CN110009727A - A kind of indoor threedimensional model automatic reconfiguration method and system with structure semantics - Google Patents
A kind of indoor threedimensional model automatic reconfiguration method and system with structure semantics Download PDFInfo
- Publication number
- CN110009727A CN110009727A CN201910177032.1A CN201910177032A CN110009727A CN 110009727 A CN110009727 A CN 110009727A CN 201910177032 A CN201910177032 A CN 201910177032A CN 110009727 A CN110009727 A CN 110009727A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- metope
- algorithm
- component
- door
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 230000011218 segmentation Effects 0.000 claims abstract description 19
- 239000003550 marker Substances 0.000 claims abstract description 12
- 238000005457 optimization Methods 0.000 claims abstract description 12
- 238000013459 approach Methods 0.000 claims abstract description 10
- 238000000605 extraction Methods 0.000 claims description 32
- 238000001914 filtration Methods 0.000 claims description 21
- 239000000284 extract Substances 0.000 claims description 18
- 238000000638 solvent extraction Methods 0.000 claims description 14
- 238000001514 detection method Methods 0.000 claims description 7
- 238000012946 outsourcing Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 3
- 230000008569 process Effects 0.000 abstract description 3
- 230000006870 function Effects 0.000 description 11
- 238000012545 processing Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000012856 packing Methods 0.000 description 3
- 238000013499 data model Methods 0.000 description 2
- 239000011521 glass Substances 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000000523 sample Substances 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- 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
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- 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
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- 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
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Geometry (AREA)
- Computer Graphics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Processing Or Creating Images (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of indoor three-dimensional model reconfiguration method and system with structure semantics, method include: to acquire indoor three dimensional point cloud by movable depth scanning device, and pre-process to three dimensional point cloud；The metope component in abstraction function space；Obtain functional space boundary；Door and window component is extracted, and is parameterized；Plane segmentation is carried out to global point cloud data, obtains global plane, the geometry and incidence relation of metope component, door and window component are optimized according to planar structure global optimization approach；The data output of parametrization is had to the indoor threedimensional model of structure semantics.The present invention is based on movable depth scanning device, the camera track after merging low precision point cloud, semantic marker carries out the indoor model quick three-dimensional reconstructing with structure semantics, has better robustness and reconstruction precision.
Description
Technical field
The present invention relates to indoor threedimensional model constructing technology field more particularly to a kind of interiors with structure semantics
Threedimensional model automatic reconfiguration method and system.
Background technique
In recent years, with structure semantics information indoor threedimensional model become indoor positioning, indoor navigation, architectural control,
Virtual reality etc. applies important data source.Currently, a large amount of research concentrates on how dividing automatically based on three-dimensional point cloud and again
Indoor part is built, and then obtains the indoor threedimensional model of geometry and semantic congruence.Laser scanning is to obtain high-precision three-dimensional point cloud
Most widely used technology, this method usually requires that different measurement websites is arranged for target environment, and then passes through algorithm pair
The three-dimensional point cloud of different station for acquiring carries out splicing and obtains global three-dimensional point cloud.Currently, the spatial digitizer of some charge machines, example
Such as depth transducer, which can largely reduce the time of data acquisition and provide complete color three dimension
Point cloud.
Currently, be substantially from three-dimensional point cloud realizing the reconstruct for automating indoor threedimensional model and extract room, ground
Plate, ceiling, the incidence relation between components and component such as door and window, therefore convert non-structured three-dimensional point cloud to and have
The indoor threedimensional model of semantic information needs a large amount of computational geometries and computer vision related algorithm to support.But currently have
The indoor three-dimensional model reconfiguration of structure semantics has the following problems: 1) indoor complex environment under, in three-dimensional data scanning process by
It is easy to produce the incomplete situation of data in occlusion issue, needs Processing Algorithm；2) it is superfluous to be easy to produce data for repeated data scanning
It is remaining, cause packing density inconsistent, handles the problems such as time-consuming；3) indoor part obtained is difficult to keep original incidence relation,
Influence model construction precision.
Therefore, the existing technology needs to be improved and developed.
Summary of the invention
The technical problem to be solved in the present invention is that in view of the above drawbacks of the prior art, providing a kind of with structure language
The indoor threedimensional model automatic reconfiguration method and system of justice, it is intended to solve three-dimensional point cloud in the prior art and be used for indoor semanteme portion
The not high problem of accuracy in part extraction.
The technical proposal for solving the technical problem of the invention is as follows:
A kind of indoor threedimensional model automatic reconfiguration method with structure semantics, kind, which comprises
Indoor three dimensional point cloud is acquired by movable depth scanning device, and three dimensional point cloud is pre-processed；
Three dimensional point cloud after pretreatment is subjected to tissue according to different functional spaces, is partitioned into independent functional space
Data, and using plane partitioning algorithm, the metope component in scene self-filtering algorithm abstraction function space；
The point cloud data for the metope component that will acquire projects to two-dimensional surface, to the metope component point cloud data after each projection
Linear fit is carried out, and functional space boundary is obtained using line segment intersection iterative algorithm；
Restore the three dimensional point cloud comprising door and window component according to the particular data frame through semantic marker, and gathers convex closure and extract calculation
Method, Octree structure algorithm and Euclid's clustering algorithm extract door and window component, and parameterize；
Plane segmentation is carried out to global point cloud data, obtains global plane, according to planar structure global optimization approach to metope portion
Part, the geometry of door and window component and incidence relation optimize；
The data output of parametrization is had to the indoor threedimensional model of structure semantics.
The indoor threedimensional model automatic reconfiguration method with structure semantics, wherein the movable depth scanning
Equipment is movable depth sensor.
The indoor threedimensional model automatic reconfiguration method with structure semantics, wherein described to three dimensional point cloud
Carry out pretreated step, comprising: it is down-sampled to carry out point cloud data by voxelization trellis algorithm；It is calculated by sparse noise filtering
Method rejects original point cloud data noise.
The indoor threedimensional model automatic reconfiguration method with structure semantics, wherein it is described will be after pretreatment
Three dimensional point cloud carries out tissue according to different functional spaces, is partitioned into independent functional space data, and using plane point
Cut algorithm, scene self-filtering algorithm abstraction function space metope component the step of, comprising:
In conjunction with the depth data frame marked by starting and ending, uniformly subdivision is carried out to the interior space, obtains independent function
Spatial data；
For the three dimensional point cloud in standalone feature space, point cloud three-dimensional planar is obtained using Region Segmentation Algorithm and plane is joined
Number；
Candidate metope is chosen, and the non-metope element in the candidate metope of algorithm rejecting is rejected by inner element；
Using plane partitioning algorithm, scene self-filtering algorithm, the metope component of functional space is obtained from the candidate metope.
