WO2023273339A1 - 一种三维模型简化方法、装置及存储介质 - Google Patents

一种三维模型简化方法、装置及存储介质 Download PDF

Info

Publication number
WO2023273339A1
WO2023273339A1 PCT/CN2022/074862 CN2022074862W WO2023273339A1 WO 2023273339 A1 WO2023273339 A1 WO 2023273339A1 CN 2022074862 W CN2022074862 W CN 2022074862W WO 2023273339 A1 WO2023273339 A1 WO 2023273339A1
Authority
WO
WIPO (PCT)
Prior art keywords
fitting
segmentation
region
vertex
module
Prior art date
Application number
PCT/CN2022/074862
Other languages
English (en)
French (fr)
Inventor
刘玉丹
王士玮
Original Assignee
广东三维家信息科技有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 广东三维家信息科技有限公司 filed Critical 广东三维家信息科技有限公司
Publication of WO2023273339A1 publication Critical patent/WO2023273339A1/zh

Links

Images

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
    • G06T17/20—Finite element generation, e.g. wire-frame surface description, tesselation
    • G06T17/205—Re-meshing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20021—Dividing image into blocks, subimages or windows
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Definitions

  • the present application relates to the technical field of home model, in particular, to a method, device and storage medium for simplification of a three-dimensional model.
  • a detailed 3D model will increase the cost of model storage, calculation and rendering, so model simplification is required. While simplifying, the topological structure and important shape features of the original model should be well maintained.
  • the surface reduction target is set by manually setting parameters, that is, the number of remaining surfaces is input.
  • the local geometric information is used to express the mesh model, which often leads to the destruction of the topology structure of the mesh model, and leads to a large difference from the original model.
  • Excessive simplification will also make it difficult to maintain local geometric features, and the quality of triangular faces will be difficult to guarantee. Very narrow and long triangles are prone to appear, and broken surfaces are prone to appear during display, making it difficult to meet the display requirements.
  • the embodiment of the present application provides a 3D model simplification method, device and storage medium, reasonably determine the area reduction target through region segmentation and adaptive segmentation, maintain the topology structure of the original model, ensure the quality of the mesh, and solve the problems in related technical fields easily
  • the topological structure of the mesh model is destroyed, local features are lost, and the quality of triangular patches is difficult to guarantee.
  • An embodiment of the present application provides a method for simplifying a 3D model, the method may include:
  • the plane feature extraction of the 3D model is carried out through region segmentation and adaptive segmentation.
  • the fitting plane and segmentation boundary obtained after segmentation represent the overall characteristics of the original grid. Simplification using these features can maintain the original model.
  • the important features of the model, so as to maintain the topology of the original model, ensure the quality of the mesh, and solve the problems that the topology of the mesh model is easily destroyed, the local features are lost, and the quality of the triangular faces is difficult to guarantee.
  • performing region segmentation on the triangular mesh model to obtain multiple block regions may include:
  • the area reduction target is determined adaptively, and the degree of segmentation and fitting can be adaptively controlled, thereby controlling the area reduction degree and avoiding artificially specifying the area reduction target.
  • performing initial region segmentation on the triangular mesh model to obtain an initial segmentation result may include:
  • E(t,P i ) ⁇
  • 2 dx area(t)
  • ni the normal vector of other triangles
  • i the number of other triangles
  • n(x) the normal vector of the initial fitting plane
  • the region fitting error for each region is expressed as:
  • the model is divided into multiple approximate planar block regions through region segmentation, and the best fitting plane for each block region is obtained. Triangular patches are clustered into one category, and for each category, the fitting plane is calculated, which is better able to grasp the plane characteristics of the model than vertex clustering.
  • performing adaptive region segmentation based on the initial segmentation result to obtain multiple block regions may include:
  • the preset number is small, it means that the segmentation is not complete, and the two block areas that are far apart and not in the same plane may be divided into one area, and the two can be separated; if the preset number is larger Large, indicating that the area approximately on the same plane is divided into multiple areas, and these areas can be merged to achieve the purpose of reasonably adjusting the block area.
  • performing a new region insertion operation may include:
  • the largest triangle and the smallest triangle are respectively used as seed regions for region growth.
  • the insertion threshold is set, the number of insertions is determined by the size of the insertion threshold, and the number of divided blocks and the degree of division can be adaptively controlled.
  • performing an adjacent area merging operation may include:
  • the merging operation as a way to realize the merging operation, it can be judged by the closeness of the normal vectors of the fitting planes. If the normal vectors of the two planes are very close, the merging operation can be performed, and the second angle of the included angle can be used. A merge threshold to determine the number of split blocks and the degree of split.
  • performing an adjacent area merging operation may include:
  • the fitting error amount is the combined region fitting error, minus the sum of the region fitting errors of the adjacent block regions ;
  • fitting error amount is not greater than the preset second merging threshold, two adjacent block regions are merged, and plane fitting is performed again.
  • the merging operation it is possible to judge whether to merge according to the change of the fitting error after the merging. If the amount of the fitting error after the merging is less than the second merging threshold, the merging operation can be performed. The number of divided blocks and the degree of division are determined using the second combination threshold of the fitting error amount.
  • the identifying the segmentation boundary of the block region, and performing vertex simplification on the segmentation boundary may include:
  • the vertex to be determined is retained to achieve vertex simplification.
  • the vertices on the boundary of the block area are selectively retained through the angle threshold, and the vertices that have little influence on the boundary contour are eliminated without affecting its topology, so as to achieve the purpose of simplification.
  • the embodiment of the present application also provides a three-dimensional model simplification device, which includes:
  • the data acquisition module is used to obtain the triangular mesh model of the furniture
  • a region segmentation module configured to perform region segmentation on the triangular mesh model to obtain multiple block regions
  • a vertex simplification module configured to identify the segmentation boundary of the block region, and perform vertex simplification on the segmentation boundary
  • the triangulation module is used to triangulate the simplified boundary vertices to obtain a simplified model.
  • the plane feature extraction of the 3D model is carried out through region segmentation and adaptive segmentation.
  • the fitting plane and segmentation boundary obtained after segmentation represent the overall characteristics of the original grid. Simplification using these features can maintain the original model.
  • the important features of the model, so as to maintain the topology of the original model, ensure the quality of the mesh, and solve the problems that the topology of the mesh model is easily destroyed, the local features are lost, and the quality of the triangular faces is difficult to guarantee.
  • the region segmentation module may include:
  • an initial area segmentation module configured to perform initial area segmentation on the triangular mesh model to obtain an initial segmentation result
  • the adaptive segmentation module is configured to perform adaptive region segmentation based on the initial segmentation result to obtain multiple block regions.
  • the initial region segmentation module may include:
  • the initial plane determination module can be configured to randomly select any preset number of selected triangles, and use the plane where each selected triangle is located as the initial fitting plane of the corresponding area;
  • the clustering module may be configured to group other triangles close to the same initial fitting plane into one class according to the proximity of other triangles in the triangular mesh model to the initial fitting plane, so as to obtain multiple clusters;
  • a fitting module configurable to refit each of said clusters to obtain a best fitting plane
  • the iterative module can be configured to repeat the clustering and fitting process until the global fitting error is minimized, so as to obtain block regions.
  • the adaptive segmentation module may include an insertion module, a first merging module, and a second merging module, and the adaptive segmentation module may also be configured to: judge the initial segmentation result to determine the predetermined Set the size of the quantity;
  • the insertion module may be configured to: perform a new area insertion operation if the preset number is small;
  • the first merging module and the second merging module may be configured to: if the preset number is larger, perform an adjacent area merging operation.
  • the plug-in module can also be configured to:
  • the largest triangle and the smallest triangle are respectively used as seed regions for region growing.
  • the first merging module may be configured to:
  • the second merging module may be configured to:
  • the fitting error amount is the combined region fitting error, minus the sum of the region fitting errors of the adjacent block regions ;
  • fitting error amount is not greater than the preset second merging threshold, two adjacent block regions are merged, and plane fitting is performed again.
  • the vertex simplification module may include:
  • An identification module configurable to identify boundary vertices on the segmentation boundary
  • the judging module may be configured to use any boundary vertex as a vertex to be determined, and determine whether the angle between the vertex to be determined and two adjacent boundary vertices is greater than a preset angle threshold;
  • the vertex to be determined is retained to achieve vertex simplification.
  • the embodiment of the present application also provides a readable storage medium, wherein computer program instructions are stored in the readable storage medium, and when the computer program instructions are read and executed by a processor, the method described in any one of the above is executed. 3D model simplification method.
  • FIG. 1 is a flow chart of a method for simplifying a three-dimensional model provided in an embodiment of the present application
  • Fig. 2 (a) to Fig. 2 (b) are schematic diagrams of topological structure changes of models before and after simplification provided by the embodiment of the present application;
  • Figure 3(a) to Figure 3(b) are schematic diagrams of the loss of local geometric features before and after the method of the related technical field provided by the embodiment of the present application;
