Non-navigational TIN-DDM automatic generalization method considering terrain shape accurate determination and feature sufficient expression

Abstract

The application provides a non-navigation TIN-DDM automatic generalization method which considers terrain form accurate determination and feature sufficient expression, and belongs to the technical fields of seabed terrain information mining and analysis and TIN-DDM multi-scale expression. The application analyzes the continuous numerical change state of the terrain feature expression from microcosmic to macroscopic of the TIN-DDM sampling point, proposes a terrain feature quantitative evaluation index of the TIN-DDM sampling point, and combines the index with the TIN-DDM sampling point generalization sequence, so that the multi-scale expression of the seabed terrain form and feature is realized, and the problems that the current non-navigation TIN-DDM automatic generalization algorithm cannot fully consider the accuracy of the seabed terrain form expression and the sufficiency of the seabed terrain feature expression are solved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of seabed terrain information mining and analysis and TIN-DDM multi-scale expression, and particularly relates to a non-navigation TIN-DDM automatic generalization method considering accurate terrain shape determination and sufficient feature expression. BACKGROUND

[0002] A digital depth model (DDM) is a digital expression of the ups and downs of a seabed terrain surface by using limited and discrete depth points. According to different data organization modes of the depth points, the DDM is divided into a regular grid DDM (GRID-DDM) and an irregular triangle network DDM (TIN-DDM). Compared with the GRID-DDM, the TIN-DDM is not subjected to any data interpolation processing, and directly uses the measured depth as a model sampling point. Therefore, the TIN-DDM has relatively prominent advantages in reflecting the terrain shape change, and the analysis conclusion of the seabed terrain shape based on the TIN-DDM is relatively more accurate. With the development of seabed observation and detection technology, high-fidelity TIN-DDM construction, analysis and expression supported by high-resolution and massive depth data are increasingly valued in seabed engineering construction, marine geology research, underwater weapon equipment performance evaluation and the like.

[0003] According to different application scenarios and object requirements, the TIN-DDM can be divided into a navigation TIN-DDM and a non-navigation TIN-DDM. Different from the navigation TIN-DDM which emphasizes the safety of ship navigation at sea, the application scenarios of the non-navigation TIN-DDM are consistent with those of a digital elevation model (DEM) on land. The basic requirement of the application object is to emphasize the effective use of navigation resources on the basis of identifying the seabed terrain shape and maintaining the seabed terrain features, assist in decision analysis of seabed engineering construction, underwater weapon equipment performance evaluation and the like, and provide important technical support for the field of marine geology and the like. Therefore, the construction, analysis and expression of the non-navigation TIN-DDM need to focus on whether the seabed terrain shape determined by the neighborhood of the sampling point is accurate and whether the sampling point reflecting the seabed terrain features is sufficient, especially in the process of TIN-DDM generalization. With the compression of the number of TIN-DDM sampling points, the maintenance of the sufficiency of the TIN-DDM terrain features by using limited sampling points is a key link to realize the multi-scale expression of the high-fidelity TIN-DDM.

[0004] The traditional navigation TIN-DDM is mainly used for the safety of ship navigation, and the automatic generalization algorithm for non-navigation TIN-DDM is less studied. Due to the similarity between non-navigation TIN-DDM and land DEM, most of the existing automatic generalization algorithms for non-navigation TIN-DDM are obtained by borrowing and improving the multi-scale expression algorithms for land TIN-DEM, such as information quantity discrimination method, triangular facet method, vector angle method, point-to-face distance method, and three-dimensional Douglas-Peucker method. It is worth noting that although the characteristics of submarine topography and land topography have no difference in geometric characteristics, they are both composed of topographic feature points and lines, but unlike land TIN-DEM which contains complete topographic control information (usually composed of typical topographic feature points), submarine topography measurement cannot directly obtain submarine topographic feature information and its related submarine topographic range due to the particularity of its measurement means and methods. Therefore, the existing automatic generalization algorithm cannot fully explore the deep-level topographic feature information in non-navigation TIN-DDM, and the correlation between spatial scale and topographic form is not deep enough. Since 2003, Smith proposed the concept of rolling ball transformation for navigation surface (NS) service, and the rigorous geometric feature measurement characteristics of rolling ball transformation have been widely used in DDM construction and multi-scale expression. Dong Jian combined with the application requirements of non-navigation TIN-DDM, proposed a fast automatic generalization algorithm for TIN-DDM considering multi-scale expression of topographic form, which used the quantitative identification characteristics of rolling ball transformation to construct the correlation model of sampling point topographic type and rolling ball radius, and realized the rolling ball transformation for continuous scale expression of non-navigation TIN-DDM topographic form by analyzing the numerical correlation between rolling ball contact point and rolling ball radius.

[0005] The algorithm proposed by Dong Jian applies the correlation model of sampling point terrain type and rolling ball radius to the non-nautical TIN-DDM terrain form continuous multi-scale expression process, and to some extent solves the problems of unclear terrain form division boundary and spatial scale cognition difference in the non-nautical TIN-DDM comprehensive application of the land TIN-DEM multi-scale expression algorithm. However, the correlation model of sampling point terrain type and rolling ball radius in the algorithm proposed by Dong Jian only realizes the TIN-DDM sampling point terrain type judgment according to the critical rolling ball radius value, cannot quantitatively describe the regional boundary (range) of the sampling point terrain type, and is difficult to reflect the continuous form change rule of the sampling point terrain type from micro to macro, thereby causing the relative uncertainty of the TIN-DDM comprehensive conclusion. In addition, the algorithm proposed by Dong Jian proposes a non-nautical TIN-DDM comprehensive idea of deleting flat (smaller undulation) terrain sampling points and retaining complex (larger undulation) terrain sampling points, which to some extent realizes the form maintenance of the overall TIN-DDM terrain, but due to the lack of concept definition of the flat terrain sampling point and the quantitative evaluation of the characteristic value of the flat terrain sampling point in the complex terrain form, it is difficult to realize the accuracy of the seafloor terrain form identification and the sufficiency of the overall seafloor terrain feature maintenance. SUMMARY

[0006] In order to overcome the problems of inaccurate seafloor terrain form identification and insufficient overall seafloor terrain feature maintenance existing in the traditional non-nautical TIN-DDM automatic comprehensive algorithm, the present application provides a non-nautical TIN-DDM automatic comprehensive method considering terrain form accurate determination and feature sufficient expression.

[0007] The technical scheme adopted by the present application to achieve the above-mentioned purpose is:

[0008] The non-nautical TIN-DDM automatic comprehensive method considering terrain form accurate determination and feature sufficient expression comprises the following three parts:

[0009] The first part is TIN-DDM sampling point terrain type determination

[0010] The range constraint of the sampling point terrain feature is reasonably set, the terrain type of the sampling point in the local terrain feature range is accurately obtained, and the overall TIN-DDM is taken as the range reference of the sampling point terrain feature to clarify the change state of the sampling point terrain type from micro to macro, and to strengthen the accuracy of the sampling point terrain type determination of the algorithm. The specific process is as follows:

[0011] The first step is the micro terrain type determination of the TIN-DDM sampling point

[0012] The Delaunay influence domain refers to the area composed of the Delaunay triangle containing the discrete point P i , and its definition is equivalent to the area containing the discrete point Pi Voronoi cell V(S, P i ) adjacent to the Voronoi cell V(S, P

[0013]

[0014] where H(S, P i ) represents the Delaunay influence domain of the discrete point P i in the finite discrete point set S; P j represents any vertex of H(S, P i ); G(S, P i , P j ) represents any Voronoi edge of the discrete point P i . The importance of the Delaunay influence domain lies in that it defines the nearest discrete point set H(S, P i ) in terms of geometric properties to the discrete point P i , and the number and location of H(S, P i ) uniquely depend on the local distribution of the finite discrete point set S at the discrete point P i . The shadow around the discrete point P i is the Delaunay influence domain H(S, P i ) of the discrete point, as shown in Fig. Figure 1

[0015] Since the TIN-DDM sampling points are not affected by the sampling points outside the Delaunay influence domain, relative to the sampling point terrain type and rolling ball radius correlation model of formula (2), there will be no positive and negative nested terrains under the constraint of the Delaunay influence domain. Therefore, relative to formula (2), the correlation model of the micro-terrain type and rolling ball radius of the TIN-DDM sampling point based on the Delaunay influence domain will change accordingly, and the change is shown in formula (3):

[0016]

[0017] In formula (2), r' maxl represents the positive critical rolling ball radius of the sampling point P l ; r" maxl represents the negative critical rolling ball radius of the sampling point P l ; Q(P l ) represents the terrain type attribute of the sampling point P l , 1 for convex terrain, -1 for concave terrain, 0 for flat terrain, 2 for positive nested terrain, and -2 for negative nested terrain.

[0018] ​

[0019] rd' maxl , rd" maxl are the positive and negative critical rolling sphere radius under the constraint of Delaunay influence domain respectively; Q(p l ) represents the topographic type attribute of p l point, 1 for concave topography, -1 for convex topography, and 0 for flat topography.

[0020] Second step, macroscopic topographic type determination of TIN-DDM sampling points

[0021] The process of TIN-DDM synthesis is to consider the global range of terrain expression. In order to further strengthen the accuracy of the topographic type determination of TIN-DDM sampling points, according to the different contact degrees of rolling sphere with TIN-DDM sampling points during the rolling process on the seafloor topographic surface, a relationship model of TIN-DDM sampling point type and critical rolling sphere radius is constructed, and the TIN-DDM sampling points are classified into detail sampling points and skeleton sampling points from a macroscopic perspective, and the classification method is shown in formula (4). Wherein, R' maxl , R" maxl represent the positive and negative critical rolling sphere radius under the constraint of Delaunay influence domain respectively; Q(p l ) represents the topographic type attribute of p l point, 1 for detail sampling point, and 2 for skeleton sampling point.

[0022]

[0023] The values of critical rolling sphere radius R' maxl or R" maxl of some TIN-DDM sampling points are infinite, that is, during the process of continuously increasing the radius of the rolling sphere, the rolling sphere is always in contact with the terrain surface, so such points can be directly used as the skeleton of the supporting terrain, so such sampling points are called skeleton sampling points. The sampling points whose positive and negative critical rolling sphere radius are not infinite, when the radius of the rolling sphere increases to a certain extent, the rolling sphere is no longer in contact with the terrain surface, so it can be considered that such sampling points play a filling role for the terrain skeleton, and such sampling points are called detail sampling points.

[0024] Second part, calculation of TIN-DDM sampling point topographic feature quantitative evaluation index considering the expression of seafloor topographic form

[0025] The analysis of the influence degree of TIN-DDM detail sampling points on topographic features is shown in Figure 2The "·" represents the TIN-DDM sampling point; the solid line segment represents the seafloor surface; A, B, C, and D are three sampling points that are longitudinally continuous in similar topography; the change in topography from left to right of A, B, C, and D represents the process of the topographic characteristics of the TIN-DDM sampling point becoming increasingly stronger; the upper and lower dashed circles on the seafloor surface represent the positive and negative critical rolling balls of the TIN-DDM sampling point; ra maxl (rb maxl , rc maxl , rd maxl ), ra maxl (rb maxl , rc maxl , rd maxl ) represent the positive and negative critical rolling ball radii of sampling points A (B, C, and D), respectively.

[0026] As can be seen from Figure 2 , when the detail sampling point is a concave topography, the process of the topographic characteristics of the sampling point becoming increasingly stronger is accompanied by a decrease in the positive critical rolling ball radius and an increase in the negative critical rolling ball radius. When the detail sampling point is a convex topography, the process of the topographic characteristics of the sampling point becoming increasingly stronger is accompanied by an increase in the positive critical rolling ball radius and a decrease in the negative critical rolling ball radius.