The indoor threedimensional model automatic reconfiguration method with structure semantics, wherein described to be calculated using plane segmentation
Method, scene self-filtering algorithm, from the candidate metope the step of metope component of acquisition functional space, comprising:
By all the points cloud data projection to horizontal plane, space convex closure is calculated, obtains convex closure polygon；
By all candidate wall projections to horizontal plane, linear fit is carried out, corresponding matching line segment is obtained, every line segment is distinguished
Ray is done using two endpoints as starting point, and is defined with endpoint that convex closure polygon intersection is odd number quantity in polygonal internal,
With convex closure polygon intersection be even number quantity endpoint in outside of polygon；
When two endpoints are in polygonal internal, which is rejected, when two-end-point is in outside of polygon, by this
Point cloud data retains, and obtains the metope component of functional space.
The indoor threedimensional model automatic reconfiguration method with structure semantics, wherein described to be calculated using plane segmentation
Method, scene self-filtering algorithm, from the candidate metope the step of metope component of acquisition functional space, further includes:
When one of endpoint is in polygonal internal, obtain the intersection point of line segment and polygon, calculate be located at polygonal internal with
It is rejected when ratio is greater than threshold value using the point cloud data as internal point cloud data positioned at line segment length ratio；Otherwise it protects
It stays.
The indoor threedimensional model automatic reconfiguration method with structure semantics, wherein the metope that will acquire
The point cloud data of component projects to two-dimensional surface, carries out linear fit to the metope component point cloud data after each projection, and adopt
The step of obtaining functional space boundary with line segment intersection iterative algorithm, comprising:
By all metope part point cloud data projections to horizontal plane, and linear fit is carried out, obtain optimum line segment, by calculating end
The distance between point determines two adjacent lines of every line segment；
Since any one line segment, its closest line segment is added, the intersection point for obtaining two line segments is put into, and iteration completes intersection point
It calculates until returning to initial segment；
When two metope fit line angles are close to 90 degree, pass through the available correct space boundary node of line intersection algorithm；
There is loss metope among the fit line of two metopes, then by the starting point of corresponding line segment vertex metope the most another, weight
Newly one metope of building；
The integral edge of functional space is obtained according to boundary extraction algorithm.
The indoor threedimensional model automatic reconfiguration method with structure semantics, wherein described according to through semantic marker
Particular data frame restore include door and window component three dimensional point cloud, and gather convex closure extraction algorithm, Octree structure algorithm
And Euclid's clustering algorithm extracts door and window component, and the step of parameterizing, comprising:
Convex closure calculating is carried out to door and window point cloud, obtains the outer profile of plane, and construct current door and window point cloud octree structure；
Obtain convex closure polygon outsourcing box, according to the density of setting in outsourcing box uniformly distributed point, judged using ray method
Point in outsourcing box carries out point cloud data filling to convex closure polygonal element whether in convex closure polygon, by filling point cloud number
According to building octree structure；
Unified door and window planar point cloud octree and filling point cloud octree structure leaf node size, compare carry out two by iteration
Variation detection between point cloud；
Based on Euclid's clustering algorithm, the door and window point cloud after extraction is subjected to point cloud data cluster, obtains unconnected gate window component；
The door and window point cloud data extracted is projected into two-dimensional surface, carries out linear fit, obtains door and window parameters of operating part.
A kind of automatic reconfiguration system of indoor threedimensional model with structure semantics, wherein the system comprises:
Data preprocessing module, for acquiring indoor three dimensional point cloud by movable depth scanning device, and to three-dimensional
Point cloud data is pre-processed；
Metope component extraction module, for three dimensional point cloud after pretreatment to be carried out group according to different functional spaces
It knits, is partitioned into independent functional space data, and using plane partitioning algorithm, scene self-filtering algorithm abstraction function space
Metope component；
The point cloud data of space boundary extraction module, the metope component for will acquire projects to two-dimensional surface, to each throwing
The metope component point cloud data of movie queen carries out linear fit, and obtains functional space boundary using line segment intersection iterative algorithm；
Door and window component extraction module, for restoring the three-dimensional point cloud comprising door and window component according to the particular data frame through semantic marker
Data, and gather convex closure extraction algorithm, Octree structure algorithm and Euclid's clustering algorithm and extract door and window component, and parameter
Change；
Global plane optimizing module obtains global plane, according to planar structure for carrying out plane segmentation to global point cloud data
Global optimization approach optimizes the geometry and incidence relation of metope component, door and window component；
Model output module, the data output for that will parameterize have the indoor threedimensional model of structure semantics.
The automatic reconfiguration system of indoor threedimensional model with structure semantics, wherein the movable depth scanning
Equipment is movable depth sensor.
Beneficial effects of the present invention: the present invention is based on movable depth scanning devices, merge low precision point cloud, semantic marker
Camera track afterwards carries out the indoor model quick three-dimensional reconstructing with structure semantics, has better robustness and rebuilds essence
Degree.
Detailed description of the invention
Fig. 1 is the process of the preferred embodiment of the indoor threedimensional model automatic reconfiguration method with structure semantics of the invention
Figure.
When Fig. 2 is the concrete application application of the indoor threedimensional model automatic reconfiguration method with structure semantics of the invention
Flow chart.
Fig. 3 is functional space internal point cloud in the indoor threedimensional model automatic reconfiguration method with structure semantics of the invention
Filter method schematic diagram.
Fig. 4 is that the functional space boundary in the indoor threedimensional model automatic reconfiguration method with structure semantics of the invention mentions
Take method schematic diagram.
Fig. 5 is the door and window component extraction side in the indoor threedimensional model automatic reconfiguration method with structure semantics of the invention
Method schematic diagram
Fig. 6 is the schematic diagram of global optimization in the indoor threedimensional model automatic reconfiguration method with structure semantics of the invention
Fig. 7 is the principle of work and power money figure of the automatic reconfiguration system of indoor threedimensional model with structure semantics of the invention.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer and more explicit, right as follows in conjunction with drawings and embodiments
The present invention is further described.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and do not have to
It is of the invention in limiting.
A kind of indoor threedimensional model automatic reconfiguration method with structure semantics provided by the invention, can be applied to intelligence
In energy terminal.Wherein, intelligent terminal can be, but not limited to be various personal computers, laptop, mobile phone, tablet computer,
Vehicle-mounted computer and portable wearable device.Intelligent terminal of the invention uses multi-core processor.Wherein, the processing of intelligent terminal
Device can be central processing unit (Central Processing Unit, CPU), graphics processor (Graphics
Processing Unit, GPU), at least one of video processing unit (Video Processing Unit, VPU) etc..