  • FIG. 4 is a flow chart of the initial region segmentation provided by the embodiment of the present application.
  • FIG. 5 is a flowchart of the insertion operation provided by the embodiment of the present application.
  • FIG. 6 is a flowchart of one of the merging operations provided by the embodiment of the present application.
  • FIG. 7 is a flowchart of another merging operation provided by the embodiment of the present application.
  • Figures 8(a) to 8(b) are the original model and the model after surface reduction provided by the embodiment of the present application;
  • Fig. 9 is a simplified process flowchart provided by the embodiment of the present application.
  • FIG. 10 is a schematic diagram of a model after region segmentation provided by the embodiment of the present application.
  • FIG. 11 is a schematic diagram of the segmentation boundary corresponding to FIG. 10;
  • Fig. 12 is a schematic diagram of the original boundary vertices of the model provided by the embodiment of the present application.
  • FIG. 13 is a simplified schematic diagram of boundary vertices provided by the embodiment of the present application.
  • Fig. 14 is a schematic diagram of the local fillet area of the model after surface reduction in Fig. 3(a) provided by the embodiment of the present application;
  • Fig. 15 is a structural block diagram of a three-dimensional model simplification device provided by an embodiment of the present application.
  • FIG. 16 is a specific structural block diagram of a three-dimensional model simplification device provided by an embodiment of the present application.
  • FIG. 1 is a flow chart of a method for simplifying a 3D model provided by an embodiment of the present application. This method is applied to the simplification of triangular mesh models. Model simplification by methods in related technical fields often only considers the local characteristics of the triangular surface, and unreasonable deletion of edges may easily lead to changes in the topology of the model, such as holes and triangle flips, thereby affecting the display effect of the model. As shown in Figure 2(a) to Figure 2(b), in order to simplify the schematic diagram of the topological structure change of the model before and after, holes appear in one corner and side of the model; for example, the rounded area is over-simplified, the deformation is serious, and the local geometric features are lost, as shown in Fig.
  • 3(a) to 3(b) are schematic diagrams showing the loss of local geometric features, and the rounded corner features of the original model are lost.
  • the quality of triangular faces is difficult to guarantee, and it is easy to appear extremely narrow and long triangles, and the phenomenon of broken faces will affect the display effect.
  • the triangular mesh model (3D model) is composed of geometric information (vertex position) and topological information (points, edges, triangle patches).
  • Triangular mesh model simplification means using appropriate algorithms to reduce the number of facets, edges and vertices of the model under the premise of keeping the geometry of the original model unchanged.
  • plane feature extraction is performed on the model through region segmentation, that is, the model is divided into multiple approximately coplanar regions.
  • the fitting plane and segmentation boundary obtained after segmentation represent the overall characteristics of the original grid, and the important characteristics of the original model can be maintained by using these characteristics for simplification.
  • the method specifically may include the following steps:
  • Step S100 Obtain the triangular mesh model of the furniture
  • Step S200 performing region segmentation on the triangular mesh model to obtain multiple block regions
  • This step can specifically include:
  • Step S210 performing initial region segmentation on the triangular mesh model to obtain an initial segmentation result
  • Step S211 Randomly select any preset number of selected triangles, and use the plane where each selected triangle is located as the initial fitting plane of the corresponding area;
  • the given goal is to divide the furniture model into k blocks, and initialize the block and fitting results: randomly select k triangles as selected triangles, assign them to k regions, and use the plane where each selected triangle is located as the The initial fitting plane for the region.
  • the size of k may be one-third of the total number of triangles, or other values, without any limitation.
  • Step S212 According to the proximity of other triangles in the triangular mesh model to the initial fitting plane, group other triangles that are close to the same initial fitting plane into one class to obtain multiple clusters;
  • This step is a clustering process: fix k initial fitting planes, and group those other triangles that are close to the same initial fitting plane into one class according to the proximity of other triangles to the initial fitting plane.
  • Use the region growing method to divide the region first select a group of triangles (selected triangles) as the seed region, and then traverse these seed regions to make them grow until all the triangles are assigned to a certain region.
  • the fitting error of the triangle patch is the normal vector deviation between the normal vector of the triangle patch t and the fitting plane Pi, defined as follows:
  • t represents other triangles, represents the fitting plane
  • n i represents the normal vector of other triangles
  • n(x) represents the normal vector of the fitting plane
  • i represents the number of other triangles.
  • Step S213 re-fitting each cluster to obtain the best fitting plane
  • This step is a fitting process. K clusters are fixed, and the best fitting plane is recalculated for each category.
  • the specific calculation process can use the least squares fitting method. This method is a related technology and will not be repeated here.
  • Step S214 Repeat the clustering and fitting process until the global fitting error is the smallest, so as to obtain the block area;
  • Step S212 and step S213 are repeated for alternate iterations until convergence.
  • the area fitting error represents the degree of approximation between the area and the fitting plane, and the smaller the error, the closer the area and the fitting plane are.
  • the region fitting error for each region is the sum of all triangle fitting errors in that region. The region fitting error for each region is expressed as:
  • Step S220 Carry out adaptive region segmentation based on the initial segmentation result to obtain multiple block regions.
  • the global fitting error can be reduced through convergent calculation.
  • the above method can obtain a set of segmentation results. However, if k is given too large or too small, the segmentation result may not be accurate enough. Therefore, the above results can be used as the initial segmentation results, and adjustments can be made on this basis to perform adaptive segmentation.
  • Adaptive segmentation includes two types of insertion operation and merging operation. If the preset number is small, a new region insertion operation is performed, and if the preset number is large, an adjacent region merging operation is performed, as shown in Figure 5 , is the insert operation flow chart, specifically:
  • Step S221 If the preset number is small, obtain the block region with the largest region fitting error
  • Step S222 obtaining the largest triangle with the largest fitting error and the smallest triangle with the smallest fitting error in the block area respectively;
  • Step S223 If the fitting error of the largest triangle is greater than a preset insertion threshold, use the largest triangle and the smallest triangle respectively as seed regions to perform region growing.
  • the segmentation is not complete, and two areas that are far apart and not in a fitting plane may be divided into the same block area, in this case a new one needs to be inserted
  • the region is about to use the triangles farthest and closest to the fitting plane as seed regions, and then grow the region, and finally become two block regions.
  • the area with the worst fitting condition that is, the block area with the largest area fitting error (the largest area fitting error).
  • the fitting error of all triangles in the block area and find the triangle with the largest and smallest fitting error (the largest triangle and the smallest triangle), that is, the triangle farthest from the fitting plane and the nearest triangle, respectively, as the seed triangle , for region growing.
  • An insertion threshold can be selected. If the fitting error of the worst-fitting area is still smaller than the insertion threshold, it means that the fitting effect of all areas is ideal, and the insertion operation will not be performed; otherwise, the insertion operation can be performed.
  • the selection of the insertion threshold is not limited here, for example, it may be one percent of the size of the bounding box of the model.
  • the adjacent area merging operation is performed, specifically:
  • the merging operation is to gather two classes together into a new class, and use a new plane to fit the new class.
  • FIG. 6 it is a flowchart of one of the merging operations, and the merging operation may specifically include the following steps:
  • Step S224 traversing adjacent block regions, and obtaining the normal vector of the fitting plane of each block region
  • Step S225 judging whether the angle between the two normal vectors is smaller than the preset first combining threshold
  • Step S226 If the angle between the two normal vectors is smaller than the preset first merging threshold, merge the two adjacent block regions, and perform plane fitting again.
  • the merge operation can be performed. Given a first merging threshold, if the included angle of the plane normal is smaller than the first merging threshold, they can be merged, otherwise they will not be merged.
  • FIG. 7 it is a flow chart of another merging operation, and the merging operation may specifically include the following steps:
  • Step S227 Traversing adjacent block regions, and calculating the increased fitting error amount after merging, the fitting error amount is the region fitting error after merging, minus the region fitting of adjacent block regions sum of errors;
  • Step S228 judging whether the fitting error amount is greater than a preset second combination threshold
  • Step S229 If the fitting error amount is not greater than the preset second merging threshold, merge two adjacent block regions and perform plane fitting again.
  • the merging operation is not performed; otherwise, the merging operation is performed. If more than one pair of regions can be merged, then the pair of regions that increases the least amount of fitting error is selected for merging. Similarly, one merging threshold can be selected, and if the fitting error caused by merging is greater than the second merging threshold, merging will be performed. It should be noted that the number of merging is not limited, and can be multiple merging.
  • An insert operation can increase a block area, and a merge operation can reduce a block area. Adjustments can be made on the basis of the original algorithm for the number of split blocks, the number of insertions can be determined by the insertion threshold, and the number of merging can be determined by the merging threshold. By setting the insertion threshold and segmentation threshold, the degree of segmentation can be adaptively controlled, avoiding artificially specifying the number of segmentation blocks, making the number of segmentation areas more reasonable and in line with the overall characteristics of the furniture model.
  • the surface reduction target is adaptively determined.
