Application of Wavelets to SAR Remote Sensing
Seisuke Fukuda · Journal of the Visualization Society of Japan · 1999
In this paper, as an application of wavelets to synthetic aperture radar (SAR) remote sensing, a wavelet-based method to reduce speckle is presented. SAR images are inherently accompanied with speckle due to coherent processing used to produce the images. Speckle in SAR images is regarded as a troublesome noise disturbing image interpretation. The presented method is based on an idea that speckle may be suppressed by reducing the amplitude of the pixels in detail images (i.e., the wavelet coefficients) in each wavelet subspace. In the method, edges are preserved by releasing the amplitude reduction around them. Experimental results with actual SAR images demonstrate the effectiveness of the method. In the latter half of the paper, the relation between the smoothing effect of the method and the parameters is fully investigated. The smoothing effect is dependent on the wavelet subspace level and the degree of amplitude reduction of wavelet coefficients. Using this relation, an approach for speckle reduction with texture preservation is derived.