Research Of The Rainfall Intensity Inversion Method By The Marine Radar Image

The monitoring of the marine environment is of great significance for maritime navigation,resource management,marine operations and national defense security.The inversion of the rainfall intensity from the marine radar images provides a convenient means for the monitoring of the rainfall on the sea,and provides a evaluation index for the credibility of radar image inversion wave parameter information.Relying on the native defense scientific research project of "the ocean wave inversion technology by shipborne X-band radar",this paper studies the method of retrieving the rainfall intensity from the navigation radar image by using a large number of X-band radar image data and rainfall data.The method of retrieving rainfall intensity from the radar image in this paper can divided into three steps: filtering the co-frequency interference noise of the original radar image,the identification of the rainfall radar image,and retrieving the rainfall intensity based on the radar image characteristic parameters,which are researched respectively.In this paper,the correlation of radar echo signals is studied by the measured radar data.It is found that there is strong correlation between adjacent sea clutter signals,and the co-channel interference noise destroies this correlation.Based on this spatial correlation between wave signals,combined with Laplace linear template,this paper proposes a spatial domain correlation method to suppress co-channel interference noise,the method is divided into noise line detection,noise point location,noise repair.First,the correlation coefficient of a large number of measured data is calculated to determine the appropriate two-stage thresholds,and the co-channel interference noise line is detected by the threshold relation;Then,an improved Laplace linear template is used to locate the noise point on the detected noise line,repair the noise point by using the weighted interpolation method based on the correlation coefficient to filter noise;Finally,the algorithm is tested by the use of real radar data,comparing with the traditional median and mean filter,which has better filtering effect on co-channel interference and can preserve the original information more completely.The prerequisite for inversion of rainfall intensity using the radar image is that the rain image has been identified.In this paper,the characteristic parameters of the radar rainfall image,the percentage of zero-intensity and the mean of echo intensity are studied by using the real radar data,and a new method is proposed,which utilizes the specific value of characteristic parameters of the rainfall radar image to select the rainfall image.The method has a higher recognition accuracy than using the zero intensity percentage or the conventionalmethod which uses the echo intensity mean and the difference coefficient.It is necessary to find the relation between the characteristic parameters of rainfall radar images and the intensity of rainfall.In this paper,a large number of real radar data and rainfall intensity data are selected for experiments to study the relationship between the percentage of zero intensity and rainfall intensity,the relationship between the mean of echo intensity and the intensity of rainfall in the shadow area of radar image.The linear relation is obtained by linear fitting,which provided the basis for the inversion of rainfall intensity.Finally,the formula of relation between the percentage of zero intensity in the shadow region and the intensity of rainfall,the formula of relationship between the mean of echo intensity in the shadow region of radar image and the rainfall intensity are respectively analyzed to define the method of solving the rainfall intensity.According to the steps of removing the co-channel interference noise,recognizing rainfall images and determining the intensity of rainfall,a specific method for inversion of rainfall intensity from radar images is designed.the real data are used to validate the method,the results show that the method has certain feasibility.