  • the degree of segmentation and fitting can be adaptively controlled, thereby controlling the degree of surface reduction and avoiding artificially specifying the surface reduction target.
  • Figure 8(a) to Figure 8(b) they are the original model and the model after surface reduction obtained by this method. After surface reduction, the number of surfaces is about one-fifth of the original.
  • the comparison of the model before and after the reduction shows that there is little difference in the effect before and after the reduction, and the geometry and topology are preserved.
  • the surface reduction target is determined adaptively. By setting the insertion threshold and segmentation threshold, the degree of segmentation and fitting can be adaptively controlled, thereby controlling the degree of surface reduction and avoiding artificially specifying the surface reduction target.
  • Step S300 Identify the segmentation boundary of the block region, and perform vertex simplification on the segmentation boundary;
  • the specific simplified process may include the following steps:
  • Step S301 Identify boundary vertices on the segmentation boundary
  • Step S302 Take any boundary vertex as a vertex to be determined, and determine whether the angle between the vertex to be determined and two adjacent boundary vertices is greater than a preset angle threshold;
  • Step S303 If the angle between the vertex to be determined and two adjacent boundary vertices is greater than the preset angle threshold, then remove the vertex to be determined;
  • Step S304 If the angle between the vertex to be determined and two adjacent boundary vertices is not greater than the preset angle threshold, then keep the vertex to be determined to realize vertex simplification.
  • Each segmented area after division is approximate to a plane and can be replaced by a plane. Then only the vertices on the boundary can determine the shape of the entire block area. The inner vertices have little influence on this area, so they can be completely eliminated.
  • the segmentation boundary is the dividing line of two adjacent block areas, that is, the dividing line of two approximate planar areas, representing the edge characteristics of the original model.
  • the segmentation boundary needs to be maintained. That is, the boundaries of the simplified mesh should contain the segmentation boundaries of the original mesh.
  • the principle of selecting vertices that need to be retained is to ensure that the boundary shape is not much different from the original one.
  • Figures 10-11 they are the schematic diagrams of the model after region segmentation and the corresponding schematic diagrams of the segmentation boundary.
  • the boundary vertices that have little influence on the shape of the segmentation boundary can be further eliminated.
  • the first step is to identify the boundary vertices on the boundary of the segmentation region: if an edge is located on the boundary of the original model, or if its two adjacent patches belong to different regions, then it is marked as the boundary of the segmentation region, and for these boundaries The vertices are relabeled;
  • the second step is to selectively retain the vertices on the boundary of the segmented area: select the included angle threshold, such as 175 degrees, retain the border vertices whose included angle is smaller than the included angle threshold, and assign new vertex labels to them;
  • the included angle threshold such as 175 degrees
  • the third step is to sort the boundary vertices of each sub-region counterclockwise or clockwise according to the label, and the simplified boundary vertices can be obtained, as shown in Figure 12- Figure 13, which are the schematic diagrams of the original boundary vertices of the model and the simplified Schematic diagram of border vertices.
  • the retained boundary vertices constitute all the boundary vertices of the simplified grid, so the boundary vertices in the simplified grid are all existing boundary vertices in the original model, which is easy to maintain the topology structure, avoids topology clutter, and ensures the simplification Consistency of topological structure before and after.
  • the genus (Genus) and manifold (Manifold) of the mesh surface remain unchanged during the simplification process, maintaining the topology.
  • Step S400 Perform triangulation on the simplified boundary vertices to obtain a simplified model.
  • Triangulation can be achieved by Delaunay triangulation, specifically:
  • Delaunay triangulation Project each block area onto the corresponding fitting plane, perform Delaunay triangulation on the two-dimensional plane, and then merge the local triangulations into the whole triangulation.
  • the Delaunay triangulation ensures that there are no narrow and long Triangular patches. Among them, Delaunay triangulation is a related technology, which will not be described in detail here.
  • the Delaunay triangulation maximizes the minimum angle of the triangle in the triangulation, so that narrow and long triangles can be avoided and the mesh quality can be guaranteed.
  • FIG 14 it is a schematic diagram of the local fillet area of the model after surface reduction corresponding to Figure 3(a). It can be seen from the figure that, different from the method in the related technical field, the local geometric features of the fillet area are preserved.
  • Use the results of region segmentation to assist mesh simplification that is, perform 3D mesh segmentation on the original model, extract the segmentation boundary, and define the edge features of the original model, so that some important geometric features of the original model can still be maintained after triangular mesh simplification , such as fillet features, so that it is maintained.
  • the embodiment of the present application also provides a three-dimensional model simplification device, as shown in Figure 15, which is a structural block diagram of the three-dimensional model simplification device, and the device may include:
  • the data acquisition module 100 can be configured to acquire the triangular mesh model of the furniture
  • the region segmentation module 200 may be configured to perform region segmentation on the triangular mesh model to obtain multiple block regions;
  • the vertex simplification module 300 may be configured to identify the segmentation boundary of the block area, and perform vertex simplification on the segmentation boundary;
  • the triangulation module 400 is configured to triangulate the simplified boundary vertices to obtain a simplified model.
  • region segmentation module 200 may include an initial region segmentation module 210 and an adaptive segmentation module 220 .
  • the initial region segmentation module 210 may include:
  • the initial plane determination module 211 can be configured to randomly select any preset number of selected triangles, and use the plane where each selected triangle is located as the initial fitting plane of the corresponding area;
  • the clustering module 212 may be configured to group other triangles close to the same initial fitting plane into one category according to the proximity of other triangles in the triangular mesh model to the initial fitting plane, so as to obtain multiple clusters ;
  • the fitting module 213 may be configured to re-fit each of the clusters to obtain the best fitting plane
  • the iterative module 214 may be configured to repeat the clustering and fitting process until the global fitting error is minimized, so as to obtain block regions.
  • the adaptive segmentation module 220 may include an insertion module 221 and a first merging module 222 and a second merging module 223 .
  • the plug-in module 221 can be configured for:
  • the largest triangle and the smallest triangle are respectively used as seed regions for region growing.
  • the first merging module 222 may be configured to:
  • the second merging module 223 may be configured to:
  • the fitting error amount is the combined region fitting error, minus the sum of the region fitting errors of the adjacent block regions ;
  • fitting error amount is not greater than the preset second merging threshold, two adjacent block regions are merged, and plane fitting is performed again.
  • Vertex reduction module 300 includes:
  • the identification module 301 may be configured to identify boundary vertices on the segmentation boundary;
  • the judging module 302 may be configured to use any boundary vertex as a vertex to be determined, and determine whether the angle between the vertex to be determined and two adjacent boundary vertices is greater than a preset angle threshold;
  • the vertex to be determined is retained to achieve vertex simplification.
  • the embodiment of the present application also provides a readable storage medium, wherein computer program instructions are stored in the readable storage medium, and when the computer program instructions are read and executed by a processor, the method described in any one of the above is executed. 3D model simplification method.
  • each block in a flowchart or block diagram may represent a module, program segment, or part of code that includes one or more Executable instructions.
  • the functions noted in the block may occur out of the order noted in the figures. For example, two blocks in succession may, in fact, be executed substantially concurrently, or they may sometimes be executed in the reverse order, depending upon the functionality involved.
  • each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations can be implemented by a dedicated hardware-based system that performs the specified function or action , or may be implemented by a combination of dedicated hardware and computer instructions.
  • each functional module in each embodiment of the present application may be integrated to form an independent part, each module may exist independently, or two or more modules may be integrated to form an independent part.
  • the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium.
  • the computer software product is stored in a storage medium, including several
  • the instructions are used to make a computer device (which may be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the methods described in the various embodiments of the present application.
  • the aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disc, etc., which can store program codes. .
  • the application provides a three-dimensional model simplification method, device and storage medium, which relate to the technical field of household models.
  • the method includes: obtaining a triangular mesh model of furniture; performing region segmentation on the triangular mesh model to obtain multiple block regions; identifying the segmentation boundaries of the block regions, and performing vertex simplification on the segmentation boundaries ; Triangulate the simplified boundary vertices to obtain a simplified model, reasonably determine the area reduction target through region segmentation and adaptive segmentation, maintain the topology of the original model, ensure the quality of the mesh, and solve the problems in related technical fields.
  • the topological structure of the mesh model is destroyed, local features are lost, and the quality of triangular patches is difficult to guarantee.
  • a three-dimensional model simplification method, device and storage medium of the present application are reproducible and can be used in various industrial applications.
  • a three-dimensional model simplification method, device and storage medium of the present application can be used in the technical field of home model.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computer Graphics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Architecture (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Hardware Design (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Geometry (AREA)
  • Processing Or Creating Images (AREA)

