The vision-based system design for the smart city environment, which is part of the ZalaZone Smart City project and presented at the Mobility and Environment - Future Shaping Automotive Research conference, addresses the potential of high-level, real-time automatic data processing and intelligent intervention in the smart city environment. The system design presented in this paper focuses on the application of artificial neural networks and machine learning techniques, with particular attention to object detection, tracking, identification, and intelligent intervention. The paper describes in detail the logical architecture of complex sensor networks and data processing, and the potential applications, with a special focus on spatial detection and the structure of hierarchical network models. The design of high-level logical units of the system will be presented, allowing optimised, automated intervention decisions in the smart city infrastructure. The paper illustrates the functioning of smart interventions through practical examples. The aim is to show how the integration of advanced artificial intelligence and sensor networks can contribute to a more efficient and intelligent urban environment.