Since that there are data is imperfect, repeated data scanning is held for the existing indoor three-dimensional model reconfiguration with structure semantics
It is also easy to produce data redundancy, causes the indoor part that packing density is inconsistent, obtains to be difficult to keep original incidence relation, influences mould
Type constructs the problems such as precision.To solve the above-mentioned problems, the present embodiment provides a kind of indoor threedimensional model with structure semantics
Automatic reconfiguration method, it is therefore intended that overcome existing three-dimensional point cloud accuracy is not high in extracting for indoor semantic component and ask
Topic obtains data using movable depth scanning device, combines multiple types of data collection, preferably realizes indoor semantic component
Rapidly extracting and three-dimensional modeling.As shown in Figure 1, specifically comprising the following steps:
Step S100, indoor three dimensional point cloud is acquired by movable depth scanning device, and to three dimensional point cloud into
Row pretreatment；
Step S200, three dimensional point cloud after pretreatment is subjected to tissue according to different functional spaces, is partitioned into independence
Functional space data, and using plane partitioning algorithm, scene self-filtering algorithm abstraction function space metope component；
Step S300, the point cloud data for the metope component that will acquire projects to two-dimensional surface, to the metope portion after each projection
Part point cloud data carries out linear fit, and obtains functional space boundary using line segment intersection iterative algorithm；
Step S400, the three dimensional point cloud comprising door and window component is restored according to the particular data frame through semantic marker, and gathered
Convex closure extraction algorithm, Octree structure algorithm and Euclid's clustering algorithm extract door and window component, and parameterize；
Step S500, plane segmentation is carried out to global point cloud data, obtains global plane, according to planar structure global optimization approach
The geometry and incidence relation of metope component, door and window component are optimized；
Step S600, the data output of parametrization is had to the indoor threedimensional model of structure semantics.
When it is implemented, as shown in Figure 2, collecting data set by depth transducer in the present embodiment and obtaining colour
Three-dimensional point cloud, and then plane segmentation is carried out to world model.It carries out space segmentation and extracts to go out based on color three dimension dot cloud
Window three-dimensional point cloud is then based on door and window three-dimensional point cloud and carries out the extraction of door and window component, carries out functional space extraction based on space segmentation,
CityGML3.0 model is finally exported, i.e., with the indoor threedimensional model of structure semantics.Specifically, the present embodiment is first using shifting
Dynamic formula depth transducer carries out data acquisition, indoor color three dimension point cloud data is obtained, due to the three dimensional point cloud of acquisition
Usually have each region density inconsistent, the problems such as point cloud data amount is excessive and noise, need using data it is down-sampled and
Point cloud filtering algorithm carries out data prediction to three dimensional point cloud.Specifically, the present embodiment uses sparse noise filtering algorithm
It is used to reject original point cloud noise, this method is closest apart from intermediate value in point cloud adjacent space by calculatingAnd standard deviationTo realize the rejecting of noise spot.Firstly the need of the size for defining contiguous space, with neighbor point numberIt is neighbouring for scale calibration
Point is for calculating apart from intermediate value and standard deviation criteria, it is assumed that distance is Gaussian Profile, therefore distance is in rangeExcept
Point will be removed.It is determined according to experiment, it is set asBe set as 50, according to this method by reject about 1% noise spot.Its
It is secondary, it is down-sampled for point cloud using voxelization trellis algorithm for data redundancy problem, reduce data redudancy, unification point
Cloud packing density.Down-sampled rate depends on voxel node size, therefore setting is specific according to demandValue.
Further, feature is modeled using movable depth scanning system interactive, the track of depth transducer is pressed
Carry out tissue according to different functional spaces, so it is divisible go out independent functional space data, and then using plane partitioning algorithm,
Scene self-filtering algorithm realizes that functional space metope component extracts.Specifically, as shown in Figure 3, in conjunction with by starting and knot
The depth data frame marked until beam carries out uniformly subdivision to the complicated interior space, standalone feature spatial data is obtained, for only
Render meritorious service can space three-dimensional point cloud data, point cloud three-dimensional planar and plane parameter is obtained using Region Segmentation Algorithm.Then by plane
Tissue is carried out according to vertical and horizontal plane, defines normal vector threshold valueWith area threshold, when in-plane and Vertical Square
It is less than normal vector threshold value to angleAnd area is greater than area thresholdWhen, as candidate metope.The candidate wall of acquisition
There is a large amount of non-metope element in face, an inner element rejects algorithm and be used to reject non-metope element in candidate metope.So
It is described afterwards referring to following algorithm 1, all the points cloud is projected into horizontal plane, calculates space convex closure, obtains convex closure polygon.To own
Candidate wall projections carry out linear fit, obtain corresponding matching line segment to horizontal plane, to every line segment to divide
Do not do ray using two endpoints as starting point, with convex closure polygon intersection be odd number quantity point in polygonal internal, with convex closure
Polygon intersection is the point of even number quantity in outside of polygon, when two o'clock in polygonal internal (a in such as Fig. 3), will
The cloud reject, when two o'clock in outside of polygon (b in such as Fig. 3), by the cloud retain.When one of point is polygon
When inside shape (c in such as Fig. 3), the intersection point of line segment and polygon is obtained, is calculated and is located at polygonal internal and is located at line segment length
Ratio, when ratio is greater than threshold valueWhen, it rejects the cloud as internal point cloud, otherwise retains.According to above method,
The metope component point cloud data of functional space can be got.
Specifically, algorithm 1: space interior element filter algorithm
Input:
Potential metope after projection
Cloud convex closure is put after projection:
Polygonal internal line segment length and total length ratio:
Based on Ransac line segment fitting function:
The line segment of fitting:
Distance calculates function:
Distance of the line segment vertex to intersection point:
Line segment intersection function:
Two intersection point of line segments:
Point is in polygonal internal discriminant function:
Ray and polygon intersection number:
Output:
Inner plane element index:;Metope index: ;
1.
2. While Non-empty do
3. ,
4. linear fit
5.
6. detecting ray and polygon intersection number
7. ,
8. If For Qi Shuo &&For even number then
9.
10. else if For Qi Shuo &&For even number then
11。
12. else
13 matching line segments and polygon intersection
14.
15. calculating separatelyWithThe distance between,
16. ,
17. if ( For Qi Shuo &&) || For Ou Shuo && ) then
18.
19. else
20.
