3ds Max Rapid manufacturing method based on three-dimensional modeling method and system for fine point cloud
FIELD
The present invention relates to the field of three-dimensional modeling digital city [0001], 3ds Max Fast Fine particularly to a method of modeling a three-dimensional point cloud based and systems.
BACKGROUND
[0002] Currently, digital city modeling there are three main ways: the use of artificial three-dimensional modeling software modeling, using laser point cloud modeling, using aerial stereo modeling. Three-dimensional modeling software such as 3ds Max, AutoCAD, SketchUp to model a traditional way, although they would have great assurance in sophistication on the model, but can not meet the needs of a wide range of rapid prototyping city. Aerial imagery in terms of cost, coverage, degree of automation on the very edge, as you can by matching the point cloud or other means to extract line features three-dimensional reconstruction of the building, but in terms of texture reconstruction, because the only look-down aerial imagery, often blocking, shooting dead phenomenon, leading to the side of the building texture information is missing, the impact reconstruction speed and quality. Existing aviation stereo pair using the modeling software products, most are also the only building a simple building, editing model modification support is relatively weak.
[0003] laser point cloud contains a wealth of information on natural and man-made objects, including vegetation, buildings, roads, green spaces and other feature information, but due to the complexity of surface features, how to extract from a number of laser point cloud information to be out of the building has been a hot research at home and abroad. The need to use of Airborne Laser Lei Dati withdrawn building model can be divided into two steps: building a building model generation detecting cassette. Building contour mapping the city and establish the basis of three-dimensional building model, building contour extraction based on the point cloud data is a major current hot and difficult point cloud data processing. The general method of extracting a building point cloud data is based on the point cloud data to raster image into a depth of DSM, to further seek regularization of nDSM, re-use segmentation algorithm and image boundary line extraction algorithms such as image analysis means to achieve profile line extraction. Templeton University Alhartthy, who designed the building three-dimensional information extraction carried out by filtering the progressive method; Dutch Delft University Vosselman, who use 3D Hough transform to obtain building plan, in order to achieve the purpose of building a three-dimensional reconstruction. Li Shukai other images extracted using DSM binding edge construction, the embodiment is divided into three steps, followed by laser ranging point analysis, analysis and building shadow boundary reconstruction. Especially in red laser dot on the proposed construction of profile-based building closest to predict the direction of the azimuth angle of the laser spot is separated from the edge of the same building, the characteristics of the laser separation of the boundary of the building in accordance with the rules of edge treatment. Simultaneously connecting edge points according to the change of direction angle packet boundary points, and then calculates the main direction of the building based on packet boundary point, the edge of the rule on the main direction of the boundary point, the final outer contour of the most buildings.
[0004] The use of the laser from point cloud point cloud data acquired by three-dimensional laser scanner, capable of scanning a complete physical and accurately reconstruct the data. But the practical application because the three-dimensional laser scan point cloud dense, solid information hidden feature extraction difficult. Three-dimensional modeling software for the carrying capacity of the laser point cloud is limited, and the laser point cloud data typically require pretreatment, hence the need for point cloud processing software Remove noise, reducing the loading point cloud. While processing point cloud point cloud generation plane sheet, the surface sheet introduced into the modeling software assisted modeling. Import and export data to and fro in point cloud processing software systems and three-dimensional modeling system, repeated manual checking, work is very cumbersome and inefficient.
[SUMMARY]
[0005] The present invention utilizes main airborne or vehicle-mounted lidar point cloud data for three-dimensional modeling urban buildings housing, in order to exhibit high-precision position of the point cloud fine modeling geometric information and modeling software, the present invention provides a kinds of technical solutions based on 3ds Max modeling fast fine point cloud.
[0006] To achieve the above object, the present invention provides the following technical solutions:
one kind of fine 3D 3ds Max flash point cloud based modeling method, comprising the steps of:
Step a, the point cloud pretreatment, including first or airborne vehicle lidar acquired point cloud point cloud data denoising, and then quickly build Κ-D clustering tree, then thinning the point cloud;
the fast clustering, including the point cloud data dENOISING Construction triangulated irregular network model to calculate the normal vector of each face irregular triangular network model, having a set of points with a plane similar to the clustering method to the point value;
the construct Κ-D tree, including the quick poly class results, for each category trees were constructed Κ-D;
step two, a thumbnail 3ds Max load point cloud, including providing all based 3ds Max point cloud according to the center point of each category in the cluster condensing a step sketch;
step three, generate a local point cloud, including when the user selects a thumbnail image point obtained in step two, the use Κ-D tree index searched point adjacent to the point, the point cloud to generate a local user is interested;
step four And a measuring point correcting user survey lines, guides the user to draw the outline of a building, comprising measuring point when the user, the user extracting corresponding feature point correction measurement point, when the user performs the survey line, the user extracts the corresponding measuring point correcting characteristic line, implementations below,
the step of providing three point cloud obtained locally in 3ds Max operation window when the user when the measuring point used to adjust the capture angle 3ds Max select building corners, the category selected by the user through the building corner where Κ-D Get tree cloud points in the neighborhood, corner detection algorithm to extract from the neighborhood of the point cloud to which all of the corner points as feature points;
Cloud points in the neighborhood of the midpoint plane cloud data obtained by fitting the plane which, as the feature line;
displayed and selected building corner nearest feature point selected by the user, the user to achieve the correction measuring point, to provide the user building corners corrected; when a building corner the user for starting the correction from a measuring line, to capture and retrieve the corresponding characteristic line at a corner of a building, to achieve the correction user survey line, draw the outline of the building edges ;
step 5 for fine modeling 3ds Max, including complete modeling based 3ds Max The resulting building outline.
Implementations [0007] Further, as the fitting planar, using a random selection of three points is calculated to be the initial value of the fitting parameter plane, and then look for other points in the point set in accordance with the initial value of the parameter, determining other comprising distance from the plane to the point, from the point within the preset threshold value also belong to the plane of the point set.
