Russian scientists have created a unique system that uses artificial intelligence (AI) to help ensure safety in hazardous industries. This system is capable of detecting violations of safety regulations, such as the absence of protective helmets on employees.
Machine vision algorithms and neural networks analyze streaming video from surveillance cameras and determine in real time when workers are without the necessary protective equipment. In the event of a violation, the system issues an alarm and blocks access to the equipment until the helmet is put on. Notifications are also sent to responsible structures for prompt response.
The system was developed by specialists from the Peoples' Friendship University of Russia (RUDN University) and the University of Artificial Intelligence. The key challenge was processing a large amount of video data. To do this, the team used pre-processing methods and labeled more than 9,000 images from surveillance cameras. The result was achieving almost 100% accuracy in detecting violations.
In addition, the system has flexibility and scalability, which allows it to be adapted to various production conditions. For example, it can be configured to monitor the presence of signal vests at airfields or railways, where special clothing is important for protecting workers. It is important to note that notifications can be sent not only to responsible persons, but also to the employees themselves who violated the rules, which allows you to quickly correct the situation.
However, according to experts, such systems can be used as a supplement to traditional safety methods, and not as a replacement for human control. Artificial intelligence is effective in processing large amounts of data and automating routine tasks, but false positives can occur, so the decision on safety should remain with the person.
The introduction of such technologies also requires well-thought-out protocols in case of unforeseen situations. Experts are confident that AI can significantly improve safety in enterprises, especially in areas such as transport, medicine and energy, where operational monitoring and rapid response are critical to preventing dangerous situations.
Read materials on the topic:
For monitoring production facilities: Cognitive Pilot presented a new unmanned robot
Russian engineers will design products on a domestic digital platform
