Research on Unmanned Aerial Vehicle Target Tracking Based on Kernel Correlation Filter and Kalman Filter

Abstract

In UAV (Unmanned Aerial Vehicle) target tracking, due to the complexity of the ground scene, it brings great interference to target detection and tracking. Therefore, a method based on kernelized correlation filter (KCF) is introduced for the tracking of UAV targets. The response peaks of inaccurate tracking are statistically counted to find the relationship between tracking anomalies and response peaks. When the tracking abnormality is judged, the Kalman filter is used to predict and update the target to achieve normal tracking. By testing the UAV videos downloaded from the Internet, it is shown that the established tracking algorithm can solve the problem of error tracking caused by complex environment, and the tracking effect is improved.