深度信息辅助的均值漂移目标跟踪算法-计算机工程与应用.pdfVIP

Computer Engineering and Applications 计算机工程与应用 2013 ，49 （23 ） 177 深度信息辅助的均值漂移目标跟踪算法 宋康康，陈 恳，郭运艳 SONG Kangkang, CHEN Ken, GUO Yunyan 宁波大学 信息科学与工程学院，浙江 宁波 315211 College of Information Science and Engineering, Ningbo University, Ningbo, Zhejiang 315211, China SONG Kangkang, CHEN Ken, GUO Yunyan. Mean Shift object tracking algorithm assisted by depth cues. Computer Engineering and Applications, 2013, 49 （23 ）：177-180. Abstract ：The background noise in the candidate object model diminishes the object color characteristic, and induces localiza- tion error. To reduce the error, according to the discriminative depth level between the object ’s and the background ’s, a Mean Shift algorithm based on depth cues assisted and corrected background-weighted histogram is proposed. The proposed algorithm can sufficiently weaken the background noisy interference in the kernel window, enhance the object ’s color feature information, and update the kernel size adaptively in due course to reduce the distractive information in the background as the object size becomes small. Experimental result shows the proposed algorithm has fewer iteration number and good localization precision of tracking. Key words ：depth cues; Mean Shift; adaptive kernel bandwidth; color histogram 摘 要：参考目标模型中混入的背景噪声会弱化目标特征的描述，导致目标跟踪定位误差。为减少误差，依据目标与背景 处于不同深度平面的特点，提出了基于深度信息辅助的和改进的背景加权直方图的Mean Shift 跟踪算法，能够有效削弱 核窗口中的背景干扰信息，突出目标的颜色特征信息，并适时自适应更新核带宽，减少因目标尺寸变小时引入较多的背景 干扰信息。实验结果表明该算法迭代次数更少，具有良好的跟踪定精度。 关键词：深度信息；均值漂移；带宽自适应；颜色直方图 文献标志码：A 中图分类号：TP39 doi ：10.3778/j.issn. 1002-8331.1202-0245 [6] 1 引言 缘特征辅助颜色特征 ，提升颜色直方图抗噪声能力，但是 Mean Shift 算法是一个自适应步长的迭代算法，已经 在发生部分遮挡时，算法就会失效。通过削弱目标颜色直 成