摘要
在多维信号检测中,通常需要采用投影法来进行降维处理,从而减少计算量,简化检测过程.本文在待投影的众多样本中只有一个样本可能为目标的情况下,提出了一种新的投影算法,它充分利用每个样本点是否为目标的概率信息来构造组合样本;推导了每个样本是否为目标的概率分配公式.作为验证,将其运用到了红外图像序列中微弱点状运动目标的检测领域,给出了其性能分析和仿真结果,并与常用的最大值投影算法进行了性能对比.结果表明,本文算法综合性能优于属本文算法特例的后者,并且对低信噪杂波比情况下的目标检测,其性能有较大的提高.
In multidimensional signal detection, there is often too much data to process. By using the concept of projection, the dimensionlity of data may be reduced, thereby simplifying the detection process. In this paper, a projection technique wherein the probabilisfic weighted sum of a set of samples is chosen as the projected sample is presented in the case of only one sample in the set of samples projected onto a single sample may contain the signal. And the probability assignment formula of each sample being signal is derived. As verification, it is used in the fields of detection of dim moving point targets in IR image sequences, the performance analysis and experimental results are also given in this paper. Meanwhile, its performance is compared to the popular technique of maximum value projection wherein the maximum value of a set of samples is chosen as the projected sample. From the comparison, it is concluded that our algorithm is superior to the latter which is belong to the special case of ours and it has a higher detection performance even at low signal-to-noise ratios than the latter.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2005年第7期1331-1333,共3页
Acta Electronica Sinica
关键词
概率加权求和投影
最大值投影
点状目标
图像序列
检测与估计
probabilistic weighted summation projection
maximum value projection
dim point target
image sequences
Detect and estimation
