CN113362461A - Point cloud matching method and system based on semantic segmentation and scanning terminal - Google Patents
Point cloud matching method and system based on semantic segmentation and scanning terminal Download PDFInfo
- Publication number
- CN113362461A CN113362461A CN202110676965.2A CN202110676965A CN113362461A CN 113362461 A CN113362461 A CN 113362461A CN 202110676965 A CN202110676965 A CN 202110676965A CN 113362461 A CN113362461 A CN 113362461A
- Authority
- CN
- China
- Prior art keywords
- dimensional
- component
- components
- building
- point cloud
- 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 45
- 230000011218 segmentation Effects 0.000 title claims abstract description 13
- 238000013461 design Methods 0.000 claims abstract description 52
- 230000009466 transformation Effects 0.000 claims description 12
- 239000013598 vector Substances 0.000 claims description 11
- 239000011159 matrix material Substances 0.000 claims description 10
- 230000001131 transforming effect Effects 0.000 claims description 9
- 238000005516 engineering process Methods 0.000 claims description 5
- 238000002372 labelling Methods 0.000 claims description 4
- 238000012937 correction Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 abstract description 5
- 238000012545 processing Methods 0.000 description 14
- 238000005259 measurement Methods 0.000 description 10
- 238000010276 construction Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 3
- 230000006872 improvement Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000000712 assembly Effects 0.000 description 1
- 238000000429 assembly Methods 0.000 description 1
- 239000003638 chemical reducing agent Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 239000003292 glue Substances 0.000 description 1
- 238000012797 qualification Methods 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 238000010408 sweeping Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/10—Constructive solid geometry [CSG] using solid primitives, e.g. cylinders, cubes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/13—Architectural design, e.g. computer-aided architectural design [CAAD] related to design of buildings, bridges, landscapes, production plants or roads
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T19/00—Manipulating 3D models or images for computer graphics
- G06T19/20—Editing of 3D images, e.g. changing shapes or colours, aligning objects or positioning parts
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- 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
- G06T2207/10004—Still image; Photographic image
- G06T2207/10012—Stereo images
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Geometry (AREA)
- General Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Graphics (AREA)
- Software Systems (AREA)
- Architecture (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Structural Engineering (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Civil Engineering (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention discloses a point cloud matching method, a point cloud matching system and a scanning terminal based on semantic segmentation, wherein the point cloud matching method based on the semantic segmentation comprises the following steps: obtaining a building model according to scanning data of a building area, and segmenting the building model to obtain a plurality of three-dimensional components; the building model is subjected to two-dimensional operation to obtain a two-dimensional indoor graph; acquiring a component connection sequence of the two-dimensional house type graph according to semantic information of the corresponding three-dimensional component; comparing the connection sequence of the members with the connection sequence of the components in a building design drawing to obtain the corresponding relation between the two-dimensional members and the components, wherein the components of the building design drawing are obtained by dividing the building design drawing; and matching and associating the two-dimensional house type drawing with the building design drawing according to the corresponding relation. The invention can effectively process the matching between the imperfect geometric primitive in the actually measured house type graph and the perfect geometric primitive in the design drawing, and can be applied to the establishment of the correlation and the comparison between the actually measured result and the CAD design graph.
Description
Technical Field
The invention relates to a point cloud matching method and system based on semantic segmentation and a scanning terminal.
Background
The actual measurement is a method for truly reflecting product quality data through field test and measurement by using a measuring tool. And according to the related quality acceptance standard, the error of the metering control engineering quality data is within the range allowed by the national housing construction standard.
In order to strengthen the quality management of house buildings, improve the quality responsibility consciousness, strengthen the quality responsibility pursuit and ensure the engineering construction quality, 8-25 months in 2014, the urban and rural construction department of housing issues a temporary solution for the quality lifelong responsibility pursuit of the main project responsible for the five responsibilities of the building engineering, so that the engineering quality problem is emphatically emphasized in the construction industry. In addition, in recent years, the real estate industry is difficult to continue the early explosive growth, the market situation is not optimistic, the competition is more intense, and developers also need to pay more attention to the quality of products to gain the favor of customers.
The actual measurement of the construction project is a method for controlling the error of the project quality data within the allowable range of the national housing construction standard according to the relevant quality acceptance standard. The development of this work can better promote the project to do good physical quality work. Through the mode of establishing a product entity quality actual measurement system and carrying out the system, the engineering quality level of each stage of the project is objectively and truly reflected, the real-time improvement and continuous improvement of the entity quality are promoted, and the aim of one-time qualification of the entity quality is further achieved. Therefore, the actual measurement system of civil engineering is playing an important role in the building market as an important component of quality control in the Chinese building market, and has become a hard index for ensuring the quality of the building engineering.
