CCTV Map Creation Method for Improving Fire Location Estimation

Abstract

Fire detection using existing CCTV cameras is a method of detecting images by applying an analysis program to the installed cameras. However, it is difficult to determine the specific location of a fire in a space, and if the same fire is detected by different cameras, it may be recognized as a different fire. To solve this problem, we developed a method for automatically detecting the location and shape of a fire according to the overlapping area of images from multiple cameras. A basic map of the space (monitoring area) was constructed for the fire, smoke, and people, who are considered mobile objects to be detected. Spatial synchronization was achieved by recognizing the location of a moving object and assigning an index to it. Additionally, we developed an algorithm that calculates the location of the fire, and the location of the fire could be specified by crossing the spatial map for each camera to extract a common index.