Textural method to evaluate speckle noise filters based on AR-2D models using P-band radar data

The performance evaluation of some speckle reduction filters for synthetic aperture radar images is addressed in this work The criteria used to evaluate the filters performance are based on the improvement of signal to noise ratio and its ability to preserve textural features. The former is achieved through equivalent number of looks (ENL) computation, while the latter is analyzed by Euclidean distance between AR-2D models, which are estimated from several samples of three land cover classes.