Analysis and Real-Time Implementation of Video Analytics Solutions

The field of video analytics has been rapidly advancing over the past two decades. However, despite all efforts, practical surveillance systems in use today are still not capable of autonomously analyzing complex events within the camera's field of view. This is a significant drawback because video recordings from millions of surveillance cameras worldwide are not being analyzed in real-time and, therefore, cannot aid in accident prevention, crime detection, or counterterrorism efforts. Currently, these recordings, at best, are archived to facilitate forensic analysis after an event.
As the importance of video surveillance systems continues to grow, especially in security applications, video analytics will play a crucial role in the future development of these systems, presenting both challenges and opportunities for technological innovation. The primary challenge lies in developing models and algorithms for analyzing high-frequency scenes.
This work consists of two parts. The first part is theoretical, providing an overview of the concept of real-time video analytics. Technologies used in video analytics are analyzed, and the structure of these systems is explained. Applications relying on these techniques, as well as algorithms used in artificial intelligence, are described.
The second part of the work is practical, focusing on the implementation of a real video analytics system that includes recording and processing video signals in real-time using Dahua equipment and artificial intelligence. This system simulates the operation of video analytics in real working conditions and demonstrates how artificial intelligence is applied in such systems. The results of measurements performed are graphically presented and extensively analyzed in the paper. Furthermore, the paper explores possibilities for further practical implementation of this solution.