Large Space Fire Image Processing of Improving Canny Edge Detector Based on Adaptive Smoothing

Abstract

To avoid the large-scale damage caused by fire occurring in large space building, there are many studies about image-processing technique for automatic real-time flame detection. Adaptive Canny edge algorithm and flame geometric features are combined to inspect flame area of video data generated by an ordinary camera monitoring. Firstly, candidate fire regions are detected using modified Canny operator method which combines adaptive smoothing to detect moving fire regions. Fire regions generally have a higher luminance contrast than neighboring regions, a luminance map is made and used to remove non-fire pixels. The proposed selection criterion is effective in improving the performances of traditional Canny operator. Experimental results showed that the proposed approach was more robust to noise, and help to separate consecutive frames flame area.