The eyes your control system doesn’t have
A plant manager running an advanced process control system told us what he actually wanted: to give the control loop “eyes.” However well-tuned, the loop optimizes what it can measure — temperature, pressure, flow — and is blind to the belt tracking off, the bearing running hot, the person in the exclusion zone. Video analytics is the perception layer that fills that gap, and it runs on top of the cameras a plant already owns rather than replacing them: one optical asset that anticipates mechanical failure and guards people safety — exactly the equipment-person interaction risk Chile’s mining-safety framework (DS 132, enforced by SERNAGEOMIN) exists to control. This brief is NTT DATA’s concrete take on scaling it — without exposing critical OT to the public internet
How video analytics gives the control system the perception it structurally lacks — for predictive maintenance and as the first line of people safety — using the cameras you already have
Why video-analytics pilots die at three walls — scalability, interoperability, and cybersecurity — and, before any of them, on who owns the budget across OT, reliability and cyber
Why scaling demands a private, segmented network — the layer operators like Codelco, BHP and Antofagasta Minerals are already building — and how to place inference at the edge versus the centre without opening the perimeter
Plant managers and control/OT leaders, plus reliability, maintenance and cybersecurity leaders in mining and heavy industry, moving video analytics from a single-camera pilot to a production program