[0027] The analysis of the degree of influence of the topographic characteristics of the skeleton sampling point is shown in Figure 3 . Among them, A and B represent concave and convex skeleton sampling points, respectively, and the process from left to right of each topography is a process of decreasing topographic characteristics. The weakening of the topographic characteristics of the concave skeleton sampling point is accompanied by an increase in the positive critical rolling ball radius; the weakening of the topographic characteristics of the convex skeleton sampling point is accompanied by an increase in the negative critical rolling ball radius.

[0028] The strength of the topographic characteristics of the detail sampling point can be represented by the ratio of r' maxl , r maxl . The smaller the value of r maxl / r' maxl of the concave detail sampling point, the weaker the topographic characteristics expressed by the sampling point, and vice versa. The smaller the value of r' maxl / r maxl of the convex detail sampling point, the weaker the topographic characteristics expressed by the sampling point, and vice versa. The strength of the topographic characteristics of the skeleton sampling point can be represented by the ratio of r' maxl , r maxlThe greater the positive critical rolling sphere radius of the concave skeleton sampling point, the weaker the feature, and vice versa. The greater the negative critical rolling sphere radius of the convex skeleton sampling point, the weaker the feature, and vice versa. The process of the sampling point terrain feature from weak to strong can represent the continuous state change trend of the sampling point terrain feature from microcosmic to macroscopic. The greater the ratio, the more the sampling point terrain feature can represent the macroscopic terrain.

[0029] In order to reflect the continuous state change of the sampling point terrain feature from microcosmic to macroscopic by the ratio of r' maxl , r" maxl , and consider the terrain feature of the skeleton sampling point and the detail sampling point in the numerical value, according to the above analysis, a terrain feature evaluation index is proposed, that is:

[0030]

[0031] In formula (5), C is any positive integer. According to the evaluation index, the index value of the terrain feature of the detail sampling point is between -∞ and 1, the index value of the terrain feature of the skeleton sampling point is between 1 and +∞, and the greater the index value of the terrain feature, the more macroscopic the terrain feature represented by the TIN-DDM sampling point.

[0032] Third part, TIN-DDM sampling point comprehensive sorting

[0033] In order to maximize the preservation of the spatial features of the TIN-DDM terrain form, while considering the requirements of the quantity of the TIN-DDM sampling point, in the non-navigation TIN-DDM automatic generalization considering the accurate determination of the terrain form and the sufficient expression of the features, the TIN-DDM sampling point expressing the microcosmic terrain feature should be deleted first, and the TIN-DDM sampling point expressing the macroscopic terrain feature should be retained first, so as to achieve the requirement of maximizing the preservation of the submarine terrain. Therefore, the sampling point with a relatively microcosmic terrain feature, that is, the sampling point with a low terrain feature evaluation index value, is regarded as the sampling point that should be deleted most in the automatic generalization process of the TIN-DDM. Therefore, the order of the terrain feature evaluation index value of the TIN-DDM sampling point is consistent with the process of the terrain generalization scale from small to large. The overall flow chart is as follows: Figure 4

[0034] ​The present application has the beneficial effects that the present application proposes a method for accurately determining the terrain type of TIN-DDM sampling points and analyzing the expression degree of terrain features of TIN-DDM sampling points, proposes a quantitative evaluation index of terrain features of TIN-DDM sampling points by analyzing the continuous numerical change state of TIN-DDM sampling points from microcosmic to macroscopic in expressing terrain features, and realizes the multi-scale expression of emphasizing the seabed terrain form and features by combining the index with the comprehensive sequence of TIN-DDM sampling points, thereby solving the problems that the current non-navigational TIN-DDM automatic generalization algorithm cannot fully consider the accuracy of seabed terrain form expression and the sufficiency of seabed terrain feature expression. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1 Delaunay influence domain of discrete point Pi.

[0036] Figure 2 Critical rolling ball radius influence degree analysis on terrain feature of detail sampling points.

[0037] Figure 3 Critical rolling ball radius influence degree analysis on terrain feature of skeleton sampling points.

[0038] Figure 4 Non-navigational TIN-DDM automatic generalization method flowchart considering accurate terrain form determination and sufficient feature expression.

