Marketability and Value of Food Products and Byproducts

Location: Healthy Processed Foods Research

Title: Rapid and nondestructive determination of oil content and distribution of potato chips using hyperspectral imaging and chemometrics

Author
item SUN, YUE - Virginia Polytechnic Institution & State University
item NAYANI, NIKHITA - Università Degli Studi Di Camerino
item Xu, Yixiang
item XU, ZHANFENG - Pepsico
item YANG, JUN - Pepsico
item FENG, YIMING - Virginia Polytechnic Institution & State University

Submitted to: ACS Food Science and Technology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/22/2024
Publication Date: 6/3/2024
Citation: Sun, Y., Nayani, N.S., Xu, Y., Xu, Z., Yang, J., Feng, Y. 2024. Rapid and nondestructive determination of oil content and distribution of potato chips using hyperspectral imaging and chemometrics. ACS Food Science and Technology. 4(6):1579-1588. https://doi.org/10.1021/acsfoodscitech.4c00196.
DOI: https://doi.org/10.1021/acsfoodscitech.4c00196

Interpretive Summary: Potato chips are a popular snack but raise health concerns due to their high oil content. Traditional methods for measuring oil content and distribution are time-consuming and involve toxic chemicals. In this study, a near-infrared hyperspectral imaging (HSI) was used for non-destructive, real-time tool determination of oil content and distribution in potato chips. The results indicate the efficacy of HSI as a rapid, non-destructive and cost-effective alternative to traditional methods for oil content determination.

Technical Abstract: Near-infrared hyperspectral imaging (HSI) was used as a non-destructive and real-time tool to determine the oil content and its distribution in potato chips. Through comparing multiple algorithmic models (PLS regression, ridge regression, random forest, gradient boosting, and support vector regression) and preprocessing methods (MSC, SNV, first-order derivatives, normalization, and baseline correction), the optimal preprocessing method and model combination (SNV-PLSR) were obtained. Furthermore, the high-precision oil content prediction model with the regression coefficient of 0.94 was achieved when further optimized the parameters of the PLSR model. The model was also used to process the hyperspectral images of potato chips at the pixel level to visualize the oil distribution of potato chips with the intent to provide a real-time, nondestructive, and rapid means of quality control for the potato chip processing industry.