[0008] The present invention further provides a corresponding rapid 3ds Max fine 3D point cloud based modeling system, comprising the following modules:
point cloud pre-processing module, for the first point cloud of airborne laser radar or vehicle will be acquired denoising cloud point, and then quickly build Κ-D clustering tree, then thinning the point cloud;
the fast clustering, including the point cloud data dENOISING triangulated irregular network model is constructed to calculate the irregular each cam surface normal vector of the triangular mesh model, having a similar method to the same plane to the point set clustering magnitude;
the construct Κ-D tree, comprising a rapid clustering results, were constructed for each category Κ-D tree;
thumbnail loading module configured to provide all the thumbnails based 3ds Max the center point of each point cloud point cloud category after pre-clustering module obtained;
Local point cloud generating module configured to include a thumbnail when the user loads a selected thumbnail resulting module point by Κ-D tree index searched point adjacent to the point, the point cloud generating local interest to the user; test user measuring point and line correction module configured to guide the user to draw the outline of buildings, including the measuring point when the user, the user extracting corresponding feature point correction measurement point, when the user performs the survey line, the user extracts the corresponding measuring point correcting characteristic line, to achieve as follows,
provided the local point cloud point cloud generating module resulting in a local window 3ds Max operation, when the user performs measurement points used to adjust the capture angle 3ds Max building corner points selected by the user selecting the categories where the building corner obtaining a KD tree cloud points in the neighborhood, corner detection algorithm to extract from the neighborhood of the point cloud to which all of the corner points as feature points;
on cloud points in the neighborhood plane midpoint cloud data obtained by fitting a plane cross line, a characteristic line;
display and select the nearest corner building feature point selected by the user, the user to achieve the correction measuring point is provided to a user after correction Buildings corners; when a user from a building corner after correction for sensing line, wherein the line capturing and retrieving a respective corner point of the next building, user achieve correction sensing line, draw the outline of the building edges;
fine construction die module, based on the modeling according to the completed 3ds Max resultant building outline.
[0009] Furthermore, the
FIELD
The present invention relates to the field of three-dimensional modeling digital city [0001], 3ds Max Fast Fine particularly to a method of modeling a three-dimensional point cloud based and systems.
BACKGROUND
[0002] Currently, digital city modeling there are three main ways: the use of artificial three-dimensional modeling software modeling, using laser point cloud modeling, using aerial stereo modeling. Three-dimensional modeling software such as 3ds Max, AutoCAD, SketchUp to model a traditional way, although they would have great assurance in sophistication on the model, but can not meet the needs of a wide range of rapid prototyping city. Aerial imagery in terms of cost, coverage, degree of automation on the very edge, as you can by matching the point cloud or other means to extract line features three-dimensional reconstruction of the building, but in terms of texture reconstruction, because the only look-down aerial imagery, often blocking, shooting dead phenomenon, leading to the side of the building texture information is missing, the impact reconstruction speed and quality. Existing aviation stereo pair using the modeling software products, most are also the only building a simple building, editing model modification support is relatively weak.
[0003] laser point cloud contains a wealth of information on natural and man-made objects, including vegetation, buildings, roads, green spaces and other feature information, but due to the complexity of surface features, how to extract from a number of laser point cloud information to be out of the building has been a hot research at home and abroad. The need to use of Airborne Laser Lei Dati withdrawn building model can be divided into two steps: building a building model generation detecting cassette. Building contour mapping the city and establish the basis of three-dimensional building model, building contour extraction based on the point cloud data is a major current hot and difficult point cloud data processing. The general method of extracting a building point cloud data is based on the point cloud data to raster image into a depth of DSM, to further seek regularization of nDSM, re-use segmentation algorithm and image boundary line extraction algorithms such as image analysis means to achieve profile line extraction. Templeton University Alhartthy, who designed the building three-dimensional information extraction carried out by filtering the progressive method; Dutch Delft University Vosselman, who use 3D Hough transform to obtain building plan, in order to achieve the purpose of building a three-dimensional reconstruction. Li Shukai other images extracted using DSM binding edge construction, the embodiment is divided into three steps, followed by laser ranging point analysis, analysis and building shadow boundary reconstruction. Especially in red laser dot on the proposed construction of profile-based building closest to predict the direction of the azimuth angle of the laser spot is separated from the edge of the same building, the characteristics of the laser separation of the boundary of the building in accordance with the rules of edge treatment. Simultaneously connecting edge points according to the change of direction angle packet boundary points, and then calculates the main direction of the building based on packet boundary point, the edge of the rule on the main direction of the boundary point, the final outer contour of the most buildings.
[0004] The use of the laser from point cloud point cloud data acquired by three-dimensional laser scanner, capable of scanning a complete physical and accurately reconstruct the data. But the practical application because the three-dimensional laser scan point cloud dense, solid information hidden feature extraction difficult. Three-dimensional modeling software for the carrying capacity of the laser point cloud is limited, and the laser point cloud data typically require pretreatment, hence the need for point cloud processing software Remove noise, reducing the loading point cloud. While processing point cloud point cloud generation plane sheet, the surface sheet introduced into the modeling software assisted modeling. Import and export data to and fro in point cloud processing software systems and three-dimensional modeling system, repeated manual checking, work is very cumbersome and inefficient.
[SUMMARY]
[0005] The present invention utilizes main airborne or vehicle-mounted lidar point cloud data for three-dimensional modeling urban buildings housing, in order to exhibit high-precision position of the point cloud fine modeling geometric information and modeling software, the present invention provides a kinds of technical solutions based on 3ds Max modeling fast fine point cloud.