Abstract

一种三维模型简化方法。三维模型简化方法包括：获取家具的三角网格模型；对三角网格模型进行区域分割，以得到多个分块区域；识别分块区域的分割边界，并对分割边界进行顶点简化；对简化后的边界顶点进行三角剖分，以得到简化模型，通过区域分割和自适应分割合理确定减面目标，维护原始模型的拓扑结构，保证网格质量，解决了网格模型拓扑结构容易被破坏、局部特征容易丢失且三角面片质量难以保证的问题。还公开了一种三维模型简化装置、可读存储介质。

Description

一种三维模型简化方法、装置及存储介质
相关申请的交叉引用
本申请要求于2021年06月30日提交中国国家知识产权局的申请号为202110733997.1、名称为“一种三维模型简化方法、装置及存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及家居模型技术领域，具体而言，涉及一种三维模型简化方法、装置及存储介质。
背景技术
精细的三维模型会增加模型存储、计算及绘制成本，因此需要进行模型简化。在简化的同时，应该较好地保持原模型的拓扑结构和重要外形特征。
相关技术领域的方法，通过人工设参的方式设置减面目标，即输入剩下多少面。简化后仅仅利用局部的几何信息对网格模型进行表达，往往会导致网格模型拓扑结构被破坏，并导致和原始模型差别较大。若简化过度还会导致局部几何特征难以保持，三角面片质量将难以保证，容易出现极狭长的三角形，在展示时容易出现破面的现象，难以满足显示的要求。
发明内容
本申请实施例提供了一种三维模型简化方法、装置及存储介质，通过区域分割和自适应分割合理确定减面目标，维护原始模型的拓扑结构，保证网格质量，解决相关技术领域的方法容易导致网格模型拓扑结构被破坏、局部特征丢失且三角面片质量难以保证的问题。
本申请实施例提供了一种三维模型简化方法，所述方法可以包括：
获取家具的三角网格模型；
对所述三角网格模型进行区域分割，以得到多个分块区域；
识别所述分块区域的分割边界，并对所述分割边界进行顶点简化；
对简化后的边界顶点进行三角剖分，以得到简化模型。
在上述实现过程中，通过区域分割和自适应分割对三维模型进行平面特征提取，分割后得到的拟合平面和分割边界代表了原始网格的整体特征，利用这些特征进行简化，可以保持原始模型的重要特征，从而保维护原始模型的拓扑结构，保证网格质量，解决相关技术领域的方法容易导致网格模型拓扑结构被破坏、局部特征丢失且三角面片质量难以保证的问题。
可选地，所述对所述三角网格模型进行区域分割，以得到多个分块区域，可以包括：
对所述三角网格模型进行初始区域分割，以得到初始分割结果；
基于所述初始分割结果进行自适应区域分割，以得到多个分块区域。
在上述实现过程中，根据拟合精度，自适应的确定减面目标，可以自适应地控制分割 和拟合程度，从而控制减面程度，避免人为指定减面目标。
可选地，所述对所述三角网格模型进行初始区域分割，以得到初始分割结果，可以包括：
随机选取任意预设数量个数的选定三角形，并将每个选定三角形所在平面作为对应区域的初始拟合平面；
根据三角网格模型中的其他三角形与所述初始拟合平面的接近程度，将逼近同一初始拟合平面的其他三角形聚成一类，以得到多个聚类；
所述其他三角形与所述初始拟合平面的接近程度利用拟合误差表示：
E(t,P i)＝∫||n(x)-n i|| 2dx＝area(t)||n(x)-n i|| 2；
其中，t表示其他三角形，
Figure PCTCN2022074862-appb-000001
示初始拟合平面；ni表示其他三角形的法向量，i表示其他三角形的个数，n(x)表示初始拟合平面的法向量；
对每个所述聚类重新进拟合以得到最佳拟合平面；
重复进行聚类和拟合过程，直到全局拟合误差最小为止，以得到分块区域；
每个区域的区域拟合误差表示为：
Figure PCTCN2022074862-appb-000002
全局拟合误差表示为：
Figure PCTCN2022074862-appb-000003
在上述实现过程中，通过区域分割将模型划分成多个近似平面的分块区域，并得到每个分块区域的最佳拟合平面，具体通过面片聚类即将连通且近似在同一平面的三角形面片聚成一类，对于每一类，都计算出拟合平面，相对于顶点聚类来说，更能够把握模型的平面特征。
可选地，所述基于所述初始分割结果进行自适应区域分割，以得到多个分块区域，可以包括：
对所述初始分割结果进行判断，以确定所述预设数量的大小；
若所述预设数量较小，则进行新区域插入操作；
若所述预设数量较大，则进行相邻区域合并操作。
在上述实现过程中，如果预设数量较小，说明分割不彻底，相差较远且不在一个平面内的两个分块区域可能被分成了一个区域，可以将两者分开；若预设数量较大，说明近似位于同一个平面的区域被分成了多个区域，可以将这些区域进行合并，达到合理调整分块区域的目的。
可选地，所述若所述预设数量较小，则进行新区域插入操作，可以包括：
若所述预设数量较小，则获取区域拟合误差最大的分块区域；
分别获取所述分块区域中拟合误差最大的最大三角形和拟合误差最小的最小三角形；
若所述最大三角形的拟合误差大于预设的插入阈值；
则将所述最大三角形和最小三角形分别作为种子区域，进行区域生长。
在上述实现过程中，设置插入阈值，插入的次数由插入阈值的大小确定，可自适应的控制分割块数和分割程度。
可选地，所述若所述预设数量较大，则进行相邻区域合并操作，可以包括：
遍历相邻的分块区域，并获取每个分块区域的拟合平面的法向量；
判断两个法向量的夹角是否小于预设的第一合并阈值；
如果两个法向量的夹角小于预设的第一合并阈值，则将相邻的两个分块区域进行合并，并重新进行平面拟合。
在上述实现过程中，作为合并操作的一种实现方式，可以通过拟合平面的法向量的接近程度来判断，若两个平面的法向量很接近，则可以进行合并操作，利用夹角的第一合并阈值来确定分割块数和分割程度。
可选地，所述若所述预设数量较大，则进行相邻区域合并操作，可以包括：
遍历相邻的分块区域，并计算合并后增大的拟合误差量，所述拟合误差量为合并后的区域拟合误差，减去相邻的分块区域的区域拟合误差之和；
判断所述拟合误差量是否大于预设的第二合并阈值；
如果所述拟合误差量不大于预设的第二合并阈值，则将相邻的两个分块区域进行合并，并重新进行平面拟合。
在上述实现过程中，作为合并操作的一种实现方式，可以通过合并后拟合误差的变化情况来判断是否合并，若合并后的拟合误差量小于第二合并阈值，则可以进行合并操作，利用拟合误差量的第二合并阈值来确定分割块数和分割程度。
可选地，所述识别所述分块区域的分割边界，并对所述分割边界进行顶点简化，可以包括：
识别分割边界上的边界顶点；
将任意边界顶点作为待判定顶点，判断所述待判定顶点与相邻的两个边界顶点之间的夹角是否大于预设的夹角阈值；
如果所述待判定顶点与相邻的两个边界顶点之间的夹角大于预设的夹角阈值，则剔除所述待判定顶点；
如果所述待判定顶点与相邻的两个边界顶点之间的夹角不大于预设的夹角阈值，则保留所述待判定顶点，以实现顶点简化。
在上述实现过程中，通过夹角阈值有选择的保留分块区域边界上的顶点，剔除对边界轮廓影响较小的顶点且不影响其拓扑结构，达到简化目的。
本申请实施例还提供一种三维模型简化装置，所述装置包括：
数据获取模块，用于获取家具的三角网格模型；
区域分割模块，用于对所述三角网格模型进行区域分割，以得到多个分块区域；
顶点简化模块，用于识别所述分块区域的分割边界，并对所述分割边界进行顶点简化；
三角化模块，用于对简化后的边界顶点进行三角剖分，以得到简化模型。
在上述实现过程中，通过区域分割和自适应分割对三维模型进行平面特征提取，分割后得到的拟合平面和分割边界代表了原始网格的整体特征，利用这些特征进行简化，可以保持原始模型的重要特征，从而保维护原始模型的拓扑结构，保证网格质量，解决相关技术领域的方法容易导致网格模型拓扑结构被破坏、局部特征丢失且三角面片质量难以保证的问题。
可选地，所述区域分割模块可以包括：
初始区域分割模块，配置成用于对所述三角网格模型进行初始区域分割，以得到初始分割结果；以及
自适应分割模块，配置成用于基于所述初始分割结果进行自适应区域分割，以得到多个分块区域。
可选地，所述初始区域分割模块可以包括：
初始平面确定模块，可以配置成用于随机选取任意预设数量个数的选定三角形，并将每个选定三角形所在平面作为对应区域的初始拟合平面；
聚类模块，可以配置成用于根据三角网格模型中的其他三角形与所述初始拟合平面的接近程度，将逼近同一初始拟合平面的其他三角形聚成一类，以得到多个聚类；
拟合模块，可以配置成用于对每个所述聚类重新进拟合以得到最佳拟合平面；
迭代模块，可以配置成用于重复进行聚类和拟合过程，直到全局拟合误差最小为止，以得到分块区域。
可选地，所述自适应分割模块可以包括插入模块以及第一合并模块和第二合并模块，所述自适应分割模块还可以配置成：对所述初始分割结果进行判断，以确定所述预设数量的大小；
所述插入模块，可以配置成用于：若所述预设数量较小，则进行新区域插入操作；
所述第一合并模块和所述第二合并模块，可以配置成用于：若所述预设数量较大，则进行相邻区域合并操作。
可选地，所述插入模块还可以配置成用于：
若所述预设数量较小，则获取区域拟合误差最大的分块区域；
分别获取所述分块区域中拟合误差最大的最大三角形和拟合误差最小的最小三角形；
若所述最大三角形的拟合误差大于预设的插入阈值，则将所述最大三角形和最小三角形分别作为种子区域，进行区域生长。
可选地，所述第一合并模块可以配置成用于：
遍历相邻的分块区域，并获取每个分块区域的拟合平面的法向量；
判断两个法向量的夹角是否小于预设的第一合并阈值；
如果两个法向量的夹角小于预设的第一合并阈值，则将相邻的两个分块区域进行合并，并重新进行平面拟合。
可选地，所述第二合并模块可以配置成用于：
遍历相邻的分块区域，并计算合并后增大的拟合误差量，所述拟合误差量为合并后的 区域拟合误差，减去相邻的分块区域的区域拟合误差之和；
判断所述拟合误差量是否大于预设的第二合并阈值；
如果所述拟合误差量不大于预设的第二合并阈值，则将相邻的两个分块区域进行合并，并重新进行平面拟合。
可选地，所述顶点简化模块可以包括：
识别模块，可以配置成用于识别分割边界上的边界顶点；
判断模块，可以配置成用于将任意边界顶点作为待判定顶点，判断所述待判定顶点与相邻的两个边界顶点之间的夹角是否大于预设的夹角阈值；
如果所述待判定顶点与相邻的两个边界顶点之间的夹角大于预设的夹角阈值，则剔除所述待判定顶点；
如果所述待判定顶点与相邻的两个边界顶点之间的夹角不大于预设的夹角阈值，则保留所述待判定顶点，以实现顶点简化。
