CN111784840B - LOD (line-of-sight) level three-dimensional data singulation method and system based on vector data automatic segmentation - Google Patents
LOD (line-of-sight) level three-dimensional data singulation method and system based on vector data automatic segmentation Download PDFInfo
- Publication number
- CN111784840B CN111784840B CN202010380448.6A CN202010380448A CN111784840B CN 111784840 B CN111784840 B CN 111784840B CN 202010380448 A CN202010380448 A CN 202010380448A CN 111784840 B CN111784840 B CN 111784840B
- Authority
- CN
- China
- Prior art keywords
- data
- boundaryobject
- boundary
- model
- segmentation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000011218 segmentation Effects 0.000 title claims abstract description 65
- 238000000034 method Methods 0.000 title claims abstract description 49
- 239000000178 monomer Substances 0.000 claims abstract description 24
- 230000008520 organization Effects 0.000 claims abstract description 11
- 238000012545 processing Methods 0.000 claims abstract description 6
- 238000000605 extraction Methods 0.000 claims description 61
- ATJFFYVFTNAWJD-UHFFFAOYSA-N Tin Chemical compound [Sn] ATJFFYVFTNAWJD-UHFFFAOYSA-N 0.000 claims description 8
- 238000007781 pre-processing Methods 0.000 claims description 8
- 238000006243 chemical reaction Methods 0.000 claims description 7
- 238000001514 detection method Methods 0.000 claims description 7
- 238000000638 solvent extraction Methods 0.000 claims description 7
- 238000013075 data extraction Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 10
- 230000000694 effects Effects 0.000 description 5
- 238000010276 construction Methods 0.000 description 3
- 238000009877 rendering Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013523 data management Methods 0.000 description 1
- 238000013499 data model Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000013439 planning Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/10—Constructive solid geometry [CSG] using solid primitives, e.g. cylinders, cubes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/20—Finite element generation, e.g. wire-frame surface description, tesselation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T19/00—Manipulating 3D models or images for computer graphics
- G06T19/20—Editing of 3D images, e.g. changing shapes or colours, aligning objects or positioning parts
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Architecture (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Processing Or Creating Images (AREA)
Abstract
The invention discloses a vector data-based automatic segmentation LOD level three-dimensional data singulation method, which belongs to the technical field of image data processing, in particular to oblique photogrammetry three-dimensional model processing, and comprises the following steps: s1, unifying a coordinate system of vector data and three-dimensional data of the inclined live-action; s2, loading vector data and acquiring attribute information of a building boundary; s3, loading and collecting real-scene three-dimensional data information; s4, extracting and dividing the monomer model; s5, respectively outputting the single building model extracted by segmentation and the model of the segmentation residual part according to the original data organization structure of the live-action three-dimensional data. According to the invention, boundary space information of a building boundary contour is obtained through the existing vector data, multi-level segmentation is carried out on the inclined live-action three-dimensional data of the data structure with the LOD level, and meanwhile, under the condition that the original data structure is not changed, the single building model can be automatically extracted in batches, and the problem of saw teeth of the segmentation edge of the single building model can be effectively solved.
Description
Technical Field
The invention relates to the technical field of image data processing, in particular to a three-dimensional model processing method based on oblique photogrammetry, and particularly relates to a three-dimensional data singulation method based on vector data automatic segmentation LOD (line-of-sight) level.
Background
The oblique photogrammetry is a high-new measuring and drawing technology developed in recent years, can acquire texture information of different angles of a measurement target to construct a large-scale, high-resolution and high-precision three-dimensional model, and provides services for various industries such as urban planning management, water conservancy and hydropower. The construction process of the live-action three-dimensional model can be summarized as three main steps of generating point cloud for images, constructing a triangular net and texture mapping, so that constructed model data are continuous irregular triangular nets, all ground objects are continuous whole objects, ground object targets are not separated, attribute information cannot be connected with the model in a hanging mode, difficulty exists in applying the data, and the model data are required to be subjected to 'monomerization' processing in order to solve the problem.
Three-dimensional model "singulation" is the partitioning of a continuous overall model into individual entities, the partitioned monomers can be selected, managed individually, attached to attributes, query statistics, and the like. In the aspect of the individualized research of a live-action three-dimensional model, the invention with the application number of 201410327662.X discloses a method for classifying and extracting monomers from buildings and vegetation by utilizing building height and image spectrum information. In the "building monomerization extraction technology research based on inclined image dense matching point clouds" of the university of the Jiujun information engineering university (Jiujun information engineering university, chen Yu), a filtering algorithm is utilized to classify and process the dense point cloud point clouds formed by image matching to extract a monomial model, the point cloud filtering algorithm generally has good classifying effect on an object, and at present, no algorithm can accurately classify all ground object targets, so that the method has a certain limitation in extracting the monomial model. In addition, in the published academic papers and patents, the single model cannot be accurately extracted in a large scale, in batches and automatically in the single-model real-scene three-dimensional model, so that the algorithm adaptability and the extraction effect are comprehensively considered in the aspect of realizing the single-model three-dimensional model.
