Research On Rainfall Detection Technology Inversion Method Based On Marine Radar Image

Ocean wave is the closest ocean phenomenon to human beings.Ocean wave monitoring is of great significance in ensuring navigation safety,preventing marine disasters and implementing military policies.Among them,rain interference is one of the common reasons that affect ocean wave monitoring.Rain itself is a natural precipitation phenomenon,which will reflect and refract the electromagnetic wave emitted by radar.At the same time,the energy of electromagnetic wave will be reduced after being absorbed by rain.In addition,rain will also change the roughness of the sea surface,which will cause the error of inversion of wave parameters.However,the detection of rainfall interference and the inversion of rainfall intensity on the navigation radar image play an important role in improving the accuracy of wave information inversion.In this paper,the original image data of X-band marine radar and the rainfall data measured by rain gauge are used to analyze the methods of recognizing rainfall interference and retrieving rainfall intensity.In this paper,the research of rainfall interference detection and rainfall intensity inversion of X-band marine radar image is divided into two steps: recognition of rain interference image in the original radar image and inversion of rainfall intensity of rainfall radar image.Assuming that the rainfall is uniform in space,this paper uses two methods to detect the rainfall interference of the original radar image.The first method is based on the difference of echo.A large number of experiments show that the mean value of echo difference between the radar image disturbed by rainfall and the radar image not disturbed by rainfall is significantly different.In this paper,the method of echo difference is improved.The distance of half main wave length is selected to calculate the mean value of echo difference in the wave monitoring area.Firstly,it is necessary to determine the threshold value to distinguish whether the radar image is affected by rainfall,and then compare the mean value of echo difference with the detection threshold value to determine whether the current radar image is affected by rainfall.The second method is to identify whether the radar image is disturbed by rainfall by convolution neural network.In this paper,the lenet-5 model of convolution neural network is improved to recognize the rainfall radar image.Firstly,the basic theory of neural network is introduced,and the composition of lenet-5 model is analyzed.By adjusting the parameters of the model and using a large number of data to train the model,the radar image can be accurately classified by inputting the Cartesian area of any radar image to determine whether it is disturbed by rainfall.After identifying the radar image disturbed by rainfall,the rainfall intensity level of the radar image can be further inverted.In this paper,a method of using the difference coefficient to retrieve the rainfall intensity is proposed.By studying the relationship between the difference coefficient and rainfall intensity in the radar image monitoring area,the fitting relationship between the two is determined by the least square method,and the error and accuracy of this method are analyzed.The feasibility of this method is proved,which can provide the rainfall intensity inversion of radar image Theoretical support.