| Tidal flat surface sediment types and grain size parameters are important indicators to characterize tidal flat changes.Interpretation of tidal flat sediments types can understand the spatial distribution of different types of tidal flat sediments;The spatial distribution of sediment grain size parameters is of great significance to study the environment of tidal flat sediment,changes in erosion and deposition,and hydrodynamics.The muddy tidal flat is wide and flat,not only affected by changes in the tidal water level,and the observation time is limited;and the sloping surface is muddy and difficult to reach,so the traditionally the measured tidal flat sediment research is time-consuming and laborious,and it is difficult to obtain the spatial distribution of tidal sediment information.Hyperspectral remote sensing technology has the advantages of abundant bands,large amount of information,and free from natural conditions.Compared with traditional methods,it has unique advantages in studying tidal flat sediments.However,from the current research,satellite hyperspectral remote sensing for tidal flat sediments research is limited by the spatial resolution of the image and the tide level at the imaging time,and it is difficult to obtain the ideal hyperspectral remote sensing data of the exposed tidal flats,so airborne remote sensing is an important technical application direction.In this paper,the tidal flat on the muddy coast in Rudong country,Jiangsu Province is selected as the study area.The natural tidal ditch in the middle of the study area is used to divide the study into two parts,the northern part mainly composed of finegrained sediment and the southern part mainly composed of coarse-grained sediment.Using airborne hyperspectral imagery and field-collected sediment sample data as data sources,based on a detailed study of airborne hyperspectral remote sensing data processing methods,the characteristic band of the sediment were determined based on the field measured spectral curve,by which the MNF transform of the airborne hyperspectral remote sensing data was compared,and the remote sensing inversion factors of the sediment parameters were determined.Using the support vector machine model,a remote sensing inversion model of grain size parameters(average grain size,sorting coefficient,and skewness)and sediment component content(sand,silt,and clay)was established.The remote sensing research of sediment grain size parameters and types was carried out to obtain the ideal space for the type and grain size parameters of the tidal flat surface sediments.The distribution results and the characteristics of the spatial distribution of sediments and their correlation with the hydrodynamic environment are analyzed.This study provides a demonstration for exploring the use of airborne hyperspectral remote sensing technology to obtain large-scale tidal flat sediment size information and sediment type distribution.It also provides important data and technical support for the development of coastal zone resources in China's coastal areas.The main research conclusions of the paper are as follows:(1)SVM support vector machine model can effectively achieve remote sensing inversion of surface sediment component content and particle size parameters of silty tidal flats.In the sediment composition model,the average relative error of the sand content was the smallest at 10.48%;the clay was the largest at 30.88%;the root-meansquare errors of the sand,silt,and clay contents were 0.37%,0.37%,and 0.51%,respectively.The model based on the component content has an accuracy of 85.15% for inversion of sediment types,and the inversion accuracy is better.In the sediment average grain size and skewness models,the average absolute errors of the model verification group were 0.31? and 0.11,the average relative errors were 7.41% and 7.04%,and the root mean square errors were 0.46? and 0.32,respectivelythe average grain size of the sediment and the inversion effect of the skewness model are ideal.(2)This paper proposes a new method to indirectly invert the sorting coefficients.Firstly,the SVM model is used to retrieve the ? value grain size at 25%(D25)and the ? value grain size at and 75%(D75)of the cumulative content on the grain size accumulation curve.According to the strong correlation between the ratio D25 / D75 and sorting coefficients,the value of the sorting coefficient is obtained.In the sediment sorting coefficient model,the average absolute errors of D25,D75,and sorting coefficients of the validation group of the model are 0.36?,0.21?,and 0.06,respectively;the average relative error of the sorting coefficient is the smallest,which is 3.58%;D25 is the largest,It was 7.32%.At the same time,the root mean square errors of the verification group were 0.27?,0.28?,and 0.23,respectively.It is seen from the comprehensive accuracy evaluation index and sorting coefficient results that the accuracy of this method for inversion of sorting coefficients is high.(3)Using the remote sensing model of sediment type based on component content method established by SVM model,combined with kriging interpolation method,the sediment type spatial distribution in the study area was obtained by inversion.Sediment types in the study area are mainly silty sand and sand,followed by sandy silt,and silt is the least.The distribution of sediment types in the study area has obvious spatial differences,the overall manifestation is that the sediment grains in the northern area are finer and the sediment grains become coarser from land to sea.The sediment types in the land adjacent area are mainly sandy silt,followed by silty sand,and the type of sediments in the middle and lower parts of intertidal zone is sand;the sediments in the south are thicker than the north,the types of sediments are mainly silty sand and sand,and a small amount of sandy silt and silt are mainly distributed in the area where the Yellow Sea Bridge roots and the tidal creek extend to the tidal flat,which is related to the reclamation of Yangkou Port in the south of the area and consistent with the hydrodynamic environment and the source of sediment material in the study area.(4)From the remote sensing inversion results of sedimentary grain size parameters,the sediments in the north area is finer than those in the south area.,the sorting coefficient is mainly sorting and poor,and the skewness is mainly near-symmetric and positive.The grains in the extending part from the tidal creek to the beach surface gradually become thinner,and the sorting in the tidal creek area is better than that in other parts;the grains in the south are coarser than those in the north,the sorting is poor,and skewness tends to be extremely positive.At the root of the Yellow Sea Bridge,the grains are relatively fine.From the root of the Yellow Sea Bridge to the beach,the grains gradually become thicker,and the skewness gradually changes from nearly symmetrical to positive.(5)Based on airborne hyperspectral remote sensing technology,the inversion of tidal flat sediment size parameters and sediment types is not limited by natural conditions and ranges.,and it can be applied to muddy tidal flats with complex sedimentary environments and artificial sampling difficulties.,and it provides a fast and effective method for study of tidal flat sedimentary characteristic and analysis of erosion and deposition. |
PDF Full Text Request