Crowd Counting and Analysis using YOLOv7 and Closed Circuit Televisions (CCTVs)
Michael Angelo D. Garcia, Daryell Joshua S. Samaniego, Charmaine C. Paglinawan · 2024
This paper introduces a Closed-Circuit Television (CCTV) system integrated with the YOLOv7 algorithm for crowd-counting applications. The system leverages the advanced object detection capabilities of YOLOv7 for accurate crowd counting with only 10% error, showing great potential for practical applications such as crowd management, public safety, and event planning. However, the system is not without limitations as it faces challenges in low-light conditions and produces false positive results. Despite that, it still shows capabilities of accurately counting up to 250 people. The research demonstrates the system’s reliability and effectiveness, offering a valuable tool for various applications such as event management and public safety.