电子学报 ›› 2005, Vol. 33 ›› Issue (7): 1331-1333.

• 论文 • 上一篇    下一篇

基于概率加权求和投影的多维信号检测技术研究

艾斯卡尔·艾木都拉1, 陈颖2, 贾振红1   

  1. 1. 新疆大学数学学科博士后科研流动站,乌鲁木齐 830046;2. 中电科技集团第10研究所,四川成都 610036
  • 收稿日期:2004-02-17 修回日期:2004-12-18 出版日期:2005-07-25

A Probabilistic Weighted Summation Projection Technique for Multidimensional Signal Detection Hamdulla·Askar1,CHEN ying2,JIA zhen-hong1

Hamdulla·Askar1, CHEN ying2, JIA zhen-hong1   

  1. 1. Postdoctoral Research Workstation of Mathematics,Xinjiang University,Urumqi,Xinjiang 830046,China; 2. China Electronics Technology Group Corporation No.10th Research Institute,Chengdu,Sichuan 610036,China
  • Received:2004-02-17 Revised:2004-12-18 Online:2005-07-25 Published:2005-07-25

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.

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

Abstract: 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 probabilistic 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.

Key words: probabilistic weighted summation projection, maximum value projection, dim point target, image sequences, Detect and estimation

中图分类号: