Fast evaluation for speckle feature in urban SAR images

Coherent speckle is an important feature in Synthetic Aperture Radar (SAR) image. Testing the property of speckle is effective for urban SAR image performance evaluation. Through correlation analysis for speckle measures, this paper proposes a measurement for speckle feature based on parametric hypothesis test. Firstly selects homogeneous areas in tested SAR image with manual mode and estimates the ENL of each area. In contrast with the speckle feature in sampled image from population, secondly utilizes the u-parametric hypothesis test to measure the quality of speckle. Finally obtains the result that accepts the null hypothesis or rejects it. Experiments of Japan Pi-SAR and East China Research Institute of Electronic Engineering (ECRIEE) SAR images in urban areas show the measurement could distinguish those tested images where speckle property is in accordance with the coherent imaging principle, which is one of the vital requirements for post processing technology of SAR imagery.