Food is one of the major necessities that provides energy required to perform our daily activities and for the survival of all living beings. It provides the required nutrients for nourishment and body growth. Basic foods that we eat everyday are easily prone to contamination by microbes and pathogens such as bacteria. Consumption of such foods containing harmful bacteria, viruses, parasites or chemical substances can cause more than 200 diseasesranging from diarrhoea to cancers¹.
Diarrheal diseases are the most common illnesses caused by the consumption of contaminated foods, which results in people falling ill, including possible occurrence of deaths in some cases every year. Hence, food safety is a prime concern globally and has been accorded utmost importance keeping in view the public health and well-being of consumers. Food safety exclusively deals with safeguarding the food supply chain from the initiation, growth or survival of hazardous microbial and chemical agents. As the food products travel across the supply chain, it is necessary to ensure utmost safety of the food products that pass through various stages such as processing, storage and distribution to protect the well-being of the end consumers.
To prevent consumers from consuming contaminated foods, rapid and accurate detection of microbiological and chemical hazards can play a crucial role to ensure food safety. Traditional food testing methods for assessing microbiological and chemical hazards include wet chemical estimation of constituents (i.e., moisture, protein, fat, etc.), physical methods (i.e., pH, colour, etc.), enzyme-linked immunosorbent assays, molecular biology-based methods (i.e., total viable counts, polymerase chain reaction), chromatography techniques and bio-sensor methods. These food testing methods have been found to be tedious, time consuming and expensive. There is an increasing necessity for well-trained analysts and established testing infrastructure in the Food Industry³. To improve the efficiency of food testing, attempts are being made to find an alternative technique to replace these traditional methods. In recent times, Hyperspectral Imaging (HSI) technique has emerged an alternative tool for food safety assessment.
Hyperspectral Imaging (HSI) is one of the most advanced, rapid, non-destructive and non-invasive methods that can be successfully used for assessing the quality and safety of various food products. Hyperspectral Imaging (HSI) has integrated conventional imaging and spectroscopy with computer vision into a single system/ model that can evaluate other food quality and safety attributes. Computer vision is a technique that provides information about external attributes of foods including size, shape, colour, texture and mechanical damages, while spectroscopic methods can detect chemical components, foreign objects and pesticide residues in food products at different wavelengths.
Particularly, meat products in the Food Industry 4.0 era are troubled by problems connected to safety and health, due to their complicated structure, high perishability and vulnerability to spoilage. Quality indicators of meat products are related to chemical, physical and sensory attributes, which are concerned with nutritive values and human health. Physical indicators, (i.e., colour, texture, pH, etc.,) that are related with microbial (i.e., total viable count), chemical (i.e., moisture content, protein, intramuscular fat, fatty acids, biogenic amine index, water-holding capacity, pH, tenderness, etc.) and sensory characteristics (i.e., marbling scores) can be taken to access the freshness and safety of fresh and processed meat products. Safety indicators of meat foods include chemical, microbiological and other associated hazardous matter.
Monitoring safety attributes of meat products in the food chain through HSI technique can enhance consumer confidence and product traceability as well. HSI has been successfully used for evaluating the microbial quality of beef under storage, total volatile basic nitrogen (TVBN) content in pork, lipid oxidation in lamb, inosinic acid in chicken and total viable count in peeled pacific white shrimp². A European Firm, IRIS Technology Group is currently working on in-line hyperspectral NIR technology for continuous monitoring and detection of foreign bodies, such as shells, stones, other arthropods, net fragments and even plastic packaging residues in seafood/ fish products⁵.
HSI technique is a chemical-free and environmentally friendly method that is successfully applied for rapidly assessing the quality and safety of various food products. HSI generates huge amount of data. However, processing of such data and extracting useful information is a challenging task, which has limited its use in real time applications in the Meat Industry. Development of efficient and emerging algorithms and chemometric methods will lower the dimensionality of data and enhance the computational process. Therefore, models that monitor food safety have been developed with improved performance and robustness by avoiding irrelevant variables and redundancies. The computational load and time of the proposed models can be reduced in the best possible way by developing multivariate models using the featured/key wavelengths representing the specific quality attributes. It further reduces the cost of the equipment also.