[0039] Figure 5 Critical rolling ball radius solving process schematic diagram. DETAILED DESCRIPTION

[0040] In order to make the problems solved by the present application model, the method scheme adopted and the effects achieved more clear, the present application will be further described in detail below in combination with the drawings and experiments. It can be understood that the specific experiments described herein are only used to explain the present application, but not limit the present application. In addition, it should be noted that, in order to facilitate the description, only the parts related to the present application are shown in the drawings, but not all the contents.

[0041] As shown in Figure 4 , a non-navigational TIN-DDM automatic generalization method considering accurate terrain form determination and sufficient feature expression includes the following steps:

[0042] Step 1): sequentially selecting sampling points P l in the TIN_DDM model, the total number of sampling points is n, and l is the position number of the selected sampling points, l = 1, 2, 3... n.

[0043] Step 2): sequentially calculating the sampling points P lthe positive critical rolling ball radius r' of the rolling ball separating from the positive surface of the sample point P maxl , and the negative critical rolling ball radius r" of the rolling ball separating from the negative surface of the sample point P maxl ; when all the sample points are calculated (l = n), end.

[0044] The solving process of the critical rolling ball radius is shown in Fig. 1; Figure 5 The solving process of the positive critical rolling ball radius is shown in Fig. 1; Figure 5 The solving process of the positive critical rolling ball radius is shown in Fig. 1; the solving principles of the negative critical rolling ball radius, the critical rolling ball radius under the constraint of the Delaunay influence domain, and the critical rolling ball radius under the global range are consistent with the solving principle of the positive critical rolling ball radius.

[0045] Figure 5 In Fig. 1, firstly, the numerical correlation between the rolling ball contact point corresponding to the upper buffer surface point on the TIN-DDM and the rolling ball radius is stored in the data chain, and the data chain is P1 to P l The upper buffer surface point in the z-axis direction of the sample point is taken as the research object, and j1 to j l indicates the number of rolling ball contact points corresponding to each upper buffer surface point in the rolling ball radius change process, indicates the rolling ball radius range boundary value of each rolling ball contact point, indicates the rolling ball contact point corresponding to each radius range; subsequently, in the area of the dashed line frame 1, the rolling ball radius range (dashed line frame 3 area) when the sample point P1 (dashed line frame 2 area) becomes the rolling ball contact point is queried and set operation is performed, and finally the rolling ball radius range when the sample point P1 becomes the rolling ball contact point is obtained. Due to the discreteness and data error of each sample point in the TIN-DEM, the rolling ball radius range corresponding to the rolling ball contact point is discontinuous, which is inconsistent with the actual situation. The maximum boundary value can be selected as the critical rolling ball radius when the rolling ball contact point state changes, and finally the positive critical rolling ball radius is obtained.

[0046] Step 3): according to the Delaunay influence domain, the TIN-DDM sample point P l is micro-topographic regionally divided, and the rolling ball radius r l of the rolling ball separating from the positive surface of the sample point P maxl and the negative critical rolling ball radius r" of the rolling ball separating from the negative surface of the sample point P maxl are calculated in turn under the limitation of the micro-topographic region range; when all the sample points are calculated (l = n), end.

[0047] Step 4): According to the division of TIN-DDM sampling point micro-topography by Delaunay influence domain, the correlation model of TIN-DDM sampling point micro-topography type and rolling ball radius is constructed, as shown in formula (3); and the rd' maxl and rd” maxl of step 2) are used to classify the micro-topography type attribute Q W (P l ) of all sampling points.

[0048]

[0049] In formula (3), Q W (P l ) represents the micro-topography type attribute of P l point, 1 for concave topography, -1 for convex topography, and 0 for flat topography.

[0050] Step 5): According to the variation law of critical rolling ball radius value, the correlation model of TIN-DDM sampling point macro-topography type and rolling ball radius is constructed, as shown in formula (4); wherein, R' maxl , R” maxl respectively represent the positive and negative critical rolling ball radius in the global range, which are solved by referring to the solving method of critical rolling ball radius in step 2), and all sampling points are required to participate in the calculation in the solving process.

