CN113223067A - Online real-time registration method for three-dimensional scanning point cloud with plane reference and incomplete - Google Patents
Online real-time registration method for three-dimensional scanning point cloud with plane reference and incomplete Download PDFInfo
- Publication number
- CN113223067A CN113223067A CN202110498865.5A CN202110498865A CN113223067A CN 113223067 A CN113223067 A CN 113223067A CN 202110498865 A CN202110498865 A CN 202110498865A CN 113223067 A CN113223067 A CN 113223067A
- Authority
- CN
- China
- Prior art keywords
- plane
- model
- standard model
- point cloud
- primitive
- 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
Images
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
- G06T7/344—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/66—Analysis of geometric attributes of image moments or centre of gravity
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- 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)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Geometry (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
- Processing Or Creating Images (AREA)
Abstract
A method of online real-time registration for an incomplete three-dimensional scan point cloud having a planar reference, comprising S1: preprocessing a standard model, inputting the standard model to obtain a plane primitive set in the standard model and relevant geometric information of each primitive; s2: preparing data, namely inputting three-dimensional scanning point cloud model data; a standard model file; a standard model feature template file; s3: calling RANSAC shape recognition algorithm to the three-dimensional scanning point cloud model data to recognize plane primitives in the three-dimensional scanning point cloud model data; s4: establishing a matching list with all plane primitives in a standard model for each plane primitive of a three-dimensionally scanned point cloud model; s5: establishing all three-plane graph element groups according to the plane graphics primitives; s6: and performing registration calculation between the three-plane combination of the point cloud model and all the possible three-plane combinations corresponding to the standard model to obtain an optimal registration homogeneous coordinate transformation matrix. The method does not need to consider the registration of all point cloud data, and only needs to consider the plane characteristics in the standard model and the actual point cloud model for identification and matching; therefore, the number of search cycles is reduced to reach the target of acceleration.
Description
Technical Field
The invention relates to the relevant technical field of applying a part point cloud model obtained by a three-dimensional scanning device to an automatic manufacturing process. And more particularly, to an algorithm for on-line real-time registration of point cloud data obtained using a three-dimensional scanning device with a pre-set standard model in an automated manufacturing workstation.
Background
In recent years, techniques for forming a digital model of a part by acquiring point cloud data of an actual part using a three-dimensional scanning apparatus have been developed. With the help of the three-dimensional scanning equipment, the three-dimensional scanning is carried out on a large batch of parts on a production line, and a real-time three-dimensional model of the parts can be obtained, so that basic conditions are provided for manufacturing automation. Taking 3C (computer, communication and consumer electronics) product production as an example, the method has the characteristics of high flexibility, high precision, high efficiency, short production period, frequent product line change and the like, and has huge demand on online automatic detection of product parts, and part data information formed after automatic detection can further support subsequent manufacturing processes such as automatic processing, automatic assembly and the like.
The point cloud model of the part obtained by three-dimensional scanning needs model reconstruction, and not only the shape data of the part appearance but also the characteristic (related to the process) data of the part appearance need to be obtained, so that the part point cloud model can be used for quality detection and subsequent processes. The method has the key problem that the standard part model and the actual point cloud part model are registered, so that after the registration is completed, the clamping posture of the actual part can be obtained, and the subsequent process can be assisted to carry out feature recognition and geometric reconstruction. However, the registration method for the point cloud data of the part and the standard part model has the following problems: 1) the data for the three-dimensional scan of the part may be incomplete; 2) due to the high precision requirement and the huge number of scanned data points, the demand of online calculation may not be met in the calculation efficiency. 3) The scanned data may include environmental (e.g., clamp) data, which is required to eliminate interference with the registration result. The method overcomes the problems, realizes the registration of the incomplete point cloud data and the standard part model, and has important significance for the online detection of large-batch high-precision parts. Given that large batches of parts require a flat datum to be set during design, manufacture and assembly, it is reasonable to assume that a flat datum must exist in a standard model of the part. The invention provides an online real-time registration algorithm for incomplete scanned point cloud parts with plane references.
There is already a lot of literature on the registration of point clouds, of which ICP (iterative approximated point) and various improved algorithms of ICP are the most widely applied. And the ICP solves the rotational translation matrix of the two point clouds and the corresponding registration error through Euclidean transformation. However, the ICP algorithm has efficiency problems when processing large-scale point clouds, and the registration result is influenced by the initial value and cannot process the problem of incomplete model. In the later period, Biber and the like propose a Normal Distribution Transform (NDT) algorithm based on a probability density model, the computational complexity is reduced, and the problem of model incompleteness cannot be solved. Researchers further disclose feature learning-based registration methods, which utilize a RANSAC (random consistent sampling) algorithm as a feature extraction tool and estimate a registration relationship through feature learning, and these methods require a large amount of training data, and if a scene is different from the training data in distribution, the registration performance will be sharply reduced. The algorithm disclosed in document [1] finds characteristic point pairs in two sets of registered point clouds, eliminates wrong pairing relation by using RANSAC algorithm to complete coarse registration, and finally realizes fine registration by improving normal abdominal transformation algorithm. Document [2] uses RANSAC for basic primitive recognition in disordered point clouds, and can efficiently recognize planes, spheres, cylinders, cones and rings from large-scale point clouds, thereby providing a basic feature extraction function for a point cloud registration algorithm based on feature matching.
In summary, in processing registration of an incomplete point cloud model with a standard model, an efficient algorithm capable of being completed on line in real time needs to be provided to meet the process beat requirement in an automatic manufacturing process, and the registered information can be used for subsequent automatic detection or processing flow.
Disclosure of Invention
In order to avoid the defects of the prior art, the invention provides an efficient registration algorithm for an incomplete point cloud model with a standard model so as to meet the requirement of online real-time operation. The invention is applicable to a scene aiming at industrial parts (comprising a plurality of plane benchmarks) with manufacturing benchmarks and standard reference model data, the actual point cloud model data is obtained by a three-dimensional scanning device, and the obtained point cloud model data is incomplete with high probability because the whole part cannot be completely scanned in the manufacturing process.
The technical scheme of the invention is as follows:
an online real-time registration method for an incomplete three-dimensional scanning point cloud with a plane reference is provided, which comprises the following steps:
s1: the standard model preprocessing, inputting the standard model, obtaining a plane primitive set and relevant geometric information of each primitive in the standard model, comprising the following substeps:
s11: inputting a standard model file, wherein the standard model file supports two formats, the first format is a CAD model file, and the other format is a discrete model format, such as point cloud data or a Facet (patch structure) model file;
s12: the input standard model file format is judged, and if the model is a CAD model, the process proceeds to S13, and if not, the process proceeds to S14.
S13: traversing a curved surface geometric object in the CAD model to obtain a plane primitive set of the curved surface geometric object;
s14: calling a RANSAC shape recognition algorithm to recognize the plane primitive in the discrete model, and acquiring a plane primitive set and fitting error data of the plane primitive set;
s15: recording the position of the geometric center of a plane primitive in the standard model, the normal of the plane, the boundary shape and the area data; calculating the plane characteristic identification degree index of each primitive and sequencing; forming standard model characteristic template file data;
s16: the characteristic template file of the standard model is examined and optimized, particularly when the characteristic template file is input into a discrete model, the identification result can be manually intervened, and plane primitives which are identified by mistake or are not obvious are eliminated;
s17: outputting a characteristic template file of the standard model as input data of subsequent registration calculation;
s2: data preparation, the following data are entered: three-dimensional scanning point cloud model data (CP)i}; a standard model file; standard model feature template file, wherein a set of standard model PLANEs { PLANE } is providedi}，PLANEi＝(Pi，Vi，Ai，Ti，Mi) Wherein P isi: the geometric center of the planar profile; vi: the normal direction of the plane points to the outer side of the entity where the plane is located; a. thei: the area of the planar profile; t isi: the fitting error of the identified plane;
s3: and calling RANSAC shape recognition algorithm to the three-dimensional scanning point cloud model data to recognize plane primitives therein, wherein the plane primitives are marked as { planej1, j, n, wherein planej＝(pj，vj，aj，tj，mj) Establishing a plane feature identification index R for each plane primitive, sorting, and updating the plane primitives { plane }jThe order of the graphics primitive is arranged from high to low according to the identification index R of the plane graphics primitive, thus ensuring the plane graphics primitive { plane }jAt least three planes which are not parallel to each other are arranged in the registration table, otherwise, the scanned data are too few or the characteristics of the scanned planes are not representative, the registration process cannot be executed, and an algorithm alarm prompt is given;
s4: plane for each plane primitive of a point cloud model for three-dimensional scanningjEstablishing a matching list { (plane) with all plane primitives in a standard modelj，PLANEi) N, and calculating a feature similarity value S for each plane pairjiAnd updating { (plane) according to the plane similarity valuej，PLANEi) The order of the points is that the points are arranged from high to low according to the plane similarity value; in order to control the computational efficiency of the algorithm, it is necessary to limit the plane primitive planejThe number of planes in the paired standard model;
s5: from plane primitive { planejAnd establishing all three-plane graph tuples (plane)m，planen，planel) Where (m, n, l) represents an identification subscript of a plane, is 1, am，planen，planelThe three are not equal and parallel to each other; calculating comprehensive plane characteristic identification value Rmnl＝RmRnRlAccording to the value of the comprehensive plane feature identification RmnlFrom high to low is a set of tri-planar graphs (plane)m，planen，planel) Sorting and updating the sequence;
s6: in the step, the optimal registration homogeneous coordinate transformation matrix is obtained by performing registration calculation between the three-plane combination of the point cloud model and all possible three-plane combinations corresponding to the point cloud model.
Firstly, sequentially setting each combination in a three-plane combination list, selecting corresponding three-plane combinations in a standard model according to the plane similarity, and calculating an optimal registration homogeneous coordinate transformation matrix between two groups of three-plane combinations; calculating the registration error value of the plane pair between all point cloud models and the standard model under the homogeneous coordinate transformation matrix; if the registration error value is smaller than a preset threshold value, the cycle is ended; otherwise, reselecting the three-plane combination according to the plane similarity in the standard model, and repeatedly calculating the optimal registration homogeneous coordinate transformation matrix between the two groups of three-plane combinations; until an optimal registered homogeneous coordinate transformation matrix is found.
The invention provides a plane primitive identification index R for the convenience of description, a typical realization of a value calculation formula of the plane primitive identification index R is related to the area of the plane outline and the number of plane primitives of a plane row in a model, the value of the plane primitive is always greater than 1, and the larger the value is, the higher the plane identification index R of the plane outline in the existing model is;
the plane pair similarity index S is provided for convenience of description by the inventor, the plane pair similarity index S between a first plane in the point cloud and a second plane in the standard model is provided, a typical realization of a calculation formula of the plane pair similarity index S value only needs to consider the area ratio between the plane pairs, and the value of the plane pair similarity index S is always greater than 0 and less than 1;
the main purpose of the plane primitive identification index R and the plane similarity index S is to provide a significance and similarity reference index for the plane feature matching process in the standard model and the point cloud model so as to accelerate the convergence speed of search iteration.
The invention has the following beneficial effects: the registration of all point cloud data is not required to be considered, and only the plane features in the standard model and the actual point cloud model are required to be considered for identification and matching; the number of search cycles can be reduced by comparing the plane identification degree index and the plane pair similarity degree index to reach an accelerated target; meanwhile, the invention only requires that the scanned point cloud comprises at least three datum plane data (the datum plane data are not parallel to each other), and all the point cloud data can adapt to the unfavorable condition of incomplete point cloud data. The result provided by the invention can be used in the subsequent manufacturing process, and the automatic feature extraction, the automatic detection analysis, the automatic programming processing and the like are realized, so that the method has higher practical value.
Drawings
FIG. 1 is a flow chart of the online real-time registration method for an incomplete point cloud model and a standard model according to the present invention;
FIG. 2 is a flowchart of step S1 of FIG. 1 according to the present invention;
FIG. 3 is a diagram of a standard model and its plane primitive set;
FIG. 4 is a schematic diagram of a three-dimensional scanning process and the resulting point cloud data;
FIG. 5 is a schematic diagram of the point cloud and the standard model after registration calculation and pose coordinate transformation.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings.
Firstly, establishing a plane primitive identification index R:
given set of PLANE primitives PLANEi1, N, per PlaneiThe data of (a) include:
Pi: the geometric center of the planar profile;
Vi: the normal direction of the plane points to the outer side of the entity where the plane is located;
Ai: the area of the planar profile;
Ti: the plane is identified as the fitting error. If the plane is from a CAD model (BREP or IGES format), then Ti0; if the planes are identified by RANSAC algorithm, TiA fitting error value identified for the plane;
Mi: represents the Plane in { PlaneiNumber of planes (including the Plane) in parallel relationship if the Plane is in { Plane }iThere are no other parallel planes in the { fraction (1) }.
The plane data set can be expanded, such as by adding specific data (PtSet) for the set of contour points for each planei) So as to realize more efficient plane corresponding relation reasoning. From the formula (1), the i-th plane primitive identification index RiIs always larger than zero, and the larger value represents the higher identification degree of the plane feature. The above formula is only one implementation of the plane primitive identification index R.
And secondly, establishing a plane pair similarity index S between the identified plane in the point cloud model and the plane in the standard model.
Set of PLANEs given a standard model { PLANEi1, wherein plan, N, where plan is equal toi＝(Pi，Vi，Ai，Ti，Mi，PeSeti) As shown in fig. 3.
Given a set of planes { plane } j 1, 1j＝(p_j，v_j，a_j，t_j，m_j，ptsetj)；
Given PLANE Pair (PLANE)i，plancej) Characteristic similarity index Sij∈[0，1]Is composed of
The formula (2) mainly considers the influence of the area on the plane similarity, and only the similarity index SijA more compact implementation of (1). Other alternatives include considering shape similarity information for planar profiles (e.g., PtSet)iAnd ptsetjDegree of match between) the index is introduced to more accurately characterize the similarity of the planar contours.
The standard model preprocessing procedure S1 is described below with reference to fig. 2.
S11: and inputting a standard model file. Two formats are supported, the first is a CAD model file such as IGES, STEP, etc., and the other is a discrete model format such as point cloud data or a Facet model file, etc.
S12: and judging the format of the input standard model file. If it is a CAD model, the process proceeds to S13, and if not, the process proceeds to S14.
S13: traversing a curved surface geometric object in the CAD model to obtain a plane primitive set of the curved surface geometric object;
s14: and calling a RANSAC shape recognition algorithm to recognize the plane primitive in the discrete model, and acquiring a plane primitive set and fitting error data of the plane primitive set.
S15: recording the position of the geometric center of a plane primitive in the standard model, the orientation of the plane, the boundary shape and area data; calculating the plane characteristic identification degree index of each primitive and sequencing; form standard model feature template file data, which can be marked as { PLANEi}，i＝1，...，N。
S16: the characteristic template file of the standard model is examined and optimized, and particularly when the characteristic template file is input into a discrete model, the identification result can be manually interfered, and plane primitives which are identified by mistake or are not obvious are eliminated.
S17: and outputting the characteristic template file of the standard model as input data of subsequent registration calculation.
Then, a registration process between the point cloud model and the standard model performed on line is introduced.
S2: and (4) preparing data. Inputting the following data, (1) three-dimensional scanning point cloud model data { CPi}; (2) a standard model file; (3) standard model feature template file, where set of PLANEs of the standard model { PLANE } is providedi}，PLANEi＝(Pi，Vi，Ai，Ti，Mi). In which (1) is obtained by the three-dimensional scanning apparatus shown in fig. 4. (2) And (3) created by the standard model preprocessing procedure of S1.
S3: calling RANSAC algorithm to the three-dimensional scanning point cloud model, identifying plane primitives therein, and recording the plane primitives as { planej1, j, n, wherein planej＝(pj，vj，aj，tj，mj) As shown in fig. 4; establishing plane feature identification index for each primitive, sorting, and recording as srt (R)j) Update the { planejThe order of which is such that it is arranged from high to low according to R. Need to guarantee the planejAnd (4) at least three planes which are not parallel to each other are provided, otherwise, the scanned data are too little or the scanned plane features are not representative, the registration process cannot be executed, and an algorithm alarm prompt is given.
S4: for each plane primitive plane of three-dimensional scanning point cloud modeljEstablishing a matching list { plane) of all plane primitives in the standard modelj，PLANEi) N, and calculating a feature similarity value S for each plane pairjiAnd sorted by plane similarity value, denoted as srt (R)ji) Update { (plane)j，PLANEi) The order of the planes is arranged according to the plane similarity value from high to low. Will { (plane)j，PLANEt) 1, N, denoted Map<plane,Vector<PLANE>>mapOfPairplane. In order to control the computational efficiency of the algorithm, it is necessary to control each set (R)ji) One recommended strategy is set (R)ji) Does not exceed 10, and is referenced to maximum and minimum similarity values (R)maxAnd Rmin) Selecting a similarity threshold value such as Rthreshold＝Rmin+0.7(Rmax-Rmin) All similarities are smaller than RthresholdThe pairs of planes of (a) are discarded.
S5: according to { planejAnd establishing all three-plane graph tuples (plane)m，planen，planel) Where (m, n, l) represents an identification subscript of a plane, is 1, am，planen，planelThe three are not equal and parallel to each other. Calculating comprehensive plane characteristic identification value Rmnl＝RmRnRlAccording to the value of the comprehensive plane feature identification RmnlFrom high to low is { (plane)m，planen，planel) The sequence is sorted and updated.
S6: in the step, the optimal registration homogeneous coordinate transformation matrix is obtained by performing registration calculation between the three-plane combination of the point cloud model and all possible three-plane combinations corresponding to the point cloud model. The algorithm pseudo code of this step is as follows:
the TriplePlaneRegistration function calculates registration calculation between two three-plane groups, and the program thereof is as follows:
another function, computeplanefeatureregistrientolerance, calculates the registration error value between two plane groups after the two plane groups are transformed by a homogeneous coordinate matrix, and is used for depicting the registration effect. The procedure is described below:
in summary, the online registration method for the three-dimensional scanning point cloud model with the plane reference but incomplete can realize rapid registration (as shown in fig. 5) by using the corresponding relation of the plane features and the registration calculation within controllable cycle times, so as to achieve the effect of online real-time operation.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (3)
1. An online real-time registration method for incomplete three-dimensional scanning point clouds with plane reference is characterized by comprising the following steps,
s1: preprocessing a standard model, inputting the standard model to obtain a plane primitive set in the standard model and relevant geometric information of each primitive;
s2: data preparation, the following data are entered: three-dimensional scanning point cloud model data (CP)i}; a standard model file; standard model feature template file, wherein a set of standard model PLANEs { PLANE } is providedi}，PLANEi＝(Pi，Vi，Ai，Ti，Mi) Wherein P isi: the geometric center of the planar profile; vi: the normal direction of the plane points to the outer side of the entity where the plane is located; a. thei: the area of the planar profile; t isi: the fitting error of the identified plane;
s3: and calling RANSAC shape recognition algorithm to the three-dimensional scanning point cloud model data to recognize plane primitives therein, wherein the plane primitives are marked as { planej1, j, n, wherein planej＝(pj，vj，aj，tj，mj) Establishing a plane feature identification index R for each plane primitive, sorting, and updating the plane primitives { plane }jThe order of the graphics primitive is arranged from high to low according to the identification index R of the plane graphics primitive, thus ensuring the plane graphics primitive { plane }jAt least three planes which are not parallel to each other are arranged in the registration table, otherwise, the scanned data are too few or the characteristics of the scanned planes are not representative, the registration process cannot be executed, and an algorithm alarm prompt is given;
s4: plane for each plane primitive of a point cloud model for three-dimensional scanningjEstablishing a matching list { (plane) with all plane primitives in a standard modelj，PLANEi) N, and calculating a feature similarity value S for each plane pairjiAnd updating { (plane) according to the plane similarity valuej，PLANEi) The order of the points is that the points are arranged from high to low according to the plane similarity value; in order to control the computational efficiency of the algorithm, it is necessary to limit the plane primitive planejThe number of planes in the paired standard model;
s5: from plane primitive { planejAnd establishing all three-plane graph tuples (plane)m，planen，planel) In which (m, n, l) represents a planeA subscript, 1,.., n, and the pane that makes up each tri-planar tuple of imagesm，planen，planelThe three are not equal and parallel to each other; calculating comprehensive plane characteristic identification value Rmnl＝RmRnRlAccording to the value of the comprehensive plane feature identification RmnlFrom high to low is a set of tri-planar graphs (plane)m，planen，planel) Sorting and updating the sequence;
s6: and performing registration calculation between the three-plane combination of the point cloud model and all the possible three-plane combinations corresponding to the standard model to obtain an optimal registration homogeneous coordinate transformation matrix.
2. The method of online real-time registration for three-dimensional scan point clouds with planar fiducials but incomplete as claimed in claim 1, wherein: the step S1 includes the following sub-steps,
s11: inputting a standard model file, wherein the standard model file supports two formats, the first format is a CAD model file, and the other format is a discrete model format;
s12: judging the format of the input standard model file, if the standard model file is a CAD model, entering S13, and if not, entering S14;
s13: traversing a curved surface geometric object in the CAD model to obtain a plane primitive set of the curved surface geometric object;
s14: calling a RANSAC shape recognition algorithm to recognize the plane primitive in the discrete model, and acquiring a plane primitive set and fitting error data of the plane primitive set;
s15: recording the position of the geometric center of a plane primitive in the standard model, the normal of the plane, the boundary shape and the area data; calculating the plane characteristic identification degree index of each primitive and sequencing; forming standard model characteristic template file data;
s16: the characteristic template file of the standard model is examined and optimized, particularly when the characteristic template file is input into a discrete model, the identification result can be manually intervened, and plane primitives which are identified by mistake or are not obvious are eliminated;
s17: and outputting the characteristic template file of the standard model as input data of subsequent registration calculation.
3. The method of online real-time registration for three-dimensional scan point clouds with planar fiducials but incomplete, according to claim 1 or 2, characterized by: the three-plane combination method in the step S6 includes the steps that firstly, each combination in a three-plane combination list is sequentially arranged, corresponding three-plane combinations are selected from a standard model according to the plane similarity, and an optimal registration homogeneous coordinate transformation matrix between two groups of three-plane combinations is calculated; calculating the registration error value of the plane pair between all point cloud models and the standard model under the homogeneous coordinate transformation matrix; if the registration error value is smaller than a preset threshold value, the cycle is ended; otherwise, reselecting the three-plane combination according to the plane similarity in the standard model, and repeatedly calculating the optimal registration homogeneous coordinate transformation matrix between the two groups of three-plane combinations; until an optimal registered homogeneous coordinate transformation matrix is found.
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110498865.5A CN113223067B (en) | 2021-05-08 | 2021-05-08 | Online registration method for incomplete three-dimensional scanning point cloud with plane reference |
| PCT/CN2022/072166 WO2022237225A1 (en) | 2021-05-08 | 2022-01-14 | Online real-time registration method for incomplete three-dimensional scanning point cloud having plane reference |
| ZA2023/09569A ZA202309569B (en) | 2021-05-08 | 2023-10-12 | Online real-time registration method for incomplete three-dimensional scanning point cloud having plane reference |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110498865.5A CN113223067B (en) | 2021-05-08 | 2021-05-08 | Online registration method for incomplete three-dimensional scanning point cloud with plane reference |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113223067A true CN113223067A (en) | 2021-08-06 |
| CN113223067B CN113223067B (en) | 2021-11-19 |
Family
ID=77091752
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110498865.5A Active CN113223067B (en) | 2021-05-08 | 2021-05-08 | Online registration method for incomplete three-dimensional scanning point cloud with plane reference |
Country Status (3)
| Country | Link |
|---|---|
| CN (1) | CN113223067B (en) |
| WO (1) | WO2022237225A1 (en) |
| ZA (1) | ZA202309569B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114882087A (en) * | 2022-07-12 | 2022-08-09 | 武汉瀚迈科技有限公司 | Real-time registration method for three-dimensional scanning point cloud with incomplete basic primitive |
| WO2022237225A1 (en) * | 2021-05-08 | 2022-11-17 | 广东三姆森科技股份有限公司 | Online real-time registration method for incomplete three-dimensional scanning point cloud having plane reference |
| CN115661220A (en) * | 2022-12-28 | 2023-01-31 | 深圳煜炜光学科技有限公司 | Point cloud data registration method, device, equipment and storage medium |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116738572B (en) * | 2023-06-19 | 2024-03-01 | 北京知元创通信息技术有限公司 | Aircraft configuration modularized assembly method and system |
| CN118134979A (en) * | 2024-05-07 | 2024-06-04 | 法奥意威(苏州)机器人系统有限公司 | Point cloud registration method and device based on single plane characteristics and electronic equipment |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160196467A1 (en) * | 2015-01-07 | 2016-07-07 | Shenzhen Weiteshi Technology Co. Ltd. | Three-Dimensional Face Recognition Device Based on Three Dimensional Point Cloud and Three-Dimensional Face Recognition Method Based on Three-Dimensional Point Cloud |
| CN106023298A (en) * | 2016-06-22 | 2016-10-12 | 山东理工大学 | Point cloud rigid registration method based on local Poisson curved surface reconstruction |
| US9472022B2 (en) * | 2012-10-05 | 2016-10-18 | University Of Southern California | Three-dimensional point processing and model generation |
| CN107610084A (en) * | 2017-09-30 | 2018-01-19 | 驭势科技（北京）有限公司 | A kind of method and apparatus that information fusion is carried out to depth image and laser spots cloud atlas |
| CN109965979A (en) * | 2017-12-27 | 2019-07-05 | 上海复旦数字医疗科技股份有限公司 | A kind of steady Use of Neuronavigation automatic registration method without index point |
| CN110443836A (en) * | 2019-06-24 | 2019-11-12 | 中国人民解放军战略支援部队信息工程大学 | A kind of point cloud data autoegistration method and device based on plane characteristic |
| CN111882658A (en) * | 2020-07-02 | 2020-11-03 | 电子科技大学 | Automatic reconstruction method for nuclear power plant kernel facility |
| CN112489207A (en) * | 2021-02-07 | 2021-03-12 | 深圳大学 | Space-constrained dense matching point cloud plane element extraction method |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9940746B2 (en) * | 2015-06-18 | 2018-04-10 | Apple Inc. | Image fetching for timeline scrubbing of digital media |
| CN104915986B (en) * | 2015-06-26 | 2018-04-17 | 北京航空航天大学 | A kind of solid threedimensional model method for automatic modeling |
| US10580114B2 (en) * | 2017-03-29 | 2020-03-03 | Zhijing George Mou | Methods and systems for real time 3D-space search and point-cloud registration using a dimension-shuffle transform |
| CN109872350A (en) * | 2019-02-18 | 2019-06-11 | 重庆市勘测院 | A kind of new point cloud autoegistration method |
| CN110442917B (en) * | 2019-07-09 | 2023-06-20 | 武汉工程大学 | Parameterized model reconstruction method based on point cloud |
| CN112001955A (en) * | 2020-08-24 | 2020-11-27 | 深圳市建设综合勘察设计院有限公司 | Point cloud registration method and system based on two-dimensional projection plane matching constraint |
| CN113223067B (en) * | 2021-05-08 | 2021-11-19 | 广东三姆森科技股份有限公司 | Online registration method for incomplete three-dimensional scanning point cloud with plane reference |
-
2021
- 2021-05-08 CN CN202110498865.5A patent/CN113223067B/en active Active
-
2022
- 2022-01-14 WO PCT/CN2022/072166 patent/WO2022237225A1/en unknown
-
2023
- 2023-10-12 ZA ZA2023/09569A patent/ZA202309569B/en unknown
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9472022B2 (en) * | 2012-10-05 | 2016-10-18 | University Of Southern California | Three-dimensional point processing and model generation |
| US20160196467A1 (en) * | 2015-01-07 | 2016-07-07 | Shenzhen Weiteshi Technology Co. Ltd. | Three-Dimensional Face Recognition Device Based on Three Dimensional Point Cloud and Three-Dimensional Face Recognition Method Based on Three-Dimensional Point Cloud |
| CN106023298A (en) * | 2016-06-22 | 2016-10-12 | 山东理工大学 | Point cloud rigid registration method based on local Poisson curved surface reconstruction |
| CN107610084A (en) * | 2017-09-30 | 2018-01-19 | 驭势科技（北京）有限公司 | A kind of method and apparatus that information fusion is carried out to depth image and laser spots cloud atlas |
| CN109965979A (en) * | 2017-12-27 | 2019-07-05 | 上海复旦数字医疗科技股份有限公司 | A kind of steady Use of Neuronavigation automatic registration method without index point |
| CN110443836A (en) * | 2019-06-24 | 2019-11-12 | 中国人民解放军战略支援部队信息工程大学 | A kind of point cloud data autoegistration method and device based on plane characteristic |
| CN111882658A (en) * | 2020-07-02 | 2020-11-03 | 电子科技大学 | Automatic reconstruction method for nuclear power plant kernel facility |
| CN112489207A (en) * | 2021-02-07 | 2021-03-12 | 深圳大学 | Space-constrained dense matching point cloud plane element extraction method |
Non-Patent Citations (2)
| Title |
|---|
| MINHWAN SHIN ET AL.: "3D LiDAR-based point cloud map registration: Using spatial location of visual features", 《 2017 2ND INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION ENGINEERING (ICRAE)》 * |
| 石力文 等: "基于点云数据与BIM的古建筑三维信息表达方法研究", 《遗产与保护研究》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022237225A1 (en) * | 2021-05-08 | 2022-11-17 | 广东三姆森科技股份有限公司 | Online real-time registration method for incomplete three-dimensional scanning point cloud having plane reference |
| CN114882087A (en) * | 2022-07-12 | 2022-08-09 | 武汉瀚迈科技有限公司 | Real-time registration method for three-dimensional scanning point cloud with incomplete basic primitive |
| CN114882087B (en) * | 2022-07-12 | 2022-10-04 | 武汉瀚迈科技有限公司 | Real-time registration method for three-dimensional scanning point cloud with incomplete basic graphic elements |
| CN115661220A (en) * | 2022-12-28 | 2023-01-31 | 深圳煜炜光学科技有限公司 | Point cloud data registration method, device, equipment and storage medium |
| CN115661220B (en) * | 2022-12-28 | 2023-03-17 | 深圳煜炜光学科技有限公司 | Point cloud data registration method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| ZA202309569B (en) | 2023-11-29 |
| CN113223067B (en) | 2021-11-19 |
| WO2022237225A1 (en) | 2022-11-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113223067B (en) | Online registration method for incomplete three-dimensional scanning point cloud with plane reference | |
| CN109064502B (en) | Multi-source image registration method based on combination of deep learning and artificial design features | |
| CN112348864B (en) | Three-dimensional point cloud automatic registration method for laser contour features of fusion line | |
| CN109919984A (en) | A kind of point cloud autoegistration method based on local feature description's | |
| Gotardo et al. | Range image segmentation into planar and quadric surfaces using an improved robust estimator and genetic algorithm | |
| CN115147437B (en) | Intelligent robot guiding machining method and system | |
| CN114743259A (en) | Pose estimation method, pose estimation system, terminal, storage medium and application | |
| CN112381862B (en) | Full-automatic registration method and device for CAD (computer-aided design) model and triangular mesh | |
| CN110490915B (en) | Point cloud registration method based on convolution-limited Boltzmann machine | |
| CN115147833A (en) | Part pose identification method and system | |
| CN114187267B (en) | Stamping part defect detection method based on machine vision | |
| CN112001244A (en) | Computer-aided iris comparison method and device | |
| CN117132630A (en) | Point cloud registration method based on second-order spatial compatibility measurement | |
| CN107316327A (en) | Knochenbruch section and knochenbruch Model registration method based on maximum public subgraph and bounding box | |
| CN110472651B (en) | Target matching and positioning method based on edge point local characteristic value | |
| CN115100277A (en) | Method for determining position and pose of complex curved surface structure part | |
| CN113313200B (en) | Point cloud precision matching method based on normal constraint | |
| CN111553410B (en) | Point cloud identification method based on key point local curved surface feature histogram and spatial relationship | |
| CN116188540A (en) | Target identification and pose estimation method based on point cloud information | |
| CN117079272A (en) | Bullet bottom socket mark feature identification method combining manual features and learning features | |
| CN116310244A (en) | Ceramic fragment three-dimensional intelligent splicing method based on contour features | |
| CN109448031A (en) | Method for registering images and system based on Gaussian field constraint and manifold regularization | |
| CN114882087B (en) | Real-time registration method for three-dimensional scanning point cloud with incomplete basic graphic elements | |
| CN113781559B (en) | Robust abnormal matching point eliminating method and image indoor positioning method | |
| CN118279599B (en) | Point cloud plane extraction method combining RANSAC and nearest point sampling algorithm |
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 |
Address after: 523000 No.3, West Zhongnan Road, Haibin District, Shangsha, Chang'an Town, Dongguan City, Guangdong Province Applicant after: Guangdong Samson Technology Co.,Ltd. Address before: 523000 No.3, West Zhongnan Road, Haibin District, Shangsha, Chang'an Town, Dongguan City, Guangdong Province Applicant before: DONGGUAN SAMSUN OPTICAL TECHNOLOGY Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |