Multiplicative noise variance evaluation in mm-band radar images using myriad estimation

In many cases it is required to evaluate the multiplicative noise relative variance in radar images, in particular, those formed by mm-band side looking aperture radars (SLARs). It is evident that. the more information about noise properties one has, the more effective the algorithms which can be designed for noise reduction in images. This evaluation, for example, is needed before image filtering by local statistical Lee or sigma filters or edge detection. These algorithms need for their normal operation the value of relative variance σ _μ², which is not always known, and therefore, should be estimated. It is often desirable to perform this operation automatically. It is of great importance to derive an appropriate model describing the real-life situation for a certain application