The actual measurement needs many indexes, mainly including: flatness, the straightness that hangs down, hollowing and fracture, ceiling and the depth of parallelism on ground are extremely poor, the bathroom is waterproof, window caulking and beat glue, door opening size, the reducer of window, socket panel's height etc. involve the safety problem in the work progress if standardize like setting up of scaffold frame, three treasures four mouths use and protect whether satisfy the standard requirement, whether the construction power consumption is standard, whether fire-fighting equipment, fire-fighting equipment satisfy the standard requirement etc..
At present, the international technology for 3D live-action scanning and restoring high-precision building space mainly obtains 3D space depth data through laser radar scanning equipment and obtains texture data through an RGB camera. Other 3D live-action scanning techniques, such as those implemented by structured light 3D sensors, are 3D live-action spatial scanning restoration techniques based on 2D image sensors and sfm (structure From motion) techniques.
After the building space scanning equipment acquires the scanning data, the scanning data has a small utilization range and few functions.
Disclosure of Invention
The invention aims to overcome the defect that the scanning data obtained by the existing actual measurement real-scale tool in the prior art has small utilization range and few functions, and provides a point cloud matching method, a point cloud matching system and a scanning terminal based on semantic segmentation, which can generate a house type graph by using the scanning data, effectively process the matching between the imperfect geometric primitives in the house type graph actually measured and the perfect geometric primitives in a design drawing, and can be applied to the establishment of the correlation and the comparison between the actual measurement result and a CAD design graph.
The invention solves the technical problems through the following technical scheme:
a point cloud matching method based on semantic segmentation is characterized by comprising the following steps:
obtaining a building model according to scanning data of a building area, and segmenting the building model to obtain a plurality of three-dimensional components, wherein each three-dimensional component comprises semantic information;
the building model is subjected to two-dimensional operation to obtain a two-dimensional house type graph, the two-dimensional house type graph comprises a plurality of two-dimensional components, and the two-dimensional components of the two-dimensional house type graph correspond to the three-dimensional components of the building model;
acquiring a component connection sequence of the two-dimensional house type graph according to semantic information of a corresponding three-dimensional component;
comparing the member connection sequence with a component connection sequence in a building design drawing to obtain the corresponding relation between the two-dimensional members and the components, wherein the components of the building design drawing are obtained by dividing the building design drawing;
and matching and associating the two-dimensional house type drawing with the building design drawing according to the corresponding relation.
Preferably, said segmenting said building model to obtain a plurality of three-dimensional elements comprises:
acquiring a continuous area and a cavity area in the scanning data;
and identifying the continuous area as a three-dimensional component of a wall surface, a ceiling or a ground according to the scanning data corresponding to the gradienter data and the compass data, and identifying the hollow area as a three-dimensional component of a window or a door according to the position and the size of the hollow area.
Preferably, the point cloud matching method includes:
decomposing the correction scanning data into a plurality of voxel units, and acquiring information of the voxel units according to point cloud coordinates contained in the voxel units;
and combining the mutually associated voxel units into the three-dimensional member according to the information of the voxel units.
Preferably, the point cloud matching method includes:
scanning the building area through a scanning terminal to obtain the scanning data, and acquiring state information of the scanning terminal when the scanning terminal scans the building area, wherein the state information is associated with the scanning data;
semantic information of the three-dimensional members is acquired according to the size of the three-dimensional members, the positions of the three-dimensional members in the building model, the relationship between the three-dimensional members and state information associated with scanning data of the three-dimensional members.
Preferably, the two-dimensionalizing the building model to obtain a two-dimensional floor plan includes:
projecting the three-dimensional member to a preset plane to obtain a two-dimensional member corresponding to the three-dimensional member, wherein the preset plane is parallel to the ground of the building model, and the two-dimensional member comprises semantic information of the corresponding three-dimensional member;
and acquiring the two-dimensional house type graph according to all the acquired two-dimensional components.
Preferably, the obtaining of the connection sequence of the members of the two-dimensional house type graph according to the semantic information of the corresponding three-dimensional members includes:
selecting a target two-dimensional component as a starting point, acquiring an adjacent two-dimensional component connected with the target two-dimensional component along a direction, and recording the connection relation between the target two-dimensional component and the adjacent two-dimensional component by utilizing semantic information;
sequentially acquiring secondary adjacent two-dimensional components connected with the adjacent two-dimensional components along the direction until the secondary adjacent two-dimensional components are the target two-dimensional components;
and generating the component connection sequence according to the connection relation recorded by the semantic information.
Preferably, the comparing the connection sequence of the members with the connection sequence of the components in a building design drawing to obtain the corresponding relationship between the two-dimensional members and the components includes:
the method comprises the steps of obtaining a component of the architectural design drawing through an image recognition technology or a labeling mode, wherein the obtaining rule of the component is matched with the segmentation rule of the three-dimensional component, the component is provided with component information, and the building information comprises name and position information.
