Research On Error Processing Methods In Multi-Beam Sounding Swath Joins

The quality control of sounding data in swath joins is an important part of the multi-beam data processing,which is of great significance to improve the quality accuracy of bathymetric data and to obtain the true submarine topography of the whole area.In this paper,a systematic and comprehensive analysis is made on the method of weakening the error of sounding data in the multi-beam swath joins.The main content of this paper including: the theory derivation and the distribution characteristics analysis of sounding points,comparison of various spatial construction methods of seabed terrain,different levels of sounding error classification and corresponding processing methods,research on abnormal value of automatic filtering sounding-robust estimation of multi-quadric method based on median parameters,research on systemic residual error weakening method based on terrain spectrum analysis.Some works and contributions are generalized as follows:? This paper makes a detailed analysis of the calculation derivation process of submarine beam footprint,so as to obtain accurate sounding point coordinates.Then,the geodetic coordinates of the sounding points are converted into spatial rectangular coordinates,which is beneficial to avoid the problem of large distortion when the data is processed with latitude and longitude as coordinates.In addition,The distribution characteristics of equiangular and isometric beams are analyzed in different signal transmission modes.? According to the spatial coordinates of the sounding points,various spatial interpolation methods are used to construct the seabed topography.Finally,the optimal interpolation method is estimated by the time loss,the visual effect of DEM and the evaluation values of residual error.The results show that the DEM of the submarine terrain should select its own optimal method synthetically by analyzing the complexity of the seabed topography and the distribution characteristics of the bathymetric points in the practical application.? The effect of multi-beam sounding error is classified from different levels.And according to each type of error is whether act on the depth value in the measurement process,the error influence of sound velocity profile measurement and attitude measurement on swath joins was analyzed.In addition,This paper focuses on the method of the abnormal value and the systematic error in the classification of errors,which provides the direction for the subsequent processing methods.? The polynomial surface fitting method to filter abnormal is not complete and is limited to close to the real submarine terrain.And the residual error obtained by the least squares is taken as the initial value of the resistance,which seriously disturbs the robustness.Therefore,a new method of multi-quadric robust estimation based on median parameters was established to filter the depth of sounding automatically.This method has been improved on the selection of known nodes of the multi-function method,which is combined the joint criterion of angle and chord with the point dispersion to extract the feature points of the region.Moreover,the integrated application based on the multivariate parameter of the median parameter has the advantage of strong robustness.The experiments show that this method can extract the feature points of the region with maximum efficiency,and there is no unusual highlight of the ups and downs,and ensures the continuous changes in the characteristics of submarine terrain from the effects of DEM.? Although some systematic measurement errors have been corrected in the data processing,the measurement error and the induced residual error of multi-beam bathymetric system still exist,which seriously affects the accuracy of the bathymetric and results in the phenomenon of crying or smiling faces in the swath joins area.Therefore,a method based on terrain spectrum analysis is proposed to weaken the influence of systematic residual error and compare it with DFT method.The experimental results show that using multiple partitions with segmentation processing and by properly decreasing the interval of known nodes will reasonably reduce the fitting precision to achieve the purpose of rapidly processing the systematic residual error.