The Research On Infrared Multi Weak Targets’ Detecting And Tracking Algorithm Based On PTS And GM-CBMeMBer Filter

With the increasing demand of military and civilian fields,the application environment of infrared imaging system is more and more complicated.The detection and tracking of dim and small infrared targets in complex background has been a hot topic in the world.This paper carries out an in-depth study starting with two aspects of detection and tracking.Firstly,we select three method from morphological filtering,partial differential equations and the multi-scale geometric transform domain for background suppression,and the three methods are simulated under the same conditions.On the basis of this,a PTS multi model background suppression is proposed.The algorithm combines the above three algorithms organically.It can switch the background suppression algorithm according to different conditions to achieve effective suppression of the background and increase the contrast of the target.And then use the decision boundary segmentation algorithm to effectively divide the target.Secondly,this paper use the Cardinality Balanced Multi-Target Multi-Bernoulli filter(CBMe MBer)for infrared multi-target tracking.This method considers each target trajectory as the corresponding Bernoulli distribution of the Bernoulli term,and each possible state of the target is regarded as a Gaussian term that obeys the Gaussian linear model.According to the parameter set and the target state set of the initial multi-Bernoulli term and the Gaussian term,the parameter set and the target estimation state set are obtained by step-by-step recursion.And the target state estimation will be achieved after elimination and merging.This paper simulates and analyzes the CBMe MBer filter algorithm based on the linear and nonlinear Gaussian mixture model.The results show that the former is better in real time,and the latter can be applied to the tracking of nonlinear moving objects.Finally,this paper combines the proposed detection algorithm and the tracking algorithms,which is applied to the dim and small multi-target detection and tracking problem based on the infrared image sequence.The simulation results show the effectiveness of the algorithm.