Preferably, the comparing the connection sequence of the members with the connection sequence of the components in a building design drawing to obtain the corresponding relationship between the two-dimensional members and the components includes:
acquiring a target assembly corresponding to a target two-dimensional component in the component connection sequence;
and traversing the two connection sequences simultaneously by taking the target two-dimensional member and the target assembly as starting points, and sequentially corresponding the two-dimensional members in the member connection sequence with the two-dimensional members in the assembly connection sequence to acquire the corresponding relation.
Preferably, the point cloud matching method includes:
selecting two-dimensional components from the two-dimensional house type graph, and respectively selecting a first position point from the two selected two-dimensional components;
acquiring components corresponding to the two-dimensional members in a building design drawing, and acquiring a second position point corresponding to the first position point in the two acquired components;
transforming the vector formed by the two first position points by a transformation matrix and then coinciding the vector formed by the two second position points;
and transforming the two-dimensional member by using the transformation matrix, wherein the transformed two-dimensional member is used for comparing with a corresponding component so as to verify the geometric constraint of the two-dimensional house type drawing and the architectural design drawing.
The invention also provides a point cloud matching system which is characterized by being used for realizing the point cloud matching method.
The invention also provides a scanning terminal, which is characterized in that the scanning terminal is used for the point cloud matching system, and the scanning terminal is used for acquiring the scanning data.
On the basis of the common knowledge in the field, the above preferred conditions can be combined randomly to obtain the preferred embodiments of the invention.
The positive progress effects of the invention are as follows:
the invention can effectively process the matching between the imperfect geometric primitive in the house type graph actually measured and the perfect geometric primitive in the design drawing, can be applied to the establishment of the correlation and the comparison between the actual measurement result and the CAD design graph, and is an effective method for accurately and quickly establishing the matching correlation between the heterogeneous data generated by different methods of the same building.
Drawings
Fig. 1 is a schematic structural diagram of a point cloud matching system according to embodiment 1 of the present invention.
Fig. 2 is a schematic structural view of a member connection sequence in embodiment 1 of the present invention.
Fig. 3 is a schematic structural diagram of a component connection sequence in embodiment 1 of the present invention.
Fig. 4 is a schematic structural diagram of a correspondence relationship in embodiment 1 of the present invention.
Fig. 5 is a schematic structural view of another connection sequence of members in embodiment 1 of the present invention.
Fig. 6 is a schematic structural diagram of another corresponding relationship in embodiment 1 of the present invention.
Fig. 7 is a flowchart of a point cloud matching method according to embodiment 1 of the present invention.
Fig. 8 is another flowchart of the point cloud matching method according to embodiment 1 of the present invention.
Detailed Description
The invention is further illustrated by the following examples, which are not intended to limit the scope of the invention.
Example 1
Referring to fig. 1 and fig. 2, the present embodiment provides a point cloud matching system, which includes a scanning terminal 11 and a processing module, where the processing module may be an intelligent terminal, a server, or a processor equipped in the scanning terminal.
In this embodiment, the scanning terminal is provided, the scanning terminal is a laser radar, the scanning terminal may also be a 3D scanning robot (such as a sweeping robot), the scanning terminal includes the processing module, and the point cloud matching system further includes an intelligent terminal 12.
The processing module is used for acquiring a building model according to scanning data of a building area and segmenting the building model to acquire a plurality of three-dimensional components, and each three-dimensional component comprises semantic information;
the processing module is used for bidimensionalizing the building model to obtain a two-dimensional house type graph, the two-dimensional house type graph comprises a plurality of two-dimensional components, and the two-dimensional components of the two-dimensional house type graph correspond to the three-dimensional components of the building model;
in addition, the intelligent terminal can also be used for two-dimensionalizing the building model to obtain the two-dimensional house type graph.
The intelligent terminal is used for acquiring the member connection sequence of the two-dimensional house type diagram according to the semantic information of the corresponding three-dimensional member;
the intelligent terminal is also used for comparing the member connection sequence with the component connection sequence in a building design drawing to obtain the corresponding relation between the two-dimensional member and the component, wherein the component of the building design drawing is obtained by dividing the building design drawing;
and the intelligent terminal is also used for matching and associating the two-dimensional house type graph with the building design graph according to the corresponding relation.
The processing module or the intelligent terminal is used for:
acquiring a continuous area and a cavity area in the scanning data;
and identifying the continuous area as a three-dimensional component of a wall surface, a ceiling or a ground according to the scanning data corresponding to the gradienter data and the compass data, and identifying the hollow area as a three-dimensional component of a window or a door according to the position and the size of the hollow area.
