Intelligent Security Surveillance System for Communities in Areas Adjacent to Abandoned or Vacant Lots with Image Processing Technology

This work presents the development and testing of an intelligent security surveillance system capable of automatically detecting people and flames, with an alert system sending notifications via Telegram Bot. The objective is to increase the efficiency of security in the communities by testing under various environmental conditions at detection ranges from 5 to 30 meters during daylight, low light, and night times, including testing gaits such as slow walking and running. Next, the system was tested in the Rung Sawang Village 1 community, Bang Khen District, Bangkok, which has an area adjacent to a large abandoned or vacant lot. The test results showed that the system was able to detect people in daylight with a maximum similarity percentage of 98.45% at a distance of 5 meters and could detect at a distance of as far as 30 meters in cases involving slow walking. Running was detectable at a distance of 20 meters. During low light hours, the system's efficiency was reduced to 72.31% similarity at a distance of 5 meters with the closest slow walking detectability at 20 meters, while running was detectable within 15 meters. As for flame detection, the system was found to only work at night due to sunlight interference during the day. Candle flames were detectable at distances of 10 meters, and piles of paper were detectable at distances of 25 meters with a similarity percentage of 78.95% at a distance of 5 meters. Next, the notification system via Telegram Bot was tested, yielding average notification speeds ranging from 6.53 to 7.19 seconds. When the system was tested in actual use by measuring user satisfaction among 30 people, it was found that the image resolution from the camera was 4.27 points, followed by the motion detection system at 4.27 points and the suitability of the installation site at 4.13 out of 5 points.