Improving smart water meter installations and maintenance with computer vision

OPTIMISE UTILITY

operations across the energy ecosystem from retail supply to production with computer vision.

Automated meter reading

Inveniam AI can automatically read meter displays during installations or maintenance checks. Instead of relying on manual reading, field workers can take pictures of the meter displays, and in just seconds – our solution returns with data after analysing the visual data. This eliminates the need for human intervention and reduces the possibility of errors associated with manual readings.

Quality assurance during the installation

With many upgrade programmes underway, we analyse images of installed digital meters to ensure they meet quality standards and are properly installed. By detecting any irregularities or deviations from installation guidelines, such as correct positioning of valves, our systems alert technicians in real-time, enabling them to address issues promptly and prevent potential problems.

Installation assistance

Our computer vision solutions can assist technicians during the installation process by providing real-time guidance and overlaying digital instructions onto physical surroundings. This can help ensure that meters are installed correctly, reducing the likelihood of errors and minimizing the need for rework.

Optimised resource use & predictive maintenance

Computer vision technology can detect malfunctions or anomalies in water meters. By continuously monitoring the meter's functioning, it can identify leaks or irregular water usage patterns that might indicate a problem. This allows for timely repairs and maintenance, reducing the potential for water wastage and saving costs in the long run. Additionally, with automated image recognition and processing capabilities, our solution documents the condition and performance of each meter over time.

Straightforward workflow

Inveniam AI integrates into existing processes and workflows

PROCESS

Our onboarding process in a few simple steps

Powerful, self-serve product and growth analytics to help you convert, engage, and retain more users.

Kick-off

DAY 1

Share visual data for review and align on use cases, detections, data structure, frequency of analysis and success measures.

Platform Access

WEEK 1

Start uploading data to our platform where you will be able to track progress across the data pipeline, review dashboards and key results.

Start Analysis

WEEK 2

We use human annotators to start analysing images if there is no pre-trained model. In parallel, we kick-off model exploration and training for each detection case.

Fully Automate

WEEK 4+

Transition to a fully automated model, with human-in-the-loop intervention used for edge cases and where 100% accuracy levels are required.