Terrain Feature-Oriented Digital Chart Sounding Generalization

Automatic generalization and multi-scale representation of digital chart are the main content of chart production integration research.As one of the key technology of automatic generalization and multi-scale representation of digital nautical chart,generalization of sounding is also the difficulty and emphasis of chart production integration,intellectualization and automatization.Generalization of sounding solves the contradictions between complicated submarine topography and the soundings after rarefying and selecting.And it realizes not only the conciseness and clarity of charts,but also the richness of expression content.Soundings generalization is a complicated process of multiple factors associated essentially,and it is difficult to describe the relationship of these factors which influence the selection of Soundings.This research aims to explore the representation and identification of submarine topography from the terrain feature.Focusing on building a comprehensive model which takes topographic feature as the main constraint condition in order to realize the generalization of sounding.The main contents and innovations of this thesis are as the followings:1)This research summarizes and analyzes the research status of geographic and sounding generalization in digital charts.The source,function and classification of sounding and the characteristics and principles of sounding generalization are investigated.It also analyzes the existing models and methods of point feature generalization.A generalization model which supports sounding generalization of digital charts is presented.2)Fast Algorithm of Contour Generation Considering Ambiguity.The modified partition rules were taken to optimize Delaunay TIN,considering the ambiguity problem caused by partition in the process of generating depth contours with navigation security.A fast algorithm of depth contours generation was proposed based on sorting the endpoints of segments in triangulation network.As the contour segment in each triangle which has isocenters was generated,the depth contours were generated by sorting the random endpoints of contour segments progressively.The experiment results show that the new algorithm can efficiently generate safety depth contours which can present seafloor terrain correctly from a large amounts of bathymetric data.3)The extended sounding tree construction.A new tree model is designed and a method of building the extended sounding tree is proposed.Firstly,soundings and depth contours are used as constraint points and edges to build a constrained Delaunay triangulation,which can divide soundings into different sounding group according to depth and find the adjacency relation between soundings and depth contours,by making use of its advantages in domain analysis.Then,building the external sounding tree by identifying the parent-child or brotherhood relation between adjacent sounding groups according to depth.The experimental results show the extended sounding tree model is logical and the tree-building algorithm is feasible.4)Sounding generalization oriented to terrain feature.From the perspective of the generalization of submarine topography,a gradual generalization model which generalizes soundings from high class to low class based on the degree of sounding classification is proposed.Submarine topography is identified and the terrain features of sounding are extracted based on the sounding tree model.Characteristic parameters of sounding are defined,and according to the mean distance,slope difference and near depth difference of sounding,different models of sounding selection are constructed.Moreover,rules and procedure of selecting important soundings,controlled soundings and background soundings are put forward respectively so as to realize sounding generalization.The experimental results show the feasibility of the method of sounding generalization from the prospective of terrain generalization.