Detection of Smoking in inside the office Using Deep Learning for CCTV Camera

It is crucial for the protection of humans in an indoor setting to have an early fire detection system. Over the course of the most recent few years, a multitude of methods that make use of image analysis and computer vision techniques have been suggested. However, the implementation of video smoke detection in an interior setting is still a difficult problem. This is due to the constraints of data for training and the absence of effective techniques. This article will propose a novel approach for detecting smoke that makes use of security cameras as its primary tool. The suggested approach may be broken down into two distinct phases. In the initial step of the process, optical flow is used to find areas of motion that occur between two consecutive frames. In the second step of the process, a deep convolutional neural network is used for the purpose of identifying smoke in areas that are in motion. In addition, simulated smoke photos are used in order to supplement the dataset in order to address the issue of insufficient data. Our data collections as well as real-world video sequences are used to evaluate the suggested technique. The results of the experiments suggest that the novel approach can be effectively used to a variety of indoor smoke recordings and has major implications for improving the precision of fire environmental monitoring. Both the information and the programming language have been uploaded to an internet repository.