CAJ | 학술논문

经典meanshift跟踪算法因其具有计算量小、实时性强等优点而得到广泛应用。但在目标发生缩放、旋转、光照变化以及目标颜色与周围背景颜色难以区分等情况下，经典meanshift算法的鲁棒性不佳，目标定位精度不高，算法稳定性不好。鉴于此，本文提出一种联合颜色和纹理直方图表示的meanshift算法，其中本文用到的局部二值模式LBP是基于分块思想改进的纹理算子，有效地提取边缘和角点等主要目标模式来更加精炼地表示目标；另外，通过有效利用目标候选区域的矩信息，来解决跟踪目标运动中尺度和方向的变化的问题。通过不同场景视频跟踪实验表明，与经典meanshift算法及改进前的联合特征表示方法相比，文中提出的改进算法在上述复杂场景下的目标跟踪具有更高的稳定性和鲁棒性。
경전meanshift근종산법인기구유계산량소、실시성강등우점이득도엄범응용。단재목표발생축방、선전、광조변화이급목표안색여주위배경안색난이구분등정황하，경전meanshift산법적로봉성불가，목표정위정도불고，산법은정성불호。감우차，본문제출일충연합안색화문리직방도표시적meanshift산법，기중본문용도적국부이치모식LBP시기우분괴사상개진적문리산자，유효지제취변연화각점등주요목표모식래경가정련지표시목표；령외，통과유효이용목표후선구역적구신식，래해결근종목표운동중척도화방향적변화적문제。통과불동장경시빈근종실험표명，여경전meanshift산법급개진전적연합특정표시방법상비，문중제출적개진산법재상술복잡장경하적목표근종구유경고적은정성화로봉성。
The meanshift tracking algorithm has been widely applied in visual tracking due to its well known merits such as the small amount of calculation and real?time.However when the scale orientation and illumination of target s changes and the color of the background is similar to the target the robustness of the traditional meanshift algorithm is poor and the accuracy of target location is not high as well as the algorithm stability is not good.For this reason an improved meanshift tracking method is proposed in the pa?per.In this algorithm a novel feature histogram is proposed which fuses the color and improved block LBP histograms as the repre?sentation of the tracked object it can extract the main information like the edge and the corner of the object.Besides with the help of moment information from the candidate object area it can solve the problem of the changes of the scale and direction of the tracked object during the tracking process. Compared with the traditional meanshift tracking algorithm and the combine feature algorithm which is changed before the proposed tracking algorithm achieves more stability and robustness in the complex condition.