| Sea surface wind field and ocean waves are two important phenomena of marine dynamic processes,and obtaining their information is of great significance to marine scientific research,disaster prevention and reduction.Synthetic aperture radar is a kind of microwave remote sensing sensor,which can detect the global ocean all-time and all-weather.It is a very important means of detecting wind field and ocean waves.At present,the retrieval of SAR wind field and wave parameters usually relies on the initial information such as wind direction and initial guess spectrum,which mostly needs to be obtained from external data sources.As a result,SAR can not become an independent observation source to detect wind field and ocean waves.In addition,obtaining initial information from external data sources will reduce the timeliness of wind field and wave retrieval.In order to solve the problem that the retrieval of wind field and wave in SAR depends on external data sources,three retrieval methods of wind field and wave parameters are established in this thesis.Each method is aimed at different SAR data sources and has its own applicable conditions.To a certain extent,it solves the problems faced by the current SAR wind field and ocean wave retrieval.The latter two methods realize the joint retrieval of wind field and wave,and improve the timeliness and convenience of wind field and wave retrieval.This thesis has completed the following innovative work:Firstly,in order to solve the problem that the wind field retrieval method of single satellite co-polarized SAR depends on external data sources,a SAR wind field retrieval method for satellite networking is proposed.This method makes the retrieval of SAR wind field independent of wind direction.It is suitable for C-band and X-band multi-incident angle networking SAR simulation data,and realizes the SAR wind field retrieval based on a single data source.The result of comparison with actual wind speed shows that the method proposed in the thesis can extract accurate sea surface wind speed from SAR data without relying on external data sources.Then,in order to solve the problem that the co-polarization SAR wind field retrieval method and the ocean wave spectrum retrieval method rely on external data sources,a mechanismdriven SAR wind field and ocean wave joint retrieval method is proposed.In this method,the retrieved wave propagation direction and wind speed are used to supplement the reference wind direction and initial guess spectrum information.It is suitable for C-band single-satellite copolarization SAR data,and realizes the joint retrieval of SAR wind field and ocean waves based on a single data source,and requires that the wave components are mainly wind wave.The results of comparison with ECMWF wind field and wave parameters show that the method proposed in the thesis can extract wind field and wave parameters that meet the accuracy requirements from the SAR data of applicable sea conditions without relying on external data sources.Finally,because the mechanism-driven SAR wind field and ocean wave joint retrieval method relies on the wind direction and initial guess spectrum,it is limited by the wave composition.In order to make the wind field and ocean wave retrieval of co-polarized SAR independent of wind direction and initial guess spectrum,a data-driven SAR wind field and ocean wave joint retrieval method CWAVE＿WIND is proposed,which can directly extract wind field and wave parameters from SAR data.In this method,the empirical model of wind field and ocean wave joint retrieval is established by using support vector regression algorithm and convolution neural network algorithm respectively,which is suitable for Sentinel-1 satellite SAR wave model data.This method can make the retrieval of SAR wind field and ocean waves completely independent of wind direction and initial guess spectrum information,and realizes the joint retrieval of SAR wind field and ocean waves based on a single data source,and is not limited by the components of ocean waves,so it is suitable for most sea conditions.The buoy verification results show that the method proposed in the thesis can extract wind field and ocean wave parameters that meet the accuracy requirements from SAR data without relying on external data sources. |
