Recently, photonics has significantly contributed to the provision of nutritious and safe food, establishing a sustainable supply chain in the food industry. The application of spectrometry devices and hyperspectral imaging provides detailed information on the nutritional value of food, such as protein levels in wheat harvests, to the health of fish stocks. This article provides an overview of how photonics-enabled hyperspectral imaging can aid in food sorting.
Image Credit: BearFotos/Shutterstock.com
What is Hyperspectral Imaging?
Hyperspectral imaging is a robust analytical technique that intertwines the abilities of spectroscopy and imaging to gather detailed information about an object’s surface structure, characteristics, and composition in a way that is otherwise unreliable by conventional imaging techniques.
With this technique, hundreds of images are collected at different wavelengths for each pixel in an image. Consequently, objects and materials are identified by analyzing their unique spectral signatures.
For instance, the human eye has only three color receptors, green, blue, and red, and hyperspectral imaging measures the continuous spectrum with a fine wavelength resolution in each pixel of the image.
The collected data is in the form of a hyperspectral cube represented in two-dimensional (2D) form. From spectral analysis of components to data acquisition, photonics, which deals with the manipulation of light, plays a crucial role.
Hyperspectral Imaging - Working Principle and Results
Hyperspectral imaging combines the working principles of spectroscopy and imaging to gather detailed information about an object. Using this technique, hundreds of images are collected at different wavelengths for each pixel in an image. Consequently, objects and their constituting materials are identified by analyzing their unique spectral signatures.
For instance, the human eye has only three color receptors: green, blue, and red, and hyperspectral imaging measures a continuous spectrum with a fine wavelength resolution in each pixel of the image. The collected data is in the form of a hyperspectral cube represented in a two-dimensional (2D) form.
The resulting images provide both spectral and spatial information regarding the chemical and physical properties of an object. While spatial information provides data on the materials’ spatial distribution and areal separation, spectral information helps in its identification and classification.
Importance of Food Sorting in the Food and Beverage Industry
Food sorting in the food and beverage industry ensures food production of satisfactory quality and is implemented at different stages in the food supply chain. Although this process was initially adopted to remove foreign objects from food, its use was later extended to include quality analysis.
Digital sorters have recently replaced conventional manual sorting systems to eliminate unwanted materials. Digital sorters consist of various components and work on different principles. Some sorters use cameras, while others use lasers, and others use a combination of both to analyze merchandise via a bottom view, a top view, or both.
Hyperspectral Imaging in the Food and Beverage Industry
Hyperspectral imaging involves the simultaneous incorporation of spectroscopic and imaging techniques to assess different components and their spatial distribution in the tested products. It is a non-invasive and non-destructive tool used in food and beverage applications.
Principal component analysis (PCA), the process of computing principal components, is used in combination with hyperspectral imaging to change the basis of the spectral data. This combination of analytical techniques can be used without interfering with the product or without wasting products because quality standards are not met.
Hyperspectral imaging, used in the food and beverage industries, offers food authentication and analysis in areas of fruits, vegetables, meat, grains, dairy products, seaweed, and powders. This analytical technique is used to:
- Analyze percentages of water and fat in the food products
- Identify defects
- Characterize product quality
- Locate contaminants
For instance, a study reported in the journal Food Control used hyperspectral imaging in the 400–1000 nm spectral range to scan minced beef adulterated with other meat. Techniques, including the partial least squares regression (PLSR), the ensemble Monte Carlo variable selection (EMCVS), and a range of spectral pre-treatments, were used to predict the amount of beef in the samples.
Good results were obtained using EMCVS on the pre-treated spectra. The obtained results were achieved in the full range and excellent prediction results were obtained with nine optimum wavelengths. Thus, this method can be used to detect and distinguish adulterated minced beef.
Similarly, another article published in Food Systems reported the use of hyperspectral imaging in the range of 400–1000 nm and with the multivariate analysis to sort Hass avocado fruits. The data obtained from the hyperspectral imaging revealed specific light bands linked to avocado ripeness and moisture.
This information can be used to create models to predict the moisture content of the avocados. Additionally, the models proposed in the present study were accurate and met the standards for moisture and dryness in avocados.
Hyperspectral imaging helps detect surface defects and classifies them based on their quality. An article published in the Journal of Food Science reported the detection of whole-surface defects using an online hyperspectral sorting device. Here, the image data of navel oranges were collected using online detection sorting equipment, and the spectral image focused on the wave peak at 1655.72 nm.
The accuracy of the detection of surface defects was achieved by adjusting the light to better see the edges of the oranges. The sorting was 100% accurate, revealing the sensitivity and efficiency of the method in detecting surface defects.
Conclusion
Overall, photonics-enabled hyperspectral imaging has great potential to revolutionize the food sorting process in the food and beverage industries. The ability of the hyperspectral imaging technique to capture detailed spectral information followed by precise analysis has immensely contributed to the quality assessment process in the food industry.
This technology enables the efficient and accurate detection of defects, ripeness, and other differentiating attributes of foods. Moreover, hyperspectral imaging has paved the way for high-throughput sorting systems in real-time, making it a reliable analytical technique for detecting food defects and quality attributes.
It is imperative to explore cost-efficient manufacturing methods to make this technology more accessible to small-scale food industries. Collaborative efforts from industry stakeholders, researchers, and policymakers can help establish regulations to ensure the safe implementation of photonics-enabled hyperspectral imaging in the food-processing industry.
More from AZoOptics: High-Precision Optical Metrology Techniques and Their Principles
References and Further Reading
What Is Hyperspectral Imaging: A Comprehensive Guide. Accessed on 30 November, 2023 at https://www.specim.com/technology/what-is-hyperspectral-imaging/
Role of Laser Emitters in Industrial Food Sorting. Accessed on 30 November, 2023 at https://www.azom.com/article.aspx?ArticleID=21007 .
Food Quality And Composition Analysis With Hyperspectral Imaging. Accessed on 30 November, 2023
Achata, E. M., Mousa, M. A., Al-Qurashi, A. D., Ibrahim, O. H., Abo-Elyousr, K. A., Aal, A. M. A., Kamruzzaman, M. (2023). Multivariate optimization of hyperspectral imaging for adulteration detection of ground beef: Towards the development of generic algorithms to predict adulterated ground beef and for digital sorting. Food Control, 109907. https://www.sciencedirect.com/science/article/abs/pii/S0956713523003079
Metlenkin, D. A., Platova, R. A., Platov, Y. T., Fedoseenko, O. V., Sadkova, O. V. (2023). Avocado fruit sorting by hyperspectral images. Food systems, 6(1), 46-52. https://www.fsjour.com/jour/article/view/228
Shang, M., Xue, L., Zhang, Y., Liu, M., & Li, J. (2023). Full‐surface defect detection of navel orange based on hyperspectral online sorting technology. Journal of Food Science. https://doi.org/10.1111/1750-3841.16569
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.