Wavelet based SAR data denoising and analysis

Real world signals are afflicted by noise, be it the communication signals or the remote sensor signals including the medical imaging sensors or surveillance sensors. In the case of signals from SAR(Synthetic Aperture Radar) sensor from remote sensing platforms, or USG or MRI from medical images, the noise is multiplicative. The coherent nature of the signal processing involved in the data generation for the latter, gives rise to a g rainy kind of signal dependent noise called speckle. Speckle noise affects the radiometric quality of the image and makes data interpretation and analysis difficult. Since the utility of any data depends on the accuracy of the data both in terms of geometry and radiometry, denoising becomes mandatory for image analysis applications. The common problem with most of these denoising filters is that they blur the images while reducing the noise. The challenge thus, lies in cleaning the image without sacrificing the geometric resolution. Wavelet based filtering techniques show promising results in this regard. In this paper the results and analysis of performing filtering on high resolution SAR data, using different mother wavelets of Daubechie´s using soft thresholding technique, are presented.