Pengembangan Sistem Deteksi Kerumunan Berbasis Edge Computing untuk Mendukung Monitoring Cerdas pada Lingkungan Smart City

Abstract

The development of smart cities requires real-time crowd monitoring systems; however, cloud-based approaches still face challenges such as high latency and bandwidth demands. This study aims to develop an edge computing-based crowd detection system to improve system efficiency and responsiveness. A quantitative experimental approach was employed to evaluate system performance based on latency, frame rate (FPS), and accuracy metrics, including precision, recall, and mean Average Precision (mAP). The results show that the system achieves a latency of 40–60 ms per frame and operates at 24 FPS, meeting real-time processing requirements. The system achieves an accuracy of 91.6% on real-time data and 95.8% on video data. However, limitations remain under complex environmental conditions such as occlusion and lighting variations. Overall, the proposed system is effective for supporting intelligent monitoring in smart city environments