Unexpected fire ,at times, is dangerous and cause loss of people and property. Live inspection by video surveillance cameras and fire detection by image processing using Open CV and Raspberry Pi is a novel and efficient method. A vision based system is adopted to increase the detection rate and lower the fault alert rate. Levels of flame-high or low, are identified by edge detection techniques using different experts by combining factors like color, shape and motion of the fire. GSM technology is incorporated which helps to share the location information to the nearest fire station .The microcontroller functions to activate the fire alarm as well as the exact sprinkler rather than activating all of them to suppress the fire .