As the industry continues to face rising fire incidents and insurance challenges, operators are rethinking how they manage risk. AI-powered X-ray and camera detection systems offer a practical, upstream solution to a growing problem: hazards entering the facility long before staff have a chance to see them.
By Lia Kiam
Insurance has become one of the most pressing operational issues facing waste and recycling facilities. In the last several years, the number and severity of facility fires have pushed insurers to reevaluate the industry entirely. Premiums and deductibles have climbed across the board, exclusions have become more common, and some carriers have stopped writing policies for waste and recycling altogether.
Underwriters are now looking far beyond suppression systems and response procedures. Increasingly, they want evidence of upstream risk prevention–proof that facilities can identify lithium-ion batteries, pressurized cylinders, and other high-risk items before they reach equipment. For many operators, fire prevention is no longer just a safety issue, it is also a core requirement for maintaining insurability.
This shift in the insurance landscape is one reason early-stage hazard detection—especially using AI-powered X-ray and camera systems—is emerging as a best practice across MRFs, transfer stations, scrap yards, WTE facilities, e-scrap operations, and more. These tools give operators a new level of visibility into inbound material and generate data that helps quantify and reduce risk in ways the industry has never been able to do before.
Why Upstream Visibility Matters More Than Ever
Most fires do not start in downstream processing lines—they start upstream, often within minutes of unloading material. Lithium-ion batteries buried inside bags, electronics, or compacted commercial loads can ignite with a single puncture, friction, or heat source. Other hazards such as propane cylinders, ammunition, and dense metal objects can trigger thermal events or damage equipment if they make it too far into the process.
Manual inspection remains important, but it can only catch what is visible at the surface. Several industry trends have made upstream detection increasingly critical:
- Embedded batteries are everywhere: From vapes to earbuds to children’s toys, small-format lithium-ion cells now arrive in nearly every type of load—and are rarely visible.
- Loads are more complex: Municipal waste composition swings daily. C&D and commercial routes bring unpredictable contamination. E-scrap often includes mixed chemistries and partially disassembled devices.
- Throughput has increased: Operators have only seconds to assess a load on a fast-moving floor.
- Insurance requires documented prevention: Carriers now ask for evidence that hazardous items are being detected and removed before processing begins.
For these reasons, many facilities are turning to imaging systems that can see through material depth and identify hazards early.
How X-Ray + Camera AI Detection Works
Modern hazard detection systems combine imaging modalities—X-ray and traditional RGB cameras—with machine learning models trained specifically on waste streams.
X-Ray Imaging
X-ray sensors reveal density and shape signatures hidden beneath layers of material, allowing operators to spot:
• Lithium-ion batteries embedded in devices
• Metal cylinders and pressurized tanks
• Dense objects that could damage downstream equipment
Unlike surface-level cameras, X-ray imaging can see through burden depth, finding hidden hazards underneath material.
Camera Imaging
High-resolution cameras complement X-ray data by spotting:
• Large, bulky items on tip floors (i.e. water heaters, gas tanks)
• Spikes in material content on conveyor lines
AI Identification
The combined imaging feed is analyzed in real time by AI models trained on millions of examples from real-world waste streams. The system identifies hazards, flags anomalies, and gives operators a clear indication of what to remove and where to find it.
Rather than replacing workers, this technology gives them clear visibility into what the human eye cannot see.
Images courtesy of Visia.
The Value of the Data Behind the Detection
One of the most impactful benefits of these systems is the dataset they generate on inbound loads, something the industry has historically lacked. Facilities gain detailed information such as:
How many hazards arrive per day, week, and season: Battery volumes often spike during holidays, move-out periods, or storm recovery. With data, supervisors can plan staffing and workflows accordingly.
Which routes or generators drive the most risk: Many facilities find that a small number of commercial accounts or municipal zones produce most of their hazardous items. This enables targeted education and corrective action.
The mix of hazard types found in inbound loads: Facilities can track lithium-ion vs. alkaline batteries, propane tanks vs. aerosol cans, and other specific categories.
Documentation for insurers: Carriers increasingly request evidence of preventive practices:
• Number and type of hazards detected
• Time and location of detection
• Safe removal logs
• Long-term trend lines
Quantified data demonstrates risk mitigation and strengthens conversations to lower premiums.
Support for regulatory and EPR reporting: As states implement new stewardship programs, facilities with structured inbound data are positioned to comply more easily. This new level of information transforms fire prevention from a reactive task into a measurable, managed process.
Integrating Detection into Daily Operations
Facilities that successfully incorporate these systems typically adopt the following best practices:
- Establish clear roles for responding to alerts: Operators must know who isolates a load, who removes the hazard, and how incidents are logged.
- Maintain a designated hot-load area: Smoldering or suspect material needs a safe, isolated zone away from equipment and structures.
- Review inbound data regularly: Supervisors can use weekly trend reports to adjust staffing, traffic flow, and operational priorities.
- Maintain equipment consistently: Like any imaging system, proper calibration and cleaning ensures reliable performance.
- Train staff thoroughly: Teams should learn how to recognize hazards, interpret alerts, and execute safe removal procedures.
- When incorporated properly, detection and data become part of everyday floor operations—not a separate layer.
Prevention, Proof, and Operational Confidence
As the industry continues to face rising fire incidents and insurance challenges, operators are rethinking how they manage risk. AI-powered X-ray and camera detection systems offer a practical, upstream solution to a growing problem: hazards entering the facility long before staff have a chance to see them.
With clearer visibility at intake and data that quantifies the risks entering the gate, facilities can improve safety, reduce downtime, strengthen insurance positions, and operate with greater confidence in an increasingly unpredictable waste stream. | WA
Lia Kiam runs Strategy and Operations for Visia, the recycling industry’s only multi-modal AI platform. Her work focuses on helping facilities improve safety and efficiency by detecting hazards like lithium-ion batteries through burden depth. Before joining Visia, she was an investor at Blackstone. Lia can be reached at [email protected] or visit .