On the other hand, selection of featured wavelengths in the laboratory is off-line and manual which has diminished the processing efficacy. Selection or development of such improper algorithm may lead to loss of crucial information related to product characteristics on an industrial level. Thereby, advanced machine learning algorithms such as deep learning and life-long machine learning techniques should be adopted for such real time applications.
Unlike traditional machine learning techniques, deep learning technique works on automatic feature learning from hyperspectral data. Lifelong learning (LL) is continuous and a more efficient learning approach that helps gain additional information and knowledge from hyperspectral data for future learning and problem solving. Deep learning and lifelong learning techniques require huge amount of data to develop more reliable and robust models and further research is necessary in this area to develop simple networks for assessing food safety at the industrial level.
Environmental conditions such as humidity and temperature play a crucial role in determining the safety as well as change in quality of food over time. These environmental as well as illumination conditions effect spectral information and the models developed under these conditions cannot yield the same accuracy in other conditions. This can be solved by creating a common database to share the HSI data and models developed by various researchers from various corners of the globe. This will be helpful to develop more robust and reliable models which can work at any environmental and illumination conditions. Most of the research is focused on spectral analysis but the acquired spatial data provides some important information which is not potentially utilized. Therefore, combination of both spatial and spectral information during image processing can develop accurate models towards food safety applications.
Future studies should focus on industrializing the laboratory prototypes that have been developed so far, for undertaking safety evaluation of meat products. Low-cost materials can be developed for fabricating hyperspectral imaging systems for industrial applications².
With the availability of high-speed and ever advancing computing systems, hardware and software systems (tools) can be developed and updated regularly to obtain stable and high signal-to-noise optical signals for rapidly processing hyperspectral images for online or real applications in the Meat Industry in near future⁴. Furthermore, the ongoing research of integrating HSI with smartphones can benefit consumers worldwide to monitor the quality and safety of fresh and processed meat products shopped at markets as well as in the case of home deliveries to help ward off concerns associated with the safety of human lives.
1. Gizaw, Z. (2019). Public health risks related to food safety issues in the food market: a systematic literature review. Environmental health and preventive medicine, 24(1), 1-21.
2. Pu, H., Wei, Q., & Sun, D. W. (2022). Recent advances in muscle food safety evaluation: Hyperspectral imaging analyses and applications. Critical Reviews in Food Science and Nutrition, 1-17.
3. Jia, W., van Ruth, S., Scollan, N., & Koidis, A. (2022). Hyperspectral imaging (HSI) for meat quality evaluation across the supply chain: Current and future trends. Current Research in Food Science, 5, 1017–1027.
4. Saha, D., & Manickavasagan, A. (2021). Machine learning techniques for analysis of hyperspectral images to determine quality of food products: A review. Current Research in Food Science, 4, 28-44.
About the Authors:
1. Naveen Kumar Mahanti
Scientist, Post-Harvest Technology Research Station,
Dr. Y.S.R Horticultural University,
Venkataramannagudem, Andhra Pradesh.
Email ID: email@example.com
2. Sai Prasanna Narakatla
Department of Chemical Engineering,
IIT, Tirupati, Andhra Pradesh.
3. Subir Kumar Chakraborty
Principal Scientist, Agro Produce Processing Division,
ICAR-Central Institute of Agricultural Engineering,
Bhopal, Madhya Pradesh – 462038.
The views/opinions expressed by authors on this website solely reflect the author(s) and do not necessarily reflect the views/opinions of the Editors/Publisher. Neither the Editors nor the Publisher can be held responsible and liable for consequences that may arise on account of errors/omissions appearing in the Articles/Opinions.