Estimation method of normal projection from point to surface based on triangular mesh point cloud
By performing PCA principal component analysis and spherical small neighborhood iterative calculation on triangular meshed point clouds, the problem that the existing technology cannot directly use point clouds for normal projection on point to surface is solved, and the point cloud-based method is realized Estimation toward the projection improves the computational efficiency.
Patent Information
- Application Number
- CN202210426965.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-04-22
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2042-04-22
AI Technical Summary
Existing algorithms cannot directly use point clouds to estimate normal projections on a point-to-surface, and rely on a given surface equation.
The point-to-surface normal projection estimation method based on triangular meshed point clouds is used to determine the point cloud surface normal vector through PCA principal component analysis, and the normal projection point is iteratively calculated using the spherical small neighborhood and the intersection of straight lines.
It realizes the estimation of point-to-surface normal projection directly using point cloud data, avoids dependence on specific surface equations, and improves calculation efficiency.
Smart Images

Figure CN114693886B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of point cloud processing in three-dimensional space, and in particular to a method for estimating normal projection of a point onto a curved surface based on a triangular meshed point cloud. Background Art
[0002] Normal projection of a point onto a free-form surface is one of the most important research issues in three-dimensional space. It has wide applications in problems such as intersection of curves and surfaces, point cloud registration, and point cloud fusion.
[0003] See Tsinghua University's June 2015 dissertation, author Song Haichuan, titled "Research on the Normal Projection Problem from Points to Free Curves and Surfaces". This paper proposes a geometric iteration algorithm for finding the normal projection from a point to a free surface based on double arc interval approximation. On the basis of finding the projection interval using a global estimation algorithm, a double arc spline is used to fit the base surface to obtain an estimated projection point that is closer to the exact projection point. When the projection point meets the user-specified tolerance, the iteration ends.
[0004] The disadvantage of this algorithm is that it relies on the given surface equation to estimate the normal projection point of the point onto the surface, and cannot directly use the point cloud to estimate the normal projection point of the point onto the surface. Summary of the invention
[0005] In order to solve the problem that the existing algorithm cannot directly use point cloud to estimate the normal projection of a point onto a surface, the present invention proposes a method for estimating the normal projection of a point onto a surface based on triangular meshed point cloud.
[0006] The technical solution of the present invention is as follows:
[0007] A method for estimating a normal projection of a point onto a surface based on a triangular meshed point cloud, the method comprising the following steps:
[0008] Step 1: Randomly select a point on the triangular mesh point cloud A as the starting projection point
[0009] Suppose the triangle mesh point cloud of any surface is A, the standard deviation statistical characteristic of A is σ; establish a three-dimensional coordinate system {S} on A; the point to be projected P is a point outside A, and the projection point sequence of P on A is The three-dimensional coordinates in the {S} coordinate system are Randomly select a point on A as the starting projection point in the projection point sequence
[0010] Step 2: δ i PCA principal component analysis is performed on the point cloud data to determine the point cloud surface The normal vector at ;
[0011] First, with r i =3σ is the radius, and the projection point is obtained from formula (1) The spherical small neighborhood δ i ,
[0012]
[0013] Then, for δ i Perform PCA principal component analysis on the midpoint cloud data and select the minimum principal component direction ξ i As a point cloud surface The direction of the normal vector at ;
[0014] Step 3: Determine the subsequent projection points on the triangular meshed point cloud A
[0015] First, connect the point to be projected P and the projection point Get line segment l i , in l i Find the replacement point of the point to be projected on Satisfy formula (2);
[0016]
[0017] in is the initial scale parameter;
[0018] Then, with ξ i Direction and point Determine a straight line If the straight line There is no intersection with A, then execute step 4; if the straight line There is an intersection with A, and this intersection is recorded as the projection point The three-dimensional coordinates in the {S} coordinate system are Jump to step 5 to execute;
[0019] Step 4: Adjust parameters Find the subsequent projection points on the triangular mesh point cloud A
[0020] Pick is the scale parameter; in l i Find the replacement point of the point to be projected on Satisfy formula (3);
[0021]
[0022] Then, with ξ i Direction and point Determine a straight line If the straight line There is no intersection with A, then repeat step 4; if the straight line There is an intersection with A, and this intersection is recorded as the projection point The three-dimensional coordinates in the {S} coordinate system are Go to step 5.
