Research On Single Vision Target Tracking Technology In Complex Environment

In recent years,the target tracking algorithm has made remarkable achievements,but due to the influence of many tracking difficulties in complex scenes in the actual tracking process,the target tracking algorithm still has great room for improvement.Therefore,this paper improves the target tracking algorithm based on correlation filtering.The main research contents are as follows:(1)In view of the problem of insufficient coping ability in complex scenes and the increasing complexity of tracking algorithm,the real-time performance of visual target tracking is declining.A multi-scale fast correlation filtering tracking algorithm based on fusion features is proposed.Firstly,the gray level features,HOG features and CNs features are fused to form a feature matrix,which makes full use of the target information and improves the ability of describing the target.Then,the principal component analysis is used to extract the significant features in real time,reconstruct the feature matrix and train the matrix effectively in dimension reduction.The position correlation filter is trained to accurately predict the position of the target,improve tracking accuracy and ensure fast tracking speed.Finally,the scale correlation filter is trained by using the fusion feature matrix to further enhance the robustness of the algorithm.The experimental results show that the improved algorithm has high tracking accuracy,good robustness and good tracking effect in complex scenes.It can stabilize the tracking target and reflect high tracking efficiency.It takes good account of the robustness and real-time performance of the tracking algorithm.(2)In complex scenes,the traditional manual feature-based correlation filtering target tracking algorithm has poor tracking effect and poor robustness in dealing with occlusion.A correlation filtering tracking algorithm based on depth feature and anti-occlusion strategy is proposed.The convolution neural network is used to extract the depth feature of the target,and then the depth feature is integrated into the classical correlation filtering tracking framework.The advantages of strong ability of describing the target by depth feature and high tracking efficiency of correlation filtering tracking algorithm are fully utilized.At the same time,the anti-occlusion strategy with high confidence update is used to update the correlation.The filter improves the tracking accuracy and robustness of the algorithm.In addition,the scale estimation module is added to further improve the tracking effect of the algorithm.The experimental results show that the tracking accuracy and success rate of the proposed algorithm are greatly improved,the tracking effect is good,and it can better deal with various complex tracking scenarios.Aiming at the difficulties and challenges in complex scene tracking,this paper makes a deep research on target tracking algorithm based on correlation filtering,and puts forward an improved method for the corresponding tracking problem.The experimental results show thatthe two algorithms proposed in this paper can better deal with various complex tracking scenarios and have good tracking performance.