CN107590825B - Point cloud hole repairing method based on SFM - Google Patents
Point cloud hole repairing method based on SFM Download PDFInfo
- Publication number
- CN107590825B CN107590825B CN201710637964.0A CN201710637964A CN107590825B CN 107590825 B CN107590825 B CN 107590825B CN 201710637964 A CN201710637964 A CN 201710637964A CN 107590825 B CN107590825 B CN 107590825B
- Authority
- CN
- China
- Prior art keywords
- point cloud
- point
- hole
- points
- boundary
- 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.)
- Active
Links
Images
Landscapes
- Processing Or Creating Images (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a point cloud hole repairing method based on SFM, which comprises the following steps: (1) respectively acquiring point cloud data by using a grating projection method and an SFM method, and extracting boundary points of three-dimensional point cloud holes by using point cloud two-dimensional phase information acquired in the grating projection method; (2) registering the point cloud data set acquired by the SFM and the point cloud data set acquired by a grating projection method; (3) extracting raster projection on a data set acquired by SFM to acquire supplementary points of a hole area in the point cloud; (4) and based on the supplementary points, further hole repairing is realized by utilizing the radial basis function. The method has the advantages of robust algorithm, high accuracy of repairing effect and capability of recovering more detailed information of the object.
Description
The technical field is as follows:
the invention relates to a point cloud hole repairing method based on SFM, belonging to the technical field of point cloud post-processing.
Background art:
in the process of three-dimensional measurement, holes often exist in three-dimensional point cloud data due to the complex shape of an object or the defects of a measurement mode. The grating projection measurement method has the characteristics of high speed, high automation degree and the like, but because auxiliary light needs to be projected during measurement, the obtained point cloud data can generate a hole area due to the shielding of an object, the reflectivity of the surface of a measured object and the like. The existence of holes seriously affects the subsequent operation of the point cloud. Therefore, the hole repairing work is indispensable.
The difficulty of hole repair is: 1. how to accurately extract hole boundaries; 2. how to obtain more information in the hole area to make the repaired point more fit to the original shape. The extraction of the hole boundary is the first task of hole repair, and its accuracy directly affects the result of hole repair. The existing methods for extracting boundary points of point cloud holes are mainly carried out on the basis of three dimensions and mainly comprise a point cloud grid-based extraction method and a method for directly extracting boundary points on point clouds. The former has complex calculation and is easy to take the internal points of the point cloud with uneven sampling as boundary points; the latter typically requires computing the relationship of points and neighborhood points and is susceptible to noise. The point cloud boundary points can be found to have obvious characteristics in the two-dimensional phase diagram through actual measurement, and the hole boundary is extracted by utilizing the two-dimensional phase, so that the need of establishing a grid or calculating neighborhood information in a three-dimensional method is avoided, and the efficiency of extracting the boundary points is obviously improved. The point cloud two-dimensional phase information is therefore used herein to extract hole boundaries.
The existing hole repairing algorithms are roughly divided into two types, one is to directly utilize neighborhood information, continuity or curvature around a point cloud hole to repair the hole, for example, a spline curve, a neural network and other methods are utilized to perform interpolation, so as to repair the hole. The algorithm is simple and easy to implement, and does not need any supplementary information. However, for point cloud holes with large curvature change or complex shapes, the repairing effect is not ideal due to excessive information loss. The other type of the method mainly carries out repairing by acquiring additional information and combining structures around the point cloud hole. Such as Yann quench et al, propose hole patching using a priori CAD model and continuity of hole area neighborhoods as constraints. M Panchetti et al propose to use two-dimensional information of the corresponding picture and curvature information between point cloud meshes to patch holes. The method has good repairing effect on complex hole regions, but the overall complexity of the algorithm is higher no matter the establishment of a prior CAD model or the registration between two-dimensional pixels and three-dimensional point clouds.
Disclosure of Invention
The technical problem is as follows: aiming at the problems that hole area information is less and complex hole repairing effect is not ideal in a hole repairing process, the invention provides a point cloud hole repairing method based on SFM by combining a mature technology for obtaining point cloud information by fusing SFM and laser scanning in three-dimensional reconstruction. The invention supplements the information of the hole part by acquiring the point cloud data of the SFM and guides the hole repairing, and the supplementary information can effectively help to recover the point cloud hole information, improve the accuracy of the hole repairing result and recover more information. And the two-dimensional phase information of the point cloud is utilized to extract the boundary points of the point cloud, so that the complexity of the boundary point positioning is reduced.
The technical scheme is as follows: in order to achieve the purpose, the invention adopts the technical scheme that:
a point cloud hole repairing method based on SFM comprises the following steps:
step 1: point cloud hole boundary acquisition: extracting a point cloud hole boundary by using point cloud two-dimensional phase information acquired in grating projection; firstly, extracting boundary points of point cloud; then removing the outer boundary points of the point cloud;
step 2: registering the two point clouds: firstly, performing coarse registration by using a RANSAC algorithm based on a coplanar four-point set, and then performing accurate registration by using an improved ICP algorithm;
and step 3: and (3) supplementary point extraction: after the registration is finished, extracting the point cloud P obtained by the SFM method by using the point cloud hole boundary obtained in the step 1sPoint cloud P obtained by raster projection methodwA supplementary point IP for the mesopore region;
and 4, step 4: filling a point cloud hole area by using a radial basis function: firstly, projecting hole boundary points to a two-dimensional plane to obtain a point cloud boundary characteristic polygon, and then performing interpolation by using the characteristics of the boundary characteristic polygon to obtain an initial patch point set FP (FP) { FPi1,2,3,. n }, and then adjusting the initial set of repairing points according to the radial basis function to obtain a final set of repairing points;
preferably: the step of extracting the point cloud hole boundary by using the two-dimensional phase information in the step 1 is as follows:
step 1.1: acquiring two-dimensional boundary points of point cloud: the size of the two-dimensional phase diagram is M x N, the p [ M ] [ N ] phase is a certain point on the phase diagram, the set of boundary points of the point cloud is set to be B, and when the p [ M ] [ N ] meets the formula, the p [ M ] [ N ] belongs to B;
wherein
Is the threshold value of the phase jump;
step 1.2: extracting external boundary points of the point cloud: on the two-dimensional phase image, the outer contour points are characterized by initial and final non-zero points of each row in B; setting the set of points on the external contour line as S, and when p [ m ] [ n ] meets the formula, p [ m ] [ n ] belongs to S;
m1, 2.. M, so the boundary point Hp of the hole is { p [ M ] [ n ] | p [ M ] [ n ] ∈ B-S };
step 1.3: calculating three-dimensional coordinates of point cloud hole boundary points: knowing the coordinates [ X ] of the point cloud datacYcZc]With corresponding point p [ m ] on the phase diagram][n]The correspondence relationship between θ and θ is as follows,
wherein θ is a phase value; a is1,a2,a3,a4,a5,a6,a7,a8Is a system parameter; rho is a scale factor; a. thecA 3 x 3 camera internal parameter matrix; and a is1,a2,a3,a4,a5,a6,a7,a8AcCan be obtained by calibration of the system; therefore, substituting the two-dimensional phase coordinates of the hole boundary points, the corresponding hole boundary three-dimensional coordinate point H can be calculated by the above two formulasi,i＝1,2,3...n}；
Preferably: the point cloud registration in step 2 comprises the following steps:
step 2.1: initial registration: in the initial registration, to increase the registration efficiency, first at PwAnd PsRespectively dividing corresponding areas with obvious characteristics; at PwIn the characteristic region, randomly selecting four non-collinear points by using RANSAC method, and selecting the four non-collinear points in PsFinding out corresponding points, and calculating to obtain a corresponding transformation matrix M; and after r iterations, selecting a transformation matrix M which enables the matching error of the two point clouds to be minimum. Obtaining an initial matching result by using M;
step 2.2: and (3) accurate registration: in exact registration, first for PwAt each point P insZhongxie OuFour points with the nearest Euclidean distance are calculated, and the gravity center of the minimum neighborhood formed by the four points is calculated to form a point pair; then, calculating parameters R and t by using a least square method; apply the final result to PsObtaining a point cloud set P after registrations'；
Preferably: the method for screening the supplementary point IP in the step 3 comprises the following steps: for the point cloud P registered in the step 2s' calculating whether it is the hole boundary point H ═ H extracted in said step 1 or notiN, i-1, 2,3.. n, if any, is considered to be a bounding box
Otherwise, not; the point cloud set added with the supplementary points is NP ═ Pω∪IP}；
Preferably: the step of filling the hole area with the radial basis function in the step 4 comprises the following steps:
step 4.1: extracting a hole characteristic polygon: and (3) extracting the point cloud hole points H ═ H extracted in the step (1) by using a least square methodiProjecting i ═ 1,2,3.. n } to a two-dimensional plane to obtain a projection matrix PM, and forming a hole characteristic polygon;
step 4.2: obtaining a rough repair point: acquiring the side length of a hole characteristic polygon, and taking the average value d of the side lengths as a step length to perform interpolation in a hole area so as to obtain a repair point on a two-dimensional plane; mapping the interpolation point back to the three-dimensional plane by using the PM acquired in the step 4.2, and acquiring a rough repair point FP ═ FP { (FP) }i,i＝1,2,3,...n}；
Step 4.3: repairing the holes: firstly, a kd tree is established for the point cloud set NP added with the information supplement points in the step 3, and then any point FP in the FP is subjected toiSelecting k neighborhood points, establishing a space curved surface by using a Radial Basis Function (RBF), and enabling fp to be subjected to gradient conditionsiAdjusting the curved surface to meet the requirement of the curvature of the surface of the point cloud; wherein each point fpiThe corresponding equation of the space curved surface is
F(x,y,z)＝p0+p1x+p2y+p3z+β
Wherein β is an error term, an
Unknown number omega of space surface equationj,p0,p1,p2,p3Can be obtained by solving the formula QA ═ B;
wherein a ═ ω1ω2… ωnp0p1p2p3]TB＝[1 2…n0 0 0 0]T，
In the experimentiIn order to be an error value,1＝2＝…＝n＝0.1，N＝{Nix，Niy，Niz，i＝1,2…，6}，Dij＝|Ni-Nj|3/2i＝1,2,...，6,j＝1,2,...6，
in the experimentiIn order to be an error value,1＝2＝...＝n＝0.1，N＝{Nix，Niy，Niz，i＝1,2...，6}，Dij＝|Ni-Nj|3/2i＝1,2,…，6,j＝1,2,...6。
has the advantages that: compared with the prior art, the SFM-based hole repairing algorithm provided by the invention has the following beneficial effects:
besides acquiring information from the hole boundary, a new method for acquiring information from the point cloud acquired by the SFM is added, so that the hole filling precision is improved; the method for searching the hole boundary in the three-dimensional space makes the algorithm complex, and the method for searching the hole boundary by using the two-dimensional phase information improves the efficiency of the algorithm.
Drawings
FIG. 1 is a complete process of the present invention.
FIG. 2 is a point cloud model diagram of a smooth ceramic cup. It can be seen from the figure that a hole is formed in the smooth ceramic cup at the middle position due to the fact that the reflectivity of the object is too high.
FIG. 3 is a point cloud model diagram of a human face. In the face model, holes are generated in the eye and eyebrow regions due to low reflectivity, and holes are generated in the nose region due to self-occlusion.
FIG. 4 shows a point cloud of a smooth ceramic cup after hole repair. The repaired point cloud picture can be more fit with the original data.
FIG. 5 is a detailed view of the hole area repair of the smooth ceramic cup, wherein there are 404 newly added hole repair points. The SFM acquired point cloud data provides 96 supplementary points. The data provides key information when the RBF function calculates the surface equation of the corresponding point, so that the hole repairing result is more accurate.
Fig. 6 is a point cloud of a face after hole patching. The repair result accurately reflects the shape of the object after three-dimensional reconstruction, and restores richer details.
Fig. 7 shows in detail the filling of the nose part. For more complex models such as human face models, the SFM-based point cloud hole repairing method can also effectively recover hole information.
Detailed Description
The present invention is further illustrated by the following description in conjunction with the accompanying drawings and the specific embodiments, it is to be understood that these examples are given solely for the purpose of illustration and are not intended as a definition of the limits of the invention, since various equivalent modifications will occur to those skilled in the art upon reading the present invention and fall within the limits of the appended claims.
Embodiments of the present invention will now be described in more detail with reference to the accompanying drawings. And the programming implementation tool selects Visual Studio2013 to respectively perform experiments on the ceramic cup and the human face.
The method comprises the following specific implementation steps:
step 1: point cloud hole boundary acquisition
Step 1.1: obtaining two-dimensional boundary points of a point cloud
The size of the two-dimensional phase diagram is M x N, p [ M ] [ N ] is a certain point on the phase diagram, the set of boundary points of the point cloud is set to be B, and when p [ M ] [ N ] meets the formula, p [ M ] [ N ] belongs to B;
wherein
Is the threshold value of the phase jump;
step 1.2: extracting external boundary points of a point cloud
On the two-dimensional phase image, the outer contour points are characterized by initial and final non-zero points of each row in B; setting the set of points on the external contour line as S, and when p [ m ] [ n ] meets the formula, p [ m ] [ n ] belongs to S;
m1, 2.. M, so the boundary point Hp of the hole is { p [ M ] [ n ] | p [ M ] [ n ] ∈ B-S };
step 1.3: calculating three-dimensional coordinates of point cloud hole boundary points
Knowing the coordinates [ X ] of the point cloud datacYcZc]With corresponding point p [ m ] on the phase diagram][n]The correspondence relationship between θ and θ is as follows,
wherein θ is a phase value; a is1,a2,a3,a4,a5,a6,a7,a8Is a system parameter; rho is a scale factor; a. thecA 3 x 3 camera internal parameter matrix; and a is1,a2,a3,a4,a5,a6,a7,a8AcCan pass throughObtaining the calibration of the system; therefore, substituting the two-dimensional phase coordinates of the hole boundary points, the corresponding hole boundary three-dimensional coordinate point H can be calculated by the above two formulasi,i＝1,2,3...n}；
Step 2: registering two point clouds
Step 2.1: initial registration
In the initial registration, to increase the registration efficiency, first at PwAnd PsRespectively dividing corresponding areas with obvious characteristics; at PwIn the characteristic region, randomly selecting four non-collinear points by using RANSAC method, and selecting the four non-collinear points in PsFinding out corresponding points, and calculating to obtain a corresponding transformation matrix M; after r iterations, selecting a transformation matrix M which enables the matching error of the two point clouds to be minimum; obtaining an initial matching result by using M;
step 2.2: precision registration
In exact registration, first for PwAt each point P insFinding four points with the nearest Euclidean distance, and solving the gravity center of a minimum neighborhood formed by the four points to form a point pair; then, calculating parameters R and t by using a least square method; apply the final result to PsObtaining a point cloud set P after registrations'；
And step 3: extracting supplemental points
For Ps' each point in the above is calculated whether it is the hole boundary point H ═ H extracted in step 1iN, i-1, 2,3.. n, if any, is considered to be a bounding box
Otherwise, not; the point cloud set added with the supplementary points is NP ═ Pω∪IP}；
And 4, step 4: filling point cloud hole area by using radial basis function
Step 4.1: extracting hole characteristic polygon
And (3) using a least square method to set the point cloud hole point H ═ H extracted in the step (1) to be HiProjecting i ═ 1,2,3.. n } to a two-dimensional plane to obtain a projection matrix PM, and forming a hole characteristic polygon;
step 4.2: obtaining rough repair points
And acquiring the side length of the hole characteristic polygon, and taking the average value d of the side lengths as a step length to perform interpolation in the hole area so as to obtain a repairing point on the two-dimensional plane. Mapping the interpolation point back to the three-dimensional plane by using the PM acquired in the step 4.1, and acquiring a rough repair point FP ═ FP { (FP) }i,i＝1,2,3,…n}；
Step 4.3: repairing a hole
Firstly, a kd tree is established for the point cloud set NP added with the information supplement points in the step 3, and then any point FP in the FP is subjected toiSelecting k neighborhood points, establishing a space curved surface by using a Radial Basis Function (RBF), and enabling fp to be subjected to gradient conditionsiAdjusting the curved surface to meet the requirement of the curvature of the surface of the point cloud; wherein each point fpiThe corresponding equation of the space curved surface is
F(x,y,z)＝p0+p1x+p2y+p3z+β
Wherein β is an error term, an
Unknown number omega of space surface equationj,p0,p1,p2,p3Can be obtained by solving the formula QA ═ B;
wherein a ═ ω1ω2… ωnp0p1p2p3]TB＝[1 2…n0 0 0 0]T，
according to the method, the data set acquired by the SFM is added into the hole repairing, and the supplementary information is added by using the data set to guide the hole repairing, so that the repaired hole area can acquire more detailed information, and the shape of the hole part can be better recovered. Adding the two-dimensional phase information into the point cloud hole extraction, and extracting the hole boundary by using the characteristics of the three-dimensional hole boundary point in the two-dimensional phase; the complexity of hole boundary extraction is reduced.
The above description is only of the preferred embodiments of the present invention, and it should be noted that: it will be apparent to those skilled in the art that various modifications and adaptations can be made without departing from the principles of the invention and these are intended to be within the scope of the invention.
Claims (4)
1. A point cloud hole repairing method based on SFM is characterized by comprising the following steps:
step 1: point cloud hole boundary acquisition: extracting a point cloud hole boundary by using point cloud two-dimensional phase information acquired in grating projection, and firstly extracting inner and outer boundary points of the point cloud; secondly, removing the outer boundary points of the point cloud;
step 2: registering the two point clouds: firstly, performing coarse registration by using a RANSAC algorithm based on a coplanar four-point set, and then performing accurate registration by using an improved ICP algorithm;
and step 3: and (3) supplementary point extraction: after the registration is finished, extracting the point cloud P obtained by the SFM method by using the point cloud hole boundary obtained in the step 1sPoint cloud P obtained by raster projection methodwA supplementary point IP for the mesopore region;
and 4, step 4: filling a point cloud hole area by using a radial basis function: firstly, projecting hole boundary points to a two-dimensional plane to obtain a point cloud boundary characteristic polygon, and then performing interpolation by using the characteristics of the boundary characteristic polygon to obtain an initial patch point set FP (FP) { FP)i1,2,3,. n }, and then adjusting the initial set of repairing points according to the radial basis function to obtain a final set of repairing points;
step 4.1: extracting a hole characteristic polygon: and (3) using a least square method to set the point cloud hole point H ═ H extracted in the step (1) to be HiI-1, 2,3.. n } is projected to a two-dimensional plane to obtain a projection momentForming a hole characteristic polygon by PM;
step 4.2: obtaining a rough repair point: firstly, obtaining the side length of a hole characteristic polygon, and taking the average value d of the side lengths as a step length to perform interpolation in a hole area so as to obtain a repair point on a two-dimensional plane; mapping the interpolation point back to the three-dimensional plane by using the PM acquired in the step 4.1, and acquiring a rough repair point FP ═ FP { (FP) }i,i＝1,2,3,...n}；
Step 4.3: repairing the holes: firstly, a kd tree is established for the point cloud set NP added with the information supplement points in the step 3, and then any point FP in the FP is subjected toiSelecting k neighborhood points, establishing a space curved surface by using a Radial Basis Function (RBF), and enabling fp to be subjected to gradient conditionsiAdjusting the curved surface to meet the requirement of the curvature of the surface of the point cloud; wherein each point fpiThe corresponding equation of the space curved surface is
F(x,y,z)＝p0+p1x+p2y+p3z+β
Wherein β is error term, unknown number omega of space surface equationj,p0,p1,p2,p3Can be obtained by solving the formula QA ═ B; wherein a ═ ω1ω2…ωnp0p1p2p3]TB＝[1 2…n0 0 0 0]T，
In the experimentiIn order to be an error value,1＝2＝...＝n＝0.1，N＝{Nix，Niy，Niz，i＝1,2...，6}，Dij＝|Ni-Nj|3/2i＝1,2,...，6,j＝1,2,...6。
2. the SFM-based point cloud hole repairing method of claim 1, wherein in step 1, the step of extracting the point cloud hole boundary using the two-dimensional phase information comprises the following steps:
step 1.1: acquiring two-dimensional boundary points of point cloud: the size of the two-dimensional phase diagram is M N, p [ M ] [ N ] is a certain point on the phase diagram, the set of boundary points of the point cloud is B, when p [ M ] [ N ] meets the formula, p [ M ] [ N ] belongs to B,
wherein
Is the threshold value of the phase jump;
step 1.2: extracting external boundary points of the point cloud: on the two-dimensional phase image, the external contour points are characterized by the initial and final non-zero points of each line in B, the set of points on the external contour is set as S, when p [ m ] [ n ] satisfies the formula, p [ m ] [ n ] belongs to S,
m1, 2.. M, so the boundary point Hp of the hole is { p [ M ] [ n ] | p [ M ] [ n ] ∈ B-S };
step 1.3: calculating three-dimensional coordinates of point cloud hole boundary points: knowing the coordinates [ X ] of the point cloud datacYcZc]TWith corresponding point p [ m ] on the phase diagram][n]The correspondence relationship between θ and θ is as follows,
wherein θ is a phase value; a is1,a2,a3,a4,a5,a6,a7,a8Is a system parameter; rho is a scale factor; a. thecFor 3 x 3 camera internal parametersA matrix; and a is1,a2,a3,a4,a5,a6,a7,a8，AcThe three-dimensional coordinate system can be obtained by calibration of the system, so that the two-dimensional phase coordinates of the hole boundary points are substituted, and the corresponding hole boundary three-dimensional coordinate point H can be calculated by the two formulasi,i＝1,2,3...n}。
3. The SFM-based point cloud hole patching method as claimed in claim 1, wherein in step 2, the point cloud registration step is as follows:
step 2.1: initial registration: in the initial registration, to increase the registration efficiency, first at PwAnd PsRespectively dividing corresponding areas with obvious characteristics; at PwIn the characteristic region, randomly selecting four non-collinear points by using RANSAC method, and selecting the four non-collinear points in PsFinding out corresponding points, and calculating to obtain a corresponding transformation matrix M; after r iterations, selecting a transformation matrix M which enables the matching error of the two point clouds to be minimum; obtaining an initial matching result by using M;
step 2.2: and (3) accurate registration: in exact registration, first for PwAt each point P insFinding four points with the nearest Euclidean distance, and solving the gravity center of a minimum neighborhood formed by the four points to form a point pair; then, calculating parameters R and t by using a least square method; apply the final result to PsObtaining a point cloud set P after registrations'。
4. The SFM-based point cloud hole repairing method according to claim 3, wherein the method for screening the supplementary point IP in the step 3 comprises the following steps: for the registered point cloud P obtained in the step 2s' for each point, calculate whether it is the hole boundary point H ═ H extracted in step 1iN } of the bounding boxes of 1,2,3
Otherwise, not; the point cloud set added with supplementary points is NP ═ last tonePω∪IP}。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710637964.0A CN107590825B (en) | 2017-07-31 | 2017-07-31 | Point cloud hole repairing method based on SFM |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710637964.0A CN107590825B (en) | 2017-07-31 | 2017-07-31 | Point cloud hole repairing method based on SFM |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107590825A CN107590825A (en) | 2018-01-16 |
| CN107590825B true CN107590825B (en) | 2020-10-02 |
Family
ID=61042042
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710637964.0A Active CN107590825B (en) | 2017-07-31 | 2017-07-31 | Point cloud hole repairing method based on SFM |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107590825B (en) |
Families Citing this family (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3740938A1 (en) | 2018-01-19 | 2020-11-25 | InterDigital VC Holdings, Inc. | Processing a point cloud |
| CN108399609B (en) * | 2018-03-06 | 2020-07-31 | 北京因时机器人科技有限公司 | Three-dimensional point cloud data repairing method and device and robot |
| CN109242828B (en) * | 2018-08-13 | 2022-04-19 | 浙江大学 | Three-dimensional defect detection method for 3D printed product based on grating projection multistep phase shift method |
| CN109147040B (en) * | 2018-08-28 | 2020-06-30 | 浙江大学 | Template-based human point cloud hole repairing method |
| CN109859114B (en) * | 2018-12-27 | 2020-10-16 | 北京大学 | Three-dimensional point cloud repairing method based on local smoothness and non-local similarity |
| CN110415331B (en) * | 2019-06-14 | 2022-08-19 | 中国电建集团昆明勘测设计研究院有限公司 | Method for rapidly detecting, judging and repairing contour or hole based on point cloud data |
| CN110349257B (en) * | 2019-07-16 | 2020-02-28 | 四川大学 | Phase pseudo mapping-based binocular measurement missing point cloud interpolation method |
| CN112102491B (en) * | 2020-08-12 | 2022-12-06 | 西安交通大学 | Skin damage surface three-dimensional reconstruction method based on surface structured light |
| CN112215763B (en) * | 2020-09-01 | 2023-02-24 | 青岛秀山移动测量有限公司 | Planar internal missing point cloud repairing method based on image data assistance |
| CN112529891B (en) * | 2020-12-21 | 2024-03-19 | 深圳辰视智能科技有限公司 | Method and device for identifying hollow holes and detecting contours based on point cloud and storage medium |
| CN113610726B (en) * | 2021-08-05 | 2022-04-26 | 华南农业大学 | Improved cubic B-spline interpolation repairing method for region limitation |
| CN114092469B (en) * | 2021-12-02 | 2022-08-26 | 四川大学 | Method and device for determining repair area of blade and readable storage medium |
| CN114581328B (en) * | 2022-03-03 | 2024-07-26 | 中冶赛迪信息技术(重庆)有限公司 | Material pile point cloud data interpolation method, device, equipment and medium |
| CN118037601B (en) * | 2024-04-07 | 2024-07-19 | 法奥意威(苏州)机器人系统有限公司 | Point cloud filling method and electronic equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102663823A (en) * | 2012-04-01 | 2012-09-12 | 北京农业信息技术研究中心 | Plant organ mesh surface hole filling method and system thereof |
| CN104361632A (en) * | 2014-11-03 | 2015-02-18 | 北京航空航天大学 | Triangular mesh hole-filling method based on Hermite radial basis function |
| CN104504663A (en) * | 2014-12-29 | 2015-04-08 | 佛山市诺威科技有限公司 | Iterative repair method for holes of false tooth triangular mesh model |
| CN106097436A (en) * | 2016-06-12 | 2016-11-09 | 广西大学 | A kind of three-dimensional rebuilding method of large scene object |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9754165B2 (en) * | 2013-07-30 | 2017-09-05 | The Boeing Company | Automated graph local constellation (GLC) method of correspondence search for registration of 2-D and 3-D data |
-
2017
- 2017-07-31 CN CN201710637964.0A patent/CN107590825B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102663823A (en) * | 2012-04-01 | 2012-09-12 | 北京农业信息技术研究中心 | Plant organ mesh surface hole filling method and system thereof |
| CN104361632A (en) * | 2014-11-03 | 2015-02-18 | 北京航空航天大学 | Triangular mesh hole-filling method based on Hermite radial basis function |
| CN104504663A (en) * | 2014-12-29 | 2015-04-08 | 佛山市诺威科技有限公司 | Iterative repair method for holes of false tooth triangular mesh model |
| CN106097436A (en) * | 2016-06-12 | 2016-11-09 | 广西大学 | A kind of three-dimensional rebuilding method of large scene object |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107590825A (en) | 2018-01-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107590825B (en) | Point cloud hole repairing method based on SFM | |
| CN110443836B (en) | Point cloud data automatic registration method and device based on plane features | |
| CN111629193B (en) | Live-action three-dimensional reconstruction method and system | |
| CN106709947B (en) | Three-dimensional human body rapid modeling system based on RGBD camera | |
| CN107123164B (en) | Three-dimensional reconstruction method and system for keeping sharp features | |
| Xie et al. | A high-accuracy method for fine registration of overlapping point clouds | |
| CN105300316B (en) | Optical losses rapid extracting method based on grey scale centre of gravity method | |
| CN109118574A (en) | A kind of fast reverse modeling method extracted based on three-dimensional feature | |
| JP5963353B2 (en) | Optical data processing apparatus, optical data processing system, optical data processing method, and optical data processing program | |
| CN112489212A (en) | Intelligent three-dimensional mapping method for building based on multi-source remote sensing data | |
| CN108335352B (en) | Texture mapping method for multi-view large-scale three-dimensional reconstruction scene | |
| CN112927360A (en) | Three-dimensional modeling method and system based on fusion of tilt model and laser point cloud data | |
| CN110363858A (en) | A kind of three-dimensional facial reconstruction method and system | |
| JP5955028B2 (en) | Image processing apparatus, image processing method, and image processing program | |
| CN109754459B (en) | Method and system for constructing human body three-dimensional model | |
| KR20180125379A (en) | Monitoring object shape and deviation from design | |
| CN107862735B (en) | RGBD three-dimensional scene reconstruction method based on structural information | |
| CN107767456A (en) | A kind of object dimensional method for reconstructing based on RGB D cameras | |
| CN106504331A (en) | Tooth modeling method based on three-dimensional model search | |
| CN108182722B (en) | Real projective image generation method for three-dimensional object edge optimization | |
| CN111179321B (en) | Point cloud registration method based on template matching | |
| CN104063860A (en) | Method for refining edge of laser-point cloud | |
| CN104748683A (en) | Device and method for online and automatic measuring numerical control machine tool workpieces | |
| CN109493372A (en) | The product point cloud data Fast global optimization method for registering of big data quantity, few feature | |
| CN114170284B (en) | Multi-view point cloud registration method based on active landmark point projection assistance |
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 |