Wavelet Analysis for RADARSAT Exploitation

This document reports on Wavelet Analysis for RADARSAT Exploitation (WARE), research software that aims to demonstrate the applications of wavelet analysis to RADARSAT ocean imagery. Wavelets are a mathematical paradigm that can be used to partition data into various frequency components, allowing the study of each component at a resolution matched to its scale. Wavelets are an efficient image processing tool for localized feature extraction and texture analysis. WARE tools have been developed using basic wavelet-based algorithms and include: (a) wind direction estimation; (b) oceanic or atmospheric front extraction; (c) oil spill detection; and (d) ship detection. WARE includes rudimentary interactive validation and has been applied to several RADARSAT ocean images that span the noted application areas. As such, the distinctive properties and advantages of wavelet analysis for target detection and recognition are demonstrated. In future, WARE could be extended to include radiometric flattening, improved algorithm efficiencies in anticipation of operationalization, and systematic testing of available wavelet families for improved algorithm performance.