21. end if
22. end if
23. end while
24. Return ,
Further, the metope component that will acquire projects to two-dimensional surface, carries out Linear Quasi to the metope point cloud after each projection
It closes, and then functional space boundary is obtained using line segment intersection iterative algorithm.Specifically, all metope point clouds project to horizontal plane,
And linear fit is carried out, obtain optimum line segment, two adjacent lines of every line segment are determined by calculating the distance between endpoint.Since any one line segment, its closest line segment is added, at the same time, obtain two line segments
Intersection point is put into, iteration completion intersection point calculation is until returning to initial segment.However, during data scanning, part wall
Face is not got, referring to a in Fig. 4, by production 1 and situation 2, in situation 1(Fig. 4 a) in, two metope fit lines
Angle is close to 90 degree, by the available correct space boundary node of line intersection algorithm, in situation 2, due to wall line 1 and wall line
There is loss metope among 2, therefore rule can not be handed over to obtain corresponding vertex according to asking, needs corresponding line segment vertex is the most another
The starting point of one metope rebuilds a metope.When the normal vector of two linesLess than certain threshold valueWhen,
It is otherwise situation 1 for the b in situation 2(Fig. 4).The integral edge of functional space can be finally obtained according to boundary extraction algorithm.
Further, it is contemplated that carry out the state of Men Weikai when movable depth scanning device mapping, window is transparent glass
Glass, therefore door and window component extracts problem and is converted to how to carry out frontier probe to cloud cavity.Referring again to Fig. 2, for door and window
The extraction of component, specifically include that plane segmentation, the detection of door and window component, plane global optimization, door and window extract, two-dimensional projection and
Door and window reconstruct.Specifically, the independent three-dimensional point cloud of door and window is constructed based on the particular data frame of interactive definition in the present embodiment,
Mentioning for door and window component is completed in conjunction with algorithm of convex hull, Octree algorithm, Euclid's clustering algorithm and point cloud change detection algorithm
It takes.Specifically, referring to Fig. 5, restore the three-dimensional point cloud comprising door and window component according to by the particular data frame of semantic marker, use
Plane partitioning algorithm obtains door and window plane.Firstly, carrying out convex closure calculating to door and window point cloud in Fig. 5 (a), the foreign steamer of plane is obtained
Exterior feature, and construct current door and window point cloud octree structure.Secondly the outer bounding box for obtaining convex closure polygon, exists according to the density of setting
Uniformly distributed point in outsourcing box, using ray method judge the point in bounding box whether in convex closure polygon, and then it is more in convex closure
The filling of filling point cloud data inside the shape of side, i.e. convex closure finishes, and filling point cloud is constructed octree structure.Finally unify door and window
Planar point cloud octree and filling point cloud octree structure leaf node size compare the variation between carrying out two o'clock cloud by iteration
Detection, by the point cloud i.e. door and window part point cloud obtained after variation detection, is based on since filling point cloud includes door and window component
Door and window point cloud after extraction is carried out a cloud and clustered, obtains unconnected gate window component by Euclid's clustering algorithm.It will finally extract
Door and window point cloud project to two-dimensional surface, carry out linear fit, obtain door and window parameters of operating part.
Further, due to point cloud segmentation algorithm be easy receptor site cloud precision influence, the segmentation result of each independent point cloud with
Global point cloud result is easy to produce inconsistent situation, global metope and separate space metope of this method according to detection, independence
The angle and distance of space metope and door and window plane, determine metope and metope, the incidence relation between metope and door and window, and then to it
Unification processing is carried out, guarantees the geometry and semantic consistency of model.Specifically, referring to Fig. 6, plane point is carried out to global point cloud
It cuts and global plane can be obtained, since cloud error influences, inconsistent situation, Tu Zhongqiang are easy to produce between global plane and metope
Face 1 and metope 2 share same global plane 1, and inconsistent by three's plane parameter that plane partitioning algorithm obtains, to guarantee
Two its consistency need to obtain global plane corresponding with current metope first.Current metope is checked by iteration first,With all global planesBetween angle,
WhenLess than certain angle thresholdWhen, determine the global plane of candidate of current metope, and then by current plane method phase
Parameter unanimously turns to, according to the following formula:
Metope is at a distance from global interplanar after calculating unification, is the complete of the metope by the plane definition when distance is minimum
Office's plane.All metopes of iteration obtain corresponding global plane.Finally by wall projections into global plane, guarantee that it is consistent
Property.The relationship between door and window component and metope is judged by identical algorithms, door and window component is projected in corresponding metope, guarantees one
Cause property.
Further, the indoor wall boundary that will acquire, door and window boundary and metope is parameterized by horizontal plane height difference,
Floor and ceiling and door and window component, and according to CityGML3.0 standard storage, exporting has the interior of semantic structure three-dimensional
Model.The present invention carries out that point cloud data is down-sampled to be handled with noise filtering first；Secondly pretreated three-dimensional point cloud is used for
The extraction of metope component constructs space boundary by the wall projections after extraction to two-dimensional surface and after carrying out linear fit；Fusion
Convex closure extracts, octree structure and realizes door and window component data reduction based on Euclid's clustering algorithm, and parameterizes；Consider
To the inconsistent situation of interplanar, plane global optimization approach is proposed to metope, door and window component geometry is optimized with incidence relation；
Finally the data after parametrization are exported as standard CityGML3.0 data model, to realize that rapidly building has structure language
The indoor model of justice, and there is better robustness and reconstruction precision.
Based on the above embodiment, the present invention also provides a kind of, and the indoor threedimensional model with structure semantics reconstructs automatically is
System, as shown in Figure 7, the system include: data preprocessing module 710, metope component extraction module 720, space boundary extraction
Module 730, door and window component extraction module 740, global plane optimizing module 750, model output module 760.
Specifically, data preprocessing module 710, for acquiring indoor three-dimensional point cloud by movable depth scanning device
Data, and three dimensional point cloud is pre-processed；
Metope component extraction module 720, for carrying out three dimensional point cloud after pretreatment according to different functional spaces
Tissue is partitioned into independent functional space data, and uses plane partitioning algorithm, scene self-filtering algorithm abstraction function space
Metope component；
The point cloud data of space boundary extraction module 730, the metope component for will acquire projects to two-dimensional surface, to each
Metope component point cloud data after projection carries out linear fit, and obtains functional space boundary using line segment intersection iterative algorithm；
Door and window component extraction module 740, for restoring the three-dimensional comprising door and window component according to the particular data frame through semantic marker
Point cloud data, and gather convex closure extraction algorithm, Octree structure algorithm and Euclid's clustering algorithm and extract door and window component, and
Parametrization；
Global plane optimizing module 750 obtains global plane, according to plane for carrying out plane segmentation to global point cloud data
Structure global optimization approach optimizes the geometry and incidence relation of metope component, door and window component；
Model output module 760, the data output for that will parameterize have the indoor threedimensional model of structure semantics.
