Bashir Bagheri Nakhjavanlo - Department of Computer and Mathematics, Firoozkooh Branch, Islamic Azad University, Firoozkooh, Iran
Monireh Ayari - Department of Computer, Karaj Branch, Islamic Azad University, Karaj, Iran
Nima Aberomand - Department of Computer Engineering, Shahr-e-Qods, Branch, Islamic Azad University, Tehran, Iran -Department of Computer Science, the University of Texas at Arlington, Texas, USA

Nowadays, jungles are burning into fire due to climatechanges. Detection and tracking any smoke will be prevent anyburning, so it needs pattern recognition in images. In thisresearch we propose a developed method for real-time andsynchronous fire and smoke tracking and detection in videos. Inthis approach, at first, we apply a pre-processing phase toenhance the image frames and then deep learning techniquebased on pyramid convolutional neural network apply for datatraining and testing based on fractal model for imagesegmentation and feature extraction. Simulation done inMATLAB platform which results indicated good results in termsof smoke and fire tracking and detection. Also we use someevaluation criteria such as accuracy, sensitivity, specificity, areaunder curve (AUC) with ۹۸.۹۱%, ۹۳.۵۴%, ۹۳.۱۷% and ۰.۸۷۱۹respectively. We use two different image and video dataset in thisresearch and both of them have good performance in terms ofaccuracy in comparison to recent methods.

component; Smoke Detection and Tracking, Fractal Model, Pyramid Convolutional Neural Network, Deep Learning