Final report December 1970-November 1971.
Abstract
Signal generated noise is called reverberation. Reverberation is considered to be a Gaussian process since the number of scatterers contributing to the process at a given time is generally large. However, for high resolution sonar, the number os scatterers in the resolved area (or volume) may be so small as to result in a non-Gaussian process. The set of sonar returns resulting from consecutive projected pulses is generally considered to be a true ensemble. However, when the mean intensity varies from return to return, the set can no longer be considered to constitute a ture ensemble. The overall objective of the study was to test the Gaussian and ensemble assumptions made about reverberation data at a carefully chosen smooth bottom site and to evaluate the performance of two signal processors - a square law detector and a multiplicative detector - operating against the recorded reverberation. Statistics of the reverberation obtained included mean intensity, variance, probability histograms, and higher order statistics. Theoretical signal-to-noise figures and ROC (receiver operating characteristics) curves were used in examining the performance of the two types of signal processors using both simple and composite target models. (Author Modified Abstract)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 12, 1972
- Accession Number
- AD0757756
Entities
Organizations
- University of Texas at Austin