本申请实施例还提供一种可读存储介质，所述可读存储介质中存储有计算机程序指令，所述计算机程序指令被一处理器读取并运行时，执行上述中任一项所述的三维模型简化方法。
附图说明
为了更清楚地说明本申请实施例的技术方案，下面将对本申请实施例中所需要使用的附图作简单地介绍，应当理解，以下附图仅示出了本申请的某些实施例，因此不应被看作是对范围的限定，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他相关的附图。
图1为本申请实施例提供的一种三维模型简化方法的流程图；
图2(a)至图2(b)为本申请实施例提供的简化前后模型拓扑结构改变示意图；
图3(a)至图3(b)为本申请实施例提供的相关技术领域的方法处理前后局部几何特征丢失示意图；
图4为本申请实施例提供的初始区域分割流程图；
图5为本申请实施例提供的插入操作流程图；
图6为本申请实施例提供的其中一种合并操作的流程图；
图7为本申请实施例提供的另一种合并操作的流程图；
图8(a)至8(b)为本申请实施例提供的原始模型和减面后模型；
图9为本申请实施例提供的简化过程流程图；
图10为本申请实施例提供的区域分割后的模型示意图；
图11为图10对应的分割边界示意图；
图12为本申请实施例提供的模型的原始边界顶点示意图；
图13为本申请实施例提供的简化后的边界顶点示意图；
图14为本申请实施例提供的图3(a)的减面后模型的局部圆角区示意图；
图15为本申请实施例提供的三维模型简化装置的结构框图；
图16为本申请实施例提供的三维模型简化装置的具体结构框图。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行描述。
应注意到：相似的标号和字母在下面的附图中表示类似项，因此，一旦某一项在一个附图中被定义，则在随后的附图中不需要对其进行进一步定义和解释。同时，在本申请的描述中，术语“第一”、“第二”等仅用于区分描述，而不能理解为指示或暗示相对重要性。
请参看图1，图1为本申请实施例提供的一种三维模型简化方法的流程图。该方法应用于三角网格模型的简化。相关技术领域的方法进行模型简化，往往只考虑三角面片局部的特征，对边进行不合理的删除，容易导致模型拓扑结构的改变，出现洞和三角形翻转等现象，从而影响模型的展示效果。如图2(a)至图2(b)所示，为简化前后模型拓扑结构改变示意图，模型一角和侧边出现洞；例如，圆角区域简化过度，变形严重，局部几何特征丢失，如图3(a)至图3(b)所示，为局部几何特征丢失示意图，原模型的圆角特征丢失。三角面片质量难以保证，容易出现极狭长的三角形，出现破面的现象，影响展示效果。
三角网格模型(三维模型)由几何信息(顶点位置)和拓扑信息(点，边，三角形面片)构成。三角网格模型简化即在保持原模型几何形状不变的前提下，采用适当的算法减少该模型的面片数、边数和顶点数。本申请通过区域分割对模型进行平面特征提取，即把模型分割成多个近似共面的区域。分割后得到的拟合平面和分割边界代表了原始网格的整体特征，利用这些特征进行简化，可以保持原始模型的重要特征。
该方法具体可以包括以下步骤：
步骤S100：获取家具的三角网格模型；
将沙发、枕头和床等家具模型看成是由多个平面构成的三维模型，将家具模型按平面进行分割，就是将家具模型划分成多个近似平面的区域，并得到每个区域的拟合平面。区域(面)和区域边界(边)就构成了家具模型的特征。所谓区域分割，在本申请中采用面片聚类，相比于顶点聚类更能把握家具模型的平面特征，把连通且近似在同一平面的三角形面片聚成一类，对于每一类，都计算出拟合平面。
步骤S200：对所述三角网格模型进行区域分割，以得到多个分块区域；
该步骤具体可以包括：
步骤S210：对所述三角网格模型进行初始区域分割，以得到初始分割结果；
首先需要进行初始化聚类和拟合平面，然后通过交替迭代的方法，逐步优化聚类和拟合平面。如图4所示，为初始区域分割流程图，具体包括以下步骤：
步骤S211：随机选取任意预设数量个数的选定三角形，并将每个选定三角形所在平面作为对应区域的初始拟合平面；
例如，给定目标是把家具模型分割成k块，并初始化分块和拟合结果：随机选取k个三角形作为选定三角形，分配给k个区域，把每个选定三角形所在的平面作为该区域的初始拟合平面。其中，k的大小可以为总的三角形个数的三分之一，也可以是其他值，不做任何限定。
步骤S212：根据三角网格模型中的其他三角形与所述初始拟合平面的接近程度，将逼近同一初始拟合平面的其他三角形聚成一类，以得到多个聚类；
该步骤为聚类过程：固定k个初始拟合平面，根据其他三角形与初始拟合平面的接近程度，把逼近同一个初始拟合平面的那些其他三角形聚成一类。使用区域增长方法来划分区域，首先选择一组三角形面片(选定三角形)作为种子区域，然后遍历这些种子区域使得它们生长，直到所有的三角形面片都被分配到某个区域。
其中，三角形与拟合平面的接近程度是可通过拟合误差来度量，拟合误差越小二者越接近。三角形面片的拟合误差是三角面片t的法向量与拟合平面P i之间的法向量偏差，定义如下：
E(t，P i)＝∫||n(x)-n i|| 2dx＝area(t)||n(x)-n i|| 2；
其中，t表示其他三角形，
Figure PCTCN2022074862-appb-000004
示拟合平面；n i表示其他三角形的法向量，n(x)表示拟合平面的法向量，i表示其他三角形的个数。
步骤S213：对每个所述聚类重新进行拟合以得到最佳拟合平面；
该步骤为拟合过程，固定k个聚类，对每一类重新计算最佳拟合平面，具体计算过程可以采用最小二乘拟合法，该方法为相关技术，在此不再赘述。
步骤S214：重复进行聚类和拟合过程，直到全局拟合误差最小为止，以得到分块区域；
重复步骤S212和步骤S213，进行交替迭代，直到收敛。
区域拟合误差代表了区域和拟合平面的逼近程度，误差越小，说明区域和拟合平面越靠近。每个区域的区域拟合误差为该区域所有三角形拟合误差的和。每个区域的区域拟合误差表示为：
Figure PCTCN2022074862-appb-000005
形；
全局拟合误差表示为：
Figure PCTCN2022074862-appb-000006
全局拟合误差：对于每个三角形t j，它被分配给了一个区域R i，每个区域R i均有一个最优拟合平面P i，把每个三角形t j到对应最优拟合平面的误差累加起来就是全局拟合误差。
步骤S220：基于初始分割结果进行自适应区域分割，以得到多个分块区域。
经过步骤S211-步骤S214的分割步骤，通过收敛计算可以降低全局拟合误差。对于事先给定的分块区域数k，上述方法可以得到一组分割结果。但是如果k给得过大或过小，都可能使分割结果不够准确。所以可以把上述结果作为初始分割结果，在此基础上进行调整，进行自适应分割。
自适应分割包括插入操作和合并操作两种，若所述预设数量较小，则进行新区域插入操作，若所述预设数量较大，则进行相邻区域合并操作，如图5所示，为插入操作流程图，具体地：
步骤S221：若所述预设数量较小，则获取区域拟合误差最大的分块区域；
步骤S222：分别获取所述分块区域中拟合误差最大的最大三角形和拟合误差最小的最小三角形；
步骤S223：若所述最大三角形的拟合误差大于预设的插入阈值，则将所述最大三角形和最小三角形分别作为种子区域，进行区域生长。
如果初始给定的分块数k过小，即分割不彻底，相差较远并不在一个拟合平面内的两个区域可能被分到了同一个分块区域，这种情况下需要插入一个新的区域即将离拟合平面最远和最近的三角形分别作为种子区域，进行区域生长，最后变成两个分块区域。
首先找到拟合情况最差的区域，也就是区域拟合误差最大的分块区域(区域拟合误差最大的)。计算出该分块区域中所有三角形的拟合误差，找到拟合误差最大和最小的三角形(最大三角形和最小三角形)，也就是离拟合平面最远的三角形和最近的三角形，分别作为种子三角形，进行区域生长。可以选定一个插入阈值，如果拟合最差的区域拟合误差仍然小于插入阈值，说明所有区域的拟合效果都达到理想情况，就不进行插入操作；否则可进行插入操作。对于插入阈值的选择在此不做限定，例如可以是模型包围盒尺寸的百分之一。
若所述预设数量较大，则进行相邻区域合并操作，具体地：
如果初始给定的分块区域数k过大，那么近似位于同一平面的三角形被分割成了多个区域，这时需要对这些区域进行合并。合并操作是把两个类聚到一起变成一个新的类，并用一个新的平面来拟合这个新类。
作为其中一种实施方式，如图6所示，为其中一种合并操作的流程图，合并操作具体可以包括以下步骤：
步骤S224：遍历相邻的分块区域，并获取每个分块区域的拟合平面的法向量；
步骤S225：判断两个法向量的夹角是否小于预设的第一合并阈值；
步骤S226：如果两个法向量的夹角小于预设的第一合并阈值，则将相邻的两个分块区域进行合并，并重新进行平面拟合。
遍历所有相邻的分块区域(区域对)，如果区域对的拟合平面的法向量很接近，那么可以进行合并操作。给定一个第一合并阈值，若平面法向的夹角小于第一合并阈值，可以合并，否则不合并。
作为另外一种实施方式，如图7所示，为另一种合并操作的流程图，合并操作具体可以包括以下步骤：
步骤S227：遍历相邻的分块区域，并计算合并后增大的拟合误差量，所述拟合误差量为合并后的区域拟合误差，减去相邻的分块区域的区域拟合误差之和；
步骤S228：判断所述拟合误差量是否大于预设的第二合并阈值；
步骤S229：如果所述拟合误差量不大于预设的第二合并阈值，则将相邻的两个分块区域进行合并，并重新进行平面拟合。
如果区域对的拟合平面不相同，那么合并这两个区域必然会带来拟合误差的增大。所以判断是否需要合并操作的另外一种方法是，遍历所有相邻区域对，计算合并这两个区域 所增大的拟合误差量，也就是计算出用一个新平面拟合这两个区域的误差，减去原来两个区域的拟合误差之和。
如果合并所带来的误差大于给定阈值(第二合并阈值)，那就不进行合并操作，否则进行合并。如果不止一对区域可以合并，那么选取增大的拟合误差量最小的那一对区域进行合并。同样，可以选取一个合并阈值，如果合并所带来的拟合误差量大于第二合并阈值，则进行合并，需要说明的是，对于合并次数并不进行限定，可以是多次合并。
合并后，对合并后的分块区域重新进行聚类和平面拟合，以得到最佳拟合平面。
进行一次插入操作，可以增加一个分块区域，一次合并操作可以减少一个分块区域。可以在原先给定分割块数算法的基础上进行调整，插入的次数可以由插入阈值来决定，合并的次数由合并阈值决定。通过给定插入阈值和分割阈值，可以自适应地控制分割程度，避免人为指定分割块数，使得分块区域的数量更加合理，符合家具模型的整体特征。
