Study Of Ocean Surface Wind Field Retrievals From L-band SAR

Ocean surface wind fields are essential factor of the ocean and the atmosphere.By the development of the remote sensing, Synthetic Aperture Radar (SAR) becomesa very important measurement on the ocean surface wind because of the capability ofhigh-resolution, working all the time and under all weather conditions. Based ontheoretical analysis, using computer simulation and real data, this thesis studied anddiscussed deeply on the ocean surface wind remote sensing by SAR.Firstly, the microwave scattering theory of ocean surface has been studied.Several important models about microwave scattering of ocean surface and theireffective scope have been introduced. We also introduced the ocean imagingmechanism of SAR. Thus, we have constructed the theoretical basement on oceansurface remote sensing by SAR.Using the physical model function which is promoted by scholar XuFeng basedon the two-scale model and can be successfully applied to a wide range of radarfrequencies, the normalized radar cross section (NRCS) of ocean surface underdifferent wind speed, wind direction, radar incident angle and polarization conditionhave been simulated by computer. The result shows that the value of NRCS isdepended on the situation of wind speed, wind direction, radar incident angle andpolarization condition. Thus is very useful at ocean surface wind retrieving by realSAR data.The ocean surface wind field is retrieved by using L-band airborne SAR data.The wind direction is estimated by the orientation of image spectrum, and the 180°directional ambiguity is removed by using the asymmetry of NRCS about upwind anddownwind directions. Because of the SAR image not been calibrated on radiation, we firstly compute the coefficient of calibration using the real wind speed bymeasurement and the SAR image data which the corresponding location is nearing thewind speed measuring point, then we use the coefficient on the average to calibratethe other region of the image, thus we can get the NRCS of every pixel of the image.Lastly, the wind speed can be retrieved by using the model function which is beeninput the NRCS obtained. From the application examples of the wind field retrieval, itis shown that the result is satisfactory and the method is effective.