CAJ | 학술논문

为解决红外目标跟踪中目标的交错、遮挡等问题，提出了一种新的基于运动估计的目标跟踪方法。建立目标的方向梯度-灰度直方图特征模型，该模型能较准确地刻画目标特征。使用最大后验概率指标在搜索区域进行目标匹配，该指标能很好地突出目标、抑制背景，并容易得到全局最优解。提出一种新的运动估计方法，即轨迹预测算法，对目标的运动进行较准确的估计。实验结果证明，该方法不仅计算复杂度低，而且能够较好地解决目标交错、遮挡等问题。
위해결홍외목표근종중목표적교착、차당등문제，제출료일충신적기우운동고계적목표근종방법。건립목표적방향제도-회도직방도특정모형，해모형능교준학지각화목표특정。사용최대후험개솔지표재수색구역진행목표필배，해지표능흔호지돌출목표、억제배경，병용역득도전국최우해。제출일충신적운동고계방법，즉궤적예측산법，대목표적운동진행교준학적고계。실험결과증명，해방법불부계산복잡도저，이차능구교호지해결목표교착、차당등문제。
A novel object tracking method based on motion estimation is proposed in this paper to solve stagger and occlu-sion problems in infrared object tracking. The gradient direction-joint gray histogram model of an object is established, which can depict characteristics of the object accurately. Then maximum posterior probability is used to match the objects in the search area, which is efficient to enhance the target, inhibit the background and easy to find the global optimal solution. A novel method based on motion estimation, trajectory prediction, is proposed to evaluate motion trajectory of object accurately. Experimental results illustrate that the proposed method not only can solve the stagger and occlusion prob-lems better, but also has lower computation complexity.