Disclosure of Invention
The invention aims to provide a method for extracting a single building model by automatically dividing multi-layer LOD-level inclined live-action three-dimensional data based on building contour vector data in order to realize automatic single extraction of the inclined live-action three-dimensional model data. The method can automatically extract building monomers in batches, quickly segment and reconstruct LOD data of model data, and simultaneously can keep the original data organization structure unchanged and solve the problem of saw teeth of the segmentation edges of the monomer models.
In order to achieve the above purpose, the technical scheme adopted by the invention is as follows:
the LOD level three-dimensional data singulation method based on vector data is automatically segmented and comprises the following steps:
s1, unifying a coordinate system of vector data and three-dimensional data of the inclined live-action;
S2, loading vector data and acquiring attribute information of a building boundary: loading vector data, constructing a boundary object BoundaryObject for each building boundary in the data, and acquiring and storing coordinate string information boundary (P1, P2, P3 …) of the boundary object and coordinates of a circumscribed circle center C of a minimum bounding rectangle and a radius R 1 of the circumscribed circle; attribute information of all building boundaries is recorded as V 1(BoundaryObject1,BoundaryObject2,BoundaryObject3 …);
S3, loading and collecting real-scene three-dimensional data information: loading live-action three-dimensional data, constructing TileObject objects for each Tile model top LOD data node of the three-dimensional data, and acquiring and storing Mesh model data, coordinates surrounding a sphere center point O and a surrounding sphere radius R 2; the attribute information of all the tiles is recorded as V 2(TileObject1,TileObject2,TileObject3 …);
S4, extracting and dividing the monomer model; the method comprises the following steps:
S41, performing intersection detection through the circle center C of the circumscribed circle in the step S2 and the xy plane where the center of the surrounding sphere in the step S3 is located, and storing the TileObject object intersected with the BoundaryObject object into the attribute of the BoundaryObject object to finish the intersection extraction at the tie level;
S42, extracting and dividing the top LOD of each Tile model in a breadth-first traversal mode; performing triangle extraction, segmentation expansion and texture extraction segmentation operations on a Mesh model data TIN network of TileObject objects in V 2 by using BoundaryObject objects in V 1 to finish segmentation extraction of a top LOD of each Tile model;
S43, pushing the real three-dimensional data to the depth direction of the Tile model, and carrying out segmentation and extraction downwards according to the Tile LOD level by LOD level until all data segmentation and extraction are completed;
S5, respectively outputting the single building model which is recorded in V 1 and is extracted by segmentation and recorded in BoundaryObject and the model of the segmentation residual part recorded in V 2 according to the original data organization structure of the live-action three-dimensional data.
As an option, the specific content of step S2 is as follows:
S21, loading vector data, and constructing a boundary object BoundaryObject for each building boundary in the data; vector data coordinate information is recorded with a set of coordinate strings (X1, Y1, X2, Y2 … … Xn, yn), vector data is loaded and coordinate string information of each vector boundary therein is acquired, identified with boundary (P1, P2, P3 …), wherein Pn (Xn, yn);
S22, traversing coordinates in the Boundary, respectively finding out the minimum value and the maximum value of the x coordinate and the minimum value and the maximum value of the y coordinate of each building Boundary, and marking the coordinates of four points as P xmin、Pxmax、Pymin、Pymax; constructing a minimum bounding Rectangle min of the boundary object through the four points, and solving coordinates P1, P2, P3 and P4 of four vertexes of the minimum bounding Rectangle; calculating the coordinate value of the circle center C of the circumscribing circle of the boundary Rectingle min and the radius R 1 of the circumscribing circle;
S23, storing information of boundary (P1, P2, P3 …), circumscribed circle center C and circumscribed circle radius R 1 into a BoundaryObject object, and recording attribute information of all building boundaries as V 1(BoundaryObject1,BoundaryObject2,BoundaryObject3 ….
As an option, the specific content of step S3 is as follows:
s31, loading live-action three-dimensional data, and organizing and managing the three-dimensional data by adopting a spatial data organization structure of BVH;
S32, constructing a TileObject object for each Tile model top LOD data node of the three-dimensional data, and acquiring and storing Mesh model data, coordinates of a surrounding sphere center point O and a surrounding sphere radius R 2; the attribute information of all tiles is recorded as V 2(TileObject1,TileObject2,TileObject3 …).
As an option, the specific content of step S41 is as follows:
Extracting a Tile model by using the BoundaryObject object in V 1; d c-o<R1+R2 judgment is carried out through a circle center C and a radius R 1 in BoundaryObject and a sphere center O and a radius R 2 of a TileObject enclosed sphere in V 2, the establishment of a formula is a condition that an object BoundaryObject and TileObject are intersected, D c-o is a distance from the circle center C in step S2 to an xy plane where the O of the sphere center in step S3 is located, and an object TileObject intersected with the object BoundaryObject is saved in an attribute of the object BoundaryObject, so that the intersecting extraction of a rule level is completed.