[0006] To achieve the above object, the present invention provides the following technical solutions:
one kind of fine 3D 3ds Max flash point cloud based modeling method, comprising the steps of:
Step a, the point cloud pretreatment, including first or airborne vehicle lidar acquired point cloud point cloud data denoising, and then quickly build Κ-D clustering tree, then thinning the point cloud;
the fast clustering, including the point cloud data dENOISING Construction triangulated irregular network model to calculate the normal vector of each face irregular triangular network model, having a set of points with a plane similar to the clustering method to the point value;
the construct Κ-D tree, including the quick poly class results, for each category trees were constructed Κ-D;
step two, a thumbnail 3ds Max load point cloud, including providing all based 3ds Max point cloud according to the center point of each category in the cluster condensing a step sketch;
step three, generate a local point cloud, including when the user selects a thumbnail image point obtained in step two, the use Κ-D tree index searched point adjacent to the point, the point cloud to generate a local user is interested;
step four And a measuring point correcting user survey lines, guides the user to draw the outline of a building, comprising measuring point when the user, the user extracting corresponding feature point correction measurement point, when the user performs the survey line, the user extracts the corresponding measuring point correcting characteristic line, implementations below,
the step of providing three point cloud obtained locally in 3ds Max operation window when the user when the measuring point used to adjust the capture angle 3ds Max select building corners, the category selected by the user through the building corner where Κ-D Get tree cloud points in the neighborhood, corner detection algorithm to extract from the neighborhood of the point cloud to which all of the corner points as feature points;
Cloud points in the neighborhood of the midpoint plane cloud data obtained by fitting the plane which, as the feature line;
displayed and selected building corner nearest feature point selected by the user, the user to achieve the correction measuring point, to provide the user building corners corrected; when a building corner the user for starting the correction from a measuring line, to capture and retrieve the corresponding characteristic line at a corner of a building, to achieve the correction user survey line, draw the outline of the building edges ;
step 5 for fine modeling 3ds Max, including complete modeling based 3ds Max The resulting building outline.
Implementations [0007] Further, as the fitting planar, using a random selection of three points is calculated to be the initial value of the fitting parameter plane, and then look for other points in the point set in accordance with the initial value of the parameter, determining other comprising distance from the plane to the point, from the point within the preset threshold value also belong to the plane of the point set.
[0008] The present invention further provides a corresponding rapid 3ds Max fine 3D point cloud based modeling system, comprising the following modules:
point cloud pre-processing module, for the first point cloud of airborne laser radar or vehicle will be acquired denoising cloud point, and then quickly build Κ-D clustering tree, then thinning the point cloud;
the fast clustering, including the point cloud data dENOISING triangulated irregular network model is constructed to calculate the irregular each cam surface normal vector of the triangular mesh model, having a similar method to the same plane to the point set clustering magnitude;
the construct Κ-D tree, comprising a rapid clustering results, were constructed for each category Κ-D tree;
thumbnail loading module configured to provide all the thumbnails based 3ds Max the center point of each point cloud point cloud category after pre-clustering module obtained;
Local point cloud generating module configured to include a thumbnail when the user loads a selected thumbnail resulting module point by Κ-D tree index searched point adjacent to the point, the point cloud generating local interest to the user; test user measuring point and line correction module configured to guide the user to draw the outline of buildings, including the measuring point when the user, the user extracting corresponding feature point correction measurement point, when the user performs the survey line, the user extracts the corresponding measuring point correcting characteristic line, to achieve as follows,
provided the local point cloud point cloud generating module resulting in a local window 3ds Max operation, when the user performs measurement points used to adjust the capture angle 3ds Max building corner points selected by the user selecting the categories where the building corner obtaining a KD tree cloud points in the neighborhood, corner detection algorithm to extract from the neighborhood of the point cloud to which all of the corner points as feature points;
on cloud points in the neighborhood plane midpoint cloud data obtained by fitting a plane cross line, a characteristic line;
display and select the nearest corner building feature point selected by the user, the user to achieve the correction measuring point is provided to a user after correction Buildings corners; when a user from a building corner after correction for sensing line, wherein the line capturing and retrieving a respective corner point of the next building, user achieve correction sensing line, draw the outline of the building edges;
fine construction die module, based on the modeling according to the completed 3ds Max resultant building outline.
[0009] Furthermore, the
Said planar implementation of fitting, using a random selection of three points is calculated to be the initial value plane fitting parameters, and then look for other points in the point set in accordance with the initial value of the parameter, comprising determining the distance from the plane other points, distance from the point within the preset threshold value also belong to the plane of the point set.
[0010] The present invention features provided technical solutions are as follows:
1. To provide a secondary city 3DMax fine 3D model building: 3ds Max by AutoDesk introduced professional modeling software, has a perfect interactive modeling engine, has a wealth of editing tool, is currently the most widely used manual modeling tools. In addition, 3ds Max provides a 3ds Max SDK development platform for creating object-oriented extension of the application can support a variety of C ++ and scripting languages available for plug-in development. Accordingly the present invention provides a method may be fused to 3ds Max, full use of its abundant fine model editing capabilities, automatic extraction algorithm results for quick editing.
. [0011] 2 propose a three-dimensional local feature point based planar sheet fitting, characteristic line extraction: extracting local fitting surface and the base sheet on the corners, the exact pickup helps modelers feature points, and tracking characteristics of the user wire drawing process, wherein the automatic positioning to the precise line to ensure high efficiency and high precision modeling.
. [0012] 3 to achieve a rapid pickup local point cloud: clustering results based on fast, avoiding all the cloud point caused by the load operation is not smooth, and can quickly acquire a local optimum point cloud data of the measuring point according to a user perspective, for fast feature point, the feature extraction line, the auxiliary artificial modeling operations.
BRIEF DESCRIPTION
[0013] FIG. 1 is a flowchart of an embodiment of the present invention.
[0014] FIG 2 is a flowchart of pre-processing point cloud embodiment of the present invention.
