Research On SAR Imaging Characteristics And Recognition Method On Internal Wave

Internal wave is a kind of wave phenomenon in stratified seawater.Its frequency is between inertial frequency and buoyancy frequency.The study of internal wave is of great significance to the theoretical research of marine science,the protection,development and utilization of marine resources,and marine military.Synthetic Aperture Radar(SAR)has the advantages of all-day,all-weather,high-resolution.It is an indirect detection method and improves the limitations of direct detection methods.SAR remote sensing images contain very rich ocean information.Any ocean phenomenon or feature that changes the roughness of the sea surface can be imaged on SAR remote sensing images,which makes SAR an indispensable means of studying internal waves.SAR remote sensing image can obtain information such as the spatial position,wavelength,and vertical parameters of internal waves.The interferometric phase can also obtain the information of the sea surface current.This paper starts with the generation mechanism of internal waves,numerically simulates the internal wave propagation model and the source-induced internal wave model.According to the principle of internal wave imaging in SAR,the In SAR images are simulated,and the effects of different radar parameters and internal wave model parameters on the imaging characteristics of internal waves are studied,and the internal wave of SAR images is extracted and identified.The main work of this paper includes the following aspects:First,the numerical simulation of internal waves is studied.Based on the propagation model of one-dimensional and two-dimensional internal solitary waves and the internal wave model formed by the disturbance of underwater moving objects,the wave heights and the distribution of the current field on the surface of internal waves are obtained by numerical simulation.According to the principle of linear superposition,the random wave model is solved,three-dimensional random waves are simulated,and then the three-dimensional simulation of the internal wave including the wave information is obtained.Secondly,according to the sea surface current field information obtained by simulation,the interference internal wave image under different parameters is simulated according to the Along-Track In SAR internal wave simulation process.In order to find the optimal parameter combination that observe the current field,the interferometric phase variation is used as an index to analyze the imaging characteristics of the parametric SAR imaging.At the same time,the variation trend of intensity information and phase information extreme points under different internal wave model parameters is observed.The sensitivity of SAR internal wave imaging features to different internal wave parameters is ana lyzed.It can provide an effective reference for selection of internal wave parameters in the iterative inversion process,and reduce the number of iterations.Finally,this paper realize the feature extraction and recognition of SAR internal wave images.The internal waves appear as bright and white bands in SAR images.Gaussian mixture model and MRF combined algorithm use SAR image statistics and pixel space constraint information to realize SAR image segmentation and enhancement of internal wave features.2D Empirical Mode Decomposition algorithm is used to extract internal wave information from ocean background information.The internal wave section of the SAR image is intercepted,and CEEMD algorithm is performed to extract the modality in which the internal wave is located and realizes the detection of internal waves.According to the distance between the dark and bright strips,the parameters such as half amplitude width and wavelength can be extracted and identified.Machine learning method is used to complete the automatic identification of internal waves.