A Model Integrating Fire Prediction and Detection for Rural-Urban Interface

Abstract

This paper proposes a model that integrates new smoke detection and fire prediction techniques for the rural-urban interface area. The model aims to predict fire risk from weather parameters, and to detect smoke using video monitoring systems. Further, the fire danger index (FDI) provided by the prediction algorithm would be utilized to enhance the certainty of smoke detection and reduce false alarm rates. Experimental results illustrate that our prediction algorithm successfully predicts fire risk on a five-point scale with mean accuracy of 94.92% and the detection algorithm more effectively detects smoke compared to other algorithms by achieving 97% average accuracy.