Hyperspectral shortwave infrared image analysis for detection of adulterants in almond powder with one-class classification method

Authors: FAQEERZADA, MOHAMMAD - Chungnam National University; SNATOSH, SANTOSH - Chungnam National University; JOSHI, RAHUL - Chungnam National University; LEE, HOONSOO - Chungbuk National University; KIM, GEONWOO - Orise Fellow; Kim, Moon; CHO, BYOUNG-KWAN - Chungnam National University

Submitted to: Sensors
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/11/2020
Publication Date: 10/16/2020
Citation: Faqeerzada, M., Snatosh, S., Joshi, R., Lee, H., Kim, G., Kim, M.S., Cho, B. 2020. Hyperspectral shortwave infrared image analysis for detection of adulterants in almond powder with one-class classification method. Sensors. 20(20), 5855. https://doi.org/doi:10.3390/s20205855.
DOI: https://doi.org/10.3390/s20205855

Interpretive Summary: Common quality analysis methods for powdered food products are typically used to screen low volumes of food samples for known adulterants. However, to address increasing problems of food fraud, the need for rapid and non-destructive methods of non-targeted detection of unexpected materials that can be used for high-throughput food inspection is growing. This study investigated shortwave-infrared hyperspectral imaging (SWIR-HSI) for nondestructive authenticity analysis of two varieties of almond powder. Samples adulterated with apricot and peanut powders at different concentrations were imaged using the SWIR-HSI system. A one-class classification technique known as DD-SIMCA and a partial least squares regression (PLSR) model were tested on the image data. DD-SIMCA showed 100% sensitivity and 89–100% specificity for detecting adulterated samples. The PLSR model produced maps visualizing the adulterant concentrations in the samples, and predicted apricot powder concentration in both almond powders very well with a low rate of error, but showed a slightly higher error rate for peanut powder concentration in one of the two almond powders. These results show that SWIR-HSI could provide an effective means for high-throughput adulterant screening of powdered foods to help the food industry and regulatory agencies prevent and detect food fraud.

Technical Abstract: Widely used techniques for analyzing the quality of powdered food products focus on targeted detection with low-throughput screening of samples. Owing to potentially significant health threats and large-scale adulterations, food regulatory agencies and industries are in need of rapid and non-destructive analytical techniques for the non-targeted detection of unexpected compounds present in products. Accordingly, shortwave-infrared hyperspectral imaging (SWIR-HSI) for high throughput authenticity analysis of almond powder was investigated in this study. Two different varieties of almond powder, adulterated with apricot and peanut powders at different concentrations, were imaged using the SWIR-HSI system. A one-class classifier technique, known as data-driven soft independent modeling of class analogy (DD-SIMCA), was used on collected data sets of pure and adulterated samples. A partial least squares regression (PLSR) model was further developed to predict adulterant concentrations in almond powder. Classification results from DD-SIMCA yielded 100% sensitivity and 89–100% specificity for different validation sets of adulterated samples. The results from the PLSR analysis yielded a high determination coefficient (R2) and low error values for each variety of almond powder adulterated with apricot, but a relatively higher error rate for the second variety of almond powder adulterated with peanut powder. PLSR-based concentration mapped images visually characterized the adulterant (apricot) concentration in the almond powder. These results demonstrate that the SWIR-HSI technique combined with the one-class classifier DD-SIMCA can be used effectively for high-throughput quality screening of almond powder regarding potential adulteration.