该方法，根据拟合精度，自适应地决定减面目标，通过设定插入阈值和合并阈值，可以自适应地控制分割和拟合程度，从而控制减面程度，避免人为指定减面目标。
如图8(a)至图8(b)所示，分别为原始模型和采用本方法得到的减面后模型，减面后，面数约为原来的五分之一。减面前后的模型对比说明，减面前后的效果差别很小，几何形状和拓扑结构都得以保持。根据拟合精度，自适应地决定减面目标，通过设定插入阈值和分割阈值，可以自适应地控制分割和拟合程度，从而控制减面程度，避免人为指定减面目标。
步骤S300：识别所述分块区域的分割边界，并对所述分割边界进行顶点简化；
如图9所示，为简化过程流程图，具体简化过程可以包括以下步骤：
步骤S301：识别分割边界上的边界顶点；
步骤S302：将任意边界顶点作为待判定顶点，判断所述待判定顶点与相邻的两个边界顶点之间的夹角是否大于预设的夹角阈值；
步骤S303：如果所述待判定顶点与相邻的两个边界顶点之间的夹角大于预设的夹角阈值，则剔除所述待判定顶点；
步骤S304：如果所述待判定顶点与相邻的两个边界顶点之间的夹角不大于预设的夹角阈值，则保留所述待判定顶点，以实现顶点简化。
分割后的每个分块区域都近似平面，可以用平面来代替。那么只需要边界上的顶点就可以决定整个分块区域的形状。内部的顶点对这个区域的影响很小，所以可以全部剔除。
对于分块区域边界上的顶点(边界顶点)：分割边界是相邻两个分块区域的分割线，即两个近似平面区域的分界线，代表了原始模型的边特征。按照新网格尽可能保持原网格特征的原则，分割边界需要被保持。也就是说，简化后网格的边界应该包含原始网格的分割边界。选取需要保留的顶点，原则是要保证边界形状和原来差别不大。如图10-图11所示，分别为区域分割后的模型示意图以及对应的分割边界示意图。
此外，在保证边界形状和原来差别不大的情况下，也可以进一步剔除对分割边界形状影响极小的边界顶点。计算分割边界上任意一个待判定顶点与相邻的两个边界顶点构成的 夹角，例如夹角为180度，即三个边界顶点在一条直线上，则将位于中间的待判定顶点剔除，并不会影响整条分割边界，因此，夹角越接近180度的顶点，对边界轮廓的形状影响越小，贡献度越低。所以可以保证在保持边界形状的前提下，剔除影响不大的边界顶点，达到简化效果。
示例地，可以通过以下步骤实现：
第一步，识别分割区域边界上的边界顶点：如果一条边位于原始模型的边界，或者它相邻的两个面片所属的区域不同，那么它就被标记为分割区域的边界，对这些边界顶点进行重新标号；
第二步，有选择地保留分割区域边界上的顶点：选定夹角阈值，例如175度，保留夹角小于夹角阈值的边界顶点，并对它们赋予新的顶点标号；
第三步，对每个子区域的边界顶点按照标号进行逆时针或者顺时针排序，可获得简化后的边界顶点，如图12-图13所示，分别为模型的原始边界顶点示意图以及简化后的边界顶点示意图。
被保留的边界顶点构成了简化网格的所有边界顶点，所以简化后网格中的边界顶点都是原始模型中已有的边界顶点，容易维护拓扑结构，避免产生拓扑杂乱的情况，保证了简化前后拓扑结构的一致。网格表面的亏格(Genus)和流型(Manifold)在简化过程中保持不变，保持了拓扑结构。
步骤S400：对简化后的边界顶点进行三角剖分，以得到简化模型。
三角剖分可以通过Delaunay三角化实现，具体地：
把每个分块区域投影到对应的拟合平面上，在二维平面上分别进行Delaunay三角化，再将局部的三角剖分合并成全部的三角剖分，Delaunay三角化保证了没有出现狭长的三角形面片。其中，Delaunay三角化为相关技术，在此不再详细赘述。
该过程中，Delaunay三角化最大化了三角剖分中三角形的最小角，从而可以避免出现狭长的三角形，保证了网格质量。
如图14所示，为对应于图3(a)的减面后模型的局部圆角区示意图，由图可以看出，区别于相关技术领域的方法，圆角区的局部几何特征得意保持。利用区域分割的结果辅助网格简化即对原始模型进行三维网格分割，提取分割边界，从而定义出原始模型的边特征，这样在进行三角网格简化后仍然能保持原始模型的一些重要几何特征，如圆角特征，使其得以保持。
本申请实施例还提供一种三维模型简化装置，如图15所示，为三维模型简化装置的结构框图，所述装置可以包括：
数据获取模块100，可以配置成用于获取家具的三角网格模型；
区域分割模块200，可以配置成用于对所述三角网格模型进行区域分割，以得到多个分块区域；
顶点简化模块300，可以配置成用于识别所述分块区域的分割边界，并对所述分割边界进行顶点简化；
三角化模块400，用于对简化后的边界顶点进行三角剖分，以得到简化模型。
如图16所示，为三维模型简化装置的具体结构框图，其中，区域分割模块200可以包括初始区域分割模块210和自适应分割模块220。
初始区域分割模块210可以包括：
初始平面确定模块211，可以配置成用于随机选取任意预设数量个数的选定三角形，并将每个选定三角形所在平面作为对应区域的初始拟合平面；
聚类模块212，可以配置成用于根据三角网格模型中的其他三角形与所述初始拟合平面的接近程度，将逼近同一初始拟合平面的其他三角形聚成一类，以得到多个聚类；
拟合模块213，可以配置成用于对每个所述聚类重新进行拟合以得到最佳拟合平面；
迭代模块214，可以配置成用于重复进行聚类和拟合过程，直到全局拟合误差最小为止，以得到分块区域。
自适应分割模块220可以包括插入模块221和第一合并模块222和第二合并模块223。
其中，插入模块221可以配置成用于：
若所述预设数量较小，则获取区域拟合误差最大的分块区域；
分别获取所述分块区域中拟合误差最大的最大三角形和拟合误差最小的最小三角形；
若所述最大三角形的拟合误差大于预设的插入阈值，则将所述最大三角形和最小三角形分别作为种子区域，进行区域生长。
第一合并模块222可以配置成用于：
遍历相邻的分块区域，并获取每个分块区域的拟合平面的法向量；
判断两个法向量的夹角是否小于预设的第一合并阈值；
如果两个法向量的夹角小于预设的第一合并阈值，则将相邻的两个分块区域进行合并，并重新进行平面拟合。
第二合并模块223可以配置成用于：
遍历相邻的分块区域，并计算合并后增大的拟合误差量，所述拟合误差量为合并后的区域拟合误差，减去相邻的分块区域的区域拟合误差之和；
判断所述拟合误差量是否大于预设的第二合并阈值；
如果所述拟合误差量不大于预设的第二合并阈值，则将相邻的两个分块区域进行合并，并重新进行平面拟合。
顶点简化模块300包括：
识别模块301，可以配置成用于识别分割边界上的边界顶点；
判断模块302，可以配置成用于将任意边界顶点作为待判定顶点，判断所述待判定顶点与相邻的两个边界顶点之间的夹角是否大于预设的夹角阈值；
如果所述待判定顶点与相邻的两个边界顶点之间的夹角大于预设的夹角阈值，则剔除所述待判定顶点；
如果所述待判定顶点与相邻的两个边界顶点之间的夹角不大于预设的夹角阈值，则保留所述待判定顶点，以实现顶点简化。
本申请实施例还提供一种可读存储介质，所述可读存储介质中存储有计算机程序指令，所述计算机程序指令被一处理器读取并运行时，执行上述中任一项所述的三维模型简化方法。
在本申请所提供的几个实施例中，应该理解到，所揭露的装置和方法，也可以通过其它的方式实现。以上所描述的装置实施例仅仅是示意性的，例如，附图中的流程图和框图显示了根据本申请的多个实施例的装置、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段或代码的一部分，所述模块、程序段或代码的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。也应当注意，在有些作为替换的实现方式中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个连续的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或动作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。
另外，在本申请各个实施例中的各功能模块可以集成在一起形成一个独立的部分，也可以是各个模块单独存在，也可以两个或两个以上模块集成形成一个独立的部分。
所述功能如果以软件功能模块的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请的技术方案本质上或者说对相关技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。
以上所述仅为本申请的实施例而已，并不用于限制本申请的保护范围，对于本领域的技术人员来说，本申请可以有各种更改和变化。凡在本申请的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本申请的保护范围之内。应注意到：相似的标号和字母在下面的附图中表示类似项，因此，一旦某一项在一个附图中被定义，则在随后的附图中不需要对其进行进一步定义和解释。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应所述以权利要求的保护范围为准。
需要说明的是，在本文中，诸如第一和第二等之类的关系术语仅仅用来将一个实体或者操作与另一个实体或操作区分开来，而不一定要求或者暗示这些实体或操作之间存在任何这种实际的关系或者顺序。而且，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者设备所 固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括所述要素的过程、方法、物品或者设备中还存在另外的相同要素。
工业实用性
本申请提供了一种三维模型简化方法、装置及存储介质，涉及家居模型技术领域。该方法包括：获取家具的三角网格模型；对所述三角网格模型进行区域分割，以得到多个分块区域；识别所述分块区域的分割边界，并对所述分割边界进行顶点简化；对简化后的边界顶点进行三角剖分，以得到简化模型，通过区域分割和自适应分割合理确定减面目标，维护原始模型的拓扑结构，保证网格质量，解决相关技术领域的方法容易导致网格模型拓扑结构被破坏、局部特征丢失且三角面片质量难以保证的问题。
此外，可以理解的是，本申请的一种三维模型简化方法、装置及存储介质是可以重现的，并且可以用在多种工业应用中。例如，本申请的一种三维模型简化方法、装置及存储介质可以用于家居模型技术领域。