[0051]

[0052] In formula (4), Q H (P l ) represents the macro-topography type attribute of P l point, 1 for detail sampling point and 2 for skeleton sampling point.

[0053] Step 6): Joint the r' maxl and r” maxl solved in step 2) and the classification results of TIN-DDM sampling point topography type in steps 4) and 5), analyze the correlation of topographic feature expression ability of the sampling points, construct the sampling point topographic feature evaluation index model for seafloor topographic form recognition, as shown in formula (5); and obtain the topographic feature expression index Ω(P l ) of the sampling points.

[0054]

[0055] In formula (5), C is any positive integer.

[0056] Step 7): The sampling point topographic feature expression index set The sequential ordering is performed, the sampling points are quantitatively thinned according to the comprehensive requirements of the user to the TIN-DDM model, and the triangular net is reconstructed according to the residual sampling points to complete the automatic generalization of the TIN-DDM.

[0057] Finally, it should be noted that the above description is only to illustrate the method of the present application, and is not limited thereto; although the present application has been described in detail, those skilled in the art should understand that the modification of the above-mentioned method scheme, or the equivalent replacement of part or all of the method features, does not make the corresponding method scheme deviate from the scope of the method scheme of the present application.

Claims

1. A non-navigational TIN-DDM automatic generalization method that takes into account the accurate determination of terrain morphology and the full expression of features, characterized in that, The method comprises the following steps: Step 1): Select sampling points P in TIN_DDM model in turn l , the total number of sampling points is n, l is the number of selected sampling points, l = 1, 2, 3... n; Step 2) : Calculate the positive critical rolling ball radius r' maxl and the negative critical rolling ball radius r" maxl of the rolling ball of the oversampling point P l in the process of its rolling ball separating from its positive surface and negative surface respectively, and end when all the sampling points are calculated.​​ Step 3) : sampling point P of TIN-DDM according to Delaunay influence domain l Micro-topographic region division is performed, and the positive critical rolling ball radius rd' maxl and the negative critical rolling ball radius rd" maxl of the rolling ball and its positive surface separation process and the rolling ball and its negative surface separation process of oversampling point P l in the range of micro-topographic region are calculated in turn under the limitation of the range of micro-topographic region; when all sampling points are calculated, the process is ended. Step 4): According to the division of TIN-DDM sampling point micro-topography by Delaunay influence domain, the correlation model of TIN-DDM sampling point micro-topography type and rolling ball radius is constructed, as shown in formula (1); and the rd maxl and rd” maxl of all sampling points are classified according to the micro-topography type attribute Q W (P l ). In formula (1), Q W (P l ) represents the micro-topographic type attribute of the P l point, 1 for concave topography, -1 for convex topography, and 0 for flat topography; Step 5): According to the critical rolling ball radius value change rule, the correlation model of TIN-DDM sampling point macro-topography type and rolling ball radius is constructed, as shown in formula (2); wherein, R' maxl , R" maxl respectively represent the positive and negative critical rolling ball radius in the global range; In equation (2), Q H (P l ) represents P l The macroscopic terrain type attribute of the point, 1 is the detailed sampling point, 2 is the skeleton sampling point; Step 6): The r' maxl and r" maxl and the classification results of the terrain type attribute of the TIN-DDM sampling points in steps 4) and 5), the correlation analysis of the terrain feature expression ability of the sampling points is performed to construct a sampling point terrain feature evaluation index model for the seafloor terrain morphology identification, as shown in formula (3); and the terrain feature expression index Ω(P l ) of the sampling point is obtained. In formula (3), C is any positive integer; Step 7): the expression index set of the terrain feature of the sampling point sought in step 6) The sequential ordering is performed according to the comprehensive requirements of the user to the TIN-DDM model, the sampling point quantitative thinning is sequentially performed, the TIN-DDM is reconstructed according to the residual sampling points, and finally the automatic generalization of the TIN-DDM is completed.

Images