[0023] Step 5: Calculate the current projection point With the previous projection point The Euclidean distance between
[0024] Current projection point With the previous projection point The Euclidean distance d i Determined using formula (4);
[0025]
[0026] like Then for the projection point Repeat steps 2 and 3;
[0027] like The iteration ends and the projection point It is the normal projection of the point P to be projected on this point cloud surface.
[0028] Beneficial effects of the present invention:
[0029] The solution of the present invention directly uses existing point cloud data to perform PCA principal component analysis when calculating the normal vector of the point cloud surface, does not rely on specific surface equations, and realizes the estimation of the normal projection of the point onto the surface based on the triangular mesh point cloud. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] Figure 1 : Flowchart of a method for estimating normal projections from point to point cloud surfaces based on triangular meshed point clouds.
[0031] Figure 2 : Schematic diagram of a method for estimating normal projections from point to point cloud surfaces based on triangular meshed point clouds. DETAILED DESCRIPTION
[0032] The present invention is further described in detail below in conjunction with the embodiments and drawings.
[0033] The present invention can directly use existing point cloud data without relying on specific surface equations, and realizes the estimation of normal projection of points onto surfaces based on triangular mesh point clouds. Figure 1 shown.
[0034] Embodiment 1:
[0035] The method for estimating the normal projection of a point-to-point cloud surface based on a triangular meshed point cloud comprises the following steps:
[0036] Step 1: Randomly select a point on the triangular mesh point cloud A as the starting projection point
[0037] like Figure 2 As shown, the standard deviation statistical characteristic of the triangular mesh point cloud A is σ = 2 mm. A three-dimensional coordinate system {S} is established on A. The point to be projected P (1676.91, -290.193, 952.94) is a point outside A. The projection point sequence of P on A is The three-dimensional coordinates in the {S} coordinate system are Randomly select a point on A as the starting projection point in the projection point sequence
[0038] Step 2: δ i PCA principal component analysis is performed on the point cloud data to determine the point cloud surface The normal vector at .
[0039] First, taking r0 = 6 mm as the radius, the projection point is obtained from formula (1): A small spherical neighborhood δ0 at;
[0040]
[0041] Then, PCA principal component analysis is performed on the point cloud data in δ0, and the minimum principal component direction ξ0 = [0.982848, -0.501762, -0.852578] is selected. T As a point cloud surface The direction of the normal vector at .
[0042] Step 3: Determine the subsequent projection points on the triangular meshed point cloud A
[0043] First, connect the point to be projected P and the projection point Get line segment l0, take Find an alternative point on l0 to be projected make Satisfies formula (2).
[0044]
[0045] Then, the direction of ξ0 is Determine a straight line straight line There is no intersection with A. Go to step 4.
[0046] Step 4: Adjust parameters Find the subsequent projection points on the triangular mesh point cloud A
[0047] Pick Find an alternative point on l0 to be projected Satisfies formula (3).
[0048]
[0049] Then, the direction of ξ0 is Determine a straight line If the straight line The intersection point with A is The three-dimensional coordinates in the {S} coordinate system are Continue to step 5.
[0050] Step 5: Calculate the current projection point With the previous projection point The Euclidean distance between .
[0051] Use formula (4) to find the current projection point With the previous projection point The Euclidean distance d0
[0052]
[0053] at this time Then for the projection point Repeat steps 2 and 3. The iterative data is shown in Table 1:
[0054] Table 1
[0055]
[0056] Projection Point With the previous projection point Euclidean distance
[0057]
[0058] at this time The iteration ends. The projection point is considered It is the normal projection of the point P to be projected on this point cloud surface.