Preferably, movable depth scanning device is movable depth sensor in the present embodiment.This system carries out first
Point cloud data is down-sampled to be handled with noise filtering；Secondly pretreated three-dimensional point cloud is used for the extraction of metope component, will mentioned
Wall projections after taking construct space boundary to two-dimensional surface and after carrying out linear fit；Merge convex closure extraction, octree structure
And door and window component data reduction is realized based on Euclid's clustering algorithm, and parameterize；In view of the inconsistent situation of interplanar,
Plane global optimization approach is proposed to metope, door and window component geometry is optimized with incidence relation；Finally by the number after parametrization
It is standard CityGML3.0 data model according to output, thus realize that rapidly building has the indoor model of structure semantics, and
With better robustness and reconstruction precision.Specific step and effect are referring to shown in the embodiment of the above method.
In conclusion the invention discloses a kind of indoor three-dimensional model reconfiguration method and system with structure semantics, side
Method includes: to acquire indoor three dimensional point cloud by movable depth scanning device, and located in advance to three dimensional point cloud
Reason；The metope component in abstraction function space；Obtain functional space boundary；Door and window component is extracted, and is parameterized；To global point cloud number
According to plane segmentation is carried out, global plane is obtained, according to planar structure global optimization approach to the geometry of metope component, door and window component
It is optimized with incidence relation；The data output of parametrization is had to the indoor threedimensional model of structure semantics.The present invention is based on shiftings
Dynamic formula depth scan equipment, the camera track after merging low precision point cloud, semantic marker carry out the indoor mould with structure semantics
Type quick three-dimensional reconstructing has better robustness and reconstruction precision.
It should be understood that the application of the present invention is not limited to the above for those of ordinary skills can
With improvement or transformation based on the above description, all these modifications and variations all should belong to the guarantor of appended claims of the present invention
Protect range.
Claims (10)
1. a kind of indoor threedimensional model automatic reconfiguration method with structure semantics, which is characterized in that the described method includes:
Indoor three dimensional point cloud is acquired by movable depth scanning device, and three dimensional point cloud is pre-processed；
Three dimensional point cloud after pretreatment is subjected to tissue according to different functional spaces, is partitioned into independent functional space
Data, and using plane partitioning algorithm, the metope component in scene self-filtering algorithm abstraction function space；
The point cloud data for the metope component that will acquire projects to two-dimensional surface, to the metope component point cloud data after each projection
Linear fit is carried out, and functional space boundary is obtained using line segment intersection iterative algorithm；
Restore the three dimensional point cloud comprising door and window component according to the particular data frame through semantic marker, and gathers convex closure and extract calculation
Method, Octree structure algorithm and Euclid's clustering algorithm extract door and window component, and parameterize；
Plane segmentation is carried out to global point cloud data, obtains global plane, according to planar structure global optimization approach to metope portion
Part, the geometry of door and window component and incidence relation optimize；
The data output of parametrization is had to the indoor threedimensional model of structure semantics.
2. the indoor threedimensional model automatic reconfiguration method according to claim 1 with structure semantics, which is characterized in that institute
Stating movable depth scanning device is movable depth sensor.
3. the indoor threedimensional model automatic reconfiguration method according to claim 1 with structure semantics, which is characterized in that institute
It states and pretreated step is carried out to three dimensional point cloud, comprising: it is down-sampled to carry out point cloud data by voxelization trellis algorithm；It is logical
It crosses sparse noise filtering algorithm and rejects original point cloud data noise.
4. the indoor threedimensional model automatic reconfiguration method according to claim 1 with structure semantics, which is characterized in that institute
It states and three dimensional point cloud after pretreatment is subjected to tissue according to different functional spaces, be partitioned into independent functional space number
According to, and using plane partitioning algorithm, scene self-filtering algorithm abstraction function space metope component the step of, comprising:
In conjunction with the depth data frame marked by starting and ending, uniformly subdivision is carried out to the interior space, obtains independent function
Spatial data；
For the three dimensional point cloud in standalone feature space, point cloud three-dimensional planar is obtained using Region Segmentation Algorithm and plane is joined
Number；
Candidate metope is chosen, and the non-metope element in the candidate metope of algorithm rejecting is rejected by inner element；
Using plane partitioning algorithm, scene self-filtering algorithm, the metope component of functional space is obtained from the candidate metope.
5. the indoor threedimensional model automatic reconfiguration method according to claim 4 with structure semantics, which is characterized in that institute
State the metope component for obtaining functional space from the candidate metope using plane partitioning algorithm, scene self-filtering algorithm
Step, comprising:
By all the points cloud data projection to horizontal plane, space convex closure is calculated, obtains convex closure polygon；
By all candidate wall projections to horizontal plane, linear fit is carried out, corresponding matching line segment is obtained, every line segment is distinguished
Ray is done using two endpoints as starting point, and is defined with endpoint that convex closure polygon intersection is odd number quantity in polygonal internal,
With convex closure polygon intersection be even number quantity endpoint in outside of polygon；
When two endpoints are in polygonal internal, which is rejected, when two-end-point is in outside of polygon, by this
Point cloud data retains, and obtains the metope component of functional space.
6. the indoor threedimensional model automatic reconfiguration method according to claim 5 with structure semantics, which is characterized in that institute
State the metope component for obtaining functional space from the candidate metope using plane partitioning algorithm, scene self-filtering algorithm
Step, further includes:
When one of endpoint is in polygonal internal, obtain the intersection point of line segment and polygon, calculate be located at polygonal internal with
It is rejected when ratio is greater than threshold value using the point cloud data as internal point cloud data positioned at line segment length ratio；Otherwise it protects
It stays.
7. the indoor threedimensional model automatic reconfiguration method according to claim 1 with structure semantics, which is characterized in that institute
The point cloud data for stating the metope component that will acquire projects to two-dimensional surface, to the metope component point cloud data after each projection into
Row linear fit, and the step of functional space boundary is obtained using line segment intersection iterative algorithm, comprising:
By all metope part point cloud data projections to horizontal plane, and linear fit is carried out, obtain optimum line segment, by calculating end
The distance between point determines two adjacent lines of every line segment；
Since any one line segment, its closest line segment is added, the intersection point for obtaining two line segments is put into, and iteration completes intersection point
It calculates until returning to initial segment；
When two metope fit line angles are close to 90 degree, pass through the available correct space boundary node of line intersection algorithm；
There is loss metope among the fit line of two metopes, then by the starting point of corresponding line segment vertex metope the most another, weight
Newly one metope of building；
The integral edge of functional space is obtained according to boundary extraction algorithm.
