Adaptive Correlation Filter Learning with Motion Smoothing for UAV Tracking

Yu-Feng Yu^1,*, Xiaoying Tan¹, Qirong Wu¹, Guoxia Xu²
1 Department of Statistics, Guangzhou University, Guangzhou, China
2 School of Communications and Information Engineering, Nanjing University of Posts and Telecommunications, Nanjing, China
* Corresponding Author: Yu-Feng Yu. Email: email
(This article belongs to the Special Issue: Advances in Deep Learning and Computer Vision for Intelligent Systems: Methods, Applications, and Future Directions)

Computer Modeling in Engineering & Sciences https://doi.org/10.32604/cmes.2026.085413

Received 11 May 2026; Accepted 23 July 2026; Published online 11 August 2026

Download PDF

Abstract

To tackle critical visual tracking difficulties arising in UAV tracking tasks, including frequent target occlusion and abrupt fast motion during high-altitude inspection, we propose an adaptive correlation filter tracking algorithm incorporating a motion smoothing module and adaptive residual regularization, named MACF. The tracker is constructed via multi-strategy fusion of two elaborately designed components at the algorithmic modeling level. First, we design a Motion Smoothing Module (MSM) that conducts weighted fusion of historical motion trends in the modeling pipeline. It suppresses search window jitter arising from instantaneous positioning errors and lowers target drift risk by providing precise spatial priors for search center prediction. Second, we embed an Adaptive Residual (AR) regularization term into the objective function of the correlation filter model, which analyzes temporal feature fluctuations to suppress contextual interference and avoid misleading caused by sporadic distractors. The synergistic effect of the motion smoothing module and adaptive residual regularization enables the filter to adapt to various feature changes, thereby maintaining tracking continuity under target occlusion, rapid motion and blurring.