[0059] Embodiment 2:
[0060] Step 1: Randomly select a point on the triangular mesh point cloud A as the starting projection point
[0061] like Figure 2As shown, the standard deviation statistical characteristic of the triangular mesh point cloud A is σ = 2 mm. A three-dimensional coordinate system {S} is established on A. The point to be projected P (1676.91, -290.193, 952.94) is a point outside A. The sequence of projection points of P on A is The three-dimensional coordinates in the {S} coordinate system are Randomly select a point on A as the starting projection point in the projection point sequence
[0062] Step 2: δ i PCA principal component analysis is performed on the point cloud data to determine the point cloud surface The normal vector at .
[0063] First, taking r0 = 6 mm as the radius, the projection point is obtained from formula (1): A small spherical neighborhood
[0064]
[0065] Then, PCA principal component analysis is performed on the point cloud data in δ0, and the minimum principal component direction ξ0 = [0.987455, -0.63331, -0.625665] is selected. T As a point cloud surface The direction of the normal vector at .
[0066] Step 3: Determine the subsequent projection points on the triangular meshed point cloud A
[0067] First, connect the point to be projected P and the projection point Get line segment l0, take Find an alternative point on l0 to be projected Satisfies formula (2).
[0068]
[0069] Then, the direction of ξ0 is Determine a straight line straight line The intersection point with A is the projection point The three-dimensional coordinates in the {S} coordinate system are If the intersection point exists, skip to step 5.
[0070] Step 5: Calculate the current projection point With the previous projection point The Euclidean distance between .
[0071] Use formula (4) to find the current projection point With the previous projection point The Euclidean distance d0
[0072]
[0073] at this time Then for the projection point Repeat steps 2 and 3. The iterative data is shown in Table 1:
[0074] Table 1
[0075]
[0076]
[0077] Projection Point With the previous projection point Euclidean distance
[0078]
[0079] at this time The iteration ends. The projection point is considered It is the normal projection of the point P to be projected on this point cloud surface.
[0080] So far, the embodiments based on the solution of the present invention do not rely on specific surface equations, and realize the estimation of the normal projection of points onto the surface based on the triangular mesh point cloud.
Claims
1. A method for estimating normal projections from points to surfaces based on triangular meshed point clouds, characterized in that: The method comprises the following steps: Step 1: Randomly select a point on the triangular mesh point cloud A as the starting projection point Suppose the triangle mesh point cloud of any surface is A, the standard deviation statistical characteristic of A is σ; establish a three-dimensional coordinate system {S} on A; the point to be projected P is a point outside A, and the projection point sequence of P on A is The three-dimensional coordinates in the {S} coordinate system are Randomly select a point on A as the starting projection point in the projection point sequence Step 2: δ i PCA principal component analysis is performed on the point cloud data to determine the point cloud surface The normal vector at ; First, with r i =3σ is the radius, and the projection point is obtained from formula (1) The spherical small neighborhood δ i , Then, for δ i Perform PCA principal component analysis on the midpoint cloud data and select the minimum principal component direction ξ i As a point cloud surface The direction of the normal vector at ; Step 3: Determine the subsequent projection points on the triangular meshed point cloud A First, connect the point to be projected P and the projection point Get line segment l i , in l i Find the replacement point of the point to be projected on Satisfy formula (2); in is the initial scale parameter; Then, with ξ i Direction and point Determine a straight line If the straight line There is no intersection with A, then execute step 4; if the straight line There is an intersection with A, and this intersection is recorded as the projection point The three-dimensional coordinates in the {S} coordinate system are Jump to step 5 to execute; Step 4: Adjust parameters Find the subsequent projection points on the triangular mesh point cloud A Pick is the scale parameter; in l i Find the replacement point of the point to be projected on Satisfy formula (3); Then, with ξ i Direction and point Determine a straight line If the straight line There is no intersection with A, then repeat step 4; if the straight line There is an intersection with A, and this intersection is recorded as the projection point The three-dimensional coordinates in the {S} coordinate system are Go to step 5. Step 5: Calculate the current projection point With the previous projection point The Euclidean distance between Current projection point With the previous projection point The Euclidean distance d i Determined using formula (4); like Then for the projection point Repeat steps 2 and 3; like The iteration ends and the projection point It is the normal projection of the point P to be projected on this point cloud surface.
