CN110189400A - A kind of three-dimensional rebuilding method, three-dimensional reconstruction system, mobile terminal and storage device - Google Patents
A kind of three-dimensional rebuilding method, three-dimensional reconstruction system, mobile terminal and storage device Download PDFInfo
- Publication number
- CN110189400A CN110189400A CN201910418032.6A CN201910418032A CN110189400A CN 110189400 A CN110189400 A CN 110189400A CN 201910418032 A CN201910418032 A CN 201910418032A CN 110189400 A CN110189400 A CN 110189400A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- cloud data
- dimensional
- global
- speckle image
- 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
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
-
- 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
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Length Measuring Devices By Optical Means (AREA)
- Image Processing (AREA)
Abstract
The present invention provides a kind of three-dimensional rebuilding method, three-dimensional reconstruction system, mobile terminal and storage devices, three-dimensional rebuilding method includes: the point cloud data (i.e. global point cloud data) for obtaining first frame 3 d measurement data, then choosing region corresponding with global point cloud data, there are the regional areas of overlay region to measure, local point cloud data is obtained to be registrated again and update global point cloud data, this process is repeated until completing the measurement in all surface region, global optimization processing finally is carried out to the global point cloud data updated after being measured, obtains point cloud model.The present invention passes through the method to break the whole up into parts and carries out three-dimensional reconstruction, and the measurement data that available high density and high-precision have both is more advantageous especially for the three-dimensional measurement of large sized object.
Description
Technical field
The invention belongs to three-dimensional sensing and fields of measurement more particularly to a kind of three-dimensional rebuilding method, three-dimensional reconstruction system, shiftings
Dynamic terminal and storage device.
Background technique
In recent years, development at full speed is occurring for 3D industry.From 3D 3D printing is measured, has formd and shown unique characteristics
Integrated industrial chain, the survey for large sized object (large fan blade, oil tank, hull, automobile panel, large-scale sculpture etc.)
For amount with modeling at present there are still certain difficulty, it is larger to be mainly reflected in testee size, and it is more tired to obtain high density data
Difficulty, high density and the high-precision of data are difficult to guarantee simultaneously.Therefore there are also to be developed for the prior art.
Summary of the invention
Technical problem to be solved by the present invention lies in provide a kind of three-dimensional rebuilding method, three-dimensional reconstruction system, it is mobile eventually
End and storage device, it is intended to solve existing measurement modeling method and be difficult to ensure the high density of data simultaneously for large sized object
With high-precision.
In order to solve the above technical problems, the invention is realized in this way, a kind of three-dimensional rebuilding method includes the following steps:
Step S1, three-dimensional reconstruction coordinate system is demarcated；Obtain the point of the first frame 3 d measurement data of testee
Cloud data, referred to as global point cloud data, the coordinate system based on the first frame 3 d measurement data are known as global coordinate system；
Step S2, to the surface progress three-dimensional measurement of some regional area of testee, binocular stereo vision principle is utilized
Carry out three-dimensional reconstruction, obtain the point cloud data of the regional area, referred to as local point cloud data, wherein the regional area with
There are overlay regions in the global point cloud data corresponding region；
Step S3, the local point cloud data is transformed under the global coordinate system, it will be described according to the overlay region
Local point cloud data is registrated with the global point cloud data, updates the global point cloud data；
Step S4, keep the testee motionless, transformation measurement visual angle measures the testee, repeats institute
Step S2 to the step S3 is stated, until completing the measurement to the testee；
Step S5, global optimization processing is carried out to the global point cloud data updated after being measured, obtains point cloud model.
Further, the step S2 specifically includes the following steps:
Step S21, three-dimensional measurement is carried out to the surface of the regional area, obtains the first speckle pattern of the regional area
Picture and the second speckle image；
Step S22, the whole Pixel-level of each pixel in first speckle image is determined in second speckle image
Corresponding points；
Step S23, according to each pixel coordinate in the whole Pixel-level corresponding points and first speckle image,
Sub-pix is carried out to second speckle image and corresponds to point search, obtains the sub-pix corresponding points in second speckle image；
Step S24, using binocular stereo vision principle, in conjunction with second speckle image sub-pix corresponding relationship into
Row three-dimensional reconstruction obtains the local point cloud data on the testee surface.