The scanning data is divided into continuous dense areas and hollow areas. The relatively flat continuous dense areas are easily obtained as walls, ceilings and floors, which can be determined from the gradienter and compass carried by the scanning terminal.
The hollow areas are distinguished according to the height from the ground, the door is arranged at the height of 0 from the floor, and the window is arranged at a certain distance from the floor. The boundary contour coordinate points of all members such as walls, doors, windows, floors and ceilings can be obtained.
Further, the processing module or the intelligent terminal is configured to:
projecting the three-dimensional member to a preset plane to obtain a two-dimensional member corresponding to the three-dimensional member, wherein the preset plane is parallel to the ground of the building model, and the two-dimensional member comprises semantic information of the corresponding three-dimensional member;
and acquiring the two-dimensional house type graph according to all the acquired two-dimensional components.
By the method of acquiring the plan view of the scan data, the Z-coordinate of all three-dimensional members is set to 0, and the two-dimensional member coordinate generated by projection can be obtained. By storing the coordinate information of the two-dimensional member in the Xml format, and the Z-axis information such as the height of a house and the height of a window, the outline of the whole room can be recorded.
The processing module or the intelligent terminal is further used for:
selecting a target two-dimensional component as a starting point, acquiring an adjacent two-dimensional component connected with the target two-dimensional component along a direction, and recording the connection relation between the target two-dimensional component and the adjacent two-dimensional component by utilizing semantic information;
sequentially acquiring secondary adjacent two-dimensional components connected with the adjacent two-dimensional components along the direction until the secondary adjacent two-dimensional components are the target two-dimensional components;
and generating the component connection sequence according to the connection relation recorded by the semantic information.
In the embodiment, the door is taken as a starting component, the counterclockwise direction is taken, and the communication relation is calculated according to the coordinates. The coordinates of the end point of the door in the counterclockwise direction are taken, the member (in this embodiment, a wall) which is common to the door and the end point is determined, and a communication relationship is established between the door and the member. And then, the other end point of the member (the wall) is used for carrying out the next round of communication relation searching. Until another end point of the starting gate is found. And recording all semantic connected nodes to obtain a semantic connected chain (namely).
Referring to fig. 2, the two-dimensional house type graph generates a member connection order recorded with semantic information.
Referring to FIG. 3, the architectural design map generation utilizes semantic record component connection order.
Referring to fig. 4, the member connection order is compared with the assembly connection order in the architectural design drawing to obtain the correspondence between the two-dimensional members and the assemblies.
Referring to fig. 5, another of the two-dimensional house type graphs generates a connection order of the building blocks recorded with semantic information. The corresponding relationship as shown in fig. 6 is obtained by comparing the two-dimensional house layout with the architectural design drawing.
The processing module or the intelligent terminal is further used for processing the architectural design drawing:
the method comprises the steps of obtaining a component of the architectural design drawing through an image recognition technology or a labeling mode, wherein the obtaining rule of the component is matched with the segmentation rule of the three-dimensional component, the component is provided with component information, and the building information comprises name and position information.
The drawings given by different design houses are slightly different. Typically the door is a quarter circle and the wall is a double line segment with the inner line segment. A window is a segment of a wall. And coordinates of members such as doors and windows on the wall on the drawing can be obtained by manual clicking identification. And generating a component connection sequence.
The processing module or the intelligent terminal is further used for comparing the specific component connection sequence with the component connection sequence in the building design drawing:
acquiring a target assembly corresponding to a target two-dimensional component in the component connection sequence;
and traversing the two connection sequences simultaneously by taking the target two-dimensional member and the target assembly as starting points, and sequentially corresponding the two-dimensional members in the member connection sequence with the two-dimensional members in the assembly connection sequence to acquire the corresponding relation.
After the corresponding relation is obtained, the semantics, the coordinate information and various data of the two-dimensional component can be obtained by the component, and the user can conveniently check the corresponding relation.
Corresponding to the semantic link chain (component connection order) of the scanned data, the semantic link chain (component connection order) in the architectural design drawing also selects a gate as a starting point, and traverses the two chains simultaneously.
If a semantic node with no match is found, the starting point selects an error, and another gate is selected as the starting point of the semantic connected chain until the two semantic connected chains are traversed to be completely matched. The relationship between the 3D point cloud component and the drawing component can be matched according to the correspondence on the chain.
The processing module or the intelligent terminal is further used for:
selecting two-dimensional components from the two-dimensional house type graph, and respectively selecting a first position point from the two selected two-dimensional components;
acquiring components corresponding to the two-dimensional members in a building design drawing, and acquiring a second position point corresponding to the first position point in the two acquired components;
transforming the vector formed by the two first position points by a transformation matrix and then coinciding the vector formed by the two second position points;
and transforming the two-dimensional member by using the transformation matrix, wherein the transformed two-dimensional member is used for comparing with a corresponding component so as to verify the geometric constraint of the two-dimensional house type drawing and the architectural design drawing.