DETAILED DESCRIPTION OF THE INVENTION
[0015] The following embodiments in conjunction with the accompanying drawings and embodiments of the aspect of the invention is described in detail.
[0016] As shown in FIG. 1, the embodiment of the present invention provides a method of embodiment from the original point cloud model building to specific implementation process is as follows:
Step one, the point cloud pretreatment;
point cloud of airborne laser radar or a vehicle is acquired a series of discrete three-dimensional coordinates of the point, which is discretely distributed irregularly, the huge amount of data and there are a large number of data redundancy, and hence the need for effective denoising index for subsequent tissue processing.
[0017] Example Point Cloud Data pretreatment flowchart Referring to Figure 2, the point cloud comprising denoising, and then quickly build Κ-D clustering tree, then the point cloud thinning.
[0018] In an embodiment, first point cloud denoising removed infinity point (such as the sky) and Journal of points (such as trees, glass, etc.), such a negative effect on the point information data processing and 3D modeling, it is necessary to these null point or point to remove interference. Construction triangulated irregular network (Triangulated Irregular Network, TIN) model, a normal vector is calculated for each small TIN model point cloud cam face after denoising, clusters having similar method to point to the same magnitude plane point set. And, for each category trees were constructed Κ-D, by using the spatial tree data structure Κ-D to organize all of the data, to find all the most recent data from a given cluster centers efficiently. Point cloud data is then thinning, thinning algorithm using embodiments TIN based on the specific embodiments may consider data characteristic feature road, buildings, ground retention feature point, thinning out the data points of the flat road, thereby reducing redundancy data, impact on the accuracy of the data is small.
[0019] Step two, 3ds Max thumbnail loading point cloud;
3ds MAX2015 support point cloud point cloud but the limited load bearing capacity, and therefore the center point of each category according to the embodiment obtained in step a clustering algorithm for clustering in the embodiment all thumbnails provide point cloud to achieve fast thumbnail indexes and navigation functions. Such as man-machine interface in 3ds Max2015 provided is provided in the lower right corner of the window a small thumbnail, reducing 3ds Max load point cloud burden, while providing local fine point cloud browser window, for editors to view and operate.
[0020] Step three, when the user selects a point in the thumbnail, generate a local user is interested in a point cloud;
in step two the resulting thumbnails to the user can quickly locate a region of interest, the clustering algorithm cloud point based on the rapid indexing, the user can quickly view the point cloud to a local interest by directing the thumbnails.
[0021] embodiment, the user profile is drawn building often requires measuring point, measuring line, the user is to select the measuring point trying to build a building corner profile building process, the user is measured from the line along a corner point cloud outline the process of building the sideline. It is accurate operation, first needs to locate a local cloud point of interest to the user. When the user selects a point in the thumbnail, you can use Κ-D tree index management techniques searched near the point of the point, to quickly establish a starting point topology cloud. Reflected in the thumbnail region is to have a rectangular box, rectangular frame is the current large screen display, mouse movement that is dragging the rectangle can change the position of the current map display, which play a role in navigation.
[0022] Step four, the correction measuring point and user survey lines, guides the user to draw the outline of a building, comprising the step of partially based on the obtained three point cloud, when the user performs measurement points, the corresponding feature point extracting amendments the user the measuring point, when the user measuring line, a user extracts the corresponding measuring point correcting characteristic line;
specific implementation, in 3ds Max operation window may be provided a local interest point cloud generated in step three users (in a similar two step operation by a user of the local window point cloud display window), the user performs the measurement points, can open 3ds Max may capture the building, the building may be selected to adjust the angle of the corner points (i.e., the initial result of the user the measuring point) in 3ds Max operation window, and then through corners where the category tree Κ-D point cloud acquired neighborhood, the neighborhood extracting from the corner point by point cloud algorithm can extract all the corner points of the area, as a feature point. Corner extraction algorithm employed prior art, the present invention is not repeated here.
[0023] In order to improve the efficiency of a subsequent drawing, it can be pre-fitted plane:
Based on the acquired through the building corner where possible category tree neighborhood Κ-D point cloud, embodiments employing RANSAC algorithm (RANdom SAmple Consensus, RANSAC) binding eigenvalue point cloud data plane fitting. Parameter estimation is performed when the plane to be fitted, may be employed criterion for a specific data set, using the criterion of this iterative cull those input data and the estimated parameters do not match, then the parameters estimated by the correct input data. The specific approach is to use randomly selected three points is calculated to be the initial value of the parameter-fit plane, and find other points in the point set in accordance with initial parameters. There needs criterion to determine whether the point is a point on the plane, embodiment, the method is used to calculate the distance from the plane point, the threshold value may be set by a person skilled in the actual pre-fitting process to fit plane, the distance set point within a preset threshold value also belong to the plane of the point set.
After the [0024] plane fitting algorithm has been processed, the feature point may be fed back to the user, the user tries to draw graphics, fitting algorithm can quickly find the result of the processing line of intersection of the plane through the building plan point cloud according to the outline of the building, as a characteristic line, to provide protection for subsequent modeling.
[0025] Fixed-point user embodiment, sensing line:
building corner (feature points) after the user selects the point, the point angle extraction algorithm according to the extracted user with the most points selected in view neighbor (3ds Max step three operating window providing a view point cloud obtained locally), highlight automatically captured and the points, for correcting the user the measuring point, after the building corner correction available to the user. The fitting algorithm then fitted plane intersecting line of the plane (characteristic line), when the building corner the user from starting a correction, to automatically capture the line of intersection of the respective planes and retrieved from the perspective of the region to the lower a corner of a building, to achieve the correction user survey line, draw the outline edges of the building, guide the user to draw the entire building outline more precisely.