Claims (17)

  1. 一种三维模型简化方法，其特征在于，所述方法包括：
    获取家具的三角网格模型；
    对所述三角网格模型进行区域分割，以得到多个分块区域；
    识别所述分块区域的分割边界，并对所述分割边界进行顶点简化；
    对简化后的边界顶点进行三角剖分，以得到简化模型。
  2. 根据权利要求1所述的三维模型简化方法，其特征在于，所述对所述三角网格模型进行区域分割，以得到多个分块区域，包括：
    对所述三角网格模型进行初始区域分割，以得到初始分割结果；
    基于所述初始分割结果进行自适应区域分割，以得到多个分块区域。
  3. 根据权利要求2所述的三维模型简化方法，其特征在于，所述对所述三角网格模型进行初始区域分割，以得到初始分割结果，包括：
    随机选取任意预设数量个数的选定三角形，并将每个选定三角形所在平面作为对应区域的初始拟合平面；
    根据三角网格模型中的其他三角形与所述初始拟合平面的接近程度，将逼近同一初始拟合平面的其他三角形聚成一类，以得到多个聚类；
    所述其他三角形与所述初始拟合平面的接近程度利用拟合误差表示：
    Figure PCTCN2022074862-appb-100001
    其中，t表示其他三角形，P i表示初始拟合平面；n i表示其他三角形的法向量，i表示其他三角形的个数，n(x)表示初始拟合平面的法向量；
    对每个所述聚类重新进拟合以得到最佳拟合平面；
    重复进行聚类和拟合过程，直到全局拟合误差最小为止，以得到分块区域；
    每个区域的区域拟合误差表示为：
    Figure PCTCN2022074862-appb-100002
    其中，R i表示分割好的第i个区域；t j表示区域R i中的第j个三角形；
    全局拟合误差表示为：
    Figure PCTCN2022074862-appb-100003
    其中，n表示所有区域的总数量。
  4. 根据权利要求3所述的三维模型简化方法，其特征在于，所述基于所述初始分割结果进行自适应区域分割，以得到多个分块区域，包括：
    对所述初始分割结果进行判断，以确定所述预设数量的大小；
    若所述预设数量较小，则进行新区域插入操作；
    若所述预设数量较大，则进行相邻区域合并操作。
  5. 根据权利要求4所述的三维模型简化方法，其特征在于，所述若所述预设数量较小，则进行新区域插入操作，包括：
    若所述预设数量较小，则获取区域拟合误差最大的分块区域；
    分别获取所述分块区域中拟合误差最大的最大三角形和拟合误差最小的最小三角形；
    若所述最大三角形的拟合误差大于预设的插入阈值，则将所述最大三角形和最小三角形分别作为种子区域，进行区域生长。
  6. 根据权利要求4所述的三维模型简化方法，其特征在于，所述若所述预设数量较大，则进行相邻区域合并操作，包括：
    遍历相邻的分块区域，并获取每个分块区域的拟合平面的法向量；
    判断两个法向量的夹角是否小于预设的第一合并阈值；
    如果两个法向量的夹角小于预设的第一合并阈值，则将相邻的两个分块区域进行合并，并重新进行平面拟合。
  7. 根据权利要求4所述的三维模型简化方法，其特征在于，所述若所述预设数量较大，则进行相邻区域合并操作，包括：
    遍历相邻的分块区域，并计算合并后增大的拟合误差量，所述拟合误差量为合并后的区域拟合误差，减去相邻的分块区域的区域拟合误差之和；
    判断所述拟合误差量是否大于预设的第二合并阈值；
    如果所述拟合误差量不大于预设的第二合并阈值，则将相邻的两个分块区域进行合并，并重新进行平面拟合。
  8. 根据权利要求4所述的三维模型简化方法，其特征在于，所述识别所述分块区域的分割边界，并对所述分割边界进行顶点简化，包括：
    识别分割边界上的边界顶点；
    将任意边界顶点作为待判定顶点，判断所述待判定顶点与相邻的两个边界顶点之间的夹角是否大于预设的夹角阈值；
    如果所述待判定顶点与相邻的两个边界顶点之间的夹角大于预设的夹角阈值，则剔除 所述待判定顶点；
    如果所述待判定顶点与相邻的两个边界顶点之间的夹角不大于预设的夹角阈值，则保留所述待判定顶点，以实现顶点简化。
  9. 一种三维模型简化装置，其特征在于，所述装置包括：
    数据获取模块，配置成用于获取家具的三角网格模型；
    区域分割模块，配置成用于对所述三角网格模型进行区域分割，以得到多个分块区域；
    顶点简化模块，配置成用于识别所述分块区域的分割边界，并对所述分割边界进行顶点简化；
    三角化模块，配置成用于对简化后的边界顶点进行三角剖分，以得到简化模型。
  10. 根据权利要求9所述的三维模型简化装置，其特征在于，所述区域分割模块包括：
    初始区域分割模块，配置成用于对所述三角网格模型进行初始区域分割，以得到初始分割结果；以及
    自适应分割模块，配置成用于基于所述初始分割结果进行自适应区域分割，以得到多个分块区域。
  11. 根据权利要求10所述的三维模型简化装置，其特征在于，所述初始区域分割模块包括：
    初始平面确定模块，配置成用于随机选取任意预设数量个数的选定三角形，并将每个选定三角形所在平面作为对应区域的初始拟合平面；
    聚类模块，配置成用于根据三角网格模型中的其他三角形与所述初始拟合平面的接近程度，将逼近同一初始拟合平面的其他三角形聚成一类，以得到多个聚类；
    拟合模块，配置成用于对每个所述聚类重新进拟合以得到最佳拟合平面；
    迭代模块，配置成用于重复进行聚类和拟合过程，直到全局拟合误差最小为止，以得到分块区域。
  12. 根据权利要求11所述的三维模型简化装置，其特征在于，所述自适应分割模块包括插入模块以及第一合并模块和第二合并模块，所述自适应分割模块还配置成：对所述初始分割结果进行判断，以确定所述预设数量的大小；
    所述插入模块，配置成用于：若所述预设数量较小，则进行新区域插入操作；
    所述第一合并模块和所述第二合并模块，配置成用于：若所述预设数量较大，则进行相邻区域合并操作。
  13. 根据权利要求12所述的三维模型简化装置，其特征在于，所述插入模块还配置成用于：
    若所述预设数量较小，则获取区域拟合误差最大的分块区域；
    分别获取所述分块区域中拟合误差最大的最大三角形和拟合误差最小的最小三角形；
    若所述最大三角形的拟合误差大于预设的插入阈值，则将所述最大三角形和最小三角形分别作为种子区域，进行区域生长。
  14. 根据权利要求12所述的三维模型简化装置，其特征在于，所述第一合并模块配置成用于：
    遍历相邻的分块区域，并获取每个分块区域的拟合平面的法向量；
    判断两个法向量的夹角是否小于预设的第一合并阈值；
    如果两个法向量的夹角小于预设的第一合并阈值，则将相邻的两个分块区域进行合并，并重新进行平面拟合。
  15. 根据权利要求12所述的三维模型简化装置，其特征在于，所述第二合并模块配置成用于：
    遍历相邻的分块区域，并计算合并后增大的拟合误差量，所述拟合误差量为合并后的区域拟合误差，减去相邻的分块区域的区域拟合误差之和；
    判断所述拟合误差量是否大于预设的第二合并阈值；
    所述拟合误差量不大于预设的第二合并阈值，则将相邻的两个分块区域进行合并，并重新进行平面拟合。
  16. 根据权利要求12至15中任一项所述的三维模型简化装置，其特征在于，所述顶点简化模块包括：
    识别模块，配置成用于识别分割边界上的边界顶点；
    判断模块，配置成用于将任意边界顶点作为待判定顶点，判断所述待判定顶点与相邻的两个边界顶点之间的夹角是否大于预设的夹角阈值；
    如果所述待判定顶点与相邻的两个边界顶点之间的夹角大于预设的夹角阈值，则剔除所述待判定顶点；
    如果所述待判定顶点与相邻的两个边界顶点之间的夹角不大于预设的夹角阈值，则保留所述待判定顶点，以实现顶点简化。
  17. 一种可读存储介质，其特征在于，所述可读存储介质中存储有计算机程序指令，所述计算机程序指令被一处理器读取并运行时，执行权利要求1至8中任一项所述的三维模型简化方法。
PCT/CN2022/074862 2021-06-30 2022-01-29 一种三维模型简化方法、装置及存储介质 WO2023273339A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110733997.1A CN113379924B (zh) 2021-06-30 2021-06-30 一种三维模型简化方法、装置及存储介质
CN202110733997.1 2021-06-30

Publications (1)

Publication Number Publication Date
WO2023273339A1 true WO2023273339A1 (zh) 2023-01-05

Family

ID=77580285

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/074862 WO2023273339A1 (zh) 2021-06-30 2022-01-29 一种三维模型简化方法、装置及存储介质

Country Status (2)