And taking the geometric center of the wall and the geometric center of the door in the two-dimensional member as a vector, and taking the geometric center of the wall and the geometric center of the door in the architectural design drawing as a vector, wherein the two have a corresponding relation.
And according to the coordinate transformation operations of scaling, translation and rotation, the two vectors are coincided, and a coordinate transformation matrix in the process is obtained. And transforming the two-dimensional component according to the transformation matrix coordinate to obtain the coordinate of the other component, and checking whether the coordinate of the other component is similar to the coordinate of the matched component. If the degree of closeness is high, the constraint check is passed.
Referring to fig. 7, with the point cloud matching system, the embodiment further provides a point cloud matching method, including:
101, performing two-dimensional transformation on the building model to obtain a two-dimensional house type graph, wherein the two-dimensional house type graph comprises a plurality of two-dimensional components, and the two-dimensional components of the two-dimensional house type graph correspond to the three-dimensional components of the building model;
102, acquiring a component connection sequence of the two-dimensional house type graph according to semantic information of a corresponding three-dimensional component;
103, comparing the connection sequence of the members with the connection sequence of the components in a building design drawing to obtain the corresponding relation between the two-dimensional members and the components, wherein the components of the building design drawing are obtained by dividing the building design drawing;
and 104, matching and associating the two-dimensional house type drawing with the building design drawing according to the corresponding relation.
Wherein step 100 comprises:
acquiring a continuous area and a cavity area in the scanning data;
and identifying the continuous area as a three-dimensional component of a wall surface, a ceiling or a ground according to the scanning data corresponding to the gradienter data and the compass data, and identifying the hollow area as a three-dimensional component of a window or a door according to the position and the size of the hollow area.
Specifically, step 101 includes:
projecting the three-dimensional member to a preset plane to obtain a two-dimensional member corresponding to the three-dimensional member, wherein the preset plane is parallel to the ground of the building model, and the two-dimensional member comprises semantic information of the corresponding three-dimensional member;
and acquiring the two-dimensional house type graph according to all the acquired two-dimensional components.
Referring to fig. 8, further, step 102 includes:
and 1023, acquiring an adjacent two-dimensional component connected with the target two-dimensional component along the direction, recording the connection relation between the target two-dimensional component and the adjacent two-dimensional component by using semantic information, judging whether the latest adjacent two-dimensional component is the target two-dimensional component of the starting point, if so, executing a step 1024, and otherwise, executing a step 1022 again.
And 1024, generating the component connection sequence according to the connection relation recorded by the semantic information.
Further, step 103 comprises:
the method comprises the steps of obtaining a component of the architectural design drawing through an image recognition technology or a labeling mode, wherein the obtaining rule of the component is matched with the segmentation rule of the three-dimensional component, the component is provided with component information, and the building information comprises name and position information.
Step 103 further comprises:
acquiring a target assembly corresponding to a target two-dimensional component in the component connection sequence;
and traversing the two connection sequences simultaneously by taking the target two-dimensional member and the target assembly as starting points, and sequentially corresponding the two-dimensional members in the member connection sequence with the two-dimensional members in the assembly connection sequence to acquire the corresponding relation.
Step 103 further comprises:
selecting two-dimensional components from the two-dimensional house type graph, and respectively selecting a first position point from the two selected two-dimensional components;
acquiring components corresponding to the two-dimensional members in a building design drawing, and acquiring a second position point corresponding to the first position point in the two acquired components;
transforming the vector formed by the two first position points by a transformation matrix and then coinciding the vector formed by the two second position points;
and transforming the two-dimensional member by using the transformation matrix, wherein the transformed two-dimensional member is used for comparing with a corresponding component so as to verify the geometric constraint of the two-dimensional house type drawing and the architectural design drawing.
The point cloud matching method, system and scanning terminal based on semantic segmentation can effectively process the matching between imperfect geometric primitives in a house type graph actually measured and perfect geometric primitives in a design drawing, can be applied to the establishment of association and comparison between an actual measurement result and a CAD design graph, and is an effective method for accurately and quickly establishing the matching association between heterogeneous data generated by different methods of the same building.
Example 2
This embodiment is substantially the same as embodiment 1 except that:
specifically, the processing module is configured to:
decomposing the correction scanning data into a plurality of voxel units, and acquiring information of the voxel units according to point cloud coordinates contained in the voxel units;
and combining the mutually associated voxel units into the three-dimensional member according to the information of the voxel units.