Further, the step S22 specifically includes the following steps:
Step S221, first speckle image and second speckle image are chosen according to relevant calculation formula identical
(the 2w of sizem+1)×(2wm+ 1) pixel region carries out related operation, the relevant calculation formula are as follows:
Wherein, IL(uL, vL) indicate selected areas in the first speckle image planar point (u, v) gray value, IR(uR, vR) table
Show the gray value of the second speckle image planar point (u, v) in selected areas,WithRespectively indicate the first speckle image, second
The average gray of speckle image selected areas, ω indicate related coefficient, WmIndicate the pixel number of certain amount；
Step S222, corresponding corresponding points are made when choosing the maximum value of related coefficient calculated value and being more than the threshold value of setting
For whole Pixel-level corresponding points.
Further, the step S23 specifically includes the following steps:
It is (2w that window size is created in first speckle imagem+1)×(2wm+ 1) reference child window；
By the non-linear space correlation function ω (s) under second order parallax model as newton-La Fuxun interative computation to
Majorized function:
Wherein,ud、vdFor
Zeroth order parallax,For single order parallax,For second order parallax,
Δ u, Δ v are the difference in the first speckle image between other pixels and central pixel point, second order parallax model are as follows:
According to preset iterative steps, and according to interative computation formulaIt is iterated fortune
It calculates, determines the correlation function value s that last time interative computation calculatesNFor end value,
Wherein, the value range of N is the integer more than or equal to 1, and n indicates the number of variable s, s0For initial value, by whole picture
It is calculated as initial value according to the end value and second order parallax model position when the related coefficient maximum value that plain relevant calculation obtains
The sub-pix corresponding points out.
Further, in the step S3, according to the overlay region by the local point cloud data and the global point cloud
Data are registrated, and are registrated based on the nearest point frame of iteration, the method for use tactful from thick to thin point to model.
Further, in the step S5, optimizing processing to global point cloud data includes: to establish unification comprising all
Registration error is shared out equally each overlay region by the solution of the error function by the error assessment function of overlapping region
Domain reduces registration error accumulation.
A kind of three-dimensional reconstruction system, comprising:
Three-dimensional measurement module, is used for: demarcating to three-dimensional reconstruction coordinate system；The first frame three-dimensional for obtaining testee is surveyed
The point cloud data of data, referred to as global point cloud data are measured, the coordinate system based on the first frame 3 d measurement data is known as the overall situation
Coordinate system；
Be also used to carry out three-dimensional measurement to the surface of testee some regional area, using binocular stereo vision principle into
Row three-dimensional reconstruction obtains the point cloud data of the regional area, referred to as local point cloud data, wherein the regional area and institute
Stating the corresponding region of global point cloud data, there are overlay regions；
Global point cloud data update module, for the local point cloud data to be transformed under the global coordinate system, root
The local point cloud data is registrated with the global point cloud data according to the overlay region, updates the global point cloud number
According to；
The data-optimized module of global point cloud, for being carried out at global optimization to the global point cloud data obtained after being measured
Reason, obtains point cloud model.
Further, the three-dimensional reconstruction system is mounted in progress three-dimensional measurement on unmanned aerial vehicle platform
A kind of mobile terminal, comprising: processor, the memory connecting with processor communication, the memory are stored with meter
Calculation machine program, the computer program realize three-dimensional rebuilding method as described above for being performed；The processor is used for
The computer program in the memory is called, to realize three-dimensional rebuilding method as described above.
A kind of storage device, the storage device are stored with computer program, the computer program can be performed with
Realize three-dimensional rebuilding method as described above.
Compared with prior art, the present invention beneficial effect is: the present invention is to carry out three based on binocular stereo vision principle
Dimension is rebuild, and obtains the point cloud data (i.e. global point cloud data) of first frame 3 d measurement data first, then selection and global point
There are the regional areas of overlay region to measure in cloud data corresponding region, obtains local point cloud data and is registrated and is updated again
Global point cloud data repeat this process until completing the measurement in all surface region.The present invention pass through the method that breaks the whole up into parts into
The measurement data that row three-dimensional reconstruction, available high density and high-precision have both is surveyed especially for the three-dimensional of large sized object
Amount, it is more advantageous.
Detailed description of the invention
Fig. 1 is a kind of flow chart of three-dimensional rebuilding method embodiment of the invention.
Fig. 2 is a kind of embodiment flow chart of acquisition part point cloud data of the invention.
