Binocular Vision-Based Fire System for University Laboratories

Abstract

With the frequent occurrence of fire accidents in college laboratories, the study of fire localization systems has become particularly important. This paper proposes a fire localization system for college laboratories based on binocular vision. The MATLAB toolbox is utilized to complete the calibration of binocular camera and the stereo correction of binocular image. Then combined with the YOLOv5 target detection algorithm for target detection and recognition of the image, to obtain the type of target and the border coordinate information. Finally, according to the principle of binocular vision ranging, the depth information of the target object is calculated to realize the accurate positioning of the fire source. The system proposed in this paper can effectively localize fires in university laboratories. The system has an important application value, which can quickly respond and take corresponding rescue measures in fire accidents and improve the safety factor of university laboratories.