注意力增强和目标模型更新的红外目标跟踪算法  

Infrared target tracking algorithm based on attention mechanism enhancement and target model update

作　　者：汲清波[1,2] 陈奎丞 侯长波 李子琦[1] 戚宇飞 Ji Qingbo;Chen Kuicheng;Hou Changbo;Li Ziqi;Qi Yufei(College of Information and Communication Engineering,Harbin Engineering University,Harbin 150001,China;Key Laboratory of Advanced Marine Communication and Information Technology,Ministry of Industry and Information Technology,Harbin Engineering University,Harbin 150001,China)

机构地区：[1]哈尔滨工程大学信息与通信工程学院,哈尔滨150001 [2]哈尔滨工程大学先进船舶通信与信息技术工业和信息化部重点实验室,哈尔滨150001

出　　处：《中国图象图形学报》2023年第9期2856-2871,共16页Journal of Image and Graphics

基　　金：国家重点研发计划项目(2018AAA0102702);黑龙江省自然科学基金项目(JJ2019LH2398);中央高校基本科研业务费专项资金资助(3072020CFT0801,3072019CF0801,3072019CFM0802)。

摘　　要：目的多数以深度学习为基础的红外目标跟踪方法在对比度弱、噪声多的红外场景下,缺少对目标细节信息的利用,而且当跟踪场景中有相似目标且背景杂乱时,大部分跟踪器无法对跟踪的目标进行有效的更新,导致长期跟踪时鲁棒性较差。为解决这些问题,提出一种基于注意力和目标模型自适应更新的红外目标跟踪算法。方法首先以无锚框算法为基础,加入针对红外跟踪场景设计的快速注意力增强模块以并行处理红外图像,在不损失原信息的前提下提高红外目标与背景的差异性并增强目标的细节信息,然后将提取的特征融合到主干网络的中间层,最后利用目标模型自适应更新网络,学习红外目标的特征变化趋势,同时对目标的中高层特征进行动态更新。结果本文方法在4个红外目标跟踪评估基准上与其他先进算法进行了比较,在LSOTB-TIR(large-scale thermal infrared object tracking benchmark)数据集上的精度为79.0%,归一化精度为71.5%,成功率为66.2%,较第2名在精度和成功率上分别高出4.0%和4.6%;在PTB-TIR(thermal infrared pedestrian tracking benchmark)数据集上的精度为85.1%,成功率为66.9%,较第2名分别高出1.3%和3.6%;在VOT-TIR2015(thermal infrared visual object tracking)和VOT-TIR2017数据集上的期望平均重叠与精确度分别为0.344、0.73和0.276、0.71,本文算法在前3个数据集的测评结果均达到最优。同时,在LSOTB-TIR数据集上的消融实验结果显示,本文方法对基线跟踪器有着明显的增益作用。结论本文算法提高了对红外目标特征的捕捉能力,解决了红外目标跟踪易受干扰的问题,能够提升红外目标长期跟踪的精度和成功率。Objective Most target tracking algorithms are designed based on visible sight scenes.However,in some cases,infrared target tracking has advantages that visible light does not have.Infrared equipment uses the radiation of an object itself to image and does not require additional lighting sources.It can display the target in weak light or dark scenes and has a certain penetration ability.However,infrared images have defects,such as unclear boundaries between targets and back grounds,blurred images,and cluttered backgrounds.Moreover,some infrared dataset images are rough,negatively impacting the training of data-driven-based deep learning algorithms to a certain extent.Infrared tracking algorithms can be divided into traditional methods and deep learning methods.Traditional methods generally take the idea of correlation filter ing as the core.Deep learning methods are mainly divided into the method of a neural network providing target features for correlation filters and the method of calculating the similarity of the image area with the framework of the Siamese network.The feature extraction ability of traditional methods for infrared targets is far inferior to that of deep learning methods.More over,the filters trained online cannot adapt to fast-moving or blurred targets,resulting in poor tracking accuracy in scenes with complex backgrounds.At present,most deep-learning-based infrared target tracking methods still lack the use of detailed information on infrared targets in infrared scenes with weak contrast and noise.Most trackers cannot effectively update the tracked target when the tracking scene has similar targets and cluttered background.This scenario results in poor robustness in long-term tracking.Therefore,an infrared target tracking algorithm based on attention and template adaptive update is proposed to solve the problems mentioned.Method The Siamese network tracking algorithm takes the target in the first frame as the template and performs similarity calculation on the search area of the subsequent fra

关 键 词：红外图像 目标跟踪 孪生网络 无锚框 高效注意力 自适应更新 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]