Noise reduction of L-band ScanSAR mode images for sea surface wind retrieval

Although L-band ScanSAR mode images are widely used for sea surface wind retrievals, some of them have serious band-like noises. These noise-like features make it difficult to obtain accurate wind vectors from SAR data. To improve the accuracy of retrieved SAR winds, it is necessary to remove the noise from the original images prior to wind retrieval. The objectives of this study are to investigate the characteristics of the noise band and to suggest the noise removal method to derive sea surface wind field from L-band SAR data. In order to reduce noise, the designed band-pass filter was applied. After noise reduction, the accuracy of retrieved wind speed was improved from 5.20 m/s to 3.92 m/s after noise reduction.