融合多特征的时空正则化相关滤波无人机跟踪  

Multi-Feature Spatial-Temporal Regularized Correlation Filter for UAV Tracking

作　　者：张微 ZHANG Wei(School of Computer Science,Baqji University of Arts and Sciences,Baqji 721000,China)

机构地区：[1]宝鸡文理学院计算机学院,陕西宝鸡721000

出　　处：《电光与控制》2022年第7期29-36,共8页Electronics Optics & Control

基　　金：陕西省教育厅科学研究计划项目(20JK0487);陕西省科技厅工业攻关项目(2022GY-071)。

摘　　要：针对无人机跟踪目标易受视角变化、遮挡、背景杂乱等因素影响的问题,提出一种融合多特征的时空正则化相关滤波无人机跟踪方法。首先,将显著性特征引入时空正则化相关滤波跟踪框架,与颜色、灰度和梯度方向直方图特征结合,提高目标外观表示的多样性;其次,利用峰值旁瓣比作为权重衡量不同特征相关响应图的峰值强度,并将加权后特征进行组合降噪,在响应层实现最终加权融合,提升目标定位精度;最后,在公开无人机视频数据集UAV123@10FPS上与12种经典跟踪器进行对比。实验结果与分析表明,所提方法在跟踪精确度和成功率上均取得较好的结果。Aiming at the problem that UAV target tracking is easily affected by factors such as the change of angle of viewocclusion and background cluttera multi-feature spatial-temporal regularized correlation filters for UAV tracking is proposed.Firstlythe saliency feature is introduced into the spatio-temporal regularized correlation filter tracking frameworkwhich is combined with the features of colorgray level and histogram of oriented gradient to improve the diversity of target appearance representation.Secondlythe peak sidelobe ratio is utilized as the weight to measure the peak strength of correlation response maps for different featuresand the weighted features are combined for noise reductionso as to realize the final weighted fusion in the response layer and improve the tracking accuracy.Finallyit is compared with 12 classic trackers on the public data set [e-mail] experimental results and analysis show that the proposed method achieves favorable results in both tracking accuracy and success rate.

关 键 词：无人机跟踪 相关滤波 多特征融合 时空正则化 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]