Food Technology · Sensors & AI
A third of all the food the world produces never gets eaten. Not because it's spoiled — often because nobody can say for certain whether it still is.
A German-Danish research consortium just built a compact sensor that answers that question directly, by smelling the one gas that decomposing meat and fish can't hide.
The Signal
The Interreg project PRECISE has built an electronic nose that detects cadaverine — a gas produced as proteins break down — to give retailers and food-service kitchens an objective, real-time read on whether meat or fish is genuinely still safe to eat.
The PRECISE electronic nose is designed as a compact, on-site device for retail and foodservice use.
In this piece: why so much edible food gets thrown away, what cadaverine actually reveals about spoilage, how PRECISE's sensor and AI pipeline works, who built it and how it's funded, what still stands between a lab device and a retail counter, and where this fits into the wider push against food waste.
A Nose Built to Stop Good Food From Becoming Waste
The German-Danish Interreg project PRECISE has developed an electronic nose designed to objectively determine the freshness of meat and fish in real time.
Instead of relying on a sniff test, a best-before date, or someone's best guess, the device measures actual biomarkers of spoilage as they form.
That distinction matters more than it might sound. Right now, most freshness calls on meat and fish are made on caution, not evidence — and caution tends to err toward throwing food out early.
Six project partners across Germany and Denmark built the device with exactly that gap in mind: a compact, on-site tool that retailers and kitchens can actually use, not a piece of lab equipment that stays in a lab.
The goal is straightforward — cut food waste and the resources wasted alongside it, by replacing guesswork with a real-time reading of what's actually happening inside the package.
↑ Return to NavigatorWhy a Third of the World's Food Never Gets Eaten
Worldwide, about one-third of all food produced is thrown away — and uncertainty about how long it will actually last is a major reason why.
In the European Union alone, roughly 59 million metric tons of food waste are generated every year. That's not a rounding error; it's a systemic drain on resources.
Meat and fish carry a heavier penalty than most other food categories when they're discarded prematurely.
Both are resource-intensive to produce in the first place, so throwing them out early doesn't just waste the product — it wastes everything that went into raising, catching, processing, and shipping it.
Multiply that across supermarket chains, restaurants, and food-service kitchens making cautious calls every single day, and the scale of the problem becomes obvious.
It's also, in principle, one of the more fixable kinds of food waste — because the uncertainty driving it is a measurement problem, not a supply problem.
↑ Return to NavigatorWhat Cadaverine Actually Tells You About a Piece of Meat
Cadaverine is exactly what its name suggests: a gas produced when proteins in meat and fish start to decompose.
It belongs to a family of compounds called biogenic amines, alongside similar spoilage markers like putrescine and histamine, which build up as bacteria break down amino acids in the tissue.
Unlike a printed expiration date, cadaverine levels reflect what's actually happening to that specific piece of meat or fish, in its specific packaging, at that specific moment.
That's the core appeal of tracking it directly. Two identical cuts, stored slightly differently, can spoil at different rates — a fixed date can't account for that, but a live gas reading can.
Detecting cadaverine reliably isn't trivial, though. It requires a sensor sensitive enough to pick up a gas at very low concentrations, without being fooled by everything else a package of meat or fish naturally smells like.
That sensitivity-and-selectivity problem is exactly what electronic nose research in this space has been chasing for years, across multiple labs and multiple sensor designs.
PRECISE's contribution is a device built specifically to isolate that one signal — cadaverine — from the noise, and turn it into a usable, real-time freshness reading.
↑ Return to Navigator
PRECISE pairs a sensor array tuned to spoilage biomarkers with data analysis and AI to turn raw gas readings into a freshness verdict.
Inside PRECISE: Sensors, Data, and AI Working Together
Roana de Oliveira Hansen, from the Mads Clausen Institute at the University of Southern Denmark and the project's lead partner, frames the device's core trick simply: it selectively responds to specific biomarkers of meat freshness.
Selectivity is the hard part. Plenty of sensors can detect that a gas is present — far fewer can confirm which gas, at what concentration, while ignoring everything else in the air around it.
Lübeck University of Applied Sciences brought the second half of the equation: expertise in sensor technology, data analysis, and artificial intelligence.
Professor Horst Hellbrück, from Lübeck, describes the approach as combining modern sensor hardware with data analysis and AI algorithms that interpret what the sensors pick up.
That combination matters because raw sensor output isn't a clean answer on its own — it's a pattern of signals that needs to be translated into something a retailer or kitchen worker can actually act on.
The AI layer is what turns "here's a gas reading" into something closer to "here's how many hours of safe shelf life are actually left."
Related electronic-nose research elsewhere has used similar sensor-plus-algorithm pipelines — metal-oxide sensors paired with classification models — to sort meat and fish samples into freshness categories with strong accuracy, which suggests PRECISE's approach sits on solid technical ground.
↑ Return to NavigatorSix Partners, One Border: Who Actually Built This
PRECISE isn't a single-lab project — it's a genuinely cross-border effort, built by six partners spanning Germany and Denmark.
