The Future of Video Monitoring: Protection, Not Surveillance. Protecting People, Equipment, and Production

Why is a pre-shift equipment check not just a formality, but a matter of safety and efficiency?

When I worked as an engineer at the Murmansk Transport Branch, I was responsible for dozens of pieces of equipment: loaders, cranes, power substations, and transshipment facilities. The productivity of the entire terminal — the ability to handle a specified volume of cargo by sea, rail, and road — depended directly on the technical readiness of this equipment.

But, as is often the case, tension constantly arose between the operations department and the maintenance department. Operations demanded: "Fix it faster — we're losing time!" Maintenance replied: "Handle the equipment more carefully!" At the heart of this struggle is the same problem: insufficient control over equipment preparation for work.

A Real-Life Story: When "Failing to Look" Led to an Accident

One striking incident occurred in winter. Before his shift, a machine operator passed his medical exam, received his trip ticket, approached the machine — and, without inspecting it, started the engine and began moving. He didn't notice that the machine was connected to a pre-heating system. As a result, he ripped the cable out along with the electrical panel from the building wall, and all of this "dangled" behind him across the terminal territory.

When his colleagues stopped him, he was genuinely surprised: "Not me! Someone else!" We had to show him the CCTV footage — only then did he admit fault.

Reactive Management is Not the Solution.

Yes, the employee was disciplined. Yes, next time he might be more careful. But in six months, something similar could happen again — to him or his colleagues. Because the system hasn't changed. People still hope that "someone else checked," and the responsibility for inspection remains a formality.

Scaling the Problem: From the Terminal to the Entire Enterprise.

Now I work at the company's head office, and I see that this problem is not local — it is systemic. At one of Nornickel's mines, we faced the same issue: electric locomotive drivers arrive for their shift, get into the cab, and drive off without performing the mandatory pre-trip inspection of cars, couplings, brakes, etc. "The mechanics checked everything" is the standard response.

We installed cameras at the locomotive parking areas. The first recordings showed that most employees do not conduct inspections. After discussions with management, some began to "cheat" the system — parking the equipment in different spots so the cameras wouldn't record the lack of inspection. Only after the chief engineer intervened did the situation begin to improve.

But even then, another problem arose: who would watch 40+ cameras daily? Doing this manually is unrealistic. This is where we decided to apply neural network-based video analytics.

The Solution: A Neural Network as an "Invisible Inspector" Before the Shift.

We set ourselves a task:

Automatically record whether an employee conducts a pre-shift equipment inspection according to a checklist — without human intervention.

Drivers have a clear regulation:

• Check lighting devices;
• Inspect the control panel;
• Test the brakes;
• Check for sand;
• Walk around the cars and inspect couplings, etc.

We trained the neural network to recognize these actions. If an employee fails to perform a certain point, the system generates a short video clip (up to 20 seconds) and sends an email notification with a link and metadata: who, when, on which shift, and exactly what they failed to do.

Result:

• We only receive "violations" — no need to watch hours of video;
• Employees know they are being watched;
• The quality of inspections has increased because control has become constant and objective;
• The response has become proactive rather than reactive.

It's like traffic cameras: you don't know when you'll be recorded, so you always follow the rules.

Benefits and Potential of the Solution

1. Routine automation — freeing people from watching videos.
2. Objectivity — no subjective assessment, only facts.
3. Preventive effect — employees start following procedures knowing they are being monitored.
4. Scalability — can be applied at any facility: mines, quarries, terminals, workshops.
5. **Flexibility** — easy to adapt to different checklists and types of equipment (LHDs, dump trucks, conveyors, etc.).

We are already planning to expand the use of video analytics to other types of equipment and processes — from PPE monitoring to compliance with traffic routes.

Important Technical Aspects of Implementation

1. Infrastructure — video streams require powerful computing resources.
2. Data security — processing must occur locally (at the mine) to avoid transferring terabytes of data to the cloud or other regions.
3. Traffic optimization — only short clips with violations are sent to the cloud (low weight, high value).
4. Integration with IT infrastructure — it is important to comply with the enterprise's information security requirements.

Conclusion: Safety Culture Starts Small — With a Pre-Shift Check

This case study shows that even a seemingly simple procedure — inspecting equipment before work — can become a growth point for the entire industrial safety system.

Technology does not replace people — it helps people be better.

Video surveillance + video analytics is not "surveillance," but a tool that:

• protects employees from accidents;
• protects equipment from breakdowns;
• protects the business from downtime and losses.

We have seen that it works. And we invite all colleagues to share their experience, implement, adapt, and scale. Because safety is not about fines and punishments. It's about culture, responsibility, and the technologies that help build it.