Application of Data Fusion Technology to Fire Detection
Jing Han · 2011
A method of data fusion based on reasoning is applied to solve the problem of fire detection.This paper analyzes the basic concepts and theories of Dempster-shafer evidence theory which is applied to multi-sensor data fusion in the field of fire detection.In the way of data processing,a method of constructing the initial probability distribution function is developed by combining the specific application background with the knowledge of expert database in order to produce the reliability distribution of each evidence.In terms of the fusion rules,the improved integration rule is used to fuse a number of evidences.The fusion results are strongly robust and effectively solve the problem of unreasonable fuses caused by invalidation of sensor data and the conflicts between evidences.