Edge detection in speckled SAR images using the continuous wavelet transform

This study presents an original method to detect edges in synthetic aperture radar images corrupted by speckle. The authors show that the continuous wavelet transform (CWT) allows to accurately detect edges embedded in speckle. Edges give rise to a conic signature in the CWT. Furthermore, the signal to noise ratio (SNR) increases with the scale of the CWT. The optimal wavelet maximizing the SNR is derived and two wavelet based detectors are proposed. Both are 2D detectors and so take advantage of information redundancy in the scale-space plane