Quantitative Research And Correction Of Incident Angle Effect For Wide Swath Sar Image

Synthetic Aperture Radar (SAR) provides an excellent capability that it couldwork in multi-polarization and multi-band, being insensitive to the weather duringday or night.However, for both space-borne and air-borne wide swath SAR, theexistence of incident angle effect have a severe impact on the accuracy of automaticinterpretation, which is not only for a moderate or remote distance interpretationbut a close one, and even those of the artificial interpretation as well. It seriouslyhinders the image information extraction.This topic, which is based on the national natural science foundation ofChina,through access to a large number of domestic and foreign literature,conductsthe thorough research to the incident angle effect for the SAR images. The mainwork of this paper is to describe the theoretical knowledge of the SAR and analyzethe impact of the scattering dependent factors. The land-cover sampling bywatershed segmentation and region-mark technology has been done after the imagepreprocessing, and then carried out the quantitative research of incident effectabout wide swath SAR image. The results show that different land-cover types havedifferent incident angle effect.Basing on the quantitative study of the incident angle effect of wide swathSAR images, this paper proposes a more thorough correction scheme—the schemeabout the distribution of land-cover types. According to the given research areabased on the data of ENVISAT ASAR in wide model, assuming the sameland-cover type has the same roughness, moisture content and physical structure,achieving the class-based correction of land-cover backscatter values based on thecosine Lambert’s law estimated through linear regression. Experimental resultsshow that there is an evident incident angle effect on radar backscatter, the higherthe moisture content of the land-cover type, the more obvious of the incident angleeffect. The correction of incident angle effect with this proposed algorithm is betterthan the absolute calibration and first order cosine correction.