FOREST FIRE DETECTION USING VIDEO SUMMARIZATION BASED ON K-MEANS CLUSTERING

  • B. PUSHPA et al.

Abstract

Fire detection is a potential threat in consideration, to prevent the loss of lives and property damage. Conventionally, many techniques have been identified so far, to discover the forest fires in the input video.  A video summarization is a process of creating and offering a meaningful abstract view of the whole video within a short period of time. For keyframe based video summarization, selection of keyframes plays an important role for the effective, meaningful and efficient summarizing process.  This paper presents a method, Forest fire detection is identifying keyframes with minimal time complexity based on K-Means clustering with calculates fitness function of all video frames by considering features such as color probability, spatio-temporal energy, intensity, and texture, support vector machine (SVM) is employed for classifying an video candidate frame into normal frame and forest fire frame. The simulation analysis is performed on the FIRESENSE dataset and the results are assessed under several dimensions. The final outcome proves the efficiency of the presented proposed model in a considerable way.

Published
2019-11-28