A Study on a Complex Flame and Smoke Detection Method Using Computer Vision Detection and Convolutional Neural Network

This study sought an effective detection method not only for flame but also the smoke generated in event of a fire. To this end, region was pre-processed using color conversion and corner method, could be detected dark channel prior optical flow. eliminates unnecessary background regions allows selection fire-related regions. Where there interest, inference conducted deep-learning-based convolutional neural network (CNN) to accurately determine whether it or smoke. Through approach, accuracy is improved by 5.5% 6% compared when fire through object model without separate pre-processing.