As an option, the specific content of step S42 is as follows:
421. Performing preliminary intersection extraction on triangles in the TIN network; obtaining the radius R i of the circumscribed circle of the triangle, the center point C i and the distance D i from the center point C in BoundaryObject to the xy plane, and recording the triangle as a set delta 0(Δ0,Δ1,Δ2,Δi … when a formula D i<Ri+R1 is established;
422. Performing further extraction, segmentation expansion and texture map segmentation extraction operations on the triangle set delta 0 obtained in the previous step; and judging the position relation between three vertexes of the triangle and the building boundary by a corner method to obtain the spatial relation between the triangle and the building boundary, and dividing and extracting the triangle intersected with the building boundary, expanding a triangular net and dividing and extracting the texture map to finish the division and extraction of the LOD on the top layer of each Tile model.
As an option, step 422 is specifically as follows:
Judging whether the points are inside the polygon or not by a corner method, and sequentially judging whether three vertexes of the triangle fall inside a building vector Boundary in BoundaryObject or not; if the triangle three points are all inside the boundary, then the triangle is contained by the polygon, then the triangle is marked and stored into the set Δ 1(Δ0,Δ1,Δ2,Δi …), and saved into the BoundaryObject object; if only one or two points of the triangle are within the polygon, then the two intersect, then the set Δ 2(Δ0,Δ1,Δ2,Δi … is marked and stored; if all three points of the triangle are outside the polygon, the triangle does not belong to the data of the segmentation extraction part; and dividing triangles in the set delta 2, adding triangles and extracting textures to expand a triangle network, respectively obtaining coordinates of vertices of the newly added triangles and texture coordinates through a line intersection principle and a linear difference value, and storing the newly added triangles and the texture coordinates into BoundaryObject objects in a Boundary according to the spatial relationship between the newly added triangles and the Boundary, and otherwise storing the newly added triangles and the Boundary into TileObject objects to finish the division and extraction of LOD on the top layer of each Tile model.
By adopting the technical scheme, the invention has the following beneficial effects:
According to the invention, boundary space information of a building boundary outline is obtained through the existing vector data, multi-level segmentation is carried out on the inclined live-action three-dimensional data of a data structure with LOD levels, and meanwhile, single building models can be automatically extracted in batches under the condition that the original data structure is not changed; the method can quickly divide the monomers in batches, solves the problem of saw teeth of the dividing edges of the monomer models, maintains the data scheduling mode of LOD of the original data, and has good performance in the aspects of dividing and extracting efficiency, quality, rendering efficiency of dividing results in subsequent application and the like.
Drawings
Fig. 1 is a flow chart of the steps of the present invention.
Fig. 2 is a diagram of the minimum boundary construction of the building of the present invention.
Fig. 3 is a diagram of the building minimum boundary circumscribing circle construction of the present invention.
Fig. 4 is a spatial data management block diagram of the present invention.
Fig. 5 is a schematic diagram of the intersection detection of the present invention.
FIG. 6 is a schematic diagram of the corner method point location determination according to the present invention.
Fig. 7 is a triangle division expansion diagram of the present invention.
Fig. 8 is an effect diagram of an example of the present invention.
FIG. 9 is a system block diagram of the singulation system of the present invention.
Detailed Description
The following is a further description of the specific embodiments of the invention with reference to the accompanying drawings.
Examples
As shown in fig. 1, the method for automatically dividing LOD-level three-dimensional data singulation based on vector data in this embodiment includes the following steps:
S1, unifying a coordinate system of vector data and oblique live-action three-dimensional data;
Acquiring coordinate system information of vector data and inclined live-action three-dimensional data (the coordinate system of the live-action three-dimensional data is generally a CGCS2000 coordinate system), if the coordinate systems are inconsistent, calculating to be under the standard of the CGCS2000 coordinate system, and carrying out coordinate conversion by adopting a two-dimensional four-parameter conversion model, wherein the conversion formula is as follows:
Wherein x 1、y1 is a converted vector data coordinate, x 2、y2 is a real-scene three-dimensional model data coordinate, deltax and Deltay are translation parameters, m is a scale transformation parameter, and alpha is a rotation parameter. In the formula, four parameters (delta x, delta y, alpha, m) are obtained by using a least square method, and the data are converted into a CGCS2000 national coordinate system by using the obtained four parameters.
S2, loading vector data and acquiring attribute information related to a building boundary;
(1) Boundary objects BoundaryObject are constructed to store attribute information for each boundary. It is known that vector data generally represent the spatial location of a geographic entity as accurately as possible by recording coordinates. Vector data coordinate information is recorded with a set of coordinate strings (X 1,Y1,X2,Y2……Xn,Yn). Reading plug-ins through self-written vector data, loading the vector data and acquiring coordinate string information of each vector boundary, identifying by boundary (P 1,P2,P3 …), wherein P 1(X1,Y1) and the like.
(2) And (3) through traversing coordinates in the Boundary, comparing and calculating, respectively finding out two points of the minimum value and the maximum value of the x coordinate and two points of the minimum value and the maximum value of the y coordinate, and marking the coordinates of the total four points as P xmin、Pxmax、Pymin、Pymax.
The minimum bounding Rectangle min of the Boundary is constructed through the four points, and coordinates P1, P2, P3 and P4 of four vertexes of the Rectangle are obtained, see FIG. 2.
(3) In preparation for the intersection detection in the subsequent step, the coordinate value of the circumscribing circle center C of the boundary Rectangle min and the radius R 1 of the circumscribing circle are required to be calculated for model extraction of the three-dimensional model at the Tile level, as shown in fig. 3.
