摘要
提出了一种特征点匹配的近距离红外目标跟踪算法,该算法利用Harris算子提取目标的特征点,然后利用Hausdorff距离匹配帧间的特征点集,为了减少噪声和杂点的干扰,还引入了特征点邻域相似性度量。该算法在目标出现尺度伸缩、位置平移、角度偏转的情况下仍有较好的匹配性能。实验证明了该算法的有效性和可行性。
An algorithm for close range infrared target tracking is proposed based on feature point matching. tance In the algorithm, Harris operator is used for detecting the feature point of the target, Hausdorff disis used for matching feature point sets between image sequences. Feature point neighborhood similarity measurement is also introduced for reducing noise disturbance. This algorithm is adaptable to target scaling change, position parallel movement or angle change because of the feature matching. Experiment demonstrates the feasibility of this algorithm.
出处
《电光与控制》
北大核心
2007年第3期8-11,共4页
Electronics Optics & Control
基金
国家自然科学基金资助项目(60572080)
