Development of a drone-based intelligent analysis system for environmental pollution and crowd density monitoring on university campuses

University campuses pose unique challenges in terms of environmental pollution and crowd management due to increasing human activity, expansive physical areas, and diverse sources of waste generation. Traditional monitoring systems often fall short in addressing these issues, as they lack the ability to deliver location-based, detailed, and real-time information. Situations such as waste accumulation and high crowd density present serious environmental and safety risks, demanding more sensitive, comprehensive, and dynamic solutions. This study presents an integrated drone-based monitoring system capable of real-time, location-aware tracking of environmental pollution and human density. The system consists of a drone that captures high-resolution imagery, a YOLOv8x model for waste detection, a YOLOv11n model for human detection, geolocation algorithms that utilize image metadata, and density maps generated using Kernel Density Estimation. Leveraging various open-source datasets, the models accurately detected waste and human objects from field-captured images. Experimental evaluations demonstrated detection accuracies of 85.87% for waste and 73.36% for humans. The detections were interactively visualized on the campus map, providing decision-makers with real-time, data-driven insights for sanitation and safety operations. The proposed system serves not only as a standalone object detection platform but also as a multi-layered decision support infrastructure that includes spatial and temporal analysis. Results indicate that the integration of UAV technology with AI-powered object detection offers a highly effective tool for environmental monitoring and operational planning in campus settings.