Operational downtime increases operating expenses and reduces production output. This issue is especially significant in the oil and gas industry, where most financial losses stem from unplanned or scheduled shutdowns. Remote monitoring solutions can be adopted to mitigate operational downtime, but marginal wellheads often lack the financial resources to implement such systems due to their limited economic viability.

To address this problem, we designed a non-intrusive system that digitalises existing analogue gauges. This system captures images of analogue gauges installed in wellheads, processes them using a computer vision model and transmits the numerical data to a remote dashboard, effectively providing existing wellheads with smart data-producing capabilities. Telegram-based alert system notifies operators when gauge readings exceed preset thresholds.

Our solution uses computer vision as it can digitalise readings without modifying existing gauge infrastructure. The camera used in our solution to enable the computer vision is housed in a protective casing that shields it from environmental effects and ensures consistent lighting for accurate readings. Initial tests conducted in a laboratory setting indicate that our solution can extract gauge readings within a 1.6% error at an accuracy rate of 94%.