CNN-FD:deep Convolutional Neural Network model for Fire Detection

Fire is one of the earliest discoveries made by mankind and has become a turning point in human evolution. Unfortunately, it is a major global disaster. Therefore, a fire system must detect a fire quickly in the shortest possible time to report it to the concerned authorities. This causes unlimited damage that is visible across social and economic levels. For the image classification task, cutting-edge approaches VGG16 [7], VGG19 [7], Inception v3 [8], and Xception [9] are extensively utilized standards. In this paper, we present an efficient deep convolutional neural network model for image fire detection that implements multipath convolutional layers of varying filter sizes (1$\times$1, 1$\times$3, 3$\times$1, and 3$\times$3) to acquire and learn features locally and globally from images. The suggested model is trained and tested using a custom dataset. The performance of the proposed model is superior to that of cutting-edge algorithms [7] –[9].