CAJ | 학술논문

在目标跟踪中，目标颜色变化或相似颜色背景干扰会导致算法的鲁棒性较差。为解决该问题，提出一种基于多级颜色累积和纹理融合的目标跟踪算法。将算法分为颜色与纹理特征提取、目标与背景特征相似度判断2个子过程，子过程交替执行，在目标颜色发生变化时，利用感兴趣区域帧差法对目标进行再定位，提取多级颜色模板，并将其累积在原模板之上。实验结果表明，该算法的平均跟踪误差约为基于颜色-纹理算法的1/2、为单视觉特征跟踪算法的1/3。
재목표근종중，목표안색변화혹상사안색배경간우회도치산법적로봉성교차。위해결해문제，제출일충기우다급안색루적화문리융합적목표근종산법。장산법분위안색여문리특정제취、목표여배경특정상사도판단2개자과정，자과정교체집행，재목표안색발생변화시，이용감흥취구역정차법대목표진행재정위，제취다급안색모판，병장기루적재원모판지상。실험결과표명，해산법적평균근종오차약위기우안색-문리산법적1/2、위단시각특정근종산법적1/3。
Aiming at several cases existing in the object tracking algorithm, such as failure happens when the object’s appearance changes or when the target and background are similar, an object tracking algorithm based on multistep color accumulation and texture fusion is proposed in this paper. The whole tracking algorithm can be divided into two parts:color and texture feature extracting and target and background’s similarity judging, in the tracking process those two parts run by turns according to different cases. When the object’s appearance changes, the Region of Interest(ROI) frame difference is used to compute the center of the target again, extracts multistep color model and accumulats it to the old one. Experimental results show that the average tracking error of this algorithm is half of the algorithm based on color-texture and one third of single visual feature tracking algorithm.