Visual Based Fire Detection System Using Speeded Up Robust Feature and Support Vector Machine

This paper proposed a fire detection system using video captured from camera. We built the system using Speeded Up Robust Feature (SURF) feature extraction on three orthogonal plane to obtain the spatial and temporal features. We used Support Vector Machine (SVM) algorithm to classify the features as the fire or non-fire object. Using SURF threshold value 0, number of cluster 5 and gaussian SVM kernel, the system generated accuracy 81,25%.