Help Desk -
9958816305, 9810335381
Email
Password

Metalloinvest Implements Neural Network to Recognise Billet Markings

Steel News - Published on Mon, 28 Sep 2020

Image Source: Metalloinvest Billet Markings
Metalloinvest’s OEMK is completing the pilot operation of an automated system for recognising markings on cast billets at Rolling Mill #1. The idea for the project was developed by OEMK employees during an innovation competition at the enterprise organised by JSA Group, part of ICS Holding, a diversified IT group. The in-house intelligent solution based on neural network technology was developed by researchers from Stary Oskol Technological Institute, a branch of the National University of Science and Technology MISiS, together with specialists from OEMK. Software engineers from Metallo-Tech LLC worked on integrating the system software into production processes.

The implemented solution uses five neural networks, each performing its own discrete task. In the process of collecting data for their training, more than 60,000 pictures of billet markings were taken.

The experience gained in implementing this project will be applied to solving similar tasks. For example, video analytics technology can be implemented to determine the granularity of incoming raw material on the conveyor, to check the quality of excavator bucket teeth and even to monitor the use of PPE by employees.

Source :

Posted By : Yogender Pancholi on Mon, 28 Sep 2020
Related News from Steel segment