System to detect fire under surveillanced area

Fire usually causes serious hazards. Therefore, to prevent catastrophes that occur in industries, buildings, and forest areas, image based fire detection has become an important issue. Especially, if the combustion at initial stage could be detected immediately, the vandalism would be reduced to a greater extent. In this proposed system, first image pre-processing is done and segmented, which includes edge detection and thresholding methods. Histogram of Oriented Gradient (HOG) algorithm and Gray Level Co-occurrence Matrix (GLCM) algorithm are used for extracting the features. Support Vector Machines (SVMs), an algorithm that is used for classification. After image processing section, detected fire level is given as input to the micro-controller unit and LCD displays the fire level like normal, mild or severe, and also a message will be sent to the concerned person. And if the fire level is mild or severe, then it is alerted by using a buzzer. Thus, the system reduces life loss and also provides detection of fire at the initial stage in several areas.