Research Of Ocean Wind Retrival From C-band Quad-polarization SAR Image

Synthetic Aperture Radar(SAR)is an effective way to obtain the sea surface wind field,with high resolution,all-weather observation.The scattering intensity of cross-pol(HV and VH)radar cross section from SAR images is independent wind direction of sea surface and radar incidence angle,it is only linear with the wind speed of sea surface,and is not saturated in the high wind speed region,so it was used to reverse the wind speed,especially in the high wind speed.At present,the retrieval of sea surface wind field by quad-polarition SAR image is still in the research stage,and whether the linear relationship between cross-polarization and sea surface wind speed can be used in other SAR systems is still the focus of discussion at present.However,Because of the speckle noise in SAR images,it is difficult to acquire the scattering characteristics of the irradiated area accurately and it increases the difficulty of the target feature extraction.The traditional de-speckle method reduces the noise,but it leads to the loss of the details of the image at the same time.In this paper,the wind retrieval of quad-polarization is utilized.Firstly,the main retrival methods of ocean wind based on C-band single-polarization are reviewed,and the advantages and disadvantages of different retrieval methods and their application status are analyzed.The relationship between the radar backscattering cross section and the wind speed is obtained and the wind speed of the sea surface is inversed.The radar incidence angle and the relative wind direction are analyzed by using the single polarization SAR image.Then the single-look-complex data of SAR image is preprocessed,which includes multi-looking processing,geometric correction,geographic coding,radiometric calibration and so on.The accuracy of sea surface wind field extraction is directly determined by the pretreatment process.In view of the problem of image noise,the BM3D algorithm is adopted,The denoising method based on the three Dimensional Matching Block(BM3D)utilize the sparse representation in the transform domain,which divides the image into a three-dimensional space.By analyzing the characteristics of the model and setting the parameters,the BM3D algorithm is applied to the SAR image.Compared with the traditional methods,we find that the BM3D algorithm,the edge stay index reached 0.4845,can be successfully applied to the S AR image,which can suppress the speckle and preserve the details of the image,verifing the effectiveness of the algorithm.However,under low wind speed,the image is affected by noise substrate.The accuracy of wind speed inversion is reduced.By analyzing the noise lever of quad-polarized SAR image,the relationship between the noise base and the radar incidence angle of SAR is studied.On this basis,the model of SAR image and the wind speed of sea surface is established.The relationship between the radar backscattering cross section and the wind speed is obtained,and the wind speed of the sea surface is inversed.The experimental results show that the error of the inversion results is less than 1.0 m/s compared with the observed data of the buoy?The error of sea surface wind direction retrieved by CMOD5.N model is less than 10°,which shows that the new model improves the accuracy of retrieving sea surface wind field,it is of great significance for the future retrieval of sea surface wind field by quad-polarization SAR images to reach the operational level.