Belt vestibule global smoke real-time detection method based on deep learning

The invention provides a belt vestibule global smog real-time detection method based on deep learning, and the method comprises the following steps: simulating the fire condition of a belt vestibule through employing artificial smog, and making a training set of a YOLOv5 smog detection model; images of the belt vestibule with dust, sunlight or walls are collected as an adversarial sample data set; training the model by using the training set and the adversarial sample data set; grouping the cameras in the whole area of the belt vestibule, and performing polling capture acquisition according to groups; sending the collected image set into a YOLOv5 smoke detection model, detecting whether a smoke label exists or not, and if the smoke label is detected, increasing the detection frequency of the camera so as to further confirm whether a fire occurs or not; and if a fire disaster occurs, smoke alarm is triggered, and alarm information is pushed to maintenance personnel in a WeChat message mode, so that countermeas