Through the steps, the scanning data can be divided into three-dimensional components with semantic information, wherein the three-dimensional components comprise walls, ceilings, floors, windows, doors, beams, pillars and the like.
Further, the processing module is configured to:
scanning the building area through a scanning terminal to obtain the scanning data, and acquiring state information of the scanning terminal when the scanning terminal scans the building area, wherein the state information is associated with the scanning data;
semantic information of the three-dimensional members is acquired according to the size of the three-dimensional members, the positions of the three-dimensional members in the building model, the relationship between the three-dimensional members and state information associated with scanning data of the three-dimensional members.
While specific embodiments of the invention have been described above, it will be appreciated by those skilled in the art that these are by way of example only, and that the scope of the invention is defined by the appended claims. Various changes and modifications to these embodiments may be made by those skilled in the art without departing from the spirit and scope of the invention, and these changes and modifications are within the scope of the invention.
Claims (10)
1. A point cloud matching method based on semantic segmentation is characterized by comprising the following steps:
obtaining a building model according to scanning data of a building area, and segmenting the building model to obtain a plurality of three-dimensional components, wherein each three-dimensional component comprises semantic information;
the building model is subjected to two-dimensional operation to obtain a two-dimensional house type graph, the two-dimensional house type graph comprises a plurality of two-dimensional components, and the two-dimensional components of the two-dimensional house type graph correspond to the three-dimensional components of the building model;
acquiring a component connection sequence of the two-dimensional house type graph according to semantic information of a corresponding three-dimensional component;
comparing the member connection sequence with a component connection sequence in a building design drawing to obtain the corresponding relation between the two-dimensional members and the components, wherein the components of the building design drawing are obtained by dividing the building design drawing;
and matching and associating the two-dimensional house type drawing with the building design drawing according to the corresponding relation.
2. The point cloud matching method of claim 1, wherein said segmenting the building model to obtain a number of three-dimensional components comprises:
acquiring a continuous area and a cavity area in the scanning data;
and identifying the continuous area as a three-dimensional component of a wall surface, a ceiling or a ground according to the scanning data corresponding to the gradienter data and the compass data, and identifying the hollow area as a three-dimensional component of a window or a door according to the position and the size of the hollow area.
3. The point cloud matching method of claim 1, wherein the point cloud matching method comprises:
decomposing the correction scanning data into a plurality of voxel units, and acquiring information of the voxel units according to point cloud coordinates contained in the voxel units;
combining the mutually associated voxel units into the three-dimensional member according to the information of the voxel units;
scanning the building area through a scanning terminal to obtain the scanning data, and acquiring state information of the scanning terminal when the scanning terminal scans the building area, wherein the state information is associated with the scanning data;
semantic information of the three-dimensional members is acquired according to the size of the three-dimensional members, the positions of the three-dimensional members in the building model, the relationship between the three-dimensional members and state information associated with scanning data of the three-dimensional members.
4. The point cloud matching method of claim 1, wherein said two-dimensionalizing the building model to obtain a two-dimensional house-type map comprises:
projecting the three-dimensional member to a preset plane to obtain a two-dimensional member corresponding to the three-dimensional member, wherein the preset plane is parallel to the ground of the building model, and the two-dimensional member comprises semantic information of the corresponding three-dimensional member;
and acquiring the two-dimensional house type graph according to all the acquired two-dimensional components.
5. The point cloud matching method of claim 1, wherein the obtaining of the component connection order of the two-dimensional house type graph according to semantic information of the corresponding three-dimensional components comprises:
selecting a target two-dimensional component as a starting point, acquiring an adjacent two-dimensional component connected with the target two-dimensional component along a direction, and recording the connection relation between the target two-dimensional component and the adjacent two-dimensional component by utilizing semantic information;
sequentially acquiring secondary adjacent two-dimensional components connected with the adjacent two-dimensional components along the direction until the secondary adjacent two-dimensional components are the target two-dimensional components;
and generating the component connection sequence according to the connection relation recorded by the semantic information.
6. The point cloud matching method of claim 5, wherein said comparing the component connection order with a component connection order in a building design drawing to obtain the correspondence between the two-dimensional components and the components comprises:
the method comprises the steps of obtaining a component of the architectural design drawing through an image recognition technology or a labeling mode, wherein the obtaining rule of the component is matched with the segmentation rule of the three-dimensional component, the component is provided with component information, and the building information comprises name and position information.
7. The point cloud matching method of claim 6, wherein said comparing the component connection order with a component connection order in a building design drawing to obtain the correspondence between the two-dimensional components and the components comprises:
acquiring a target assembly corresponding to a target two-dimensional component in the component connection sequence;
and traversing the two connection sequences simultaneously by taking the target two-dimensional member and the target assembly as starting points, and sequentially corresponding the two-dimensional members in the member connection sequence with the two-dimensional members in the assembly connection sequence to acquire the corresponding relation.
