Fire detecting for dense bus ducts based on data fusion

Abstract

A method based on data fusion is proposed for detection of fire in dense bus ducts in buildings. The environmental temperature data, infrared image data and gas concentration data are used in the fire warning system. The features of temperature data and gas concentration data are extracted by statistical method, the features of infrared image are extracted by Faster RCNN (Region Convolutional Neural Networks), then three types of features are fused and inputted into the SVM (Support Vector Machine) model to detect fire early. The experimental results show that this method has high recognition rate and good robustness.

Keywords

• Dense bus duct
• Data fusion
• Fire detecting
• SVM