摘要
针对复杂背景下单帧红外图像中的弱小目标检测的问题,提出一种基于蜕化决策树的检测算法。该算法通过提取红外小目标梯度对比度、邻域灰度分布等多个统计特征,充分利用了各个特征的互补作用,达到了高检测概率和低虚警率。通过对实测数据进行仿真,结果表明该方法能有效检测红外弱小目标。
Aimming at the detection of infrared small target under complicated background,the paper proposes a kind of degenerate decision tree arithmetic.Through picking up infrared character,grad contrast,distribution of gray level in the neighborhood,etc,it utilizes complementarities fully between each infrared character,attaining high detection ratio and lower false alarm rates.Simulation result indicates that it can detect infrared target with high efficiency.
出处
《激光与红外》
CAS
CSCD
北大核心
2010年第9期1017-1022,共6页
Laser & Infrared