[0026] Step five, 3ds Max geological modeling;
According to building outlines steps above income, combined with 3ds Max powerful model editing features, users can take advantage of 3dsMax finely edited and perfect model. From the creation of the standard model, modified to stake out and create complex models, to advanced patch modeling, improved modeling capabilities to help users easily express the true reality of the scene.
When [0027] In particular embodiments, the above process can be implemented automatically run computer software technology, the system may provide a modular manner.
[0028] Embodiments of the invention further provides a corresponding point cloud based 3
[0010] The present invention features provided technical solutions are as follows:
1. To provide a secondary city 3DMax fine 3D model building: 3ds Max by AutoDesk introduced professional modeling software, has a perfect interactive modeling engine, has a wealth of editing tool, is currently the most widely used manual modeling tools. In addition, 3ds Max provides a 3ds Max SDK development platform for creating object-oriented extension of the application can support a variety of C ++ and scripting languages available for plug-in development. Accordingly the present invention provides a method may be fused to 3ds Max, full use of its abundant fine model editing capabilities, automatic extraction algorithm results for quick editing.
. [0011] 2 propose a three-dimensional local feature point based planar sheet fitting, characteristic line extraction: extracting local fitting surface and the base sheet on the corners, the exact pickup helps modelers feature points, and tracking characteristics of the user wire drawing process, wherein the automatic positioning to the precise line to ensure high efficiency and high precision modeling.
. [0012] 3 to achieve a rapid pickup local point cloud: clustering results based on fast, avoiding all the cloud point caused by the load operation is not smooth, and can quickly acquire a local optimum point cloud data of the measuring point according to a user perspective, for fast feature point, the feature extraction line, the auxiliary artificial modeling operations.
BRIEF DESCRIPTION
[0013] FIG. 1 is a flowchart of an embodiment of the present invention.
[0014] FIG 2 is a flowchart of pre-processing point cloud embodiment of the present invention.
DETAILED DESCRIPTION OF THE INVENTION
[0015] The following embodiments in conjunction with the accompanying drawings and embodiments of the aspect of the invention is described in detail.
[0016] As shown in FIG. 1, the embodiment of the present invention provides a method of embodiment from the original point cloud model building to specific implementation process is as follows:
Step one, the point cloud pretreatment;
point cloud of airborne laser radar or a vehicle is acquired a series of discrete three-dimensional coordinates of the point, which is discretely distributed irregularly, the huge amount of data and there are a large number of data redundancy, and hence the need for effective denoising index for subsequent tissue processing.
[0017] Example Point Cloud Data pretreatment flowchart Referring to Figure 2, the point cloud comprising denoising, and then quickly build Κ-D clustering tree, then the point cloud thinning.
[0018] In an embodiment, first point cloud denoising removed infinity point (such as the sky) and Journal of points (such as trees, glass, etc.), such a negative effect on the point information data processing and 3D modeling, it is necessary to these null point or point to remove interference. Construction triangulated irregular network (Triangulated Irregular Network, TIN) model, a normal vector is calculated for each small TIN model point cloud cam face after denoising, clusters having similar method to point to the same magnitude plane point set. And, for each category trees were constructed Κ-D, by using the spatial tree data structure Κ-D to organize all of the data, to find all the most recent data from a given cluster centers efficiently. Point cloud data is then thinning, thinning algorithm using embodiments TIN based on the specific embodiments may consider data characteristic feature road, buildings, ground retention feature point, thinning out the data points of the flat road, thereby reducing redundancy data, impact on the accuracy of the data is small.
[0019] Step two, 3ds Max thumbnail loading point cloud;
3ds MAX2015 support point cloud point cloud but the limited load bearing capacity, and therefore the center point of each category according to the embodiment obtained in step a clustering algorithm for clustering in the embodiment all thumbnails provide point cloud to achieve fast thumbnail indexes and navigation functions. Such as man-machine interface in 3ds Max2015 provided is provided in the lower right corner of the window a small thumbnail, reducing 3ds Max load point cloud burden, while providing local fine point cloud browser window, for editors to view and operate.
[0020] Step three, when the user selects a point in the thumbnail, generate a local user is interested in a point cloud;
in step two the resulting thumbnails to the user can quickly locate a region of interest, the clustering algorithm cloud point based on the rapid indexing, the user can quickly view the point cloud to a local interest by directing the thumbnails.
[0021] embodiment, the user profile is drawn building often requires measuring point, measuring line, the user is to select the measuring point trying to build a building corner profile building process, the user is measured from the line along a corner point cloud outline the process of building the sideline. It is accurate operation, first needs to locate a local cloud point of interest to the user. When the user selects a point in the thumbnail, you can use Κ-D tree index management techniques searched near the point of the point, to quickly establish a starting point topology cloud. Reflected in the thumbnail region is to have a rectangular box, rectangular frame is the current large screen display, mouse movement that is dragging the rectangle can change the position of the current map display, which play a role in navigation.
[0022] Step four, the correction measuring point and user survey lines, guides the user to draw the outline of a building, comprising the step of partially based on the obtained three point cloud, when the user performs measurement points, the corresponding feature point extracting amendments the user the measuring point, when the user measuring line, a user extracts the corresponding measuring point correcting characteristic line;
specific implementation, in 3ds Max operation window may be provided a local interest point cloud generated in step three users (in a similar two step operation by a user of the local window point cloud display window), the user performs the measurement points, can open 3ds Max may capture the building, the building may be selected to adjust the angle of the corner points (i.e., the initial result of the user the measuring point) in 3ds Max operation window, and then through corners where the category tree Κ-D point cloud acquired neighborhood, the neighborhood extracting from the corner point by point cloud algorithm can extract all the corner points of the area, as a feature point. Corner extraction algorithm employed prior art, the present invention is not repeated here.
