Detecting Food Adulteration in Real time
To ensure consumer protection against fraudulent activities, authentication of food and the detection of adulterants in various food items requires be taken into consideration. The methods used in industries to detect food adulteration are quite expensive and complex. The specialised infrastructure is required for quality evaluation methods. These methods demand intensive manual labour and are sometimes quite tedious and inefficient. Hence, artificial Intelligence (AI) can be used as a platform to develop a low- cost automated system that could be used by the end-user to detect adulteration in fruits, vegetables and dairy products.
Almost all foods are subject to food adulteration including dairy products, grains, seafood, oils, alcoholic drinks, honey, etc. Also, the fruits and vegetables being sold in the market are not pure as they are injected with harmful chemicals and pesticides. There are multiple ways of impairing the quality of the food. Food is considered to be adulterated if the valuable constituents of the food are removed or if the poor quality of food products are concealed with an actual food product.
Prohibited colouring agents are common food adulterants even though they cause many types of health hazards. Also, the permissible food dyes are used in large quantities to attract customers. There are other colouring agents like Metanil Yellow and Rhodamine B etc. that are widely used in confectionery products, dried fruits, wines, bitter sodas, juices, sauces, pastes, and spices. Food dyes including Allura red and sunset yellow are used in Strawberry jelly and wine.
Every civilised society is using food preservatives but such practice can pose a threat to public health. Secure and effective preservative production for perishable food products is a subject of intense study. For example, a suitable mixture of potassium lactate and sodium diacetate is observed as an acceptable preservative under refrigeration conditions. Salt is found to be an effective meat preservative but it can cause hypertension. Safety and efficiency of preservatives are the fundamental criteria that have to be considered for long-term food preservation. However, malpractice like the addition of harmful preservatives to food is often reported.
It seems that traditional food safety methods are not enough to control the issue. Therefore, innovative and advanced ways have to be developed that may be used by the common public or less trained people in this field to keep a check on the quality. The methods should be user-friendly and affordable tools should be designed to evaluate the food quality and achieve the desired aim.
Need for Ai-based detection
Artificial Intelligence (AI) can be used as an opportunity in the food industry. It has a major role to support our food system as it can help in precision farming and many other applications in food production and food consumption. It can also be used as a quality control measure in the food sector. AI is changing the way one thinks about food production, quality, delivery etc. and the era of intelligent mobile apps has a big contribution to this transition.
The researchers at the Indian Institute of Technology, Hyderabad have been working on a project to develop a smartphone-based system that is equipped with sensors to detect the amount of adulteration in milk. Initially, they have developed a system to measure the acidity of milk through an indicator paper that changes colour based on the level of adulteration. Besides, they have also developed algorithms that can be incorporated on a mobile phone to detect the colour change accurately.
The artificial brains can be used efficiently to create food databases and analyse them. It has the potential to create a healthier and more affordable food industry for workers as well as consumers. AI can be used as a platform to develop a low-cost automated system that could be used by the end-user to detect adulteration in fruits, vegetables and dairy products.
The different modern methods in this area like electronic tongues, electronic noses, computer vision, spectroscopy and spectral imaging, and so on, have been widely used to detect food quality.
The electronic nose (E-nose) and electronic tongue (E-tongue) are devices that work the same as human nose and taste organs and are composed of an array of sensors. These systems have broad applications in the food adulteration detection system as the complex data sets from E-nose and E-tongue signals coupled with multi- variate statistics constitute fast and effective instruments for classifying discriminating, recognising and identifying samples, as well as predicting the concentrations level of various compounds.
The sensory devices such as a spectrophotometer, thermometer etc. are used to examine solid products. Further, machine learning (ML) and deep learning (DL) models can be applied for classification and pattern recognition based upon which the prediction is done about the presence of adulterants in food. Also, the other parameters such as odour, taste, the flavour of the food, aroma appearance and texture of the food can be analysed from the predictions of the model. Thus, to build a computer-based food adulteration detection system, one needs to use IOT devices to sense the data and ML models for predictions based on the collected data.
Work in progress
There has been a practical application of ML at Amazon that uses these algorithms to predict the quality of groceries. It grades different types of products and prevents the wastage of fruits and vegetables by providing consistent results. It predicts if the fruit quality is good or bad. The different fruits stored in the warehouse are scanned through a set of cameras and sensors to inspect their quality. The researchers at the Indian Institute of Technology (IIT) Hyderabad have been working on a project to develop a smartphonebased system that is equipped with sensors to detect the amount of adulteration in milk. Initially, they have developed a system to measure the acidity of milk through an indicator paper that changes colour based on the level of adulteration. Besides this, they have also developed algorithms that can be incorporated on a mobile phone to detect the colour change accurately.
An innovative kit has been developed to detect the adulteration of milk by the National Dairy Research Institute (NDRI), Karnal. The paper strip-based tests have been developed which can rapidly detect adulteration of milk containing neutralisers, urea, glucose, hydrogen peroxide, sucrose and maltodextrin. The test involves dipping a strip in the milk sample for a short duration followed by immediate visualisation of the colour of the strip. A smart portable kitchen gadget has been developed to check the freshness of raw meat, poultry or fish. It's a wireless device designed by Swiss scientists that detects these non-veg products if they are fresh, spoiled or in the stage of getting spoiled and the results are displayed on the smartphone. It contains a sensor that collects the gases emitted by meat to examine its freshness and helps to avoid food wastage and of course takes care of food safety.
Challenges
There is a lot of work to be done to improve the accuracy of the detection system for food adulteration. It has been observed that the datasets are not readily available online because the researchers do not provide any links to the same. As there is a lack of annotated datasets and the creation of labeled datasets for different foodstuffs is a time-consuming task.
In the future, these resources can be provided to utilise these by other researchers so that they can focus only on enhancing the efficiency of the system by developing new food adulteration detection methods. It has also been analysed that deep learning approaches for food adulteration detection are in demand.
Hence, researchers can experiment with these approaches to achieve improved results. And, also there is a need to build online systems which can perform food adulteration detection. Research can be done to develop a low-cost smartphone-system to detect food adulteration which can serve as an aid to end-users for their quality satisfaction. This field needs to be integrated with a realtime system for food adulteration detection.