论文部分内容阅读
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 multi-dimensional signal detection, it is usually necessary to use the projection method to reduce the dimensionality, so as to reduce the computational complexity and simplify the detection process.In this paper, a new sample is proposed in which only one sample of many samples to be projected may be a target Projection algorithm, which makes full use of each sample point is the probability of the target information to construct a combination of samples; derived each sample is a probability distribution formula for the target. As a verification, it is applied to the infinitesimal point moving target in the infrared image sequence , The performance analysis and simulation results are given and compared with the commonly used maximum projection algorithm.The results show that the proposed algorithm is superior to the latter one in the special case of this algorithm, The target detection in the case of Bobby, its performance has greatly improved.