[0023] In order to improve the efficiency of a subsequent drawing, it can be pre-fitted plane:
Based on the acquired through the building corner where possible category tree neighborhood Κ-D point cloud, embodiments employing RANSAC algorithm (RANdom SAmple Consensus, RANSAC) binding eigenvalue point cloud data plane fitting. Parameter estimation is performed when the plane to be fitted, may be employed criterion for a specific data set, using the criterion of this iterative cull those input data and the estimated parameters do not match, then the parameters estimated by the correct input data. The specific approach is to use randomly selected three points is calculated to be the initial value of the parameter-fit plane, and find other points in the point set in accordance with initial parameters. There needs criterion to determine whether the point is a point on the plane, embodiment, the method is used to calculate the distance from the plane point, the threshold value may be set by a person skilled in the actual pre-fitting process to fit plane, the distance set point within a preset threshold value also belong to the plane of the point set.
After the [0024] plane fitting algorithm has been processed, the feature point may be fed back to the user, the user tries to draw graphics, fitting algorithm can quickly find the result of the processing line of intersection of the plane through the building plan point cloud according to the outline of the building, as a characteristic line, to provide protection for subsequent modeling.
[0025] Fixed-point user embodiment, sensing line:
building corner (feature points) after the user selects the point, the point angle extraction algorithm according to the extracted user with the most points selected in view neighbor (3ds Max step three operating window providing a view point cloud obtained locally), highlight automatically captured and the points, for correcting the user the measuring point, after the building corner correction available to the user. The fitting algorithm then fitted plane intersecting line of the plane (characteristic line), when the building corner the user from starting a correction, to automatically capture the line of intersection of the respective planes and retrieved from the perspective of the region to the lower a corner of a building, to achieve the correction user survey line, draw the outline edges of the building, guide the user to draw the entire building outline more precisely.
[0026] Step five, 3ds Max geological modeling;
According to building outlines steps above income, combined with 3ds Max powerful model editing features, users can take advantage of 3dsMax finely edited and perfect model. From the creation of the standard model, modified to stake out and create complex models, to advanced patch modeling, improved modeling capabilities to help users easily express the true reality of the scene.
When [0027] In particular embodiments, the above process can be implemented automatically run computer software technology, the system may provide a modular manner.
[0028] Embodiments of the invention further provides a corresponding point cloud based 3
ds Max Rapid fine three-dimensional modeling system comprising the following modules:
point cloud pre-processing module, for the first data point cloud of airborne or vehicle-mounted lidar acquired point cloud denoising is performed, and then fast clustering build Κ- D tree, then thinning the point cloud;
the fast clustering, including the point cloud data dENOISING construct triangulated irregular network model to calculate the normal vector of each face of the TIN model, having similar to clustering method to the magnitude of the set point the same plane;
the construct Κ-D tree, comprising a fast clustering results for each category trees were constructed Κ-D;
thumbnail loading module, based on 3ds Max thumbnails, all according to the center point of the point cloud for each category resulting after pre-clustering module point cloud;
local point cloud generating module configured to include a thumbnail when the user loads the resulting module select a point in the thumbnail by Κ-D tree index searched point adjacent to the point, the point cloud generating local interest to the user;
the user the measuring point and the measuring line correction module configured to guide the user to draw the outline of buildings, including when the user Line measuring points, extracting corresponding feature point correcting the user the measuring point, when the user performs the survey line, extracts the corresponding characteristic line correction user measuring points, for the following manner,
provided the local point cloud generation resulting module local point clouds in 3ds Max operation window, when user measuring point used when adjusting the angle 3ds Max capture select building corner, the category selected by the user through the building corner where KD neighborhood tree acquiring point cloud extracted from the corner detection algorithm neighborhood point cloud wherein all of the corner points as feature points;
Cloud points in the neighborhood of the midpoint plane cloud data obtained by fitting the plane which, as the feature line; displayed and selected building corner nearest feature point selected by the user, the user to achieve the correction measuring point, to provide the user building corners corrected; when a building corner the user for starting the correction from a measuring line, to capture and retrieve the corresponding characteristic line at a corner of a building, to achieve the correction user survey line, draw the outline of the building edges ;
geological modeling means for modeling according to the completed based 3ds Max resultant building outline.
Specific Example [0029] described herein is merely illustrative for spirit of the invention. Those skilled in the art of the present invention can be made to the specific embodiments described various modifications or additions, or a similar alternative embodiment, but without departing from the spirit of the invention or exceed defined in the appended claims range.
