| On the air battlefield,the rapid and accurate recognition of targets has a very large impact on the deployment and decision-making of the entire battlefield,and is also the key to victory in war.Therefore,effective target recognition technology is essential for air defense operations.The modern battlefield environment is complex and changeable.Target recognition by one sensor alone is far from meeting the needs of modern battlefields.The fusion of information extracted by multiple sensors can establish a more complete target recognition model,which is conducive to improving target recognition.rate.However,due to the influence of its quality and external factors,the information obtained by the sensor may be incomplete or uncertain.For uncertain problems,D-S(Dempster-Shafer)evidence theory can be effectively solved,so it is widely used in the field of target recognition.However,when encountering high conflict information,directly using D-S evidence theory may affect the accuracy of target recognition or even draw the conclusion that the facts do not match,which in turn leads to the inability to identify the correct target.Therefore,based on D-S evidence theory and practical problems,this paper focus on the air target recognition,the main research contents are as follows:1.A multi-sensor air target recognition method based on improved evidence support is proposed.In the air target recognition,the relationship between each sensor and the evidence is calculated and then converted into a support coefficient,so that the evidence with a high support coefficient plays a larger role in the synthesis,and accordingly reduces the conflict evidence for the entire target Identify the impact of the system.Especially when dealing with high-conflict evidence,the method in this paper can reasonably allocate conflicts to obtain results that are in line with reality.Compared with other methods,its recognition results are better.In short,the method proposed in this section has a strong advantage.On the one hand,it can effectively solve the problem of conflicts between evidences.On the other hand,it improves the recognition fusion accuracy,so that the recognition ability of the target recognition system can be effectively improved.2.A multi-sensor air target recognition method based on Pignistic probability distance function and Deng entropy is proposed.Conflict information will have a huge impact on the fusion effect of D-S theory.To eliminate the source of conflict and further improve the ability of the target recognition system,this chapter proposes a fusion method of multi-sensor to eliminate conflict information.First,by introducing the Pignistic probability distance function and Deng entropy,the degree of conflict between evidences and the uncertainty of evidence can be obtained to obtain the evidence correction coefficient.Then,the revised evidence is fused and identified.Finally,two experiments show the effectiveness and accuracy of the algorithm when dealing with low-conflict evidence,and verify its effectiveness and superiority in dealing with high-conflict evidence.3.A multi-sensor air target recognition method based on corrected conflict evidence is proposed,which fully considers the correlation between the evidence provided by different sensors and their respective credibility.The credibility of the evidence is converted into a support factor to modify the evidence.If there is evidence of high conflict between the evidences in the group,the evidence with the least support is replaced by the weighted average;on the contrary,there is no evidence of high conflict between the evidences in the group,which is unchanged.The weighted average of all target focal elements is obtained to obtain synthetic evidence,and then fusion recognition is performed.In the process of target recognition,evidence with a high support factor plays a greater role,and the accuracy of target recognition is therefore greatly improved.The proposed method overcomes the shortcomings of the traditional method of dealing with evidence conflicts and greatly improves the recognition ability of the system,especially for the case of high evidence conflicts.This method is superior to other existing methods. |
