Object Tracking Algorithm Based On Fusion Of Scale Invariant Feature

Infrared target tracking has always been a hot topic in the field of visual tracking technology at home and abroad.The main task of infrared target tracking is to search and locate the targets which interests us in infrared video,and obtain the moving trajectories of the targets.After nearly a hundred years of research and development,infrared target tracking has achieved remarkable achievements and extensive application in military investigation,missile guidance,security detection and medical imaging.In reality,Infrared video images and targets are changing all the time,for example,the size of the target,its speed,the rotation of the target,noise and scene clutter interference and so on.And these factors can reduce the accuracy and stability of target tracking,which has brought great difficulties in the infrared target tracking.In order to eliminate the impact of these factors on target tracking as much as possible,researchers have done a lot.And a number of classic target tracking algorithms have been proposed,such as gray-related matching,feature matching and other target tracking algorithm.Aiming at the problems existing in target tracking,the target tracking algorithm based on SURF-BRISK feature matching and the target tracking algorithm based on scale space correlation matching are studied in this paper.The target tracking algorithm which based on SURF-BRISK feature matching uses SURF for feature detection,BRISK is used to feature description and calculates the Hamming distance for feature matching to obtain target location.When the target size is large and the detail information is obvious,the algorithm can track the target stably.But when the target size is small,it is difficult to track the target.Scale spatial correlation matching algorithm establishes a multi-scale target template firstly,and the similarity coefficient of target template and image is calculated to obtain the target location.The algorithm can achieve accurate tracking of the target when the target size is small,but it is easy to fail when the target size changes greatly.Aiming at the limitation of the above two algorithms,an improved SURF-BRISK combination algorithm is proposed which is the target tracking algorithm based on SURF-BRISK feature fusion of scale spatial correlation matching.When the target moves in the field of view,target positions are obtained by feature matching and correlation matching firstly.Then the offsets are calculated between the two target positions and the target position of the previous frame.And the sum of the deviations and the proportional value of their respective deviations are calculated.According to the proportional value,the corresponding weight is assigned to the obtained target position,and the sum of the weight and the proportional is one.Finally,the two target positions are weighted to obtain the position of the current frame target.Experimental results have proved that the algorithm can track the target in the case of target size change and occlusion.