The detection of weak signal patterns in radar ocean intensity images
Detection of weak patterns in radar ocean RCS images is complicated by the fact that signals and noise are interactive rather than additive and the ambient noise background is non Gaussian or even strongly non Gaussian at low grazing angles. This paper addresses this difficult problem with the aid of two simplifying assumptions: (1) the signal modulation is weak, and (2) departure from Gaussianity is small. In situations where this departure is large, an approach is suggested for reducing this non Gaussianity. The relevant weak signal detection theory, based on the Likelihood ratio, is reviewed and adapted for use in the analysis. The approach to this problem, similar to that previously used for complex images, is facilitated by approximating the multivariate probability distributions as a composite integral involving underlying processes which are assumed to be Gaussian. This formulation, subject to the approximations in the analysis, permits derivation of an ideal detection statistic (which determines the form of optimum receiver) and a signal/noise ratio which characterizes detection performance in the weak signal limit. Implications for image processing are discussed and directions for future analysis are suggested.
- Research Organization:
- Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
- Sponsoring Organization:
- USDOE, Washington, DC (United States)
- DOE Contract Number:
- W-7405-ENG-48
- OSTI ID:
- 279551
- Report Number(s):
- UCRL-ID-124927; ON: DE96012636
- Resource Relation:
- Other Information: PBD: 15 Jun 1996
- Country of Publication:
- United States
- Language:
- English
Similar Records
Improved deconvolution of very weak confocal signals
Improved deconvolution of very weak confocal signals