Multi-sensor Information Fusion Applied Research In The Detection Of The Fire

The major task of fire detecting technology today is advance warning and reducing mis-warning. However advance warning and reducing mis-warning is just a contradiction, Which brings forward higher demand for fire detecting technology. While historic fire detecting technology (single unit fire detector)can not satisfy the requirement. So the study of fire detecting technology shows more important.The project which the thesis based on is to do some researches on the multi-sensor data fusion technology, then to design and implement fire detecting system. The theoretic in this thesis consists of three parts: basal principle and system structure of data fusion technology; Principle and method of fire detecting and fire detecting system based on data fusion system. In this paper, after analyzing the three layer structure of data fusion system, then a kind of fire detection's three-layer structure based on data fusion is proposed. Also according to fire's non-structure character, proposes to use artificial intelligence method in fire detecting three layer structure, mainly including expert system, neural network and fuzzy control technology. The aim is to design a kind of quickly, exactly and effectively fire detecting system. At the same time, considering slow-varying character of fire signal, feedback cybernetic thought and holding time of fire signal is introduced into fire detecting three-layer structure realizing method in this paper. Thus better reaches the aim of early warning and reducing mis-warning.In order to adapt fire detecting to different environment, at last simulates all kinds of fire and disturbing environment. Simulating model can give fire probability in different environment and proves the feasibility and validity of the fire detecting system proposed in this paper.