IoT Based Forest Fire Detection Using Deep Learning Techniques- A Survey

Real hazards to infrastructure, environmental systems, and human life come from forest fires. According to a survey, forest fire will damage the forest to a half by 2030. Forest fire can be prevented by taking actions by having fire detection devices. Most of the institutions and research development centers are attracted towards the fire detection systems. There are many sensors for detecting the fire but it also has many demerits like expensive, requires large space for installation and other problems. Due to the emergence of a digital camera, detection of fire has become easier. The use of sensors has many limitations but it is resolved by using image processing techniques. Because certain things share characteristics with fire, making accurate forest fire detection algorithms difficult, false alert rates may be significant. In this study, we may examine several ways for detecting forest fires and provide a brand-new, four-stage video-based method for detecting forest fires using image processing. To find moving areas, a background-subtraction technique is used first. Secondly, RGB color space is used to identify possible fire zones. Third, because candidate zone might have moving fire-like things, extraction of characteristics is employed to differentiate between the actual fire as well as fire-like items. The category of the fire can be determined by using the convolutional neural network method. Further incorporate devices for fire detection alerts into this system.