According to the rectangular property, the coordinate value of the center point C (Cx, cy) of the rectangle can be obtained through two points P1, P3 or P2, P4 of the diagonal end point of the rectangle, and the circle center O of the circumscribed circle coincides with the center C of the rectangle according to the circumscribed circle property of the rectangle, so that a circle center calculation formula is obtained:
Wherein Cx and Cy are x and y coordinates of the center of the circle; p1x and P1y are the x coordinate and y coordinate of the P1 point; p3x and P3y are the x coordinate and y coordinate of the P3 point.
Obtaining a calculation formula of the radius of the circumscribed circle:
(4) And storing the information of boundary, circumscribed circle center C and circumscribed circle radius R 1 into BoundaryObject object, and finally recording the attribute information of all building boundaries as V 1(BoundaryObject1,BoundaryObject2,BoundaryObject3 …).
S3, loading and collecting real-scene three-dimensional data information;
and loading the live-action three-dimensional data, and organizing and managing the three-dimensional data by adopting a spatial data organization structure of BVH (bounding volume hierarchical structure), as shown in fig. 4.
And constructing TileObject objects for each Tile model top-layer LOD data node of the three-dimensional data to store Mesh model data, coordinates surrounding a sphere center point O and information surrounding a sphere radius R 2. The attribute information of all tiles is recorded as V 2(TileObject1,TileObject2,TileObject3 …).
S4, extracting and dividing the model; the method comprises three links:
1) The primary intersection detection extraction mainly completes the intersection extraction of the tie level, the BoundaryObject object in V 1 is used for extracting the TileObject object in V 2, the center C and the radius R 1 in BoundaryObject are used for judging Dc-O < R 1+R2 with the center O and the radius R 2 of a TileObject surrounding sphere in V 2, the establishment of the formula is the condition that the BoundaryObject object and the TileObject are intersected, the Dc-O is the distance from the center C in the step S2 to the center O of the sphere in the step S3 on the xy plane, and the TileObject object intersected with the BoundaryObject object is saved in the attribute of the BoundaryObject object. Through the step, the Tile model data which cannot be segmented is eliminated, and the next step segmentation extraction operand is reduced.
2) And adopting a breadth-first traversal mode, namely preferentially extracting and dividing the top LOD of each Tile model. The extraction segmentation and texture extraction segmentation of the TIN net triangle in the Mesh model data of the BoundaryObject object in V 1 to the TileObject object in V 2 are further used on the basis of the previous step. The specific operation steps are as follows:
(1) Performing preliminary extraction on triangles in the TIN network, obtaining the radius R i of the circumscribed circle of the triangle and the center point C i and the distance D i of the center point C in the xy plane in BoundaryObject by referring to the method in the step S2, and recording the triangles as delta 0(Δ0,Δ1,Δ2,Δi … when the formula D i<Ri+R1 is established), thereby completing the intersection extraction of the triangles, as shown in figure 5.
(2) And (3) further extracting the triangle set delta 0 obtained in the last step, expanding the segmentation triangle and extracting the segmentation of the texture map. In the extraction process of the triangle level, the method of constructing the circumcircle in claim 3 is also adopted to perform preliminary intersection extraction, so that the position relationship between three vertexes of the triangle and the building boundary is judged by a corner method, and then the spatial relationship between the triangle and the building boundary is obtained. And (3) dividing and expanding a triangle net and dividing and extracting a texture map for the triangle intersected with the building boundary, thereby solving the problem of saw tooth of the dividing edge of the monomer model.
Whether the point is inside the polygon is judged by the corner method, see fig. 6, and the formula is as follows:
Where wn is the number of circles, V i is the vertex of the polygon, and PV i is the vector of P to V i. The number of circles is 0, the point is outside the polygon, whereas the point is inside the polygon.
Sequentially judging whether three vertexes of the triangle fall inside a building vector boundary in BoundaryObject, if the three points of the triangle are all inside the boundary, the triangle is contained by the polygon, and as shown in fig. 7 (1), the triangle is marked and stored in delta 1(Δ0,Δ1,Δ2,Δi …), and the triangle is stored in a BoundaryObject object; only one or two points in the triangle are within the polygon, the two intersecting, as shown in fig. 7 (2) and fig. 7 (3), are marked and stored as Δ 2(Δ0,Δ1,Δ2,Δi …. If all three points are outside the polygon, the triangle does not belong to the data of the segmentation extraction section. For the intersecting situation, in order to avoid the situation that saw teeth appear at the dividing edge, the triangle in delta 2 is divided, triangle and texture extraction are added, and then a triangle net is expanded, as shown in fig. 7 (1) and fig. 2), coordinates and texture coordinates of newly added triangle vertices are respectively obtained through a line intersecting principle and a linear difference value, and are stored in a BoundaryObject object in a Boundary according to the spatial relationship between the newly added triangle and the Boundary, otherwise, the triangle is stored in a TileObject object, and the dividing extraction of the top LOD of each Tile model is completed.
3) Advancing to the depth direction of the Tile model of the live-action three-dimensional data, and carrying out segmentation and extraction downwards according to the Tile LOD hierarchy until all data segmentation and extraction are completed.
And S5, respectively outputting the single building model which is recorded in V 1 and is extracted by segmentation and recorded in BoundaryObject and the model of the segmentation residual part recorded in V 2 according to the original data organization structure of the live-action three-dimensional data.
The following will be specifically exemplified:
referring to fig. 8, live-action three-dimensional data of a certain area and vector building contour data of a corresponding area are prepared, model data and vector data are loaded respectively, and then automatic segmentation extraction of a monomer model and relevant application test on the monomer model are performed. Referring to fig. 8 (a), in order to load data to complete the automatic single model segmentation and extraction effect, it can be seen that the present invention completes the automatic single building model batch segmentation and extraction according to the vector building contour data, and effectively solves the problem of saw tooth of the single model segmentation edge. Referring to fig. 8 (b), for the data structure diagrams before and after the segmentation and extraction of the monomer model, it can be seen that the present invention implements multi-level segmentation of the oblique live-action three-dimensional data, and completes data output without changing the original data structure. Referring to fig. 8 (c), an application effect diagram for attribute query is selected for supporting highlight points for the monomer model.
As described above, the method for the singulation is characterized in detail as follows:
1. According to the invention, boundary space information of a building boundary contour is obtained through the existing vector data, multi-level segmentation is carried out on the inclined live-action three-dimensional data of the data structure with the LOD level, and meanwhile, under the condition that the original data structure is not changed, the single building model can be automatically extracted in batches, and the problem of saw teeth of the segmentation edge of the single building model can be effectively solved.
2. Because the vector data can provide high-precision building boundary information and the advanced oblique photography technology at the present stage can obtain a high-precision real-scene three-dimensional data model, the method provided by the invention can automatically obtain high-precision monomer building models in batches, has better time efficiency and monomer model quality than the traditional manual modeling monomer, and meanwhile, the result data keeps the original LOD data structure, so that rendering and rendering are smoother in a large three-dimensional scene.
3. According to the invention, the extracted monomer building model and the model of the segmented residual part are respectively segmented according to the original data organization structure of the live-action three-dimensional data, so that the extraction of the monomer model is realized, and the method can be further used for the application of attribute information hooking, inquiry and the like, and the integrity of the whole three-dimensional scene data is not influenced. The method provides a solution for the application of using the real-scene three-dimensional monomer building model in a large-scale scene, and further expands the application of the real-scene three-dimensional data in various industries.
Based on the foregoing method of singulation, a system based on the method of singulation will be described below, with reference to the foregoing examples of singulation methods for details.
As shown in fig. 9, the system for automatically dividing LOD-level three-dimensional data into monomers based on vector data according to this embodiment includes the following:
And a coordinate conversion module: the coordinate system is used for unifying vector data and inclined live-action three-dimensional data;
Boundary data loading preprocessing module: the method is used for loading vector data and acquiring attribute information of a building boundary, and specifically comprises the following steps: loading vector data, constructing a boundary object BoundaryObject for each building boundary in the data, and acquiring and storing coordinate string information boundary (P1, P2, P3 …) of the boundary object and coordinates of a circumscribed circle center C of a minimum bounding rectangle and a radius R 1 of the circumscribed circle; attribute information of all building boundaries is recorded as V 1(BoundaryObject1,BoundaryObject2,BoundaryObject3 …);
The three-dimensional data loading preprocessing module comprises: the method is used for loading and collecting the three-dimensional data information of the live-action, and specifically comprises the following steps: loading live-action three-dimensional data, constructing TileObject objects for each Tile model top LOD data node of the three-dimensional data, and acquiring and storing Mesh model data, coordinates surrounding a sphere center point O and a surrounding sphere radius R 2; the attribute information of all the tiles is recorded as V 2(TileObject1,TileObject2,TileObject3 …);
The data extraction and segmentation module: extracting and dividing the monomer model; the method comprises the following processing flows:
1) Intersecting detection is carried out through the circle center C of the circumscribed circle of the boundary data loading preprocessing module and the xy plane where the sphere center O of the surrounding sphere of the three-dimensional data loading preprocessing module is located, and TileObject objects intersected with BoundaryObject objects are stored in the attribute of BoundaryObject objects, so that intersecting extraction at the tie level is completed;
2) Extracting and dividing the top LOD of each Tile model in a breadth-first traversal mode; performing triangle extraction, segmentation expansion and texture extraction segmentation operations on a Mesh model data TIN network of TileObject objects in V 2 by using BoundaryObject objects in V 1 to finish segmentation extraction of a top LOD of each Tile model;
3) Advancing to the depth direction of a Tile model of the live-action three-dimensional data, and carrying out segmentation and extraction downwards according to the Tile LOD level by LOD level until all data segmentation and extraction are completed;
and the data output module is used for: the method is used for respectively outputting a single building model extracted by segmentation recorded in BoundaryObject in V 1 and a model of the rest part of segmentation recorded in TileObject in V 2 according to the original data organization structure of the live-action three-dimensional data.
The foregoing description is directed to the details and illustrations of the preferred embodiments of the invention, but these descriptions are not intended to limit the scope of the invention claimed, and all equivalent changes or modifications that may be accomplished under the teachings of the invention should be construed to fall within the scope of the invention as defined by the appended claims.
Claims (8)
1. The method for automatically dividing LOD level three-dimensional data into monomers based on vector data is characterized by comprising the following steps:
s1, unifying a coordinate system of vector data and three-dimensional data of the inclined live-action;
S2, loading vector data and acquiring attribute information of a building boundary: loading vector data, constructing a boundary object BoundaryObject for each building boundary in the data, and acquiring and storing coordinate string information boundary (P1, P2, P3 …) of the boundary object and coordinates of a circumscribed circle center C of a minimum bounding rectangle and a radius R 1 of the circumscribed circle; attribute information of all building boundaries is recorded as V 1(BoundaryObject1,BoundaryObject2,BoundaryObject3 …);
S3, loading and collecting real-scene three-dimensional data information: loading live-action three-dimensional data, constructing TileObject objects for each Tile model top LOD data node of the three-dimensional data, and acquiring and storing Mesh model data, coordinates surrounding a sphere center point O and a surrounding sphere radius R 2; the attribute information of all the tiles is recorded as V 2(TileObject1,TileObject2,TileObject3 …);
S4, extracting and dividing the monomer model; the method comprises the following steps:
S41, performing intersection detection through the circle center C of the circumscribed circle in the step S2 and the xy plane where the center of the surrounding sphere in the step S3 is located, and storing the TileObject object intersected with the BoundaryObject object into the attribute of the BoundaryObject object to finish the intersection extraction at the tie level;
S42, extracting and dividing the top LOD of each Tile model in a breadth-first traversal mode; performing triangle extraction, segmentation expansion and texture extraction segmentation operations on a Mesh model data TIN network of TileObject objects in V 2 by using BoundaryObject objects in V 1 to finish segmentation extraction of a top LOD of each Tile model;
S43, pushing the real three-dimensional data to the depth direction of the Tile model, and carrying out segmentation and extraction downwards according to the Tile LOD level by LOD level until all data segmentation and extraction are completed;
S5, respectively outputting the single building model which is recorded in V 1 and is extracted by segmentation and recorded in BoundaryObject and the model of the segmentation residual part recorded in V 2 according to the original data organization structure of the live-action three-dimensional data.
2. The method for automatically partitioning LOD-level three-dimensional data singulation based on vector data according to claim 1, wherein the specific content of step S1 is as follows: and acquiring coordinate system information of the vector data and the three-dimensional data of the inclined live-action, and if the coordinate systems are inconsistent, performing coordinate conversion by adopting a four-parameter conversion model, and directly converting the vector data into coordinates of the three-dimensional model data.
3. The method for automatically partitioning LOD-level three-dimensional data singulation based on vector data according to claim 1, wherein the specific content of step S2 is as follows:
S21, loading vector data, and constructing a boundary object BoundaryObject for each building boundary in the data; vector data coordinate information is recorded with a set of coordinate strings (X1, Y1, X2, Y2 … … Xn, yn), vector data is loaded and coordinate string information of each vector boundary therein is acquired, identified with boundary (P1, P2, P3 …), wherein Pn (Xn, yn);
S22, traversing coordinates in the Boundary, respectively finding out the minimum value and the maximum value of the x coordinate and the minimum value and the maximum value of the y coordinate of each building Boundary, and marking the coordinates of four points as P xmin、Pxmax、Pymin、Pymax; constructing a minimum bounding Rectangle min of the boundary object through the four points, and solving coordinates P1, P2, P3 and P4 of four vertexes of the minimum bounding Rectangle; calculating the coordinate value of the circle center C of the circumscribing circle of the boundary Rectingle min and the radius R 1 of the circumscribing circle;
S23, storing information of boundary (P1, P2, P3 …), circumscribed circle center C and circumscribed circle radius R 1 into a BoundaryObject object, and recording attribute information of all building boundaries as V 1(BoundaryObject1,BoundaryObject2,BoundaryObject3 ….
4. The method for automatically partitioning LOD-level three-dimensional data singulation based on vector data according to claim 1, wherein: the specific content of the step S3 is as follows:
s31, loading live-action three-dimensional data, and organizing and managing the three-dimensional data by adopting a spatial data organization structure of BVH;
S32, constructing a TileObject object for each Tile model top LOD data node of the three-dimensional data, and acquiring and storing Mesh model data, coordinates of a surrounding sphere center point O and a surrounding sphere radius R 2; the attribute information of all tiles is recorded as V 2(TileObject1,TileObject2,TileObject3 …).
5. The method for automatically partitioning LOD-level three-dimensional data singulation based on vector data according to claim 1, wherein: the specific content of the step S41 is as follows:
Extracting a Tile model by using the BoundaryObject object in V 1; d c-o<R1+R2 judgment is carried out through a circle center C and a radius R 1 in BoundaryObject and a sphere center O and a radius R 2 of a TileObject enclosed sphere in V 2, the establishment of a formula is a condition that an object BoundaryObject and TileObject are intersected, D c-o is a distance from the circle center C in step S2 to an xy plane where the O of the sphere center in step S3 is located, and an object TileObject intersected with the object BoundaryObject is saved in an attribute of the object BoundaryObject, so that the intersecting extraction of a rule level is completed.
6. The method for automatically partitioning LOD-level three-dimensional data singulation based on vector data according to claim 1, wherein: the specific content of the step S42 is as follows:
421. Performing preliminary intersection extraction on triangles in the TIN network; obtaining the radius R i of the circumscribed circle of the triangle, the center point C i and the distance D i from the center point C in BoundaryObject to the xy plane, and recording the triangle as a set delta 0(Δ0,Δ1,Δ2,Δi … when a formula D i<Ri+R1 is established;
422. Performing further extraction, segmentation expansion and texture map segmentation extraction operations on the triangle set delta 0 obtained in the previous step; and judging the position relation between three vertexes of the triangle and the building boundary by a corner method to obtain the spatial relation between the triangle and the building boundary, and dividing and extracting the triangle intersected with the building boundary, expanding a triangular net and dividing and extracting the texture map to finish the division and extraction of the LOD on the top layer of each Tile model.
7. The method for automatically partitioning LOD-level three-dimensional data singulation based on vector data according to claim 6, wherein: the specific content of step 422 is as follows:
Judging whether the points are inside the polygon or not by a corner method, and sequentially judging whether three vertexes of the triangle fall inside a building vector Boundary in BoundaryObject or not; if the triangle three points are all inside the boundary, then the triangle is contained by the polygon, then the triangle is marked and stored into the set Δ 1(Δ0,Δ1,Δ2,Δi …), and saved into the BoundaryObject object; if only one or two points of the triangle are within the polygon, then the two intersect, then the set Δ 2(Δ0,Δ1,Δ2,Δi … is marked and stored; if all three points of the triangle are outside the polygon, the triangle does not belong to the data of the segmentation extraction part; and dividing triangles in the set delta 2, adding triangles and extracting textures to expand a triangle network, respectively obtaining coordinates of newly added triangle vertexes and texture coordinates through a line intersection principle and a linear difference value, storing the newly added triangle vertexes and the texture coordinates into BoundaryObject objects in the Boundary according to the spatial relationship between the newly added triangle and the Boundary, and otherwise storing the newly added triangle and the Boundary into TileObject objects to finish the division and extraction of the top LOD of each Tile model.
8. An LOD level three-dimensional data singulation system based on vector data automatic segmentation is characterized by comprising the following contents:
And a coordinate conversion module: the coordinate system is used for unifying vector data and inclined live-action three-dimensional data;
Boundary data loading preprocessing module: the method is used for loading vector data and acquiring attribute information of a building boundary, and specifically comprises the following steps: loading vector data, constructing a boundary object BoundaryObject for each building boundary in the data, and acquiring and storing coordinate string information boundary (P1, P2, P3 …) of the boundary object and coordinates of a circumscribed circle center C of a minimum bounding rectangle and a radius R 1 of the circumscribed circle; attribute information of all building boundaries is recorded as V 1(BoundaryObject1,BoundaryObject2,BoundaryObject3 …);
The three-dimensional data loading preprocessing module comprises: the method is used for loading and collecting the three-dimensional data information of the live-action, and specifically comprises the following steps: loading live-action three-dimensional data, constructing TileObject objects for each Tile model top LOD data node of the three-dimensional data, and acquiring and storing Mesh model data, coordinates surrounding a sphere center point O and a surrounding sphere radius R 2; the attribute information of all the tiles is recorded as V 2(TileObject1,TileObject2,TileObject3 …);
The data extraction and segmentation module: extracting and dividing the monomer model; the method comprises the following processing flows:
Intersecting detection is carried out through the circle center C of the circumscribed circle of the boundary data loading preprocessing module and the xy plane where the sphere center O of the surrounding sphere of the three-dimensional data loading preprocessing module is located, and TileObject objects intersected with BoundaryObject objects are stored in the attribute of BoundaryObject objects, so that intersecting extraction at the tie level is completed;
Extracting and dividing the top LOD of each Tile model in a breadth-first traversal mode; performing triangle extraction, segmentation expansion and texture extraction segmentation operations on a Mesh model data TIN network of TileObject objects in V 2 by using BoundaryObject objects in V 1 to finish segmentation extraction of a top LOD of each Tile model; the method comprises the steps of,
Advancing to the depth direction of a Tile model of the live-action three-dimensional data, and carrying out segmentation and extraction downwards according to the Tile LOD level by LOD level until all data segmentation and extraction are completed;
and the data output module is used for: the method is used for respectively outputting a single building model extracted by segmentation recorded in BoundaryObject in V 1 and a model of the rest part of segmentation recorded in TileObject in V 2 according to the original data organization structure of the live-action three-dimensional data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010380448.6A CN111784840B (en) | 2020-05-08 | 2020-05-08 | LOD (line-of-sight) level three-dimensional data singulation method and system based on vector data automatic segmentation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010380448.6A CN111784840B (en) | 2020-05-08 | 2020-05-08 | LOD (line-of-sight) level three-dimensional data singulation method and system based on vector data automatic segmentation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111784840A CN111784840A (en) | 2020-10-16 |