The lineup includes the University of Southern Denmark, Lübeck University of Applied Sciences, the KIN Food Institute, the Fraunhofer Institute for Silicon Technology (ISIT), Flensburg University of Applied Sciences, and the startup AmiNIC.
That mix is worth noticing. It pairs academic sensor and AI research with a dedicated food-science institute, a materials and microfabrication specialist in Fraunhofer ISIT, and a startup positioned to actually commercialize the result.
The project is funded through the Interreg Germany-Denmark Program, with European Union funding of just over 1.88 million euros.
Interreg funding exists specifically to support cross-border cooperation projects like this one — research that benefits from combining expertise on both sides of a national line rather than staying siloed in one country.
It's a structure that makes sense here: sensor fabrication, data science, food safety expertise, and commercialization rarely all live under one institutional roof.
↑ Return to NavigatorFrom Lab Prototype to Retail Counter
The stated destination for this device is deliberately practical: retail and foodservice settings, not just research benches.
That's a meaningful design constraint. A device meant to sit at a supermarket meat counter or in a restaurant kitchen has to be compact, fast, and simple enough for staff to use without special training.
It also has to be robust — real retail environments are noisier, messier, and less controlled than a university sensor lab.
The publicly available project material doesn't specify a commercial launch date, unit cost, or which retailers might pilot it first — details that matter but simply aren't part of what's been disclosed yet.
What is clear is the intended handoff: AmiNIC's presence in the consortium signals that turning this from a validated prototype into an actual product is very much part of the plan, not an afterthought.
Getting from a working lab device to something sitting on a counter usually takes exactly the unglamorous steps that don't make headlines — manufacturing at scale, cost-down engineering, and real-world validation across many types of meat and fish.
↑ Return to NavigatorWhat This Device Won't Solve — and Why That's Fine
It's worth being honest about scope here: an electronic nose that detects cadaverine addresses one specific point in a very long food supply chain.
It doesn't fix over-ordering, inefficient logistics, inconsistent cold-chain handling, or the confusing mix of "best before" and "use by" labeling that already misleads plenty of shoppers.
What it does address is narrower and, arguably, more tractable: the moment someone has to decide, right now, whether a specific piece of meat or fish is still good.
That's the decision point where caution currently costs the most — and where an objective reading could shift outcomes without asking anyone to change their broader behavior.
Tools like this tend to work best as one piece of a larger waste-reduction strategy, not a single fix for a genuinely multi-causal problem.
Judged on that narrower, realistic scope, PRECISE's device looks like a solid, well-targeted contribution rather than an overreaching claim to have solved food waste outright.
↑ Return to NavigatorManohar Insight
What I find genuinely encouraging about this project is how unglamorous its target is. Nobody is chasing a dramatic new material or a headline-grabbing discovery here.
They're chasing one specific, boring, expensive uncertainty — "is this still good?" — and building a device narrow enough to answer it well.
That's usually where the most practically useful engineering happens: not in the exotic breakthrough, but in the patient work of making an existing measurement objective, portable, and fast enough to actually use.
I'd also point out how sensible the consortium structure is. Sensor fabrication, AI, food science, and commercialization are four different disciplines, and PRECISE didn't try to force one institution to cover all of them.
The detail I'd watch closest going forward is the transition from prototype to product. A lot of promising food-tech sensors have stalled exactly at that step — proven in the lab, never quite cheap or rugged enough for a real supermarket floor.
If PRECISE clears that hurdle, the payoff isn't exotic either — it's simpler supply chains, less wasted protein, and fewer edible meals thrown out on a guess instead of a reading.
Sometimes the most valuable technology isn't the one that does something new. It's the one that finally lets us stop guessing about something we should have been measuring all along.
↑ Return to NavigatorJoin the Signal
Would you trust a sensor reading over your own nose when deciding if meat or fish is still safe to eat — or does food safety still feel like it needs a human judgment call? Share your take in the comments.
#FoodTech #FoodWaste #ElectronicNose #Sustainability #Sensors #ArtificialIntelligence #KMManoharInsights
Explore More
AI Searched 150 Million Materials and Found Two That Actually WorkAnother case of AI narrowing an impossibly large search space down to something a lab can actually test.
Scientists Create AI Skin Patch That Acts Like an Instant Personal DoctorA related biosensor story — reading biomarkers directly off the body instead of off a package.
A Single Molecule Just Became a Quantum Sensor — And It Could Change How We See ProteinsA look at just how far biomarker-sensing technology can be pushed, down to the single-molecule scale.
A Breakthrough in Nanoscale Cleaning & Microbiology: Light-Driven Nanorobots That Hunt BacteriaA different angle on food-safety-adjacent microbiology, this time hunting bacteria rather than measuring their byproducts.
Scientists Found Microplastics in Every Human Artery They TestedA sobering reminder of why objective measurement, not assumption, matters so much across the food chain.
Visit us at: kmmanoharinsights.netlify.app
About the Author
K M Manohar is an independent Sci-Tech writer and research curator who explains breakthrough discoveries that are shaping the future of technology.
Comments
Post a Comment