A Machine Learning approach for Smart Community Tracking System Using Yolov3 Algorithm

A Machine Learning approach for Smart Community Tracking System Using Yolov3 Algorithm

Publication Date : 2024-11-27
Author(s) :

Meena V, Mega Nathan S, Mithulashwaran C, Pavamana K
Conference Name :

International Conference on Recent Trends in Computing & Communication Technologies (ICRCCT’2K24)
Abstract :

The machine learning method for Effective management of public spaces, especially in scenarios involving large crowds, is crucial for ensuring public safety and efficient use of resources. This project focuses on the development of a real time person counting and tracking system, specifically designed to assist with crowd control and safety in public areas. The system leverages advanced deep learning techniques, combining the Convolutional Autoencoder for crowd analysis with YOLOv3 (You Only Look Once version 3), a powerful object detection algorithm, to accurately identify, track, and monitor individuals in real time. The YOLOv3 algorithm serves as the core component for object detection, enabling the system to recognize and follow individuals across consecutive frames of video footage. Its ability to detect multiple objects in a single pass ensures efficient tracking of people in dynamic environments, even in cases of partial occlusion or varying lighting conditions. YOLOv3’s high speed processing and accuracy make it ideal for real time applications, where quick and reliable detection of people is essential. In parallel, the Convolutional Autoencoder is used to enhance the tracking capabilities by analysing crowd density and patterns within a given area. By learning spatial features from video frames, the autoencoder supports the system in estimating the number of people in crowded environments, where individual detection may be challenging due to overlaps or dense packing of people. This combination of techniques ensures both precise detection of individual people and the ability to monitor overall crowd movement and density. The system further enhances its utility by recording entry and exit times, thus offering a comprehensive log of people flow in and out of the monitored area. This feature is especially valuable for authorities or organizations responsible for managing large public spaces, such as airports, train stations, shopping malls, and event venues, where the real time data on crowd size and movement can inform better decision making. The recorded data is stored for further analysis, providing insights that can be used to optimize the use of space, improve crowd management strategies, and enhance overall safety protocols. By integrating these cutting edge deep learning techniques, this project provides a reliable, scalable solution for crowd monitoring, offering real time insights into people flow, occupancy levels, and movement patterns. The system’s ability to operate autonomously, without the need for human intervention, makes it a valuable tool for improving the safety and efficiency of public space management.

No. of Downloads :

3