The Research On The Moving Target Tracking System Based On The IMM-UKF

With the development of computer technology, tracking Technologies of the moving objectshave been applied to the practice. It is an advanced technique including a multiplicity of subjects.The relevant application has related to video Monitoring, medical image analysis, trackingcontrol, many fields. Object dodging and Speed Processes is a key problem for the technology ofthe tracking of the moving objects. This paper focuses on doing the thorough research to theproblem of the maneuvering motion target tracking. Motion object tracking system based on theIMM-UKF.In this paper, we applied the monocular visual technology to monitoring object’s distance.Firstly, EXIF is a specification for the image file format used by digital cameras. Secondly, it isused to correct Camera internal parameters. Finally, the distance between the front recognitionobject and robot is obtained through the image information. The altitude of the objects and theangle of the camera is gotten at the same time. The improved distance-measuring method basedon monocular vision is not constrained by camera axis, height and the placement height of themeasured object. The advantage of this method is quick, accurate, and portable for measuring thedistance of the object which altitude is given. The algorithm avoided complex self-calibrationprocedure; Experimental results show that the improved method is available, fast, accurate andflexible.The research on the object tracking is related to improving the accuracy of the UKF filteringalgorithm of interact multiple models. An interacting multiple model algorithm based on theUKF was proposed to improve the accuracy of interacting multiple model (IMM). The targetmotion is divided into two parts, namely global motion and local motion, which are separatelypredicted according to their own dynamic behavior. CA (Constant Accelerator) and CT (ConstantTurn) target model of interaction solve the state of the target maneuver adaptive estimation. TheUKF simulation results show IMM algorithm for CV, CT maneuvering target tracking canachieve good results.IMM model of UKF algorithm theory is tested by simulation with Matlab program. Motion control system of the camera using MSP430 is designed. The results of experimentsshow that IMM model method has high performance and UKF filter has high accuracy.