8. the indoor threedimensional model automatic reconfiguration method according to claim 1 with structure semantics, which is characterized in that institute
It states and the three dimensional point cloud comprising door and window component is restored according to the particular data frame through semantic marker, and gather convex closure and extract calculation
Method, Octree structure algorithm and Euclid's clustering algorithm extract door and window component, and the step of parameterizing, comprising:
Convex closure calculating is carried out to door and window point cloud, obtains the outer profile of plane, and construct current door and window point cloud octree structure；
Obtain convex closure polygon outsourcing box, according to the density of setting in outsourcing box uniformly distributed point, judged using ray method
Point in outsourcing box carries out point cloud data filling to convex closure polygonal element whether in convex closure polygon, by filling point cloud number
According to building octree structure；
Unified door and window planar point cloud octree and filling point cloud octree structure leaf node size, compare carry out two by iteration
Variation detection between point cloud；
Based on Euclid's clustering algorithm, the door and window point cloud after extraction is subjected to point cloud data cluster, obtains unconnected gate window component；
The door and window point cloud data extracted is projected into two-dimensional surface, carries out linear fit, obtains door and window parameters of operating part.
9. a kind of automatic reconfiguration system of indoor threedimensional model with structure semantics, which is characterized in that the system comprises:
Data preprocessing module, for acquiring indoor three dimensional point cloud by movable depth scanning device, and to three-dimensional
Point cloud data is pre-processed；
Metope component extraction module, for three dimensional point cloud after pretreatment to be carried out group according to different functional spaces
It knits, is partitioned into independent functional space data, and using plane partitioning algorithm, scene self-filtering algorithm abstraction function space
Metope component；
The point cloud data of space boundary extraction module, the metope component for will acquire projects to two-dimensional surface, to each throwing
The metope component point cloud data of movie queen carries out linear fit, and obtains functional space boundary using line segment intersection iterative algorithm；
Door and window component extraction module, for restoring the three-dimensional point cloud comprising door and window component according to the particular data frame through semantic marker
Data, and gather convex closure extraction algorithm, Octree structure algorithm and Euclid's clustering algorithm and extract door and window component, and parameter
Change；
Global plane optimizing module obtains global plane, according to planar structure for carrying out plane segmentation to global point cloud data
Global optimization approach optimizes the geometry and incidence relation of metope component, door and window component；
Model output module, the data output for that will parameterize have the indoor threedimensional model of structure semantics.
10. according to the indoor threedimensional model automatic reconfiguration system as claimed in claim 9 with structure semantics is weighed, feature exists
In the movable depth scanning device is movable depth sensor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910177032.1A CN110009727B (en) | 2019-03-08 | 2019-03-08 | Automatic reconstruction method and system for indoor three-dimensional model with structural semantics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910177032.1A CN110009727B (en) | 2019-03-08 | 2019-03-08 | Automatic reconstruction method and system for indoor three-dimensional model with structural semantics |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110009727A true CN110009727A (en) | 2019-07-12 |
| CN110009727B CN110009727B (en) | 2023-04-18 |
Family
ID=67166659
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910177032.1A Active CN110009727B (en) | 2019-03-08 | 2019-03-08 | Automatic reconstruction method and system for indoor three-dimensional model with structural semantics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110009727B (en) |
Cited By (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110532602A (en) * | 2019-07-19 | 2019-12-03 | 中国地质大学（武汉） | A kind of indoor autodraft and modeling method based on plan view image |
| CN110533778A (en) * | 2019-08-09 | 2019-12-03 | 中国科学院自动化研究所 | Large Scale Graphs picture point cloud parallel distributed gridding reconstruction method, system, device |
| CN110555284A (en) * | 2019-09-12 | 2019-12-10 | 杭州群核信息技术有限公司 | Cloud rapid modeling and design method based on spatial position |
| CN110715618A (en) * | 2019-09-29 | 2020-01-21 | 北京天远三维科技股份有限公司 | Dynamic three-dimensional scanning method and device |
| CN110751706A (en) * | 2019-10-29 | 2020-02-04 | 网易（杭州）网络有限公司 | Method, device and equipment for processing graphics in game and storage medium |
| CN110969701A (en) * | 2019-11-29 | 2020-04-07 | 司空定制家居科技有限公司 | Method and device for modeling house and storage medium |
| CN111210501A (en) * | 2019-10-12 | 2020-05-29 | 香港理工大学深圳研究院 | Indoor modeling method and device and terminal equipment |
| CN111507919A (en) * | 2020-04-16 | 2020-08-07 | 北京深测科技有限公司 | Denoising processing method for three-dimensional point cloud data |
| CN111652137A (en) * | 2020-06-03 | 2020-09-11 | 上海眼控科技股份有限公司 | Illegal vehicle detection method and device, computer equipment and storage medium |
| CN111737802A (en) * | 2020-06-24 | 2020-10-02 | 广州华建工智慧科技有限公司 | Method for automatically separating beam from column based on three-dimensional model and semantic description information |
| CN111915730A (en) * | 2020-07-20 | 2020-11-10 | 北京建筑大学 | Method and system for automatically generating indoor three-dimensional model from point cloud in consideration of semantics |
| CN112164145A (en) * | 2020-10-30 | 2021-01-01 | 武汉大学 | Method for rapidly extracting indoor three-dimensional line segment structure based on point cloud data |
| CN112001026B (en) * | 2020-10-28 | 2021-02-09 | 深圳大学 | Method, terminal and storage medium for building information model construction |
| CN112381948A (en) * | 2020-11-03 | 2021-02-19 | 上海交通大学烟台信息技术研究院 | Semantic-based laser stripe center line extraction and fitting method |
| CN112571415A (en) * | 2020-12-03 | 2021-03-30 | 哈尔滨工业大学（深圳） | Robot autonomous door opening method and system based on visual guidance |
