Efficient Fire Detection for Uncertain Surveillance Environment
To efficient CNN based system for fire detection in videos captured in uncertain surveillance scenarios
To detect fire, researchers have presented both traditional and learned representation based fire detection methods. In literature, the traditional methods use either color or motion characteristics for fire detection. For instance, [9-16] used color features for fire detection by exploring different color models including HSI , YUV , YCbCr , RGB , and YUC . The major issue with these methods is their high rate of false alarms. Several attempts have been made to solve this issue by combing the color information with motion and analyses of fire’s shape and other characteristics
We propose an efficient CNN based system for fire detection in videos captured in uncertain surveillance scenarios. Our approach uses light-weight deep neural networks with no dense fully connected layers, making it computationally inexpensive. Experiments are conducted on benchmark fire datasets and the results reveal the better performance of our approach compared to state-of-the-art. Considering the accuracy, false alarms, size, and running time of our system, we believe that it is a suitable candidate for fire detection in uncertain IoT environment for mobile and embedded vision applications during surveillance.