[Items] sovereignty
A fine fast 3dS Max dimensional modeling method based on the point cloud, which is characterized in that it comprises the following steps: a step, pretreatment cloud point, before the point cloud data including airborne or vehicle-mounted lidar acquired point- cloud denoising, and then quickly build Κ-D clustering tree, then thinning the point cloud; the fast clustering, including the point cloud data dENOISING triangulated irregular network model is constructed to calculate TIN each normal vector of the triangular face mesh model having a similar method to the same plane to the set point value of clustering; Κ-D tree of the construct, comprising the fast clustering results for each category were constructed K0 -D tree; two step, a point cloud thumbnail 3ds Max load, comprising providing a thumbnail based on the entire point cloud 3ds Max according to the center point of each category in a clustering step; step three, generate a local point cloud, including when the user selects a thumbnail image point obtained in step two, the use Κ-D tree index searched point adjacent to the point, the point cloud to generate a local user is interested; step four, the correction measuring point and user survey line, the guide use User profile the drawing of a building, comprising measuring point when the user, the user extracting corresponding feature point correction measurement point, when the user performs the survey line, extracts the corresponding user characteristic line correction measurement points, for the following manner, provided in the operation window 3ds Max step three resulting partial cloud point, when the user when the measuring point employed to adjust the capture angle 3ds Max select building corner, the category selected by the user through the building corner where Κ-D neighborhood tree acquiring point cloud, by angle point extraction algorithm extracts from the point cloud to the neighborhood where all the corner points as feature points; cloud points in the neighborhood of the midpoint plane cloud data obtained by fitting the plane which, as the feature line; display and select user building corners nearest the selected feature points, for correcting the user the measuring point, there is provided a building corner point corrected to a user; building corners when the user carried out starting from a sensing line correction, to capture the corresponding characteristic line and retrieving the next corner of a building, to achieve user sensing line correction, edge contouring building; step 5 for fine 3ds Max modeling, including those based on 3ds M ax to complete the modeling of income according to building outlines. The rapid 3dsMax fine 3D point cloud based modeling method of claim 1, wherein: The implementation is a planar fitting, using a random selection of three points is calculated to be the initial value plane fitting parameters, and then look for other points in the point set in accordance with the initial value of the parameter, comprising determining the distance from the plane to the other points , from within a preset threshold point also belongs to the plane of the point set. 3.- 3ds Max Rapid kinds of fine 3D point cloud based modeling system, characterized by comprising the following modules: point cloud pre-processing module, for the first point cloud data acquired by airborne laser radar performs point cloud denoising , and then quickly build Κ-D clustering tree, then thinning the point cloud; the fast clustering, including the point cloud data dENOISING triangulated irregular network model is constructed, triangulated irregular network model is calculated each cam surface normal vector, having a similar method to the same plane to the point set clustering magnitude; Κ-D tree of the construct, comprising the fast clustering results for each category trees were constructed Κ-D ; thumbnail loading module configured to provide all the thumbnails based 3ds Max point cloud according to the center point of each category obtained after pre-clustering module point cloud; local point cloud generating module, a thumbnail image when a user comprising loading a selected thumbnail resulting module point by Κ-D tree index searched neighboring points to the point, to generate a local cloud point of interest to the user; measuring point and the measuring line user correction module configured to guide the user through built Draw object outline, including when the user performs measurement points, extracting corresponding feature point correcting the user the measuring point, when the user performs the survey line, extracts the corresponding characteristic line correction user measuring points, for the following manner, provided the local point cloud in 3ds Max operation window the resulting point cloud generating module partially, when the user when adjusting the angle 3ds Max capture select building corner, the category selected by the user through the building corner where KD neighborhood tree acquiring point cloud measuring point employed, by extracting corner points neighbor algorithm is extracted from the point cloud to which all of the corner points as feature points; cloud points in the neighborhood of the midpoint plane cloud data obtained by fitting the plane which, as the feature line; select and display the selected user building corners nearest feature point, to achieve user-point correction, there is provided a building corner point corrected to a user; building corners when the user carried out starting from a sensing line correction, to capture and retrieve the corresponding characteristic line the corner of a building, to achieve the correction user survey line, draw the outline of the building edges; fine modeling module, based on the resulting construction according to 3ds Max Complete contour modeling. The point cloud based 3ds Max Rapid fine 3D modeling system of claim 3, wherein:
Patent Abstract The present invention provides a point cloud based 3ds? Max Rapid fine three-dimensional modeling method and system, comprising a pretreatment point cloud, point cloud data including the first airborne or vehicle-mounted lidar acquired point cloud denoising is performed, and then build KD fast clustering tree, then point clouds thinning; based on 3ds? The center point of each category by Max after all clustering thumbnail point cloud; generating a local point cloud, when the user selects the building corner points, cloud points in the neighborhood by acquiring the category where a KD tree using corner extraction neighbor algorithm is extracted from the point cloud to which all of the corner points as feature points, cloud points in the neighborhood of the midpoint plane cloud data obtained by fitting the plane which, as the feature line; correction feature point according to the user the measuring point, when a user survey line, search contouring edges of the building, amendments user survey line; based on 3ds? Max modeling is completed according to building outlines.
[IPC classification] G06T19 / 00, G06T5 / 00
[] Publication No. CN105354883
[Application Number] CN201510831919
[inventor] Huang Yuchun, Sun Hui, Ding Yalan, Yuan Ye
[the applicant] Wuhan University
[Publication date] February 24, 2016 date
[date of application] November 25, 2015
point cloud pre-processing module, for the first data point cloud of airborne or vehicle-mounted lidar acquired point cloud denoising is performed, and then fast clustering build Κ- D tree, then thinning the point cloud;
the fast clustering, including the point cloud data dENOISING construct triangulated irregular network model to calculate the normal vector of each face of the TIN model, having similar to clustering method to the magnitude of the set point the same plane;
the construct Κ-D tree, comprising a fast clustering results for each category trees were constructed Κ-D;
thumbnail loading module, based on 3ds Max thumbnails, all according to the center point of the point cloud for each category resulting after pre-clustering module point cloud;
local point cloud generating module configured to include a thumbnail when the user loads the resulting module select a point in the thumbnail by Κ-D tree index searched point adjacent to the point, the point cloud generating local interest to the user;
the user the measuring point and the measuring line correction module configured to guide the user to draw the outline of buildings, including when the user Line measuring points, extracting corresponding feature point correcting the user the measuring point, when the user performs the survey line, extracts the corresponding characteristic line correction user measuring points, for the following manner,
provided the local point cloud generation resulting module local point clouds in 3ds Max operation window, when user measuring point used when adjusting the angle 3ds Max capture select building corner, the category selected by the user through the building corner where KD neighborhood tree acquiring point cloud extracted from the corner detection algorithm neighborhood point cloud wherein all of the corner points as feature points;
Cloud points in the neighborhood of the midpoint plane cloud data obtained by fitting the plane which, as the feature line; displayed and selected building corner nearest feature point selected by the user, the user to achieve the correction measuring point, to provide the user building corners corrected; when a building corner the user for starting the correction from a measuring line, to capture and retrieve the corresponding characteristic line at a corner of a building, to achieve the correction user survey line, draw the outline of the building edges ;
geological modeling means for modeling according to the completed based 3ds Max resultant building outline.
