Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{198375,
        author = {Prof. Dnyaneshwar Gurjar and Ms.Saraswati Shinde and Ms.Pranjali  Kapse and Ms.Gauri Kadam and Ms.Ritu Jadhav},
        title = {AI-Driven Drone-Based Crowd Management Using Image Processing and Machine Learning},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {12430-12435},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=198375},
        abstract = {Effective crowd management has become a critical challenge in today’s rapidly urbanizing world, particularly during large-scale public events, festivals, protests, and emergencies. Traditional surveillance systems relying on static closed-circuit television (CCTV) cameras and manual monitoring are limited by fixed viewpoints, delayed response times, and human error. This paper presents an AI-driven drone-based crowd management system that integrates real-time aerial imagery with advanced machine learning algorithms for intelligent crowd monitoring, density estimation, and anomaly detection. The proposed system employs Convolutional Neural Networks (CNNs), specifically the ResNet50 architecture, along with YOLOv8-based real-time object detection, OpenCV-based image preprocessing, and an automated alert generation module. The system was trained and evaluated on multiple benchmark datasets including the UCF Crowd Dataset, ShanghaiTech, JHU-CROWD++, and a custom drone-captured dataset comprising over 5,000 annotated frames. Experimental results demonstrate that the ResNet50 model achieved the highest accuracy of 98.3%, with a precision of 98.1%, recall of 98.6%, and F1-score of 98.3%, significantly outperforming baseline CNN models. The system provides real-time crowd density heat maps, behavioral classification, and authority alerts through an interactive web dashboard. Practical deployment is validated through a Flask-based Crowd Monitor application demonstrating live camera feed monitoring with capacity-based alert generation. This study contributes to the domains of computer vision, intelligent surveillance, and smart city management by offering a scalable, autonomous, and data-driven solution for public safety.},
        keywords = {Crowd Management, Deep Learning, Drone Surveillance, Density Estimation, Machine Learning, Real-Time Monitoring, ResNet50, Smart City, YOLOv8.},
        month = {April},
        }