摘要
提出了一种基于相关的自适应阈值区域匹配、匹配区域分层加权和模板自适应更新的目标跟踪方法,该方法能在序列图像中自适应地寻找最佳的动态阈值与更新模板;采用区域分层加权后经相关图形分析及实验验证,能使相关的目标最佳匹配区域更小,匹配中心更加突出且易于匹配,从而缩小了匹配范围,提高了匹配精度。同时提出了一种区域分层快速搜索算法,该算法能较大地提高匹配速度。实验结果表明该方法匹配精度高、匹配速度快、能克服不同光强背景及噪声干扰等影响,具有较强的实用价值。
To adaptively search the optimum dynamic threshold and update the template in image sequence, a target tracking method is developed. The correlation-based method consists of adaptive threshold region matching, layered-weighting in the matching region, and updating the templete adaptively. Experiment shows that the easier matching and higher matching accuracy are reached due to the smaller target matching region and more distinguished matching center obtained in the method. A region layer-based search algorithm is also presented. The fast matching speed with higher accuracy and the stronger noise resistant ability are obtained by the method in the different light intensity backgrounds.
出处
《航空学报》
EI
CAS
CSCD
北大核心
2006年第4期670-675,共6页
Acta Aeronautica et Astronautica Sinica
基金
国家自然科学基金(60475002)
江西省自然科学基金(0511071)
江西省测控中心开放基金(KG200204011)资助项目
关键词
信息处理技术
自适应阈值
分层加权
模板自适应更新
分层快速搜索
information processing technology
adaptive threshold
layered weighting
templete adaptive update
layered fast search
