RSS
Логотип
Баннер в шапке 1
Баннер в шапке 2

MTUSI Intelligent security system of employees at work

Product
The name of the base system (platform): PyTorch
Developers: Moscow Technical University of Communications and Informatics (MTUSI)
Date of the premiere of the system: 2023/02/014
Technology: Video Analytics Systems

Main article: Video analytics (terms, applications, technologies)

2023: System Development

Scientists MTUSI have developed an intelligent system aimed at ensuring the safety of employees in production. According to the authors, the decision was made thanks to the use of the framework. machine learning PyTorch The university announced this on February 14, 2023.

In the industrial industry, more and more attention is paid to the safety of workers: it accounts for a significant part of all expenses of the enterprise. Accidents lead not only to injuries, but also to the loss of the company's image, a decrease in income.

Despite the accelerated development of information technology, a significant part of the tasks of ensuring safety at work is carried out manually. According to the researchers, this approach has obvious drawbacks: important information often escapes from the human field of view, and the costs of such activities are high.

{{quote 'We offer a program devoid of these imperfections: photo and video materials are processed automatically, then safety violations are detected and a report is provided, - explained the Dean of the Faculty of Information Technologies of MTUSI, Ph.D. Mikhail Gorodnichev. }}

The presented solution will allow you to recognize whether the safety technique is violated by workers using machines at work. Photos and videos captured on the camera are transmitted to the program along with markup for the equipment. After processing the material, a status appears confirming or refuting the presence in the danger zone.

File:Aquote1.png
The person responsible for security has constant access to the system's reports: thanks to this, you can view detailed information, see the screenshot and data of the offender, - said Ksenia Polyantseva, senior lecturer at the Department of Mathematical Cybernetics and Information Technologies, co-author of the work.
File:Aquote2.png

The application is developed in Python, the PyTorch framework and the KeyPoint R-CNN neural network are used to recognize people and key points on their bodies (wrists, shoulders, knees, head, etc.).

All research in the field of artificial intelligence is conducted at the Faculty of Information Technologies of MTUSI.