Real time visual traffic map for vehicle density estimation using IP-CCTV networks

• Main Authors: Bautista, John Carl B., Fernan, Adrian Giuseppe Francis M., Gacuya, Zendrel G., Perez, Eldrine Jay
• Format: text
• Language: English
• Published: Animo Repository 2021
• Subjects: Traffic monitoring; Traffic monitoring—Equipment and supplies; Traffic cameras; Electrical and Electronics; Systems and Communications
• Online Access: https://animorepository.dlsu.edu.ph/etdb_ece/1; https://animorepository.dlsu.edu.ph/cgi/viewcontent.cgi?article=1002&context=etdb_ece
• Institution: De La Salle University

Closed Circuit Television (CCTV) systems are being used to monitor traffic behavior. Multiple cameras are being used to capture footage and the video information is analyzed to extract useful information. In creating an effective traffic management, knowing the road traffic density in real time is essential. Vehicle detection and traffic density estimation can be achieved using video monitoring systems. The purpose of designing an IP-CCTV system is to be able to simplify the process of monitoring and to provide a robust and reliable traffic system. The IP-CCTV system consists of eight cameras with four Raspberry Pis. Two cameras are processed by one Raspberry Pi. The system is tested during daytime to achieve higher vehicle detection accuracy. A Graphical User Interface (GUI) displays the video feed of cameras, hourly traffic report, and the map notification system. All Raspberry Pi can send and receive data, they can also create the visual traffic map and store it in their directories while Raspberry Pi 1 will upload the image to the GUI. By default, the map will not display any color if there is light traffic, or no vehicles are present. For moderate traffic, the map will display yellow and red for heavy traffic. Due to the recent Covid-19 pandemic, we created a miniature model of the system instead of an actual setup inside the campus. The system accurately detects 93% of the vehicles during daytime. On average, 31% of the vehicles were detected under poor lighting conditions. The accuracy of the notification system yielded 84%.