Comparative analysis of deep learning algorithms in fire detection

With technology advancing at a dizzying pace, deep learning applications, one of the methods of machine learning, have become applicable in many areas of our lives. Applications such as license plate recognition and optical character recognition have become indispensable in our daily lives. In parallel with the ongoing technological developments today, the development of technologies that will be intertwined with our lives in the near future, such as suspicious situation detection from security cameras or autonomous vehicles, is increasing noticeably. It has been observed that the performance and accuracy rates of this technology have reached high values. In this study, it is aimed to detect forest fires early and accurately before they reach devastating consequences, using mostly images of forest fires from data sets obtained from Kaggle. Different deep learning algorithms in the literature; Trained through transfer learning through the MATLAB program. Thus, it was possible to compare the deep learning algorithm or algorithms that detect forest fires in the shortest time and most accurately in terms of performance rates. At this point, generally high performances above 90% have been achieved.