Specific Example [0029] described herein is merely illustrative for spirit of the invention. Those skilled in the art of the present invention can be made to the specific embodiments described various modifications or additions, or a similar alternative embodiment, but without departing from the spirit of the invention or exceed defined in the appended claims range.
[Items] sovereignty
A fine fast 3dS Max dimensional modeling method based on the point cloud, which is characterized in that it comprises the following steps: a step, pretreatment cloud point, before the point cloud data including airborne or vehicle-mounted lidar acquired point- cloud denoising, and then quickly build Κ-D clustering tree, then thinning the point cloud; the fast clustering, including the point cloud data dENOISING triangulated irregular network model is constructed to calculate TIN each normal vector of the triangular face mesh model having a similar method to the same plane to the set point value of clustering; Κ-D tree of the construct, comprising the fast clustering results for each category were constructed K0 -D tree; two step, a point cloud thumbnail 3ds Max load, comprising providing a thumbnail based on the entire point cloud 3ds Max according to the center point of each category in a clustering step; step three, generate a local point cloud, including when the user selects a thumbnail image point obtained in step two, the use Κ-D tree index searched point adjacent to the point, the point cloud to generate a local user is interested; step four, the correction measuring point and user survey line, the guide use User profile the drawing of a building, comprising measuring point when the user, the user extracting corresponding feature point correction measurement point, when the user performs the survey line, extracts the corresponding user characteristic line correction measurement points, for the following manner, provided in the operation window 3ds Max step three resulting partial cloud point, when the user when the measuring point employed to adjust the capture angle 3ds Max select building corner, the category selected by the user through the building corner where Κ-D neighborhood tree acquiring point cloud, by angle point extraction algorithm extracts from the point cloud to the neighborhood where all the corner points as feature points; cloud points in the neighborhood of the midpoint plane cloud data obtained by fitting the plane which, as the feature line; display and select user building corners nearest the selected feature points, for correcting the user the measuring point, there is provided a building corner point corrected to a user; building corners when the user carried out starting from a sensing line correction, to capture the corresponding characteristic line and retrieving the next corner of a building, to achieve user sensing line correction, edge contouring building; step 5 for fine 3ds Max modeling, including those based on 3ds M ax to complete the modeling of income according to building outlines. The rapid 3dsMax fine 3D point cloud based modeling method of claim 1, wherein: The implementation is a planar fitting, using a random selection of three points is calculated to be the initial value plane fitting parameters, and then look for other points in the point set in accordance with the initial value of the parameter, comprising determining the distance from the plane to the other points , from within a preset threshold point also belongs to the plane of the point set. 3.- 3ds Max Rapid kinds of fine 3D point cloud based modeling system, characterized by comprising the following modules: point cloud pre-processing module, for the first point cloud data acquired by airborne laser radar performs point cloud denoising , and then quickly build Κ-D clustering tree, then thinning the point cloud; the fast clustering, including the point cloud data dENOISING triangulated irregular network model is constructed, triangulated irregular network model is calculated each cam surface normal vector, having a similar method to the same plane to the point set clustering magnitude; Κ-D tree of the construct, comprising the fast clustering results for each category trees were constructed Κ-D ; thumbnail loading module configured to provide all the thumbnails based 3ds Max point cloud according to the center point of each category obtained after pre-clustering module point cloud; local point cloud generating module, a thumbnail image when a user comprising loading a selected thumbnail resulting module point by Κ-D tree index searched neighboring points to the point, to generate a local cloud point of interest to the user; measuring point and the measuring line user correction module configured to guide the user through built Draw object outline, including when the user performs measurement points, extracting corresponding feature point correcting the user the measuring point, when the user performs the survey line, extracts the corresponding characteristic line correction user measuring points, for the following manner, provided the local point cloud in 3ds Max operation window the resulting point cloud generating module partially, when the user when adjusting the angle 3ds Max capture select building corner, the category selected by the user through the building corner where KD neighborhood tree acquiring point cloud measuring point employed, by extracting corner points neighbor algorithm is extracted from the point cloud to which all of the corner points as feature points; cloud points in the neighborhood of the midpoint plane cloud data obtained by fitting the plane which, as the feature line; select and display the selected user building corners nearest feature point, to achieve user-point correction, there is provided a building corner point corrected to a user; building corners when the user carried out starting from a sensing line correction, to capture and retrieve the corresponding characteristic line the corner of a building, to achieve the correction user survey line, draw the outline of the building edges; fine modeling module, based on the resulting construction according to 3ds Max Complete contour modeling. The point cloud based 3ds Max Rapid fine 3D modeling system of claim 3, wherein:
Patent Abstract The present invention provides a point cloud based 3ds? Max Rapid fine three-dimensional modeling method and system, comprising a pretreatment point cloud, point cloud data including the first airborne or vehicle-mounted lidar acquired point cloud denoising is performed, and then build KD fast clustering tree, then point clouds thinning; based on 3ds? The center point of each category by Max after all clustering thumbnail point cloud; generating a local point cloud, when the user selects the building corner points, cloud points in the neighborhood by acquiring the category where a KD tree using corner extraction neighbor algorithm is extracted from the point cloud to which all of the corner points as feature points, cloud points in the neighborhood of the midpoint plane cloud data obtained by fitting the plane which, as the feature line; correction feature point according to the user the measuring point, when a user survey line, search contouring edges of the building, amendments user survey line; based on 3ds? Max modeling is completed according to building outlines.
[IPC classification] G06T19 / 00, G06T5 / 00
[] Publication No. CN105354883
[Application Number] CN201510831919
[inventor] Huang Yuchun, Sun Hui, Ding Yalan, Yuan Ye
[the applicant] Wuhan University
[Publication date] February 24, 2016 date
[date of application] November 25, 2015