Study On Water Depth Inversion In Shallow Water With Satellite Multi-spectral Image

Water depth is one of the important elements of marine exploration in coastal areas such as coastal zone and island,and it is also one of the main contents of marine surveying and mapping.Whether it is port,pier,waterway,anchorage and so on,or ship safety navigation,marine scientific research,marine engineering construction,Environmental assessment,etc.,need water depth information as a basic guarantee.Compared with traditional side-scan sonar and multi-beam measurement methods,water depth remote sensing measurement method has the advantages of fast,large-scale,quasi-synchronous and high spatial resolution measurements,which can make up for the measurement of the vessel can not be directly reached the coral reef intensive and controversial areas such as water depth data deficiencies,and it is also a new technical method for the depth of water to create a new path.In this paper,a class of water bodies in shallow water waters around Lingshui Bay in Hainan Island and a class of water bodies in shallow water waters around Dekui Island,using high resolution WorldView-2 multi-spectral image and four-band stereo image on the image,medium-resolution Landsat-8 multi-spectral satellite images and chart data.After the pretreatment of remote sensing images such as water and land boundary separation and tide correction of sea chart data,the linear ratio of multi-band and two-source two realities is carried out respectively,and Exponential,logarithmic,quadratic polynomial and exponential mathematical model of the nonlinear model of the fitting,based on these basic models,the traditional global depth remote sensing inversion method and the geographically adaptive inversion depth remote sensing inversion model are studied respectively,and the depth value in the range of 30 m was extracted.Research indicates:(1)The mid-resolution satellite imagery can extract the land-sea boundary line quickly and efficiently according to the normalized water body exponent algorithm,andthe edge of the land and water body is highlighted by the Laplacian operator 's edge enhancement.(2)In the correlation analysis,the correlation between the green band and the water depth control point is the highest,and the correlation is improved by the band ratio operation.It shows that the spectral information of water in two bands is combined to eliminate or weaken the influence caused by the difference of water body quality and sediment type.(3)The average absolute error and the standard error of the depth of the traditional global water depth remote sensing inversion method are very large,which indicates that there is some difficulty in expressing the real water depth information,whether it is a kind of water body or a second kind water body.(4)The geographically adaptive inversion depth remote sensing inversion method,the combination of the basic model parameters and the image source in different water depths are adaptive changes,and the depth accuracy of the inversion is greatly improved.(5)On the whole,the depth accuracy of the geographically adaptive inversion model is greatly improved,and the error is less than 1.2 m.It is pointed out that the problem of inhomogeneity of the sediment and the nature of the water in the traditional global model can be effectively mitigated by subdividing the study area,and the water depth extraction can be carried out by means of remote sensing method for shallow and controversial shallow waters will be a viable way.(6)The verification accuracy of the results of remote sensing inversion is in relation to the chart,because the chart also has a certain error range in the drawing process,indicating that the depth of the inversion will increase relative to the actual water depth error.However,there are several problems in the remote sensing inversion of water depth,which needs to be further improved or solved.(1)Although it is assumed that the remote sensing images obtained at different times do not change much in the same study area,the water quality and the type of the sediments do not change greatly,but the synchronization will still bring about the error of the water depth inversion.(2)In the process of data preprocessing,the atmospheric parameters,sea surface condition parameters and relevant parameters of the water body can not be obtainedwhen the image is imaged,resulting in some defaults or selection of approximate values in the FLAASH atmospheric correction,resulting in errors.(3)The remote sensing inversion model of water depth fitted in one region has yet to be developed in other areas of study.In summary,the improved geographically adaptive depth remote sensing inversion method proposed to solve the problem of water quality and sediment type difference in multi-spectral water depth remote sensing inversion,which greatly improved Depth of water depth inversion for multi-spectral water depth remote sensing technology to promote and provide support.However,there are some shortcomings in the deep water depth inversion,and the depth of inversion can not reach the measured depth.The optimization of the model needs to be further improved.