Key Technology Research Of Digital Estuary Platform Construction

This is a big era with big change and big challenge, Information Technology has remodeled our life, our thought, and has greatly influenced each domain of whole society, this is an era of information. More and more Information is demanded in the progress, especially in scientific research activity, multi-discipline integration research call for deep integration,process and share of multi-source Data, the traditional information process methods can't satisfies us now.Paper map and ocean survey data are important data sources nowadays, its datatype is various and complex, its quantity is also large, the data integration and process job is tedious and difficult only depend on manual process. In this paper, through the improvement of paper map automatic vectorization algorithm and ocean survey data process method, the data integration and process is quickened, large amounts of time is saved, this is very helpful to the share of data.The process method of ocean data have its own feature, and is very different from ordinary data process method, in this paper, we are not restricted of theory, method and algorithms that proved effectively, we put our focus on the ocean data's feature, through analysis of ocean data's essential characteristics and application of up to time computer technology achievement, new methods are recommended. Main contributions of the dissertation include:1 In the isoline automatic vectorization domain, a novel picture-text separation algorithm which use neural network to separate text from isoline is present, the process result is fine; and we present a new isoline nonlinear Hermite interpolation tracing method, the method use nonlinear Hermite interpolation and energy function to optimize the position of key points, it acquire a big tracing step, overcome the linear interpolation's shortcome, the information used to pin the key points is more than linear interpolation method, this is accord with the holistic automatic vectorization tendency of isolines.2 In the CTD data process domain, after analyse of CTD data's characteristics, we present a unique CTD wild data eliminatation method, the process efficiency of CTD data is greatly improved.3 Finally, in the marine sedimentology data process domain, we give a End Member unmixing program, in the study of the sediment from East China Sea shelf, we get many information hidden in the grain size data using this program.