Real-Time Monitoring for Crowd Counting Using Video Surveillance and GIS

Real-Time Monitoring for Crowd Counting Using Video Surveillance and GIS

1Mr. R. Krishna Nayak

Associate Professor, Department of Computer Science and Engineering Vignan’s Institute of management and Technology for Women, Hyd.

Email: ramavath.krishna12@gmail.com

2Ojasvi Pattanaik

UG Student, Department of Computer Science and Engineering Vignan’s Institute of Management and Technology for Women, Hyd.

Email: ojasvipattanaik22@gmail.com

3Sai Medha Sri

UG Student, Department of Computer Science and Engineering Vignan’s Institute of Management and Technology for Women, Hyd.

Email: medhagaddam23@gmail.com

4Rupavath Akhila

UG Student, Department of Computer Science and Engineering Vignan’s Institute of Management and Technology for Women Email: rupavathakhila80@gmail.com

Abstract-- In the context of smart city development, real- time people counting plays a pivotal role in enhancing public safety, optimizing urban planning, and managing communal resources. Conventional systems frequently grapple with latency and inaccuracy, particularly when dealing with occlusions or non-human objects in crowded scenes.

To overcome these problems, we present the YOLO-PC, a novel approach that leverages the YOLO(You Only Look Once) object detection concept for real-time and accurate people counting.

This solution combines the detection capabilities of YOLO with enhanced video processing and filtering techniques to track individuals reliably even in dynamic and complex environments.YOLO-PC employs robust filtering to minimize false detections caused by overlapping entities or irrelevant items. Furthermore, the system is augmented by Geographic Information System (GIS) tools, which support spatial and temporal data visualization across multiple surveillance feeds.

This solution combines the detection capabilities of YOLO with enhanced video processing and filtering techniques to track individuals reliably even in dynamic and complex environments.YOLO-PC employs robust filtering to minimize false detections caused by overlapping entities or irrelevant items. Furthermore, the system is augmented by Geographic Information System (GIS) tools, which support spatial and temporal data visualization across multiple surveillance feeds.

Empirical validation through urban surveillance datasets confirms that YOLO-PC performs with high accuracy in real time, effectively managing dense crowd scenarios. This integration enables timely data-driven interventions by urban authorities, contributing significantly to emergency

response strategies and city operations. The research demonstrates that YOLO-PC is a scalable and dependable framework that supports the evolving needs of smart urban environments, while also laying a foundation for further exploration into model efficiency and collaborative multi- camera systems.

Keywords- Real-time surveillance, People counting, YOLOv8, Crowd detection, Multi-camera systems, Computer vision, Deep learning.