Fire is a terrifying weapon, with nearly unlimited destructive power. Fire accidents are a major cause of human suffering and material loss and the one that perhaps are predicted the least accurately. Most existing work in fire occurrence prediction focuses on prediction of wildfires in forests and those caused by volcanic eruptions. Surprisingly prediction of fire occurrence in residential and official buildings has not been fully explored because the factors that influence fires are too many. The idea behind this research is to provide an alert to fire stations in the event of fire in hospitals, official and commercial buildings by the use of Image processing and Artificial Intelligence techniques that might significantly reduce the death toll and loss of property caused by fire accidents. The Digital snapshots of the building can be taken (1,600 x 1,200 pixels at 1MB image per second) continuously using Closed circuit digital photography (CCDP) and these snapshots are then automatically sent to the server for storage as timed and dated JPEG files. The digital images are converted from RGB to XYZ color space and then segmented by utilizing anisotropic diffusion to identify the presence of fires. Subsequently, Radial Basis Function Neural Network is trained with the color space values of the segmented fire regions and is employed in the design of this novel system. The proposed intelligent system will thus aid in alerting the fire stations with the help of a Global System for Mobile Communications in event of any fire to take immediate actions before fire spreads quickly and causes traumatizing loss.
分享
引用