| CN112712584A (en) * | 2019-10-25 | 2021-04-27 | 阿里巴巴集团控股有限公司 | Wall line determining method, space modeling method, device and equipment |
| CN112927334A (en) * | 2021-02-22 | 2021-06-08 | 中铁二院工程集团有限责任公司 | Three-dimensional model fast voxelization method based on GPU |
| CN113177477A (en) * | 2021-04-29 | 2021-07-27 | 湖南大学 | Target detection and identification method based on three-dimensional point cloud analysis |
| CN113628330A (en) * | 2021-08-23 | 2021-11-09 | 北京京东乾石科技有限公司 | Finite space three-dimensional model construction method and system |
| WO2022016311A1 (en) * | 2020-07-20 | 2022-01-27 | 深圳元戎启行科技有限公司 | Point cloud-based three-dimensional reconstruction method and apparatus, and computer device |
| CN113989376A (en) * | 2021-12-23 | 2022-01-28 | 贝壳技术有限公司 | Method and device for acquiring indoor depth information and readable storage medium |
| CN114898066A (en) * | 2022-06-14 | 2022-08-12 | 中国电子工程设计院有限公司 | Non-convex indoor space identification reconstruction method and system |
| CN115239951A (en) * | 2022-06-08 | 2022-10-25 | 广东领慧建筑科技有限公司 | Wall surface segmentation and identification method and system based on point cloud data processing |
| CN115661378A (en) * | 2022-12-28 | 2023-01-31 | 北京道仪数慧科技有限公司 | Building model reconstruction method and system |
| WO2023024482A1 (en) * | 2021-08-23 | 2023-03-02 | 奥比中光科技集团股份有限公司 | Interior structured reconstruction method and apparatus, and computer-readable storage medium |
| CN117152374A (en) * | 2023-07-25 | 2023-12-01 | 北京建筑大学 | Digital twin establishing method of efficient cable dome structure based on three-dimensional laser point cloud |
| CN117422810A (en) * | 2023-12-19 | 2024-01-19 | 深圳大学 | Structure and parameter guided indoor element regularization and relationship reasoning method and terminal |
| CN117475110A (en) * | 2023-12-27 | 2024-01-30 | 北京市农林科学院信息技术研究中心 | Semantic three-dimensional reconstruction method and device for blade, electronic equipment and storage medium |
| CN117496086A (en) * | 2023-10-12 | 2024-02-02 | 南京林业大学 | Indoor geometric reconstruction method for semantic perception |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106709481A (en) * | 2017-03-03 | 2017-05-24 | 深圳市唯特视科技有限公司 | Indoor scene understanding method based on 2D-3D semantic data set |
| CN107862738A (en) * | 2017-11-28 | 2018-03-30 | 武汉大学 | One kind carries out doors structure three-dimensional rebuilding method based on mobile laser measurement point cloud |
| CN107978017A (en) * | 2017-10-17 | 2018-05-01 | 厦门大学 | Doors structure fast modeling method based on wire extraction |
| US20180275277A1 (en) * | 2017-03-22 | 2018-09-27 | Here Global B.V. | Method, apparatus and computer program product for mapping and modeling a three dimensional structure |
| CN109087388A (en) * | 2018-07-12 | 2018-12-25 | 南京邮电大学 | Object dimensional modeling method based on depth transducer |
-
2019
- 2019-03-08 CN CN201910177032.1A patent/CN110009727B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106709481A (en) * | 2017-03-03 | 2017-05-24 | 深圳市唯特视科技有限公司 | Indoor scene understanding method based on 2D-3D semantic data set |
| US20180275277A1 (en) * | 2017-03-22 | 2018-09-27 | Here Global B.V. | Method, apparatus and computer program product for mapping and modeling a three dimensional structure |
| CN107978017A (en) * | 2017-10-17 | 2018-05-01 | 厦门大学 | Doors structure fast modeling method based on wire extraction |
| CN107862738A (en) * | 2017-11-28 | 2018-03-30 | 武汉大学 | One kind carries out doors structure three-dimensional rebuilding method based on mobile laser measurement point cloud |
| CN109087388A (en) * | 2018-07-12 | 2018-12-25 | 南京邮电大学 | Object dimensional modeling method based on depth transducer |
Non-Patent Citations (1)
| Title |
|---|
| 熊汉江等: "基于2D-3D语义传递的室内三维点云模型语义分割", 《武汉大学学报(信息科学版)》 * |
Cited By (45)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110532602B (en) * | 2019-07-19 | 2022-11-18 | 中国地质大学（武汉） | Indoor automatic drawing and modeling method based on plane graph image |
| CN110532602A (en) * | 2019-07-19 | 2019-12-03 | 中国地质大学（武汉） | A kind of indoor autodraft and modeling method based on plan view image |
| CN110533778A (en) * | 2019-08-09 | 2019-12-03 | 中国科学院自动化研究所 | Large Scale Graphs picture point cloud parallel distributed gridding reconstruction method, system, device |
| CN110555284A (en) * | 2019-09-12 | 2019-12-10 | 杭州群核信息技术有限公司 | Cloud rapid modeling and design method based on spatial position |
| CN110715618A (en) * | 2019-09-29 | 2020-01-21 | 北京天远三维科技股份有限公司 | Dynamic three-dimensional scanning method and device |
| CN111210501A (en) * | 2019-10-12 | 2020-05-29 | 香港理工大学深圳研究院 | Indoor modeling method and device and terminal equipment |
| CN112712584A (en) * | 2019-10-25 | 2021-04-27 | 阿里巴巴集团控股有限公司 | Wall line determining method, space modeling method, device and equipment |
| CN112712584B (en) * | 2019-10-25 | 2024-05-24 | 阿里巴巴集团控股有限公司 | Space modeling method, device and equipment |
| CN110751706B (en) * | 2019-10-29 | 2024-03-22 | 网易（杭州）网络有限公司 | Method, device, equipment and storage medium for processing graphics in game |
| CN110751706A (en) * | 2019-10-29 | 2020-02-04 | 网易（杭州）网络有限公司 | Method, device and equipment for processing graphics in game and storage medium |
| CN110969701B (en) * | 2019-11-29 | 2023-12-29 | 司空定制家居科技有限公司 | Method, device and storage medium for modeling house |
| CN110969701A (en) * | 2019-11-29 | 2020-04-07 | 司空定制家居科技有限公司 | Method and device for modeling house and storage medium |
| CN111507919A (en) * | 2020-04-16 | 2020-08-07 | 北京深测科技有限公司 | Denoising processing method for three-dimensional point cloud data |
| CN111507919B (en) * | 2020-04-16 | 2023-07-14 | 北京深测科技有限公司 | Denoising processing method for three-dimensional point cloud data |
| CN111652137A (en) * | 2020-06-03 | 2020-09-11 | 上海眼控科技股份有限公司 | Illegal vehicle detection method and device, computer equipment and storage medium |
| CN111737802B (en) * | 2020-06-24 | 2023-02-24 | 广州华建工智慧科技有限公司 | Method for automatically separating beam from column based on three-dimensional model and semantic description information |
| CN111737802A (en) * | 2020-06-24 | 2020-10-02 | 广州华建工智慧科技有限公司 | Method for automatically separating beam from column based on three-dimensional model and semantic description information |
| CN111915730A (en) * | 2020-07-20 | 2020-11-10 | 北京建筑大学 | Method and system for automatically generating indoor three-dimensional model from point cloud in consideration of semantics |
| CN111915730B (en) * | 2020-07-20 | 2023-05-19 | 北京建筑大学 | Method and system for automatically generating indoor three-dimensional model by taking semantic slave point cloud into consideration |
