The Research On Infrared Target Tracking Based On Particle Filters

Object tracking technology, which widely used in ATR system and InfraredSurveillance system, plays the key role in the national defense and military. Duringrelated technology, infrared object tracking technology distract object or interests areafrom the background in infrared images, which obtained by infrared detector. As thebest solution in Bayesian, particle filters used to tackle the problem of object trackingunder the states which is non linear and non Gaussian.The article does some research on the infrared object tracking with particlefilters. The main work are as following:It introduces the tendency of the infrared object tracking research and relatedworks in details.Generally speaking,Infrared object tracking system is non linear andnon Gaussian system,and particle filters is the best candidate for the tracking problem.The article also analyses the basic feature of infrared images and infrared objects andintroduces how to weigh the distance between object and background in a well means.It proposes a new algorithm for particle filters since some unanticipated casehappened in resampling process.During simulation experiment,compared with thegeneral particle filters,better performance of the new approach has been proved in thischapter.As the best solution for the tracking problems in non-linear/non-guassiansystem,particle filters needs a big sampling set so as to cover all the cases of thetarget information in next image frame,and it hardly to deal with the tracking problemof real-time system because of the big set.To cut the problem,it proposes a newsolution,which named based on hypothesis testing in auto-adapt particle filters forinfrared target tracking, in the real-time system.It takes the hypothesis testing theoryinto particle filters framework,and proposes the best policy of number of particle inthis chapter.The new approach also has been proved with good real-time feature androbustness during simulation experiment on matlab.Finally,it introduces the frame construction of infrared object tracking system.