8. The point cloud matching method of claim 1, wherein the point cloud matching method comprises:
selecting two-dimensional components from the two-dimensional house type graph, and respectively selecting a first position point from the two selected two-dimensional components;
acquiring components corresponding to the two-dimensional members in a building design drawing, and acquiring a second position point corresponding to the first position point in the two acquired components;
transforming the vector formed by the two first position points by a transformation matrix and then coinciding the vector formed by the two second position points;
and transforming the two-dimensional member by using the transformation matrix, wherein the transformed two-dimensional member is used for comparing with a corresponding component so as to verify the geometric constraint of the two-dimensional house type drawing and the architectural design drawing.
9. A point cloud matching system for implementing the point cloud matching method of any one of claims 1 to 8.
10. A scanning terminal for use in the point cloud matching system of claim 9, the scanning terminal being configured to acquire the scan data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110676965.2A CN113362461B (en) | 2021-06-18 | 2021-06-18 | Point cloud matching method and system based on semantic segmentation and scanning terminal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110676965.2A CN113362461B (en) | 2021-06-18 | 2021-06-18 | Point cloud matching method and system based on semantic segmentation and scanning terminal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113362461A true CN113362461A (en) | 2021-09-07 |
| CN113362461B CN113362461B (en) | 2024-04-02 |
Family
ID=77534889
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110676965.2A Active CN113362461B (en) | 2021-06-18 | 2021-06-18 | Point cloud matching method and system based on semantic segmentation and scanning terminal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113362461B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114372313A (en) * | 2022-01-07 | 2022-04-19 | 上海盎维信息技术有限公司 | Image processing method and system for actual measurement and laser scanner |
| CN115130191A (en) * | 2022-07-28 | 2022-09-30 | 盎维云(深圳)计算有限公司 | Positive and reverse BIM fusion modeling engine |
| CN117236079A (en) * | 2023-11-14 | 2023-12-15 | 元宇智数(深圳)科技有限公司 | Engineering drawing labeling method, device, system and medium based on laser scanning |
| CN118587461A (en) * | 2024-08-06 | 2024-09-03 | 元宇智数(深圳)科技有限公司 | Model matching method, device and storage medium for house type |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101944239A (en) * | 2009-07-08 | 2011-01-12 | 富士通株式会社 | Method and device for segmenting 3D model and image processing system with device |
| CN104331699A (en) * | 2014-11-19 | 2015-02-04 | 重庆大学 | Planar fast search and comparison method of three-dimensional point cloud |
| CN105469388A (en) * | 2015-11-16 | 2016-04-06 | 集美大学 | Building point cloud registration algorithm based on dimension reduction |
| CN106780735A (en) * | 2016-12-29 | 2017-05-31 | 深圳先进技术研究院 | A kind of semantic map constructing method, device and a kind of robot |
| US20170358087A1 (en) * | 2016-06-10 | 2017-12-14 | The Board Of Trustees Of The Leland Stanford Junior University | Systems and Methods for Performing Three-Dimensional Semantic Parsing of Indoor Spaces |
| CN109828284A (en) * | 2019-03-18 | 2019-05-31 | 上海盎维信息技术有限公司 | The method and device of actual measurement actual quantities based on artificial intelligence |
| CN110419049A (en) * | 2017-03-17 | 2019-11-05 | 奇跃公司 | Room layout estimation method and technology |
| CN111047548A (en) * | 2020-03-12 | 2020-04-21 | 腾讯科技（深圳）有限公司 | Attitude transformation data processing method and device, computer equipment and storage medium |
| CN112102411A (en) * | 2020-11-02 | 2020-12-18 | 中国人民解放军国防科技大学 | Visual positioning method and device based on semantic error image |
-
2021
- 2021-06-18 CN CN202110676965.2A patent/CN113362461B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101944239A (en) * | 2009-07-08 | 2011-01-12 | 富士通株式会社 | Method and device for segmenting 3D model and image processing system with device |
| CN104331699A (en) * | 2014-11-19 | 2015-02-04 | 重庆大学 | Planar fast search and comparison method of three-dimensional point cloud |
| CN105469388A (en) * | 2015-11-16 | 2016-04-06 | 集美大学 | Building point cloud registration algorithm based on dimension reduction |
| US20170358087A1 (en) * | 2016-06-10 | 2017-12-14 | The Board Of Trustees Of The Leland Stanford Junior University | Systems and Methods for Performing Three-Dimensional Semantic Parsing of Indoor Spaces |
| CN106780735A (en) * | 2016-12-29 | 2017-05-31 | 深圳先进技术研究院 | A kind of semantic map constructing method, device and a kind of robot |