| CN111784840B true CN111784840B (en) | 2024-04-19 |
Family
ID=72753147
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010380448.6A Active CN111784840B (en) | 2020-05-08 | 2020-05-08 | LOD (line-of-sight) level three-dimensional data singulation method and system based on vector data automatic segmentation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111784840B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112700531B (en) * | 2020-12-18 | 2023-05-16 | 武汉大学 | Hierarchical household display method for building inclination model fused with vector house type diagram |
| CN112785708B (en) * | 2021-03-15 | 2024-04-12 | 广东南方数码科技股份有限公司 | Method, equipment and storage medium for building model singulation |
| CN116150290B (en) * | 2022-05-07 | 2023-09-08 | 西宁市测绘院 | Method for hooking service attribute of geographic entity three-dimensional data based on forced offset operation of central point |
| CN114756798B (en) * | 2022-06-13 | 2022-10-18 | 中汽创智科技有限公司 | Point cloud rendering method and system based on Web end and storage medium |
| CN115600307B (en) * | 2022-12-01 | 2023-03-10 | 北京飞渡科技有限公司 | Method for generating single building from Mesh model of urban scene |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR3009471A1 (en) * | 2013-12-20 | 2015-02-06 | Thomson Licensing | METHOD FOR PROCESSING AN IMAGE SEQUENCE, CORRESPONDING COMPUTER PROGRAM AND PROCESSING DEVICE |
| CN106097456A (en) * | 2016-06-06 | 2016-11-09 | 王洪峰 | Oblique photograph outdoor scene three dimensional monolithic model method based on self-adapting cluster algorithm |
| CN108198194A (en) * | 2017-12-29 | 2018-06-22 | 武汉理工大学 | Oblique photography single object extraction method based on lateral multi-plane cutting |
| CN109492606A (en) * | 2018-11-26 | 2019-03-19 | 上海同繁勘测工程科技有限公司 | Multispectral vector picture capturing method and system, three dimensional monolithic method and system |
-
2020
- 2020-05-08 CN CN202010380448.6A patent/CN111784840B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR3009471A1 (en) * | 2013-12-20 | 2015-02-06 | Thomson Licensing | METHOD FOR PROCESSING AN IMAGE SEQUENCE, CORRESPONDING COMPUTER PROGRAM AND PROCESSING DEVICE |
| CN106097456A (en) * | 2016-06-06 | 2016-11-09 | 王洪峰 | Oblique photograph outdoor scene three dimensional monolithic model method based on self-adapting cluster algorithm |
| CN108198194A (en) * | 2017-12-29 | 2018-06-22 | 武汉理工大学 | Oblique photography single object extraction method based on lateral multi-plane cutting |
| CN109492606A (en) * | 2018-11-26 | 2019-03-19 | 上海同繁勘测工程科技有限公司 | Multispectral vector picture capturing method and system, three dimensional monolithic method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111784840A (en) | 2020-10-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111784840B (en) | LOD (line-of-sight) level three-dimensional data singulation method and system based on vector data automatic segmentation | |
| CN109242862B (en) | Real-time digital surface model generation method | |
| CN111932671A (en) | Three-dimensional solid model reconstruction method based on dense point cloud data | |
| WO2022227910A1 (en) | Virtual scene generation method and apparatus, and computer device and storage medium | |
| CN116310192A (en) | Urban building three-dimensional model monomer reconstruction method based on point cloud | |
| CN107833273B (en) | Oblique photography three-dimensional model objectification application method based on three-dimensional simulation model | |
| Laing et al. | Scan to BIM: the development of a clear workflow for the incorporation of point clouds within a BIM environment | |
| CN112598796A (en) | Method for building and automatically updating three-dimensional building information model based on generalized point cloud | |
| Zhao | Application of 3D CAD in landscape architecture design and optimization of hierarchical details | |
| CN114926602B (en) | Building singleization method and system based on three-dimensional point cloud | |
| CN114612488A (en) | Building-integrated information extraction method, computer device, and storage medium | |
| CN109118588B (en) | Automatic color LOD model generation method based on block decomposition | |
| CN108765567A (en) | A kind of system for building urban design digitlization sand table | |
| CN116152461B (en) | Geological modeling method, device, computer equipment and computer readable storage medium | |
| CN107945217A (en) | A kind of image characteristic point suitable for automatic assembling is to rapid screening method and system | |
| CN115861527A (en) | Method and device for constructing live-action three-dimensional model, electronic equipment and storage medium | |
| CN115937461A (en) | Multi-source fusion model construction and texture generation method, device, medium and equipment | |
| CN109064482B (en) | Method and device for automatically acquiring house outline in three-dimensional oblique photography scene | |
| Borisov et al. | An automated process of creating 3D city model for monitoring urban infrastructures | |
| Zhao et al. | Completing point clouds using structural constraints for large-scale points absence in 3D building reconstruction | |
| Zhao et al. | A 3D modeling method for buildings based on LiDAR point cloud and DLG | |
| CN115186347B (en) | Building CityGML modeling method combining house type plan view and inclination model | |
| CN113838199B (en) | Three-dimensional terrain generation method | |
| Sugihara et al. | Automatic Generation of 3D Building Models with Efficient Solar Photovoltaic Generation | |
| CN114998544B (en) | Building entity model light-weight method considering visual main body |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |