Research Of Ocean Surface Wind Fields Retrieval Algorithm In The Dynamic Environment Based On Marine Radar Images

The ocean breeze is most common.Winds spawn waves,directly affecting the size of the wind waves,and winds are destructive,which impact the safety of shipping,offshore exploration and other human activities.Modern remote sensing equipment in space oceanography attracted people's attention and scholars from various countries began using scatterometer,synthetic aperture radar(airborne or spaceborne),altimeter and other equipment to monitor sea surface wind field.However,the disadvantages of these devices are very clear,such as low resolution,low oversampling rate,subjecting to weather and other issues.In recent years,marine radar is appliedin-depth in remote sensing fields,especially in the monitoring of sea surface wind field.Marine radar has advantages of the large measurement area,real-time,high resolution.Research on the use of marine radar monitoring wind fields dates back to the 1960's.At present,the more sophisticated navigation radar-based field retrieval in two ways: sea surface wind field retrieval algorithm based on optical flow motion and wind field retrieval algorithm based on gradient and BP neural network.With research of the two methods in years,the retrieval results are satisfactory.But both methods are shore radar-based or inshore radar-based images The radar position are fixed,and image datas are not pure,because the images include coastal shipping,the building in addition to noise disturbance.Therefore the data which can be really taken advantage of is the only part of the image.Moreover,coastal sea level is not only undergoing interaction of wind field but also effected by vulnerable hills,seabed topography,tidal.These factors is adverse are for wind retrieval research.And if wind field retrieval methods are wanted to show practical value,it should be allowed to monitor wind field in real time during the mobile platform sailing.Therefore,the paper carried out to study wind retrieval method on the dynamic environment using marine radar images from mobile platforms(ships,drilling platforms,etc.)Firstly,the paper describes the technology and which are needed for surface wind field retrieval,and then introduces the theory of wave generation and radar imaging mechanism to further improve radar image generation theory considering the speed of the ship which can provide simulated radar image data for the proposed wind fields retrieval methods on dynamic environment.The sea surface wind field retrieval methods based on shore or near the sea surface radar are briefly descriped: sea surface wind field retrieval algorithm based on optical flow motion and wind field retrieval algorithm based on gradient and BP neural network.By studying wind velocity retrieval algorithm based on BP neural network andcombined Hatten's thoery and geophysical model function(GMF),the paper builds relational model between wind speed and radar echo image values to retrieve wind based on the method of least squares theory.By studying the wind field retrieval algorithm based on optical flow motion to complete the radar image filtered,image moving average and images reduced to obtain images with a wind field stripes.Thenthe wind direction with 180 degree fuzzy is gained,and the final wind direction is determined by comparing the correlation of two successive sub-block images based on image matching technique.Finally,the proposed wind retrieval method is tested using simulation radar images and compared with traditional wind retrieval method,the results show that the proposed method is effective.