Multi-sensor Data Fusion Methodbased On UAV Target Recognition

With the rapid development of science and technology in China,UAV target recognition technology has gradually been applied to crop growth monitoring,tactical target detection,rescue search and other practical applications.Multi-sensor data fusion is one of the key technologies of UAV target recognition,especially in target recognition application with high-efficiency and adaptability in complex environment.In order to improve the accuracy of UAV target recognition,this paper takes multi-sensor data fusion as the core and studies the flight attitude and fusion algorithm of UAV carrying multi-sensor in the process of data acquisition and data processing,mainly completing the following three tasks:(1)Through the analysis and research of data fusion method,multi-sensor technology,Kalman filter,quaternion method and D-S evidence theory,the basic method of multi-sensor data fusion is formed by combining decision-level data fusion with D-S evidence theory method of reducing evidence conflict,which combined with a method of UAV optimization,the basic method of multi-sensor data fusion based on UAV target recognition is constructed.(2)For the sensor data is not accurate and the conflict between data problems due to the complexity of UAV flight environment,using kalman filtering processing flight attitude movement,and using quaternions to eliminate trigonometric function calculations in the analytical solution of motion equation,this paper proposes a posture optimization algorithm based on extended kalman filter.For the problem of high conflict evidence in data fusion,this paper studied the fusion method of high conflict evidence.Based on Dempster’s combination rules,allocating reasonable weight of evidence by using Jousselme distance to express the degree of conflict,and a data fusion optimization algorithm based on improved D-S evidence theory was proposed.(3)For the optimization algorithm of flight attitude and data fusion conduct simulation experiments.By comparing the fitting degree of the attitude Angle curve between the original model and optimization method,the effectiveness of the attitude optimization algorithm is verified.By comparing the target recognition results of the data fusion optimization algorithm based on the improved D-S evidence theory with the five classic data fusion methods such as Yager,Dempster and Murphy,the effectiveness of the data fusion optimization algorithm based on the improved D-S evidence theory is verified...