NIT Rourkela Patent: AI food testing system detects spice adulteration in seconds, pilot industry trials soon

National Institute of Technology (NIT) Rourkela has obtained a patent for a food-testing system designed to identify adulteration in spices and quantify the extent of contamination within seconds, marking a new intervention in the growing concern around food quality monitoring in India. As per the NIT Rourkela press release, the technology has been developed as a faster alternative to conventional laboratory procedures that are often expensive, chemical-intensive and time consuming.

The press release noted that food adulteration, particularly in powdered spices, remains a recurring issue in Indian markets where low-cost substitutes are frequently mixed to increase bulk. Existing verification methods largely depend on chromatography and molecular testing, processes that require trained manpower and substantial turnaround time before a result is obtained. In contrast, the NIT Rourkela system has been designed to function as a rapid screening mechanism capable of being used in industrial quality control units as well as food testing laboratories.

The institute stated that the patented model combines Fourier Transform Infrared Spectroscopy (FTIR) with machine learning tools that read the infrared absorption signatures of food samples and compare them against learned adulteration patterns. Instead of merely indicating whether a product is pure or impure, the system estimates the level of adulteration, a feature researchers say could help processors and regulators make quicker compliance decisions.

Also read Rice research needs fortification too, say scientists at agriculture universities

NIT Rourkela: Patent focuses faster food checks

As per the details shared in the institute’s press release, the patent has been granted under the title “Method and System for Detecting and Quantifying Adulteration in Food Stuff”. The work was carried out by Sushil Kumar Singh, Late Poonam Singha and MTech graduate Rishabh Goyal from the department of food process engineering.The associated study was also published in the journal ‘Food Chemistry.’

The official NIT Rourkela release quoted Singh saying that the objective was to address the “absence of a fast and reliable spice adulteration detection system” in routine food processing environments. He noted that the team combined standard rapid detection equipment with what he described as “novel machine learning approaches” to improve decision-making speed. The release said the system is intended to reduce delays caused by lengthy sample preparation and repeated chemical analysis.

A demonstration study cited by the institute focused on coriander powder, where sawdust was used as the adulterant. In that case, the machine learning-assisted model reportedly detected contamination with an accuracy of about 92%. Researchers indicated that the same framework could be expanded to other spices and food materials if trained under varied datasets.

Also read Education ministry plans Rs 14 crore grants for Prime Minister Research Chairs, Rs 4-6.5 crore fellowships

NIT Rourkela: Pilot testing planned

Beyond laboratory validation, NIT Rourkela said it is now looking at pilot-scale collaborations with food companies to test whether the model can be embedded into existing production lines for real-time screening. Such deployment, if successful, would allow raw material and processed spice batches to be checked without waiting for external lab reports.

The institute also indicated that future experiments will attempt to widen the detection range beyond spices.