Computer Vision-Based Early Fire Detection Using Machine Learning

Thousands of hectares of land are destroyed every year by fire. There is more carbon monoxide generated by these fires than from all the traffic. The early detection of possible danger areas and early detection of fires can greatly reduce response times and firefighting costs as well as the possibility of damage. By using image processing technology, a fire could be detected early and people would be alerted. The sensor method detects fires through smoke, but it has limited applications, and is only suitable for certain areas. In this proposed work, limitations will be reduced and the technology will be optimized. In this proposed system, a Haar Cascade classifier is used for fire detection. Pycharm IDE is used for implementing this work. The system uses the webcam as a source of input for capturing the video feed from the surrounding environment. Detection of the fire will be exact and accurate without any delay with the proposed work.