Synthetic Aperture Sonar Processing with MMSE Estimation of Image Sample Values
Abstract
MMSE (minimum mean-square error) target sample estimation using non-orthogonal basis functions is applied to objectimaging with SAS/multi-aspect data. Theoretical and simulation results are impressive: Accurate images from echoes atthree aspects, without a significant increase in SNR. The MMSE process creates images with new features that can beapplied to image-based compensation of receiver rotation/location measurement errors. The simulation, however, indicatesthat a weighted sum of current image features is insufficient to reliably compensate for such errors when the gradient of theweighted sum with respect to location/rotation parameters is used in an optimization procedure.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 2016
- Accession Number
- AD1034518
Entities
People
- Richard A. Altes