Fig. 3 is a kind of embodiment functional module structure figure of three-dimensional reconstruction system of the invention.
Fig. 4 is a kind of Principle of Communication figure of three-dimensional reconstruction mobile terminal embodiment of the invention.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
A kind of three-dimensional rebuilding method embodiment provided by the invention, process is as shown in Figure 1, include the following steps S1-S5:
Step S1, three-dimensional reconstruction coordinate system is demarcated；Obtain the point of the first frame 3 d measurement data of testee
Cloud data, referred to as global point cloud data, the coordinate system based on the first frame 3 d measurement data are known as global coordinate system.
It is illustrated by taking photo taking type three-dimensional measurement as an example, includes two cameras and a projection dress in three-dimensional reconstruction system
It sets, forms the three-dimensional measurement unit of an active binocular, be used for real-time reconstruction three dimensional point cloud.Three-dimensional reconstruction coordinate system mark
Surely include the calibration to two cameras (left camera and right camera), specifically obtain inner parameter, the lens distortion ginseng of two cameras
Rigid body translation between several and two camera coordinates systems obtains corresponding rigid body translation, that is, completes corresponding component in this step
Calibration.In calibration process, the point cloud data of the first frame 3 d measurement data of testee, i.e. global point cloud data are obtained, is made
For the registration benchmark of subsequent three-dimensional measurement.
Step S2, to the surface progress three-dimensional measurement of some regional area of testee, binocular stereo vision principle is utilized
Carry out three-dimensional reconstruction, obtain the point cloud data of the regional area, referred to as local point cloud data, wherein the regional area with
There are overlay regions in the global point cloud data corresponding region.
Preferably, three-dimensional measurement can be carried out using surface of UAV flight's three-dimensional test system to testee.It passes
System is assisted using industrial machinery arm, is unable to measure complete pattern for large-sized object, and is used sticking sign point or set up and survey
Network method is measured, then can reduce measurement efficiency and increases measurement cost.Therefore preferably method is using based on unmanned aerial vehicle platform
Three-dimension measuring system come such issues that solve, carry unmanned aerial vehicle platform, camera enabled flexibly to capture general camera
The viewpoint that can not be captured, so that three dimensional point cloud is more abundant.The three-dimensional reconstruction based on unmanned aerial vehicle platform of the present embodiment
The mode of external trigger can be used in system, enables camera and projection arrangement synchronized projection and acquisition operation, to reduce three
Tie up the error rebuild.Specifically, obtaining the process of the point cloud data of the regional area as shown in Fig. 2, including the following steps
S21-S24:
Step S21, three-dimensional measurement is carried out to the surface of the regional area, obtains the first speckle pattern of the regional area
Picture and the second speckle image.
It specifically uses left camera and right camera with different visual angles, the surface of the regional area is carried out simultaneously respectively
Shooting, obtains left speckle image (the first speckle image) and right speckle image (the second speckle image).
Step S22, the whole Pixel-level of each pixel in first speckle image is determined in second speckle image
Corresponding points.Specific method includes step S221, S222.
Step S221, first speckle image and second speckle image are chosen according to relevant calculation formula identical
(the 2w of sizem+1)×(2wm+ 1) pixel region carries out related operation, the relevant calculation formula are as follows:
Wherein, IL(uL, vL) indicate selected areas in the first speckle image planar point (u, v) gray value, IR(uR, vR) table
Show the gray value of the second speckle image planar point (u, v) in selected areas,WithRespectively indicate the first speckle image, second
The average gray of speckle image selected areas, ω indicate related coefficient, WmIndicate the pixel number of certain amount；
Step S222, corresponding corresponding points are made when choosing the maximum value of related coefficient calculated value and being more than the threshold value of setting
For whole Pixel-level corresponding points.
In this step, the related coefficient of selection need to meet two conditions, first is that calculated value is more than the threshold value of setting；Meeting
Under conditions of premise one, which is maximum value, and the maximum value corresponding points are as whole Pixel-level corresponding points.
Step S23, according to each pixel coordinate in the whole Pixel-level corresponding points and first speckle image,
Sub-pix is carried out to second speckle image and corresponds to point search, obtains the sub-pix corresponding points in second speckle image.
Specifically, to correspond to point searching method as follows for sub-pix:
It is (2w that window size is created in first speckle imagem+1)×(2wm+ 1) reference child window；
By the non-linear space correlation function ω (s) under second order parallax model as newton-La Fuxun interative computation to
Majorized function:
Wherein,ud、vdFor
Zeroth order parallax,For single order parallax,For second order parallax,
Δ u, Δ v are the difference in the first speckle image between other pixels and central pixel point, second order parallax model are as follows:
According to preset iterative steps, and according to interative computation formulaIt is iterated fortune
It calculates, determines the correlation function value s that last time interative computation calculatesNFor end value,
Wherein, the value range of N is the integer more than or equal to 1, and n indicates the number of variable s, s0For initial value, by whole picture
It is calculated as initial value according to the end value and second order parallax model position when the related coefficient maximum value that plain relevant calculation obtains
The sub-pix corresponding points out.
Step S24, using binocular stereo vision principle, in conjunction with second speckle image sub-pix corresponding relationship into
Row three-dimensional reconstruction obtains the local point cloud data on the testee surface.
The present invention carries out related operation to left image and right image using speckle correlation technique, to determine in right image
The whole Pixel-level corresponding points of each pixel in left image；According to the whole Pixel-level corresponding points, spatial correlation function and left figure
The pixel coordinate of each left image as in carries out sub-pix corresponding points search arithmetic to the right image, it is corresponding to obtain sub-pix
Point；Using binocular stereo vision principle, three-dimensional reconstruction is carried out in conjunction with sub-pix corresponding relationship, obtains testee regional area
Point cloud data.This method can be realized single image quick three-dimensional reconstructing, be particularly suitable for the dynamic measurement of three-dimensional scenic.
Step S3, the local point cloud data is transformed under the global coordinate system, it will be described according to the overlay region
Local point cloud data is registrated with the global point cloud data, updates the global point cloud data.
Specifically, be registrated local point cloud data with the global point cloud data using overlay region as reference point, and
Based on the nearest point frame of iteration, the method for use tactful from thick to thin point to model is registrated.
Step S4, keep the testee motionless, transformation measurement visual angle measures the testee, repeats institute
Step S2 to the step S3 is stated, until completing the measurement to the testee.
It is every to have measured a regional area, when selecting next measured zone, therefore, to assure that this measured zone with
There are overlay regions in the region surveyed, and are registrated so as to subsequent with the global point cloud data of last update.
Step S5, global optimization processing is carried out to the global point cloud data updated after being measured, obtains point cloud model.
Specifically, optimizing processing to global point cloud data includes: to establish the unified error comprising all overlapping regions
Registration error is shared out equally each overlapping region, reduces registration error by evaluation function by the solution of the error function
Accumulation, to keep three-dimensional measurement model more accurate.
Three-dimensional rebuilding method provided by the invention is to carry out three-dimensional reconstruction based on binocular stereo vision principle, obtains first
Then the point cloud data (i.e. global point cloud data) of first frame 3 d measurement data chooses region corresponding with global point cloud data
There are the regional areas of overlay region to measure, and obtains local point cloud data and is registrated and is updated global point cloud data again, weight
This multiple process is until complete the measurement in all surface region.The present invention passes through the method that breaks the whole up into parts and carries out three-dimensional reconstruction, can be with
The measurement data that high density and high-precision have both is obtained, it is more advantageous especially for the three-dimensional measurement of large sized object.Into
One step, for each measuring node, the present invention uses speckle projection correlation technique, establishes the whole pixel of left and right speckle image
Corresponding points, and realize sub-pix using newton-La Fuxun iterative optimization method and correspond to point location, it is final to realize single node depth
Data reconstruction, furthermore by can be realized single image quick three-dimensional reconstructing for speckle related operation, additionally it is possible to be suitable for three-dimensional
The dynamic of scene measures；For multi-view depth data, the present invention realizes multi-view depth number using iterative closest point approach
According to matching with merge, and final output three-dimensional reconstruction data.Three-dimensional rebuilding method of the invention also can for large sized object
High density and high-precision measurement data are obtained, realizes flexible, stable three-dimensional measurement.
The present invention also provides a kind of three-dimensional reconstruction systems, comprising:
Three-dimensional measurement module 1, is used for: being demarcated based on testee to three-dimensional reconstruction coordinate system, obtains world coordinates
System；The point cloud data of the first frame 3 d measurement data of testee under the global coordinate system, i.e. global point cloud are obtained simultaneously
Data.
Be also used to carry out three-dimensional measurement to the surface of testee some regional area, using binocular stereo vision principle into
Row three-dimensional reconstruction obtains the point cloud data of the regional area, i.e., local point cloud data, wherein the regional area with it is described
There are overlay regions in global point cloud data corresponding region.
Global point cloud data update module 2, for the local point cloud data to be transformed under the global coordinate system, root
The local point cloud data is registrated with the global point cloud data according to the overlay region, updates the global point cloud number
According to.
The data-optimized module 3 of global point cloud, for carrying out global optimization to the global point cloud data obtained after being measured
Processing, obtains point cloud model.
The measurement data that three-dimensional reconstruction system provided by the invention, available high density and high-precision have both, especially
It is more advantageous for the three-dimensional measurement of large sized object.
The present invention also provides a kind of mobile terminals, as shown in Fig. 2, the mobile terminal includes:
As shown in Fig. 2, the mobile terminal includes: processor (processor) 10, memory (memory) 20, communication
Interface (Communications Interface) 30 and bus 40；Wherein,
The processor 10, memory 20, communication interface 30 complete mutual communication by the bus 40；
The communication interface 30 is for the information transmission between the communication equipment of the mobile terminal；
The processor 10 is used to call the computer program in the memory 20, to execute above-mentioned each method embodiment
Provided method, for example, step S1, three-dimensional reconstruction coordinate system is demarcated based on testee, obtains global seat
Mark system；Obtain point cloud data (the i.e. global point cloud number of the first frame 3 d measurement data of testee under the global coordinate system
According to)；Step S2, to the surface progress three-dimensional measurement of some regional area of testee, binocular stereo vision principle is utilized to carry out
Three-dimensional reconstruction obtains the point cloud data (i.e. local point cloud data) of the regional area, wherein the regional area and described complete
There are overlay regions in office's point cloud data corresponding region；Step S3, the local point cloud data is transformed into the global coordinate system
Under, the local point cloud data is registrated with the global point cloud data according to the overlay region, updates the global point
Cloud data；Step S4, keep the testee motionless, transformation measurement visual angle measures the testee, repeats institute
Step S2 to the step S3 is stated, until completing the measurement to the testee；Step S5, complete to what is updated after being measured
Office's point cloud data carries out global optimization processing, obtains point cloud model.
The present invention also provides a kind of storage devices, wherein the storage device is stored with computer program, the computer
Program can be performed to realize the three-dimensional rebuilding method
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Made any modifications, equivalent replacements, and improvements etc., should all be included in the protection scope of the present invention within mind and principle.
Claims (10)
1. a kind of three-dimensional rebuilding method, which comprises the steps of:
Step S1, three-dimensional reconstruction coordinate system is demarcated；Obtain the point cloud number of the first frame 3 d measurement data of testee
According to referred to as global point cloud data, the coordinate system based on the first frame 3 d measurement data are known as global coordinate system；
Step S2, to the surface progress three-dimensional measurement of some regional area of testee, binocular stereo vision principle is utilized to carry out
Three-dimensional reconstruction obtains the point cloud data of the regional area, referred to as local point cloud data, wherein the regional area with it is described
There are overlay regions in global point cloud data corresponding region；
Step S3, the local point cloud data is transformed under the global coordinate system, according to the overlay region by the part
Point cloud data is registrated with the global point cloud data, updates the global point cloud data；
Step S4, keep the testee motionless, transformation measurement visual angle measures the testee, repeats the step
Rapid S2 to the step S3, until completing the measurement to the testee；
Step S5, global optimization processing is carried out to the global point cloud data updated after being measured, obtains point cloud model.
2. three-dimensional rebuilding method according to claim 1, which is characterized in that the step S2 specifically includes the following steps:
Step S21, three-dimensional measurement is carried out to the surface of the regional area, obtain the regional area the first speckle image and
Second speckle image；
Step S22, determine that the whole Pixel-level of each pixel in first speckle image is corresponding in second speckle image
Point；
Step S23, according to each pixel coordinate in the whole Pixel-level corresponding points and first speckle image, to institute
It states the second speckle image progress sub-pix and corresponds to point search, obtain the sub-pix corresponding points in second speckle image；
Step S24, using binocular stereo vision principle, three are carried out in conjunction with the corresponding relationship of the sub-pix of second speckle image
Dimension is rebuild, and the local point cloud data on the testee surface is obtained.
3. three-dimensional rebuilding method according to claim 2, which is characterized in that the step S22 specifically includes the following steps:
Step S221, same size is chosen to first speckle image and second speckle image according to relevant calculation formula
(2wm+1)×(2wm+ 1) pixel region carries out related operation, the relevant calculation formula are as follows:
Wherein, IL(uL, vL) indicate selected areas in the first speckle image planar point (u, v) gray value, IR(uR, vR) indicate institute
The gray value of second speckle image planar point (u, v) in favored area,WithRespectively indicate the first speckle image, the second speckle
The average gray of image selected areas, ω indicate related coefficient, WmIndicate the pixel number of certain amount；
Step S222, corresponding corresponding points are as whole when choosing the maximum value of related coefficient calculated value and being more than the threshold value of setting
Pixel-level corresponding points.
4. three-dimensional rebuilding method according to claim 2, which is characterized in that the step S23 specifically includes the following steps:
It is (2w that window size is created in first speckle imagem+1)×(2wm+ 1) reference child window；
By the non-linear space correlation function ω (s) under second order parallax model as the to be optimized of newton-La Fuxun interative computation
Function:
Wherein,ud、vdFor zeroth order
Parallax,For single order parallax,For second order parallax, Δ u,
Δ v is the difference in the first speckle image between other pixels and central pixel point, second order parallax model are as follows:
According to preset iterative steps, and according to interative computation formulaIt is iterated operation, really
Determine the correlation function value S of last time interative computation calculatingNFor end value,
Wherein, the value range of N is the integer more than or equal to 1, and n indicates the number of variable s, s0It is for initial value, whole pixel is related
Position when the related coefficient maximum value being calculated is as initial value, according to the end value and the calculating of second order parallax model
Sub-pix corresponding points.
5. three-dimensional rebuilding method as described in claim 1, which is characterized in that, will according to the overlay region in the step S3
The part point cloud data is registrated with the global point cloud data, is based on the nearest point frame of iteration, using from thick to thin
What the method for point to the model of strategy was registrated.
6. three-dimensional rebuilding method as described in claim 1, which is characterized in that in the step S5, to global point cloud data into
Row optimization processing include: establish it is unified include all overlapping regions error assessment function, by the solution of the error function,
Registration error is shared out equally into each overlapping region, reduces registration error accumulation.
7. a kind of three-dimensional reconstruction system characterized by comprising
Three-dimensional measurement module, is used for: demarcating to three-dimensional reconstruction coordinate system；Obtain the first frame three-dimensional measurement number of testee
According to point cloud data, referred to as global point cloud data, the coordinate system based on the first frame 3 d measurement data be known as world coordinates
System；
It is also used to utilize binocular stereo vision principle to carry out three the surface progress three-dimensional measurement of some regional area of testee
Dimension is rebuild, and the point cloud data of the regional area is obtained, referred to as local point cloud data, wherein the regional area and described complete
There are overlay regions in office's point cloud data corresponding region；
Global point cloud data update module, for transforming to the local point cloud data under the global coordinate system, according to institute
It states overlay region and is registrated the local point cloud data with the global point cloud data, update the global point cloud data；
The data-optimized module of global point cloud, for carrying out global optimization processing to the global point cloud data obtained after being measured,
Obtain point cloud model.
8. three-dimensional reconstruction system as claimed in claim 7, which is characterized in that the three-dimensional reconstruction system is to be mounted in unmanned plane
Three-dimensional measurement is carried out on platform.
9. a kind of mobile terminal characterized by comprising processor, the memory being connect with processor communication, the memory
It is stored with computer program, the computer program realizes three-dimensional as claimed in any one of claims 1 to 6 for being performed
Method for reconstructing；The processor is used to call the computer program in the memory, to realize such as any one of claim 1-6
The three-dimensional rebuilding method.
10. a kind of storage device, which is characterized in that the storage device is stored with computer program, the computer program energy
It is enough performed to realize three-dimensional rebuilding method as claimed in any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910418032.6A CN110189400B (en) | 2019-05-20 | 2019-05-20 | Three-dimensional reconstruction method, three-dimensional reconstruction system, mobile terminal and storage device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910418032.6A CN110189400B (en) | 2019-05-20 | 2019-05-20 | Three-dimensional reconstruction method, three-dimensional reconstruction system, mobile terminal and storage device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110189400A true CN110189400A (en) | 2019-08-30 |
| CN110189400B CN110189400B (en) | 2023-04-14 |
Family
ID=67716846
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910418032.6A Active CN110189400B (en) | 2019-05-20 | 2019-05-20 | Three-dimensional reconstruction method, three-dimensional reconstruction system, mobile terminal and storage device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110189400B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110675440A (en) * | 2019-09-27 | 2020-01-10 | 深圳市易尚展示股份有限公司 | Confidence evaluation method and device for three-dimensional depth data and computer equipment |
| CN110895823A (en) * | 2020-01-10 | 2020-03-20 | 腾讯科技（深圳）有限公司 | Texture obtaining method, device, equipment and medium for three-dimensional model |
| CN110910493A (en) * | 2019-11-29 | 2020-03-24 | 广州极飞科技有限公司 | Three-dimensional reconstruction method and device and electronic equipment |
| CN111340942A (en) * | 2020-02-25 | 2020-06-26 | 电子科技大学 | Three-dimensional reconstruction system based on unmanned aerial vehicle and method thereof |
| CN112489207A (en) * | 2021-02-07 | 2021-03-12 | 深圳大学 | Space-constrained dense matching point cloud plane element extraction method |
| CN112967399A (en) * | 2021-03-31 | 2021-06-15 | 东莞中国科学院云计算产业技术创新与育成中心 | Three-dimensional time sequence image generation method and device, computer equipment and storage medium |
| US20210223048A1 (en) * | 2020-04-29 | 2021-07-22 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for updating point cloud |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170216981A1 (en) * | 2016-02-03 | 2017-08-03 | The Boeing Company | Aligning parts using multi-part scanning and feature based coordinate systems |
| US20180158200A1 (en) * | 2016-12-07 | 2018-06-07 | Hexagon Technology Center Gmbh | Scanner vis |
| US20180253821A1 (en) * | 2017-03-02 | 2018-09-06 | Topcon Corporation | Point cloud data processing device, point cloud data processing method, and point cloud data processing program |
| CN109087342A (en) * | 2018-07-12 | 2018-12-25 | 武汉尺子科技有限公司 | A kind of three-dimensional point cloud global registration method and system based on characteristic matching |
| US20190271780A1 (en) * | 2016-11-18 | 2019-09-05 | Dibotics | Methods and systems for vehicle environment map generation and updating |
| US20200386862A1 (en) * | 2017-09-04 | 2020-12-10 | Commonwealth Scientific And Industrial Research Organisation | Method and system for use in performing localisation |
-
2019
- 2019-05-20 CN CN201910418032.6A patent/CN110189400B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170216981A1 (en) * | 2016-02-03 | 2017-08-03 | The Boeing Company | Aligning parts using multi-part scanning and feature based coordinate systems |
| US20190271780A1 (en) * | 2016-11-18 | 2019-09-05 | Dibotics | Methods and systems for vehicle environment map generation and updating |
| US20180158200A1 (en) * | 2016-12-07 | 2018-06-07 | Hexagon Technology Center Gmbh | Scanner vis |
| US20180253821A1 (en) * | 2017-03-02 | 2018-09-06 | Topcon Corporation | Point cloud data processing device, point cloud data processing method, and point cloud data processing program |
| US20200386862A1 (en) * | 2017-09-04 | 2020-12-10 | Commonwealth Scientific And Industrial Research Organisation | Method and system for use in performing localisation |
| CN109087342A (en) * | 2018-07-12 | 2018-12-25 | 武汉尺子科技有限公司 | A kind of three-dimensional point cloud global registration method and system based on characteristic matching |
Non-Patent Citations (4)
| Title |
|---|
| 李云雷等: "形貌视觉测量中立体拼接靶标的设计及应用", 《仪器仪表学报》 * |
| 李健等: "采用彩色光度立体法的动态物体全局三维数字化", 《陕西科技大学学报(自然科学版)》 * |
| 汤其剑等: "数字散斑三维重建中散斑特性分析", 《中国激光》 * |
| 蔡勇等: "多摄像机视觉检测大范围布置方法及其数据拼接", 《中国机械工程》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110675440A (en) * | 2019-09-27 | 2020-01-10 | 深圳市易尚展示股份有限公司 | Confidence evaluation method and device for three-dimensional depth data and computer equipment |
| CN110675440B (en) * | 2019-09-27 | 2022-07-12 | 深圳市易尚展示股份有限公司 | Confidence evaluation method and device for three-dimensional depth data and computer equipment |
| CN110910493A (en) * | 2019-11-29 | 2020-03-24 | 广州极飞科技有限公司 | Three-dimensional reconstruction method and device and electronic equipment |
| CN110910493B (en) * | 2019-11-29 | 2021-05-14 | 广州极飞科技股份有限公司 | Three-dimensional reconstruction method and device and electronic equipment |
| CN110895823A (en) * | 2020-01-10 | 2020-03-20 | 腾讯科技（深圳）有限公司 | Texture obtaining method, device, equipment and medium for three-dimensional model |
| CN110895823B (en) * | 2020-01-10 | 2020-06-05 | 腾讯科技（深圳）有限公司 | Texture obtaining method, device, equipment and medium for three-dimensional model |
| US11989894B2 (en) | 2020-01-10 | 2024-05-21 | Tencent Technology (Shenzhen) Company Limited | Method for acquiring texture of 3D model and related apparatus |
| CN111340942A (en) * | 2020-02-25 | 2020-06-26 | 电子科技大学 | Three-dimensional reconstruction system based on unmanned aerial vehicle and method thereof |
| US20210223048A1 (en) * | 2020-04-29 | 2021-07-22 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for updating point cloud |
| US11828606B2 (en) * | 2020-04-29 | 2023-11-28 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for updating point cloud |
| CN112489207A (en) * | 2021-02-07 | 2021-03-12 | 深圳大学 | Space-constrained dense matching point cloud plane element extraction method |
| CN112967399A (en) * | 2021-03-31 | 2021-06-15 | 东莞中国科学院云计算产业技术创新与育成中心 | Three-dimensional time sequence image generation method and device, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110189400B (en) | 2023-04-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110189400A (en) | A kind of three-dimensional rebuilding method, three-dimensional reconstruction system, mobile terminal and storage device | |
| US9972067B2 (en) | System and method for upsampling of sparse point cloud for 3D registration | |
| CN108198145B (en) | Method and device for point cloud data restoration | |
| CN111275750B (en) | Indoor space panoramic image generation method based on multi-sensor fusion | |
| CN110335343B (en) | Human body three-dimensional reconstruction method and device based on RGBD single-view-angle image | |
| CN108520537B (en) | Binocular depth acquisition method based on luminosity parallax | |
| CN103337069B (en) | High-quality three-dimensional color image acquisition methods and device based on multiple camera | |
| CN110288659B (en) | Depth imaging and information acquisition method based on binocular vision | |
| CN110992487B (en) | Rapid three-dimensional map reconstruction device and reconstruction method for hand-held airplane fuel tank | |
| CN111091076B (en) | Tunnel limit data measuring method based on stereoscopic vision | |
| CN107170000B (en) | Stereopsis dense Stereo Matching method based on the optimization of global block | |
| CN113361365B (en) | Positioning method, positioning device, positioning equipment and storage medium | |
| CN111127540B (en) | Automatic distance measurement method and system for three-dimensional virtual space | |
| WO2024045632A1 (en) | Binocular vision and imu-based underwater scene three-dimensional reconstruction method, and device | |
| CN113393577B (en) | Oblique photography terrain reconstruction method | |
| CN105551020A (en) | Method and device for detecting dimensions of target object | |
| CN103886595A (en) | Catadioptric camera self-calibration method based on generalized unified model | |
| CN113743155A (en) | Method for constructing multi-detail level model of object and computer equipment | |
| CN111429571B (en) | Rapid stereo matching method based on spatio-temporal image information joint correlation | |
| CN113034571A (en) | Object three-dimensional size measuring method based on vision-inertia | |
| CN116051766A (en) | External planet surface environment reconstruction method based on nerve radiation field | |
| CN107274448B (en) | Variable weight cost aggregation stereo matching algorithm based on horizontal tree structure | |
| Sergiyenko et al. | Multi-view 3D data fusion and patching to reduce Shannon entropy in Robotic Vision | |
| CN115880448B (en) | Three-dimensional measurement method and device based on binocular imaging | |
| CN115100254B (en) | Point cloud registration method for dual quaternion description under plane feature constraint |
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 |