In line with IPI-4 guidelines, PKP Polskie Linie Kolejowe S.A. requires video surveillance systems (VSS) at passenger service facilities to meet specific image analysis standards. Below are the key requirements for intelligent video content analysis (VCA) algorithms:
- Support for at least two video streams,
- Compliance with the ONVIF standard,
- Advanced image analysis,
- Integration with the central security management platform (PSIM).
The solutions proposed by Everest5 in the field of video analysis meet all PKP PLK requirements and offer additional monitoring capabilities based on artificial intelligence. Thanks to advanced neural networks, the system ensures extremely precise detection of objects and events, making it highly effective in public transport environments such as stations and railway platforms. Advanced algorithms are redefining railway safety standards!
The Neurotracker detector recognises the presence of people in the video and generates an alarm when someone crosses the safety line, provided that no train is in the field of view. The system operates dynamically, automatically deactivating the rule when a train is detected by a dedicated algorithm. This solution guarantees a high level of safety and significantly reduces the risk of potential accidents.
Advantages of neural networks over traditional machine learning:
Technologies based on neural networks offer significantly greater effectiveness than traditional machine learning algorithms. The implementation of this technology has made it possible to reduce the number of false alarms by up to 90% compared to previous solutions. Earlier systems often misinterpreted dynamic conditions, such as shadows or reflections, as potential threats. Neural networks are better able to distinguish real events from interference, which translates into more precise results.
The solutions implemented at Warszawa Zachodnia station utilise indispensable NVIDIA graphics cards, which enable efficient handling of complex AI algorithms. The high computational power of GPUs allows for faster image analysis and an increased number of cameras to be handled simultaneously without compromising detection quality. This is crucial for high-traffic locations, such as railway stations.
Potential for further algorithm improvement
There is potential for further refinement of the algorithms for the VMS platform, allowing them to be tailored to the specific conditions of a given location. As a result, the effectiveness of the event recognition can be further improved, and the number of false alarms reduced to an absolute minimum. This means that the system can be continuously improved and better adapted to the requirements of specific environments, such as railway platforms.
Ideal solution for stations and platforms
The solution implemented at Warszawa Zachodnia Station by Everest5 is a system consisting of over 1,600 video cameras. It is a modern video management system (VMS) that combines advanced image analysis capabilities with intuitive monitoring infrastructure management. With its modular architecture and wide range of functions, this platform performs excellently in dynamic environments such as railway stations and platforms. Advanced artificial intelligence algorithms, integration with a variety of devices, and the ability to adapt to specific requirements make the implemented VMS a reliable tool supporting safety and operational efficiency.
Advanced video analysis using AI – Everest5