Country Link
CN (1) CN113379924B (zh)
WO (1) WO2023273339A1 (zh)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116186864A (zh) * 2023-04-24 2023-05-30 中南冶勘资源环境工程有限公司 一种基于bim技术的深基坑模型快速建模方法及系统
CN116580048A (zh) * 2023-07-12 2023-08-11 武汉峰岭科技有限公司 一种提取房屋倾斜模型上直角房屋的轮廓线的方法及系统
CN116740102A (zh) * 2023-06-15 2023-09-12 广州大学 一种基于测地循环的流形人脸网格分割与提取方法
CN116977530A (zh) * 2023-07-11 2023-10-31 优酷网络技术（北京）有限公司 三维模型的处理方法、装置、电子设备及介质
CN117058668A (zh) * 2023-10-10 2023-11-14 中冶武勘智诚(武汉)工程技术有限公司 一种三维模型减面评估方法及装置
CN117113478A (zh) * 2023-07-28 2023-11-24 中水淮河规划设计研究有限公司 一种bim模型轻量化方法
CN117251906A (zh) * 2023-09-01 2023-12-19 深圳图为技术有限公司 一种流程工厂的三维设备模型轻量化方法和系统
CN117315375A (zh) * 2023-11-20 2023-12-29 腾讯科技（深圳）有限公司 虚拟部件分类方法、装置、电子设备及可读存储介质
CN118097069A (zh) * 2024-04-29 2024-05-28 中铁四局集团有限公司 山区地形网格模型的植被过滤方法、装置及计算机设备

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113379924B (zh) * 2021-06-30 2023-06-09 广东三维家信息科技有限公司 一种三维模型简化方法、装置及存储介质
CN113838212A (zh) * 2021-09-22 2021-12-24 杭州趣村游文旅集团有限公司 一种数字乡村三维模型的区块拼接方法
CN114399583B (zh) * 2021-12-03 2024-07-19 聚好看科技股份有限公司 一种基于几何的三维模型拼接方法及装置
CN114494641B (zh) * 2022-01-06 2023-04-28 广州市城市规划勘测设计研究院 一种三维模型轻量化方法及装置
CN117115391B (zh) * 2023-10-24 2024-01-12 中科云谷科技有限公司 模型更新方法、装置、计算机设备及计算机可读存储介质
CN117557740B (zh) * 2024-01-10 2024-04-09 四川见山科技有限责任公司 三维模型分割层级切换方法、装置、电子设备及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2002342785A (ja) * 2001-05-15 2002-11-29 Ricoh Co Ltd 三角形メッシュ簡単化装置およびプログラム
CN101877147A (zh) * 2010-06-29 2010-11-03 浙江大学 三维三角形网格模型的简化算法
CN106504330A (zh) * 2016-09-21 2017-03-15 中国科学院自动化研究所 基于最小角消除的三角形网格曲面的重新网格化方法
CN109785443A (zh) * 2018-12-21 2019-05-21 博迈科海洋工程股份有限公司 一种针对大型海洋工程装备的三维模型简化方法
CN113379924A (zh) * 2021-06-30 2021-09-10 广东三维家信息科技有限公司 一种三维模型简化方法、装置及存储介质

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9373192B2 (en) * 2013-06-12 2016-06-21 Google Inc. Shape preserving mesh simplification
CN107085865B (zh) * 2017-05-12 2020-10-16 杭州电子科技大学 应用于有限元分析的四边形分割方法
CN108961411B (zh) * 2018-07-02 2023-04-18 南京大学 一种保持外观特征的复杂三维建筑物模型简化方法
CN109801299A (zh) * 2019-01-22 2019-05-24 中国科学院大学 基于二次曲面拟合的模型的交互式分割方法、系统、装置
CN111696111B (zh) * 2020-06-15 2023-04-18 重庆大学 一种基于ssdf衰减图聚类的3d模型网格分割方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2002342785A (ja) * 2001-05-15 2002-11-29 Ricoh Co Ltd 三角形メッシュ簡単化装置およびプログラム
CN101877147A (zh) * 2010-06-29 2010-11-03 浙江大学 三维三角形网格模型的简化算法
CN106504330A (zh) * 2016-09-21 2017-03-15 中国科学院自动化研究所 基于最小角消除的三角形网格曲面的重新网格化方法
CN109785443A (zh) * 2018-12-21 2019-05-21 博迈科海洋工程股份有限公司 一种针对大型海洋工程装备的三维模型简化方法
CN113379924A (zh) * 2021-06-30 2021-09-10 广东三维家信息科技有限公司 一种三维模型简化方法、装置及存储介质

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
DAVID COHEN-STEINER ; PIERRE ALLIEZ ; MATHIEU DESBRUN: "Variational shape approximation", ACM TRANSACTIONS ON GRAPHICS, vol. 23, no. 3, 1 August 2004 (2004-08-01), US , pages 905 - 914, XP058213756, ISSN: 0730-0301, DOI: 10.1145/1015706.1015817 *
JIA, NING: "Point Cloud Data Segmentation Based on Quadric Surface Approach", INFORMATION SCIENCE AND TECHNOLOGY, CHINESE MASTER’S THESES FULL-TEXT DATABASE, 15 January 2009 (2009-01-15), pages 1 - 64, XP093019703, ISSN: 1674-0246 *
JIN YONG,WU QING-BIAO,LIU LI-GANG: "Efficient Algorithm for Surface Simplification Based OU Variational Mesh", JOURNAL OF SOFTWARE, vol. 5, no. 22, 15 May 2011 (2011-05-15), pages 1097 - 1105, XP093019724, ISSN: 1001.2011, DOI: 10.3724/SP.J.1001.2011.03750 *
MARTIN SKRODZKI; ERIC ZIMMERMANN; KONRAD POLTHIER: "Variational Shape Approximation of Point Set Surfaces", ARXIV.ORG, 4 November 2020 (2020-11-04), pages 1 - 37, XP081806996, DOI: 10.1016/j.cagd.2020.101875 *

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116186864B (zh) * 2023-04-24 2023-07-07 中南冶勘资源环境工程有限公司 一种基于bim技术的深基坑模型快速建模方法及系统
CN116186864A (zh) * 2023-04-24 2023-05-30 中南冶勘资源环境工程有限公司 一种基于bim技术的深基坑模型快速建模方法及系统
CN116740102B (zh) * 2023-06-15 2024-04-30 广州大学 一种基于测地循环的流形人脸网格分割与提取方法
CN116740102A (zh) * 2023-06-15 2023-09-12 广州大学 一种基于测地循环的流形人脸网格分割与提取方法
CN116977530A (zh) * 2023-07-11 2023-10-31 优酷网络技术（北京）有限公司 三维模型的处理方法、装置、电子设备及介质
CN116580048A (zh) * 2023-07-12 2023-08-11 武汉峰岭科技有限公司 一种提取房屋倾斜模型上直角房屋的轮廓线的方法及系统
CN116580048B (zh) * 2023-07-12 2023-09-26 武汉峰岭科技有限公司 一种提取房屋倾斜模型上直角房屋的轮廓线的方法及系统
CN117113478A (zh) * 2023-07-28 2023-11-24 中水淮河规划设计研究有限公司 一种bim模型轻量化方法
CN117251906A (zh) * 2023-09-01 2023-12-19 深圳图为技术有限公司 一种流程工厂的三维设备模型轻量化方法和系统
CN117058668A (zh) * 2023-10-10 2023-11-14 中冶武勘智诚(武汉)工程技术有限公司 一种三维模型减面评估方法及装置
CN117058668B (zh) * 2023-10-10 2024-02-02 中冶武勘智诚(武汉)工程技术有限公司 一种三维模型减面评估方法及装置
CN117315375B (zh) * 2023-11-20 2024-03-01 腾讯科技（深圳）有限公司 虚拟部件分类方法、装置、电子设备及可读存储介质
CN117315375A (zh) * 2023-11-20 2023-12-29 腾讯科技（深圳）有限公司 虚拟部件分类方法、装置、电子设备及可读存储介质
CN118097069A (zh) * 2024-04-29 2024-05-28 中铁四局集团有限公司 山区地形网格模型的植被过滤方法、装置及计算机设备

Also Published As

Publication number Publication date
CN113379924B (zh) 2023-06-09
CN113379924A (zh) 2021-09-10

Similar Documents

Publication Publication Date Title
WO2023273339A1 (zh) 一种三维模型简化方法、装置及存储介质
US20210173966A1 (en) Method for constructing curve of robot processing path of part with small curvature based on point cloud boundary
US7760201B2 (en) Transfer of attributes between geometric surfaces of arbitrary topologies with distortion reduction and discontinuity preservation
Ji et al. A novel simplification method for 3D geometric point cloud based on the importance of point
Xiao et al. CAD mesh model segmentation by clustering
CN109685080A (zh) 基于霍夫变换与区域生长的多尺度平面提取方法
CN102938161B (zh) 一种基于Mean Shift的三维形状自动分割方法
Weber et al. Sharp feature preserving MLS surface reconstruction based on local feature line approximations
Branch et al. Automatic hole-filling of triangular meshes using local radial basis function
CN109215112B (zh) 一种单侧点云模型的标注方法
Pound et al. A patch-based approach to 3D plant shoot phenotyping
CN111275724A (zh) 一种基于八叉树和边界优化的机载点云屋顶平面分割方法
CA2925435A1 (en) Methods and systems for generating polycube segmentations from input meshes of objects
Yuan et al. Feature preserving multiresolution subdivision and simplification of point clouds: A conformal geometric algebra approach
CN112669462B (zh) 一种适用于扫描点云的模型处理方法及系统
AU2022292160A1 (en) Pointcloud processing, especially for use with building intelligence modelling (bim)
Zhang et al. Resolving topology ambiguity for multiple-material domains
Zhu et al. 3D reconstruction of plant leaves for high-throughput phenotyping
CN111260596A (zh) 一种具有抗噪的快速图像超像素自动生成方法，设备以及可读存储介质
US7733338B2 (en) Reduction of a mesh with preservation of flow lines
CN115937460A (zh) 基于最优传输的保特征表面重建方法
Ji et al. Point cloud segmentation for complex microsurfaces based on feature line fitting
EP3091512A1 (en) Point cloud simplification
CN112231800B (zh) Bim图形的优化方法、装置以及计算机存储介质
CN114612538A (zh) 一种面向文物三维曲面纹样展平的方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22831173

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 22831173

Country of ref document: EP

Kind code of ref document: A1