A Review of Methods and Sensors for Detecting Forest Fire

Forest fires are a real hazard to human lives, biotic components and abiotic components. Predictions says, by 2030 almost half of the world's jungle will be destroyed by forest fire. The early fire detection is only an optimized way to decrease damages caused by forest fires. That’s why the forest fire prediction is gaining attention. Forecasting the origin and progression of forest fires, could yield substantial advantages for public health and safety, the financial sector, and the natural world. This might aid in identifying high-risk areas; for instance, with limited resources, professionals could opt to focus on studying certain places. A number of factors have led to the use of image processing in this study, including the quick advancement of digital camera technology, the camera's excellent ability to cover vast areas, the quicker reaction times of image processing techniques compared to current sensor systems, and the lower overall cost of image processing systems compared to sensor systems. Using satellite images, forest fires can be predicted as soon as possible, informing the fire department to operate rescue missions and stop fires. Old methodology uses Monitoring and Results Management (MRM), with University of California Irvine (UCI) machine learning repository, and Heat map to make more predictive power while recently used methodology, Random Forest regression uses Optimal Mass Transport (OMT) and NSD.