| WO2022016311A1 (en) * | 2020-07-20 | 2022-01-27 | 深圳元戎启行科技有限公司 | Point cloud-based three-dimensional reconstruction method and apparatus, and computer device |
| CN112001026B (en) * | 2020-10-28 | 2021-02-09 | 深圳大学 | Method, terminal and storage medium for building information model construction |
| CN112164145A (en) * | 2020-10-30 | 2021-01-01 | 武汉大学 | Method for rapidly extracting indoor three-dimensional line segment structure based on point cloud data |
| CN112381948A (en) * | 2020-11-03 | 2021-02-19 | 上海交通大学烟台信息技术研究院 | Semantic-based laser stripe center line extraction and fitting method |
| CN112381948B (en) * | 2020-11-03 | 2022-11-29 | 上海交通大学烟台信息技术研究院 | Semantic-based laser stripe center line extraction and fitting method |
| CN112571415B (en) * | 2020-12-03 | 2022-03-01 | 哈尔滨工业大学（深圳） | Robot autonomous door opening method and system based on visual guidance |
| CN112571415A (en) * | 2020-12-03 | 2021-03-30 | 哈尔滨工业大学（深圳） | Robot autonomous door opening method and system based on visual guidance |
| CN112927334A (en) * | 2021-02-22 | 2021-06-08 | 中铁二院工程集团有限责任公司 | Three-dimensional model fast voxelization method based on GPU |
| CN113177477A (en) * | 2021-04-29 | 2021-07-27 | 湖南大学 | Target detection and identification method based on three-dimensional point cloud analysis |
| CN113628330B (en) * | 2021-08-23 | 2024-04-12 | 北京京东乾石科技有限公司 | Three-dimensional model construction method and system for limited space |
| WO2023024482A1 (en) * | 2021-08-23 | 2023-03-02 | 奥比中光科技集团股份有限公司 | Interior structured reconstruction method and apparatus, and computer-readable storage medium |
| CN113628330A (en) * | 2021-08-23 | 2021-11-09 | 北京京东乾石科技有限公司 | Finite space three-dimensional model construction method and system |
| CN113989376A (en) * | 2021-12-23 | 2022-01-28 | 贝壳技术有限公司 | Method and device for acquiring indoor depth information and readable storage medium |
| CN115239951A (en) * | 2022-06-08 | 2022-10-25 | 广东领慧建筑科技有限公司 | Wall surface segmentation and identification method and system based on point cloud data processing |
| CN115239951B (en) * | 2022-06-08 | 2023-09-15 | 广东领慧数字空间科技有限公司 | Wall surface segmentation recognition method and system based on point cloud data processing |
| CN114898066A (en) * | 2022-06-14 | 2022-08-12 | 中国电子工程设计院有限公司 | Non-convex indoor space identification reconstruction method and system |
| CN114898066B (en) * | 2022-06-14 | 2022-12-30 | 中国电子工程设计院有限公司 | Non-convex indoor space identification reconstruction method and system |
| CN115661378A (en) * | 2022-12-28 | 2023-01-31 | 北京道仪数慧科技有限公司 | Building model reconstruction method and system |
| CN117152374A (en) * | 2023-07-25 | 2023-12-01 | 北京建筑大学 | Digital twin establishing method of efficient cable dome structure based on three-dimensional laser point cloud |
| CN117152374B (en) * | 2023-07-25 | 2024-04-26 | 北京建筑大学 | Digital twin establishing method of efficient cable dome structure based on three-dimensional laser point cloud |
| CN117496086A (en) * | 2023-10-12 | 2024-02-02 | 南京林业大学 | Indoor geometric reconstruction method for semantic perception |
| CN117496086B (en) * | 2023-10-12 | 2024-05-07 | 南京林业大学 | Indoor geometric reconstruction method for semantic perception |
| CN117422810A (en) * | 2023-12-19 | 2024-01-19 | 深圳大学 | Structure and parameter guided indoor element regularization and relationship reasoning method and terminal |
| CN117422810B (en) * | 2023-12-19 | 2024-05-17 | 深圳大学 | Structure and parameter guided indoor element regularization and relationship reasoning method and terminal |
| CN117475110A (en) * | 2023-12-27 | 2024-01-30 | 北京市农林科学院信息技术研究中心 | Semantic three-dimensional reconstruction method and device for blade, electronic equipment and storage medium |
| CN117475110B (en) * | 2023-12-27 | 2024-04-05 | 北京市农林科学院信息技术研究中心 | Semantic three-dimensional reconstruction method and device for blade, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110009727B (en) | 2023-04-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110009727A (en) | A kind of indoor threedimensional model automatic reconfiguration method and system with structure semantics | |
| Ochmann et al. | Automatic reconstruction of fully volumetric 3D building models from oriented point clouds | |
| Zhang et al. | Online structure analysis for real-time indoor scene reconstruction | |
| Kim et al. | Fully automated registration of 3D data to a 3D CAD model for project progress monitoring | |
| US10762704B2 (en) | Method for establishing a deformable 3D model of an element, and associated system | |
| CN110363849A (en) | A kind of interior three-dimensional modeling method and system | |
| Qian et al. | Automatic unstructured all-hexahedral mesh generation from B-Reps for non-manifold CAD assemblies | |
| CN106504331A (en) | Tooth modeling method based on three-dimensional model search | |
| CN112784403A (en) | Numerical simulation method for establishing jointed rock mass discrete element model based on point cloud data | |
| Pound et al. | A patch-based approach to 3D plant shoot phenotyping | |
| Tabib et al. | Learning-based hole detection in 3D point cloud towards hole filling | |
| Hu et al. | Geometric feature enhanced line segment extraction from large-scale point clouds with hierarchical topological optimization | |
| CN107545602B (en) | Building modeling method under space topological relation constraint based on LiDAR point cloud | |
| Guo et al. | Line-based 3d building abstraction and polygonal surface reconstruction from images | |
| Palma et al. | Detection of geometric temporal changes in point clouds | |
| US8948498B1 (en) | Systems and methods to transform a colored point cloud to a 3D textured mesh | |
| Zhao et al. | Completing point clouds using structural constraints for large-scale points absence in 3D building reconstruction | |
| Wu et al. | [Retracted] Intelligent City 3D Modeling Model Based on Multisource Data Point Cloud Algorithm | |
| Cantzler | Improving architectural 3D reconstruction by constrained modelling | |
| Zhang et al. | Model reconstruction from cloud data | |
| Lu et al. | Image-based 3D reconstruction for Multi-Scale civil and infrastructure Projects: A review from 2012 to 2022 with new perspective from deep learning methods | |
| Huang | Geometric feature extraction and model reconstruction from unorganized points for reverse engineering of mechanical objects with arbitrary topology | |
| Denker et al. | On-line reconstruction of CAD geometry | |
| Wuttke et al. | Quality preserving fusion of 3d triangle meshes | |
| Acosta et al. | Geodesic-based manifold learning for parameterization of triangular meshes |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |