CN110021039A - The multi-angle of view material object surface point cloud data initial registration method of sequence image constraint - Google Patents
The multi-angle of view material object surface point cloud data initial registration method of sequence image constraint Download PDFInfo
- Publication number
- CN110021039A CN110021039A CN201811356769.1A CN201811356769A CN110021039A CN 110021039 A CN110021039 A CN 110021039A CN 201811356769 A CN201811356769 A CN 201811356769A CN 110021039 A CN110021039 A CN 110021039A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- cloud data
- point
- image
- angle
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 64
- 239000000463 material Substances 0.000 title claims abstract description 22
- 230000009466 transformation Effects 0.000 claims description 16
- 239000011159 matrix material Substances 0.000 claims description 15
- 230000000007 visual effect Effects 0.000 claims description 8
- 238000013507 mapping Methods 0.000 claims description 5
- 238000005457 optimization Methods 0.000 claims description 4
- 238000000354 decomposition reaction Methods 0.000 claims description 3
- 238000002271 resection Methods 0.000 claims description 2
- 238000004422 calculation algorithm Methods 0.000 abstract description 22
- 230000008569 process Effects 0.000 abstract description 8
- 238000005516 engineering process Methods 0.000 abstract description 4
- 238000004364 calculation method Methods 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 8
- 238000001514 detection method Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 238000013519 translation Methods 0.000 description 5
- 238000005070 sampling Methods 0.000 description 4
- 238000013459 approach Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 241000287196 Asthenes Species 0.000 description 1
- 241001269238 Data Species 0.000 description 1
- 206010051379 Systemic Inflammatory Response Syndrome Diseases 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000007634 remodeling Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
- G06T7/337—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving reference images or patches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/38—Registration of image sequences
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Length Measuring Devices By Optical Means (AREA)
- Image Processing (AREA)
Abstract
The present invention requires excessively stringent and is registrated the lower problem of efficiency for initial position of the existing multi-angle of view registration algorithm of point clouds for multi-angle of view sampled data in surface in kind, a kind of multi-angle of view material object surface point cloud data initial registration method for proposing sequence image constraint, belongs to product reverse Engineering Technology field.Position of the video camera in point cloud data local coordinate system is positioned according to perspective projection principle, point cloud data is transformed into corresponding camera coordinate system, and using image characteristic point and its corresponding match point as control point of the same name, by point cloud data preliminary registration to the position of approximation overlapping in sequence image reconstruction process.The initial registration method, without strict demand, can obtain the point cloud data initial registration of approximate global optimum as a result, improving the robustness and computational efficiency of method for registering to the initial position of point cloud data with lesser calculation amount.
Description
Technical field
The present invention provides a kind of multi-angle of view material object surface point cloud data initial registration method of sequence image constraint, belongs to production
Product reverse Engineering Technology field.
Background technique
Active optical three-dimensional measurement technology is widely used in solving the neck such as reverse-engineering, computer vision, biomedicine
Digitization modeling problem in domain.Since the visibility at single visual angle limits, it usually needs will using three-dimensional point cloud registration technique
Several partly overlapping clouds unification under different perspectives are into the same coordinate system to obtain complete mathematical model.Three-dimensional point cloud
The precision of registration determines the quality of testee three-dimensionalreconstruction and the efficiency and precision of point cloud post-processing.
Existing technical literature is retrieved and is found, Besl etc. is at academic conference " IEEE Computer Society "
In the paper " A Method for Registration of 3-D Shapes " delivered on (1992,14 (2): P239-256)
Iteration closest approach algorithm (ICP) is proposed, is a kind of point-to-point registration Algorithm based on geometrical model, it is the smallest with Euclidean distance
Point is used as corresponding points, by minimizing the corresponding points square distance and obtain rigid between two 3-D data sets that two points are concentrated
Property transformation.ICP algorithm after 1992 propose, gradually derives numerous variants, the master between these variants from Besl and McKay
Distinguishing is the difference that mode is chosen with eka-element and registration strategies.Sappa etc. is in " In Proceedings of th 9th
International Symposium on Intelligent Robotic Systems " deliver on (SIRS ' 01,2001)
Based on fast in paper " Range Image Registration by using an Edge-Based Representation "
The marginal point of the fast direct selected point cloud depth map of edge cutting techniques reduces closest approach search range as sampling point set, but
Object curved surface features are had ignored, the less point cloud of two width overlapping region inward flange feature of accuracy registration is not suitable for.Li et al. exists
(2015,65 (11): paper " the A modified ICP delivered on 0000) " Pattern Recognition Letters "
algorithm based on dynamic adjustment factor for registration of point cloud
Based on the stl file Searching point cloud being made of the facet data and Vector Message list of triangle gridding in and CAD model "
Corresponding points, and by dynamic adjust rigid transformation parameters effectively reduce the number of iterations.Above-mentioned algorithm mentions to a certain extent
High ICP algorithm convergence rate, but exist equally have big difference because of initial position with ICP algorithm and lead to nearest point search not
The defect of locally optimal solution that is accurate and converging to mistake.Rusu etc. is in academic conference " IEEE International
Conference on Robotics and Automation " paper " Fast that delivers on (2009:P1848-1853)
A kind of FPFH feature is proposed in point feature histograms (FPFH) for 3D registration ", describes point
The local geometric information of any sampling point neighborhood in cloud data set.Initial FPFH characteristic point is adopted by unanimity of samples method
Sample retains satisfactory characteristic point and is used for subsequent registration, this method reduce the number of iterations, achieves preferably
Registration effect, but need to solve the geological informations such as sampling point method arrow, curvature, complicate algorithm, can not guarantee convergence speed simultaneously
Degree and registration accuracy.The paper that Al-Manasir and Fraser is delivered on " Iaprs Dresden " (2006,21 (115))
It is proposed in " Automatic Registration of Terrestrial Laser Scanner Data via Imagery "
The autoegistration methods of territorial laser scanning data based on image information, the encoding target for being placed on measured object surface are constituted
Determine the control point of the same name of rigid transformation between subject to registration cloud.By additional feature information, a cloud initial position is reduced
Influence to registration result, and improve registration Algorithm convergence rate, but the problem of bringing is the point cloud data on measured object surface
It is destroyed.Han etc. is in " IEEE Geoscience&Remote Sensing Letters " (2012,10 (4): P746-750)
On lead in the paper " LiDAR Point Cloud Registration by Image Detection Technique " delivered
Detection and matching sequence image character pair point are crossed, is established indirectly according to the mapping relations of its image coordinate and space coordinate of the same name
Control point three-dimensional coordinate corresponding relationship is to realize point cloud registering.Since there are certain errors for characteristic point detection and matching, work as spy
Sign point not can guarantee correct registration and registration accuracy when less.Wang Ruiyan etc. is in " mapping journal " (2016,45 (1): P96-
102.) registration process is divided into two stages in the paper " in conjunction with quick cloud algorithm of image information " delivered on, it is first
The relative rotation transformation that image information solves video camera is first passed through, translation transformation is then iterated to calculate, improves the steady of algorithm
Property, but it is only used for the video camera occasion coaxial with establishing is scanned.
It for point cloud registering problem, is easily achieved in view of ICP algorithm and registration accuracy is higher, therefore by ICP algorithm application
It is very reasonable in the solution for the point cloud registering that there are problems that greater overlap region, but the registration result of ICP algorithm is to first
Initial value is more sensitive, poor initial motion parameter be easy increase the number of iterations, even result in registration process converge on it is undesirable
Locally optimal solution, it is therefore desirable to initial motion parameter is optimized.In addition, ICP algorithm time complexity is mainly nearest
Point search is dominated, if closest approach search range can be reduced, calculation amount just can be effectively reduced.
In conclusion existing during multi-angle of view cloud data registration to the first of multi-angle of view sampled data in surface in kind at present
Beginning status requirement is excessively stringent and is registrated the lower technical problem of efficiency, therefore, mentions for multi-angle of view material object surface point cloud data
Those skilled in the art's skill urgently to be resolved is had become for preferable initial registration parameter and reduction registration process time loss
Art problem.
Summary of the invention
In order to solve the above technical problems, the technical scheme adopted by the invention is that a kind of multi-angle of view of sequence image constraint is real
Object surface point cloud data initial registration method, this method to the initial position of cloud without strict demand, can be with lesser calculation amount
The point cloud initial registration of approximate global optimum is obtained as a result, carrying out accuracy registration according to initial registration result, can be significantly improved more
The robustness and computational efficiency of visual angle material object surface cloud data registration process.Its implementation are as follows:
A kind of multi-angle of view material object surface point cloud data initial registration method of sequence image constraint, it is characterised in that successively wrap
Containing following steps: (1) successively acquisition testee under different perspectives type face point cloud data and each visual angle under it is corresponding
Two dimensional image；(2) position of the video camera in point cloud data local coordinate system is positioned, point cloud data is locally sat from it
Mark system is transformed into camera coordinate system；(3) estimated outside all video cameras based on sequence image increment type method of movement method for reconstructing
Parameter；(4) according under camera coordinate system point cloud data and all external parameters of cameras point cloud data is initially matched
It is quasi-.
For realization goal of the invention, the multi-angle of view material object surface point cloud data initial registration method that the sequence image constrains,
It is characterized by: being measured with calibrated video camera to calibration object, being demarcated based on perspective projection principle in step (2)
The world coordinates of identification point establishes mapping relations one by one with the image coordinate recognized and realizes that video camera is sat in point cloud local on object
Positioning in mark system, is transformed into camera coordinate system for point cloud data according to the transformation matrix of solution.
For realization goal of the invention, the multi-angle of view material object surface point cloud data initial registration method that the sequence image constrains,
In step (3), external parameters of cameras Μ={ M is estimated using increment type method of movement method for reconstructingi∈ SE (3) | i=1,2 ...,
NIThe step of be specifically: 1) read sequence image Σ={ Ii| i=1,2 ..., NI, the characteristic point of every piece image is detected, and
Set of characteristic points are denoted asWherein, xijIndicate image IiThe image coordinate of characteristic point, fijTable
Show xijThe feature descriptor of point；2) according to the feature descriptor f of character pair pointijBetween Euclidean distance minimum principle and right
Epipolar geometric constraint successively carries out Feature Points Matching to arbitrary neighborhood two images in Σ, then Feature Points Matching is combined into collection3) two images I before being chosen from Σ1And I2Feature Points Matching to setInitialize external parameters of cameras M1、M2And three-dimensional point set X(1), and by X(1)Addition point
Collect Ω；4) increase piece image Ii, estimate external parameters of cameras Mi；5) pass through triangulation reconstruction image Ii, the point set that newly extends
For X(i-1), Ω ← Ω ∪ X(i-1)；6) bundle adjustment nonlinear optimization external parameters of cameras Μ and three-dimensional point coordinate Ω is utilized；
7) step 4)~6 are repeated), until all image procossings are completed.
For realization goal of the invention, the multi-angle of view material object surface point cloud data initial registration method that the sequence image constrains,
It is characterized in that, initializing external parameters of cameras M in step (3)1、M2And three-dimensional point set X(1)The step of are as follows: (1) to I1And I2
Feature Points Matching to setEssential matrix is calculated using 8 methods；(2) with I1Take the photograph
Camera coordinate system is world coordinate system, decomposes essential matrix by singular value decomposition method and obtains I2External parameters of cameras M2；(3)
According to image I1And I2The picpointed coordinate of Feature Points Matching pair calculates three-dimensional point coordinate by triangulation method.
For realization goal of the invention, the multi-angle of view material object surface point cloud data initial registration method that the sequence image constrains,
Newly-increased image I is calculated in step (3)iExternal parameters of cameras MiThe step of are as follows: (1) from Ii-1And IiFeature Points Matching pairMiddle selection can find the characteristic point of corresponding space coordinate in point set Ω
Pairing, is denoted as(2) it is based onMiddle IiCharacteristic point image coordinate establish space resection side with corresponding space coordinate
Journey, linear solution camera perspective projection matrix；(3) it decomposes camera perspective projection matrix and finds out external parameters of cameras Mi。
For realization goal of the invention, the multi-angle of view material object surface point cloud data initial registration method that the sequence image constrains,
It is characterized by: in step (4) based on external parameters of cameras calculated result in step (3), accurate rotation transformation being used only and joins
Point cloud data under several pairs of camera coordinate systems is initially registered.
Compared with prior art, the invention has the following advantages that
(1) search control point of the same name can be effectively reduced in the point cloud initial registration process rebuild based on the sequential image feature point method of movement
Calculating cost, and subject to registration cloud can be adjusted to the position of approximate overlapping；
(2) external parameters of cameras estimation problem is converted by point cloud registering problem, two dimension is rebuild based on Epipolar geometry constraint increment
Image realizes that the initial registration of multi-angle of view point cloud enhances the robustness of registration Algorithm to cloud initial position without strict demand；
(3) using based on the sequential image feature point method of movement rebuild point cloud initial registration algorithm determine initial registration parameter as
The initial value of ICP algorithm can significantly improve the iteration efficiency of ICP algorithm, and with the increase of point cloud data scale, registration effect
Rate is in increasing trend.
Detailed description of the invention
Fig. 1 is the flow chart of the multi-angle of view material object surface point cloud data initial registration method of sequence image constraint of the present invention；
Fig. 2 is camera coordinate system and point cloud data local coordinate system positional diagram；
Fig. 3 is video camera Attitude estimation schematic diagram；
Fig. 4 is automobile head sampled data and sequence image schematic diagram in embodiment one；
Fig. 5 is the detection of two images characteristic point and matching result schematic diagram in embodiment one；
Fig. 6 is two groups of point cloud data initial registration result schematic diagrams in embodiment one；
Fig. 7 is two groups of point cloud data accuracy registration result schematic diagrams in embodiment two.
Specific embodiment
Below in conjunction with the accompanying drawings and embodiment the invention will be further described.
Fig. 1 is the program flow of the multi-angle of view material object surface point cloud data initial registration method of invention sequence image constraint herein
Cheng Tu realized using C++ programming language, and the main process of institute's support method of the present invention includes point cloud data to camera coordinates
The lower conversion of system utilizes point cloud data under sequence image method of movement method for reconstructing estimation video camera posture and camera coordinate system
Initial registration, the solution for the completion initial registration parameter that can more optimize.
Fig. 2 is camera coordinate system and point cloud data local coordinate system positional relationship, if the local coordinate system of point cloud data
For OW-XWYWZW, camera coordinate system OC-XCYCZC, the rigid transformation relationship between the two coordinate systems is denoted as rotation transformation
RWWith translation transformation tW.Calibration object is measured with calibrated video camera, the three-dimensional coordinate of an identification point is (X, Y, Z) in space,
As the image coordinate in plane is (u, v), perspective projection relationship be can be described as
It arranges and eliminates ZCObtain relational expression
In formula, mijFor the element of camera perspective projection matrix.Based on the matching double points coordinate of picture point and object point using minimum two
Multiplication solves above-mentioned equation, and then decomposites RW、tW.If coordinate of any point in two coordinate systems is respectively X in spaceW
And XC, then the coordinate under local coordinate system is transformed under camera coordinate system using formula (3):
XC=RWXW+tW (3)
In sequence image increment type reconstruction process, key step includes: (1) initialization camera motion information and object
Three-dimensional structure: X(1)={ XWj, Ω={ X(1), Μ={ M1,M2, k ← 2, wherein XWjIndicate that rebuilding obtained spatial point sits
Mark；(2) new image is added, carry out SIFT feature detection with preceding piece image and matches, and by Feature Points Matching to being denoted as(3) from C(k)Middle selection can find corresponding space in point set Ω and sit
Target Feature Points Matching pair, is denoted asThen another part Feature Points Matching is to being denoted as(4) basisSpace is established with Ω
Point-picture point mapping relations one by one use least square method to solve video camera transformation matrix M(k+1), Μ ← Μ ∪ M(k+1)；
(5) restore the new video camera transformation matrix P that image is added(k+1)=KM(k+1), rebuild using principle of triangulationIn feature
Newly-increased reconstruction point is denoted as X by point(k)；(6)Ω←Ω∪X(k)；(7) bundle adjustment parameter optimization, and update point set Ω with
And external parameters of cameras Μ；(8)k←k+1；(9) (2)~(8) are repeated, until all image procossings are completed；(10) it exports
External parameters of cameras Μ after optimization.Fig. 3 is that the schematic diagram for realizing external parameters of cameras estimation is constrained based on Epipolar geometry, if I1
With I2It is video camera in P1、P2Image captured by two positions, any one spatial point X is in I in theorem in Euclid space1With I2On projection
Point is respectively x1j、x2k.Video camera is by P1Position moves to P2The relative rotation of position, translation transformation are denoted as R respectivelyC、tC, then by
Known to Epipolar geometry constraint:
Wherein,Essential matrix of the E between two views, have next property, and only with RC、tC
It is related, meet
E=[tC]x RC (5)
Wherein, [tC]xFor translation vector tCAntisymmetric matrix.The t for applying singular value decomposition method to obtain to formula (5)CWith it is true
There are uncertain scale factor λ between translation vector, therefore image information can only provide video camera between two camera sites
Accurate relative rotation matrices.Therefore, video camera shoots I1、I2When two images, spatial point X is in corresponding two camera coordinates
Coordinate under system is denoted as X respectivelyC1And XC2, then meet
XC2=RCXC1+λtC (6)
Since λ can not be determined, the point cloud data under camera coordinate system is carried out using only accurate rotation transformation parameter initial
Registration.
Embodiment one: Fig. 4 is the automobile head model multi-angle of view surface sampling obtained using optical grating projection formula three-dimensional measurement
Data and sequence image schematic diagram, the point cloud number under 5 visual angles is respectively 150423,156056,191188,191238,
177400.Fig. 5~6 show the multi-angle of view material object surface point cloud data initial registration using sequence image proposed in this paper constraint
Method illustrates only the effect picture under the first two visual angle to the effect picture of cloud data registration to obtain preferable visual effect.
From Fig. 6 registration result it is found that subject to registration cloud can be adjusted to the position of approximate overlapping by the initial registration parameter that context of methods solves
It sets.5 groups of point cloud datas are successively initially registered, and runing time is respectively 6.649s, 6.738s, 7.145s, 7.402s, are missed
Difference is respectively 10.9741mm, 7.5126mm, 7.1157mm, 3.5361mm.
At the beginning of embodiment two: Fig. 7 show the multi-angle of view material object surface point cloud data for constraining sequence image proposed in this paper
The initial registration parameter that beginning method for registering solves as ICP algorithm initial value to the point cloud data at visual angle 1 and 5 in Fig. 4 into
The effect picture of row registration, two groups of point cloud data initial position differences are larger, and being registrated total time used is 154.001s, and result is just
Really, error 0.1575mm.
By embodiment, it can be concluded that, the present invention can obtain the initial registration of approximate global optimum with lesser calculating cost
As a result, this method is not strict with the initial position of point cloud data, the robustness of registration Algorithm is improved, is imitated in registration
Also there is apparent advantage in terms of rate.
The above described is only a preferred embodiment of the present invention, being not that the invention has other forms of limitations, appoint
What those skilled in the art changed or be modified as possibly also with the technology contents of the disclosure above equivalent variations etc.
Imitate embodiment.But without departing from the technical solutions of the present invention, according to the technical essence of the invention to above embodiments institute
Any simple modification, equivalent variations and the remodeling made, still fall within the protection scope of technical solution of the present invention.
Claims (6)
1. a kind of multi-angle of view material object surface point cloud data initial registration method of sequence image constraint, it is characterised in that successively include
Following steps: (1) successively acquisition testee under different perspectives type face point cloud data and each visual angle under corresponding two
Tie up image；(2) position of the video camera in point cloud data local coordinate system is positioned, by point cloud data from its local coordinate
System is transformed into camera coordinate system；(3) estimate to join outside all video cameras based on sequence image increment type method of movement method for reconstructing
Number；(4) according under camera coordinate system point cloud data and all external parameters of cameras point cloud data is initially registered.
2. the multi-angle of view material object surface point cloud data initial registration method of sequence image constraint as described in claim 1, special
Sign is: in step (2), being measured with calibrated video camera to calibration object, will be demarcated on object based on perspective projection principle
The world coordinates of identification point establishes mapping relations one by one with the image coordinate recognized and realizes video camera in point cloud local coordinate system
In positioning, point cloud data is transformed by camera coordinate system according to the transformation matrix of solution.
3. the multi-angle of view material object surface point cloud data initial registration method of sequence image constraint as described in claim 1, special
Sign is: in step (3), estimating external parameters of cameras Μ={ M using increment type method of movement method for reconstructingi∈ SE (3) | i=1,
2,…,NIProcess, step are as follows: 1) read sequence image Σ={ Ii| i=1,2 ..., NI, detect the feature of every piece image
Point, and set of characteristic points are denoted asWherein, xijIndicate image IiThe image coordinate of characteristic point,
fijIndicate xijThe feature descriptor of point；2) according to the feature descriptor f of character pair pointijBetween Euclidean distance minimum principle with
And Epipolar geometry constraint, Feature Points Matching is successively carried out to arbitrary neighborhood two images in Σ, then Feature Points Matching is combined into collection3) two images I before being chosen from Σ1And I2Feature Points Matching to setInitialize external parameters of cameras M1、M2And three-dimensional point set X(1), and by X(1)Addition point
Collect Ω；4) increase piece image Ii, estimate external parameters of cameras Mi；5) pass through triangulation reconstruction image Ii, the point set that newly extends
For X(i-1), Ω ← Ω ∪ X(i-1)；6) bundle adjustment nonlinear optimization external parameters of cameras Μ and three-dimensional point coordinate Ω is utilized；
7) step 4)~6 are repeated), until all image procossings are completed.
4. the multi-angle of view material object surface point cloud data initial registration method of sequence image constraint as claimed in claim 3, special
Sign is, initializes external parameters of cameras M1、M2And three-dimensional point set X(1)The step of are as follows: (1) to I1And I2Feature Points Matching
To setEssential matrix is calculated using 8 methods；(2) with I1Camera coordinate system be
World coordinate system decomposes essential matrix by singular value decomposition method and obtains I2External parameters of cameras M2；(3) according to image I1With
I2The picpointed coordinate of Feature Points Matching pair calculates three-dimensional point coordinate by triangulation method.
5. the multi-angle of view material object surface point cloud data initial registration method of sequence image constraint as claimed in claim 3, special
Sign is, calculates newly-increased image IiExternal parameters of cameras MiThe step of are as follows: (1) from Ii-1And IiFeature Points Matching pairMiddle selection can find the characteristic point of corresponding space coordinate in point set Ω
Pairing, is denoted as(2) it is based onMiddle IiCharacteristic point image coordinate establish space resection side with corresponding space coordinate
Journey, linear solution camera perspective projection matrix；(3) it decomposes camera perspective projection matrix and finds out external parameters of cameras Mi。
6. the multi-angle of view material object surface point cloud data initial registration method of sequence image constraint as described in claim 1, special
Sign is: in step (4), based on external parameters of cameras calculated result in step (3), accurate rotation transformation parameter pair is used only
Point cloud data under camera coordinate system is initially registered.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811356769.1A CN110021039A (en) | 2018-11-15 | 2018-11-15 | The multi-angle of view material object surface point cloud data initial registration method of sequence image constraint |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811356769.1A CN110021039A (en) | 2018-11-15 | 2018-11-15 | The multi-angle of view material object surface point cloud data initial registration method of sequence image constraint |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110021039A true CN110021039A (en) | 2019-07-16 |
Family
ID=67188549
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811356769.1A Pending CN110021039A (en) | 2018-11-15 | 2018-11-15 | The multi-angle of view material object surface point cloud data initial registration method of sequence image constraint |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110021039A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110443840A (en) * | 2019-08-07 | 2019-11-12 | 山东理工大学 | The optimization method of sampling point set initial registration in surface in kind |
| CN110473258A (en) * | 2019-07-24 | 2019-11-19 | 西北工业大学 | Monocular SLAM system initialization algorithm based on dotted line Unified frame |
| CN111457930A (en) * | 2020-04-02 | 2020-07-28 | 武汉中海庭数据技术有限公司 | High-precision mapping positioning method combining vehicle-mounted L idar and unmanned aerial vehicle |
| CN111899303A (en) * | 2020-07-14 | 2020-11-06 | 中国人民解放军63920部队 | Novel feature matching and relative positioning method considering space inverse projection constraint |
| CN112184662A (en) * | 2020-09-27 | 2021-01-05 | 成都数之联科技有限公司 | Camera external parameter initial method and system applied to unmanned aerial vehicle image stitching |
| CN112508999A (en) * | 2020-11-20 | 2021-03-16 | 西北工业大学深圳研究院 | Space target motion state identification method based on cooperative observation image sequence |
| CN113077523A (en) * | 2021-03-31 | 2021-07-06 | 商汤集团有限公司 | Calibration method, calibration device, computer equipment and storage medium |
| CN113362446A (en) * | 2021-05-25 | 2021-09-07 | 上海奥视达智能科技有限公司 | Method and device for reconstructing object based on point cloud data |
| CN113538531A (en) * | 2020-04-14 | 2021-10-22 | 山东理工大学 | Method for determining point cloud data registration sequence of morphology constraint |
| CN113538532A (en) * | 2020-04-14 | 2021-10-22 | 山东理工大学 | Morphology-constrained point cloud data hierarchical registration method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102982548A (en) * | 2012-12-11 | 2013-03-20 | 清华大学 | Multi-view stereoscopic video acquisition system and camera parameter calibrating method thereof |
| CN104376552A (en) * | 2014-09-19 | 2015-02-25 | 四川大学 | Virtual-real registering algorithm of 3D model and two-dimensional image |
| CN104463894A (en) * | 2014-12-26 | 2015-03-25 | 山东理工大学 | Overall registering method for global optimization of multi-view three-dimensional laser point clouds |
| CN105184856A (en) * | 2015-09-02 | 2015-12-23 | 泰山学院 | Two-phase human skin three-dimensional reconstruction method based on density matching |
| CN108090960A (en) * | 2017-12-25 | 2018-05-29 | 北京航空航天大学 | A kind of Object reconstruction method based on geometrical constraint |
| US20180158200A1 (en) * | 2016-12-07 | 2018-06-07 | Hexagon Technology Center Gmbh | Scanner vis |
-
2018
- 2018-11-15 CN CN201811356769.1A patent/CN110021039A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102982548A (en) * | 2012-12-11 | 2013-03-20 | 清华大学 | Multi-view stereoscopic video acquisition system and camera parameter calibrating method thereof |
| CN104376552A (en) * | 2014-09-19 | 2015-02-25 | 四川大学 | Virtual-real registering algorithm of 3D model and two-dimensional image |
| CN104463894A (en) * | 2014-12-26 | 2015-03-25 | 山东理工大学 | Overall registering method for global optimization of multi-view three-dimensional laser point clouds |
| CN105184856A (en) * | 2015-09-02 | 2015-12-23 | 泰山学院 | Two-phase human skin three-dimensional reconstruction method based on density matching |
| US20180158200A1 (en) * | 2016-12-07 | 2018-06-07 | Hexagon Technology Center Gmbh | Scanner vis |
| CN108090960A (en) * | 2017-12-25 | 2018-05-29 | 北京航空航天大学 | A kind of Object reconstruction method based on geometrical constraint |
Non-Patent Citations (1)
| Title |
|---|
| YULAN GUO ET AL.: "An Accurate and Robust Range Image Registration Algorithm for 3D Object Modeling", 《IEEE TRANSACTIONS ON MULTIMEDIA》 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110473258A (en) * | 2019-07-24 | 2019-11-19 | 西北工业大学 | Monocular SLAM system initialization algorithm based on dotted line Unified frame |
| CN110473258B (en) * | 2019-07-24 | 2022-05-13 | 西北工业大学 | Monocular SLAM system initialization algorithm based on point-line unified framework |
| CN110443840A (en) * | 2019-08-07 | 2019-11-12 | 山东理工大学 | The optimization method of sampling point set initial registration in surface in kind |
| CN111457930A (en) * | 2020-04-02 | 2020-07-28 | 武汉中海庭数据技术有限公司 | High-precision mapping positioning method combining vehicle-mounted L idar and unmanned aerial vehicle |
| CN111457930B (en) * | 2020-04-02 | 2021-11-23 | 武汉中海庭数据技术有限公司 | High-precision mapping positioning method by combining vehicle-mounted Lidar and unmanned aerial vehicle |
| CN113538531A (en) * | 2020-04-14 | 2021-10-22 | 山东理工大学 | Method for determining point cloud data registration sequence of morphology constraint |
| CN113538532A (en) * | 2020-04-14 | 2021-10-22 | 山东理工大学 | Morphology-constrained point cloud data hierarchical registration method |
| CN111899303A (en) * | 2020-07-14 | 2020-11-06 | 中国人民解放军63920部队 | Novel feature matching and relative positioning method considering space inverse projection constraint |
| CN111899303B (en) * | 2020-07-14 | 2021-07-13 | 中国人民解放军63920部队 | Novel feature matching and relative positioning method considering space inverse projection constraint |
| CN112184662A (en) * | 2020-09-27 | 2021-01-05 | 成都数之联科技有限公司 | Camera external parameter initial method and system applied to unmanned aerial vehicle image stitching |
| CN112184662B (en) * | 2020-09-27 | 2023-12-15 | 成都数之联科技股份有限公司 | Camera external parameter initial method and system applied to unmanned aerial vehicle image stitching |
| CN112508999A (en) * | 2020-11-20 | 2021-03-16 | 西北工业大学深圳研究院 | Space target motion state identification method based on cooperative observation image sequence |
| CN112508999B (en) * | 2020-11-20 | 2024-02-13 | 西北工业大学深圳研究院 | Space target motion state identification method based on collaborative observation image sequence |
| CN113077523A (en) * | 2021-03-31 | 2021-07-06 | 商汤集团有限公司 | Calibration method, calibration device, computer equipment and storage medium |
| CN113077523B (en) * | 2021-03-31 | 2023-11-24 | 商汤集团有限公司 | Calibration method, calibration device, computer equipment and storage medium |
| CN113362446A (en) * | 2021-05-25 | 2021-09-07 | 上海奥视达智能科技有限公司 | Method and device for reconstructing object based on point cloud data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110021039A (en) | The multi-angle of view material object surface point cloud data initial registration method of sequence image constraint | |
| Ishikawa et al. | Lidar and camera calibration using motions estimated by sensor fusion odometry | |
| CN112927360A (en) | Three-dimensional modeling method and system based on fusion of tilt model and laser point cloud data | |
| CN109242954B (en) | Multi-view three-dimensional human body reconstruction method based on template deformation | |
| Li et al. | A 4-point algorithm for relative pose estimation of a calibrated camera with a known relative rotation angle | |
| CN110443840A (en) | The optimization method of sampling point set initial registration in surface in kind | |
| CN104596502A (en) | Object posture measuring method based on CAD model and monocular vision | |
| Hoegner et al. | Evaluation of methods for coregistration and fusion of rpas-based 3d point clouds and thermal infrared images | |
| CN105574812B (en) | Multi-angle three-dimensional data method for registering and device | |
| CN109272574A (en) | Linear array rotary scanning camera imaging model building method and scaling method based on projective transformation | |
| CN112907631B (en) | Multi-RGB camera real-time human body motion capture system introducing feedback mechanism | |
| Khoshelham et al. | Generation and weighting of 3D point correspondences for improved registration of RGB-D data | |
| CN110310331A (en) | A kind of position and orientation estimation method based on linear feature in conjunction with point cloud feature | |
| CN111060006A (en) | Viewpoint planning method based on three-dimensional model | |
| CN112465849B (en) | Registration method for laser point cloud and sequence image of unmanned aerial vehicle | |
| CN106125907A (en) | A kind of objective registration method based on wire-frame model | |
| CN116385507A (en) | Multi-source point cloud data registration method and system based on different scales | |
| Wan et al. | A study in 3d-reconstruction using kinect sensor | |
| Peng et al. | Binocular-vision-based structure from motion for 3-D reconstruction of plants | |
| CN110030979B (en) | Spatial non-cooperative target relative pose measurement method based on sequence images | |
| CN110443837B (en) | Urban area airborne laser point cloud and aerial image registration method and system under constraint of linear features | |
| Habib et al. | Linear features in photogrammetric activities | |
| Das et al. | Extrinsic calibration and verification of multiple non-overlapping field of view lidar sensors | |
| CN112767459A (en) | Unmanned aerial vehicle laser point cloud and sequence image registration method based on 2D-3D conversion | |
| Jurjević et al. | 3D data acquisition based on OpenCV for close-range photogrammetry applications |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190716 |