Portable AI Food Analyser Detects Food Contamination in Real-Time

The increasingly fragmented food ingredient industry faces the challenges of rampant adulteration, quality inconsistencies and taste inaccuracies. Detecting fraudulent ingredients, modifications and contamination is extremely difficult, with international food supply chains adding to the complexity. Current solutions such as laboratory analysis, track and trace, have met with limited success and can lack accuracy, or be costly, tedious or complex.

This technology being offered is a rapid food analyser built on an Internet of Things (IoT) platform and housed in a portable device. It uses advanced chemometrics, big data, artificial intelligence (AI) modelling and sensor technology to rapidly authenticate, identify adulteration and predict characteristics such as taste profiles—without destroying the sample.

This technology is available in a proprietary portable device that requires 5g of the sample food product to capture its metabolomic signatures. Then a proprietary algorithm based on a combination of chemometrics and AI analyses the food product at the molecular level in a non-destructive manner. It digitally matches the sample to an encrypted cloud-based database and identifies the quality, adulteration, geographic origins, and even taste profile of the food product.

In tests conducted, the device can accurately match and detect all the origins of raw coffee beans. It has also successfully graded and practiced quality control on raw cacao, and determined the quality and freshness of milk without chemical analysis. 

The technology is applicable for:

• Dried plant-based products, such as tea, coffee, grains, herbs, spices, cacao
• Liquids, such as milk and edible oils
• Non-organic materials

The technology can be used by processors, wholesalers, manufacturers and quality assurance (QA) at any point in the supply chain and retrieve results in real-time. Some examples include:

• Procurement: Digitally match required quality instead of using physical samples
• Acceptance check: Verification by supplier before shipping
• Ingredient check: Buyer conducts quality check prior to production
• End-product check: Conduct quality check prior to shipment
• TasteMap: Predict taste profiles without tasting
• Nutritional analysis: Predict content, such as fat, sugar
• Quality control: Identify ingredient freshness and inconsistencies
• Blend optimisation: Recommend blends to achieve target taste profiles and quality
• Adulteration screening: Identify non-targeted additives

Food contamination is estimated to cost the global food economy US$10-15 billion a year, and affects about 10% of all commercially sold food products, according to the Grocery Manufacturers Association.

In addition, this technology is suitable for the following industries that are on a growth path:

• US$623 billion grain industry with a compound annual growth rate (CAGR) of 3.6%
• US$102 billion coffee industry with a CAGR of 4.3%
• US$116 billion traditional Chinese medicine (TCM) and herb industry with a CAGR of 15%
• US$40 billion spice/cacao market with a CAGR of 4.9% 

• Speed and Accuracy: Identifies the quality, adulteration, geographic origins, and taste profile of food ingredients in seconds. In contrast, visually inspection is inaccurate, while tasting requires preparation time
• Portability: Technology is housed in a portable device allowing analysis to be done on site and in real-time
• Non-destructive testing: Analyses the food product in a non-destructive manner, without the need to prepare a sample
• Knowledge management: Knowledge is retained in the AI model generated, along with the training data from the quality control (QC) manager, even if the QC manager leaves the company
• Significant cost savings for producers