| CN110419049A (en) * | 2017-03-17 | 2019-11-05 | 奇跃公司 | Room layout estimation method and technology |
| CN109828284A (en) * | 2019-03-18 | 2019-05-31 | 上海盎维信息技术有限公司 | The method and device of actual measurement actual quantities based on artificial intelligence |
| CN111047548A (en) * | 2020-03-12 | 2020-04-21 | 腾讯科技（深圳）有限公司 | Attitude transformation data processing method and device, computer equipment and storage medium |
| CN112102411A (en) * | 2020-11-02 | 2020-12-18 | 中国人民解放军国防科技大学 | Visual positioning method and device based on semantic error image |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114372313A (en) * | 2022-01-07 | 2022-04-19 | 上海盎维信息技术有限公司 | Image processing method and system for actual measurement and laser scanner |
| CN115130191A (en) * | 2022-07-28 | 2022-09-30 | 盎维云(深圳)计算有限公司 | Positive and reverse BIM fusion modeling engine |
| CN117236079A (en) * | 2023-11-14 | 2023-12-15 | 元宇智数(深圳)科技有限公司 | Engineering drawing labeling method, device, system and medium based on laser scanning |
| CN117236079B (en) * | 2023-11-14 | 2024-03-15 | 元宇智数(深圳)科技有限公司 | Engineering drawing labeling method, device, system and medium based on laser scanning |
| CN118587461A (en) * | 2024-08-06 | 2024-09-03 | 元宇智数(深圳)科技有限公司 | Model matching method, device and storage medium for house type |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113362461B (en) | 2024-04-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Romero-Jarén et al. | Automatic segmentation and classification of BIM elements from point clouds | |
| Mill et al. | Combined 3D building surveying techniques–terrestrial laser scanning (TLS) and total station surveying for BIM data management purposes | |
| Huber et al. | Using laser scanners for modeling and analysis in architecture, engineering, and construction | |
| Murali et al. | Indoor Scan2BIM: Building information models of house interiors | |
| CN113362461A (en) | Point cloud matching method and system based on semantic segmentation and scanning terminal | |
| Hong et al. | Semi-automated approach to indoor mapping for 3D as-built building information modeling | |
| Riveiro et al. | Automatic morphologic analysis of quasi‐periodic masonry walls from LiDAR | |
| Dore et al. | Semi-automatic generation of as-built BIM façade geometry from laser and image data | |
| Previtali et al. | Towards automatic indoor reconstruction of cluttered building rooms from point clouds | |
| Murphy et al. | Developing historic building information modelling guidelines and procedures for architectural heritage in Ireland | |
| US20240249488A1 (en) | Tracking an ongoing construction by using fiducial markers | |
| Gómez‐García‐Bermejo et al. | Automated registration of 3D scans using geometric features and normalized color data | |
| Lo Brutto et al. | Integrating topographic, photogrammetric and laser scanning techniques for a scan-to-bim process | |
| Sun et al. | Data processing workflows from low-cost digital survey to various applications: Three case studies of Chinese historic architecture | |
| Pöchtrager et al. | Automated reconstruction of historic roof structures from point clouds–Development and examples | |
| Pöchtrager et al. | Digital reconstruction of historic roof structures: Developing a workflow for a highly automated analysis | |
| CN114372313A (en) | Image processing method and system for actual measurement and laser scanner | |
| Kaushik et al. | Accelerated patch-based planar clustering of noisy range images in indoor environments for robot mapping | |
| Andreasson et al. | Improving plane extraction from 3d data by fusing laser data and vision | |
| Sammartano et al. | Parametric generation in HBIM workflows for slam-based data: Discussing expectations on suitability and accuracy | |
| Xiong et al. | Reality capture and site-scanning techniques for material reuse planning | |
| CN113436244B (en) | Model processing method and system for actual measurement actual quantity and laser radar | |
| Riveiro et al. | Laser scanning for the evaluation of historic structures | |
| Sareen et al. | Hierarchical data clustering approach for segmenting colored three-dimensional point clouds of building interiors | |
| Kuzmić et al. | Application of Terrestrial Laser Scanning Technology for the Purpose of Creating 3D Models of Objects |
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 | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Country or region after: China Address after: Room 390, Building 17, No. 2723 Fuchunwan Avenue, Chunjiang Street, Fuyang District, Hangzhou City, Zhejiang Province, 311400 Applicant after: Angrui (Hangzhou) Information Technology Co.,Ltd. Address before: 201703 No.206, building 1, no.3938 Huqingping Road, Qingpu District, Shanghai Applicant before: UNRE (SHANGHAI) INFORMATION TECHNOLOGY Co.,Ltd. Country or region before: China |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |