基于稀疏编码直方图的稳健红外目标跟踪

发布时间：2018-01-16 12:24

本文关键词：基于稀疏编码直方图的稳健红外目标跟踪　出处：《光学学报》2017年11期 　论文类型：期刊论文

【摘要】：充分利用红外图像信息建立有效的观测模型是实现稳健红外目标跟踪的基础。影响红外目标跟踪结果的因素除可见光目标跟踪也会面临的干扰因素之外,还有诸如边缘和纹理信息缺失、信噪比低和背景噪声影响等特有因素。提出基于稀疏编码直方图(HSC)特征和扰动感知模型(DAM)的红外目标跟踪方法,使用K-奇异值分解算法得到过完备字典,利用该字典计算得到每个像素点的稀疏编码,并组成HSC对目标进行表达,同时通过引入DAM增强算法抗背景干扰能力。该方法充分利用了红外图像中运动目标的结构特性,能够有效去除背景干扰。与其他跟踪器相比,在VOT-TIR2015数据集上,该方法的精确度和成功率指标分别获得3.8%和4.4%的提升,具有较高的研究价值和实用价值。
[Abstract]:Making full use of infrared image information to establish an effective observation model is the basis of robust infrared target tracking. There are also missing information such as edges and textures. This paper presents an infrared target tracking method based on sparse coded histogram (HSC) features and disturbance perception model (DAM). The over-complete dictionary is obtained by using the K-SVD algorithm, and the sparse encoding of each pixel is obtained by using the dictionary, and the target is represented by HSC. At the same time, DAM is introduced to enhance the ability of anti-background interference. This method makes full use of the structural characteristics of moving targets in infrared images, and can effectively remove background interference, compared with other trackers. The accuracy and success rate of this method are improved by 3.8% and 4.4% respectively on the VOT-TIR2015 dataset, which has high research value and practical value.
【作者单位】：河北工业大学控制科学与工程学院;
【基金】：国家自然科学基金(61203076) 天津市自然科学基金(13JCQNJC03500) 河北省自然科学基金(F2017202009)
【分类号】：TN219;TP391.41
【正文快照】： 红外目标跟踪广泛应用于军事及生产生活中,具有重要的价值。红外目标跟踪与可见光目标跟踪相比,除了需要解决目标形变、发生遮挡、运动模糊等问题之外,还需要解决更加独特的问题,比如信噪比低、纹理信息缺乏、背景复杂混乱[1]。这些特有的影响因素使红外目标跟踪更具有挑战性

【相似文献】