Service for digitizing instrument readings
Computer vision
Customer
An industrial company with outdated equipment
Solution
Service based on computer vision technology, that "translates" the data of counters, levers, switches into digital form
Opportunities
Reading numbers, symbols on analog, digital displays
The ability to "see" the position of the lever, arrow, switch
Tasks
Find a solution for reading meter readings and sensors to connect them to the IoT 4.0 system
Find the ability to remotely monitor the operation of equipment and track malfunctions in its work without dismantling and disassembly
To create the service, we used classical and neural network methods of machine learning. Classical methods were used to pre-process and reduce the video format. Then the image was processed by neural network
The data is structured in the form of reports, graphs
How it works
01/
We install video cameras in production
The video stream is transmitted to a remote computer 24/7
The neural network segments the image, isolates the data
Step 01
Data collection
Step 02
Image analysis and markup
Step 03
Computer model training
Step 04
Software development
Step 05
Implementation of the computer vision model
Development process
02/
Quote
03/
«Our client needed technical solution that would accelerate the digitalization of production without replacing equipment. At the first stage, we forced his counters and sensors "talk". The second stage is the analysis of the received readings, which will optimize production»
Andrey Ragimov
Project Manager
Development time
04/
2-4 months
depending on the type of equipment and the depth of data processing
Difficulties
Launching neural networks on an edge device with low computing power
05/
Project results
Our client received a ready-made solution, which he was satisfied with. He decided to take further steps to introduce it into production on his own.
06/
Application of the service
07/
Nuclear power station
Drilling rigs
«Our solution can be adapted to almost any production site with outdated equipment. We exclude employees of enterprises from the chain, who now have to check the operation of equipment manually, in uncomfortable and sometimes dangerous conditions.»
Similar projects
08/
Neural network for recognizing counter readings
The bot that counts pipes on the photo
Contact us