Research On Fire Monitoring Technology For Unmanned Aerial Vehicles Based On Multi-sensor Information Fusion

Nowadays,the requirement of fire monitoring system is increasing,Acquisition information of the traditional single sensor is one-sidedness and locality,while acquisition information and data fusion processing of the multi-sensor can obtain more comprehensive and accurate environmental information.So,information fusion of the multi-sensor has high practical value in fire monitoring.This paper mainly focuses on some basic algorithms of multi-sensor information fusion technology,and applies them to fire monitoring.A fire monitoring system based on quad-rotor UAVs(Unmanned Aerial Vehicles)is designed.Firstly,the main characteristic parameters of fire are analyzed,and the types of sensors available and detection algorithms are given.Secondly,the algorithms of space-time registration and outlier elimination in multi-sensor information fusion technology are studied,and an outlier detection algorithm based on innovation change is designed.The simulation results show the effectiveness of the algorithm.Thirdly,an improved evidence theory algorithm for conflict is proposed to solve data decision-making in information fusion technology.By analyzing the consistency of evidence and the importance of determining the focal element,the algorithm solves the problem of one vote rejection and excessive evidence conflict in evidence theory,and reduces the uncertainty of decision results.For the decision-making problem of multi-sensor confidence,an improved evidence theory algorithm based on sensor confidence is proposed.The grey correlation is used to obtain the sensor confidence,and the target is judged by combining the focal element information obtained by the sensor and the sensor confidence.The theoretical analysis and simulation experiments show that the algorithm has a good decision effect.Finally,a fire monitoring system based on quad-rotor UAVs is developed,and the above algorithms are applied to fire monitoring.The algorithm of the improved evidence conflict is compared with several classical DS evidence theory algorithms and single sensor algorithm.The experimental results show that the proposed algorithm has great advantages in accuracy and robustness has higher practicability.It is suitable for complex and changeable disaster environment.