"AI" Doesn't Automatically Mean Camera-Based Technology

In waste and recycling tech, "AI" usually gets associated with camera-based sorting. That's a fair assumption, most of the AI applications in this industry are built around visual recognition, cameras scanning objects on a belt and sorting them by what they look like. But that's not what Litech does.

How magnetic induction spectroscopy works

Our system is built on electromagnetic induction, the same principle behind a metal detector, just far more precise. As material passes through, the sensor picks up the electromagnetic signal each object gives off. No lens, no image, no light needed.

This is where AI comes in. Litechs AI differentiates the signal each object produces and tells different metals and alloys apart. The process is simple in principle: an object gives off a signal, and AI turns that signal into a unique material fingerprint.

How AI reads each object's material fingerprint

Different materials produce different fingerprints. A battery, a laughing gas bottle, and a laptop each give off a distinct signal, shaped by what they're made of, not what they look like. That's the difference our AI is trained to tell apart.

In practice, this means a hazard item's signature can be matched and identified before it becomes a fire risk, without a camera ever being involved. The object doesn't need to be visible, upright, or unobstructed. It just needs to pass through the sensor's field.

A safer alternative to cameras and X-ray

Camera-based systems depend on how an object looks: its shape, color, position on the belt, or whether something else is sitting on top of it. That works well for sorting by material type when objects are visible and separated. But hazard detection is a different problem. Batteries and gas cartridges often end up buried, crushed, or hidden inside other waste, exactly where a camera can't help.

Our sensor reads what an object is made of, regardless of how it's oriented, what it's covered in, or where it sits in the material stream. This becomes even more obvious in mixed waste streams. E-waste, scrap metal, and household waste often move through the same line, tangled together, with hazardous items sealed inside plastic bags or wrapped in other material. A camera can only judge what's visible on the surface, so anything inside a bag is effectively invisible to it.

Our sensor doesn't depend on sight. Because it reads the material itself, it can detect a battery or metal object even when it's fully enclosed in a bag or buried under other waste.

No cameras, no radiation

This approach is the first of its kind in the industry: no cameras, no radiation, just magnetic induction and AI reading the signal underneath. Since there's no X-ray or radiation involved, it's safer for workers and the facility, and it doesn't require the extra costly licensing that radiation-based systems do. On top of that, the design is simple: it slides in right underneath the conveyor belt, no rebuild of the line required.

For facilities handling municipal waste, e-waste, or scrap metal, that means fire risk detection that doesn't depend on visibility or line of sight, and doesn't come with the overhead of a radiation permit.

Installing a radiation-based system is rarely simple. X-ray units typically require regulatory approval, radiation safety protocols, and often structural changes to shield the surrounding area, all before the system is even running.

Litech's sensor works differently. Because it's based on magnetic induction rather than radiation, there's no licensing process, no shielding requirements, and no need to redesign the line around it. It installs directly underneath the existing conveyor belt, working with the infrastructure a facility already has rather than replacing it. For operators, that means avoiding the kind of extended shutdown that typically comes with retrofitting a radiation-based system.

Making fire prevention simple

Fire prevention in waste facilities shouldn't require complicated technology to understand. That's the goal behind how we've built this: making it simple to see what's actually inside the material stream, before it becomes a problem, using a principle that's been trusted and understood for decades, just applied with far more precision.

Curious how this applies to your material stream? Get in touch.