Smarter Safety in Longwall Mining: Real-Time Spark Detection with Machine Vision
Smarter Tech. Safer Mines
Longwall mining is among the most demanding industrial environments when it comes to safety and operational resilience. The risks of fire and equipment damage are ever-present. That’s why MRA Automation has developed a groundbreaking Axo33 Spark Detection System that uses machine vision, not traditional thermal imaging, to detect sparks at the moment they appear and before they escalate into hazards.
Built for real-time monitoring and seamless integration, Axo33’s system enhances mine safety while minimising cost and deployment costs by leveraging existing camera infrastructure already in place on-site.
Why Traditional Spark Detection Falls Short
Conventional spark detection systems in mining have typically relied on FLIR thermal cameras. While useful in some scenarios, these systems struggle with dust, confined spaces and obstructions that are common in underground operations. They are also expensive, slow to react and susceptible to false alarms that drain operator attention and trust.
Manual monitoring isn't the answer either as operators face fatigue, inconsistent lighting, poor visibility and the unpredictable nature of longwall operations. This often leads to missed events, delayed responses and higher risk in high-stakes areas such as belt transfer points, longwall shearers and hydraulic supports.
A New Approach: Machine Vision for Spark Detection
Axo33’s Spark Detection System offers a smarter, faster and more cost-effective alternative to traditional thermal-based monitoring. Instead of relying on heat signatures, it uses advanced machine vision algorithms to detect visible sparks in real time. The system connects directly to existing site cameras, eliminating the need for costly new equipment.
Key monitoring zones, such as longwall drums, conveyors, and boot ends, are continuously scanned. If a spark is detected, alerts are instantly pushed to the site’s control system, enabling a rapid response that can prevent escalation.
The system is fully configurable, adaptable to a wide range of mining environments, and integrates seamlessly with PLCs and Ethernet/IP controllers. It supports event logging, alarm triggering, and visual dashboards — all without disrupting your existing workflows.
Detected spark event
Detected spark event during camera washdown spray
What You Can Expect:
- Real-Time Detection & Response
Sparks are identified in milliseconds, allowing automatic control system actions to kick in faster, protecting people and equipment. - Reduced False Alarms
Intelligent masking, rising-edge logic, and configurable regions of interest (ROIs) reduce noise and avoid alert fatigue. - Seamless PLC Integration
Detection events are written directly to PLC tags, supporting automation, real-time control, and fast operator feedback. - Reliable Logging & Insight
Every spark detection includes a timestamped image and metadata, viewable via dashboard for auditing, operator training, and incident analysis.
Ready to future-proof your mine safety systems?
MRA’s Axo33 Spark Detection System is a smarter, faster and more scalable way to reduce fire risk, avoid downtime, and meet compliance with confidence.
Contact MRA Automation to learn more or request a demonstration.
MRA’s Axo33 Spark Detection System has significantly enhanced our operation’s safety. It gives us greater visibility into events as they happen, allowing us to respond more effectively and confidently.
— Mike Hutchinson, Automation Coordinator, Longwall Operations