A gap measurement method and system based on point cloud data

Abstract

The invention relates to the field of industrial measurement technology and discloses a gap measurement method and system based on point cloud data. The method comprises the following steps: performing local curvature mutation estimation on acquired pre-processed point cloud data to obtain a curvature point cloud feature set; performing boundary feature estimation on the pre-processed point cloud data to obtain a boundary point cloud feature set; then fusing the curvature point cloud feature set with the boundary point cloud feature set to obtain gap feature point cloud data; projecting the gap feature point cloud data onto a projection coordinate system; performing linear fitting on the gap feature point cloud data projected onto the projection coordinate system to obtain a fitting line, and calculating the distance from the gap feature point cloud data to the fitting line to obtain step difference data; then calculating gap data using the neighbor point distance and the step difference data, and using the step difference data and the gap data as measurement data of a target gap; and solving the problems of unclear features and poor robustness in existing gap measurement methods.

Description

Technical Field

[0001] The present invention relates to the field of industrial measurement technology, and in particular to a gap measurement method and system based on point cloud data. Background Art

[0002] In the field of industrial measurement, especially in component assembly quality control, measuring the gaps between components is a key step in ensuring product quality. Currently, the gaps between components are often measured by obtaining component point cloud data. Most gap measurement methods based on point cloud data rely on higher-quality data. The higher the quality of the acquired point cloud data, the more accurate the data obtained by the measurement method. However, in the actual image acquisition process, it is inevitable that the point cloud data will be incomplete and the boundary points of the gaps will be missing. Although the missing data can be extracted through feature extraction to improve the data quality, there are often limitations when processing gap data with incomplete boundaries, the algorithm has poor robustness, the features are not obvious, or even the features fail. Therefore, the existing gap measurement methods have the problems of unclear features and poor robustness. Summary of the Invention

[0003] The present invention provides a gap measurement method and system based on point cloud data to solve the problems of unclear features and poor robustness in existing gap measurement methods.

[0004] In order to achieve the above object, the present invention is implemented through the following technical solutions:

[0005] In a first aspect, the present invention provides a gap measurement method based on point cloud data, comprising:

[0006] Acquire original point cloud data of components on both sides of the target gap, and preprocess the original point cloud data of components on both sides of the target gap to obtain preprocessed point cloud data of components on both sides of the target gap;

[0007] The local curvature mutation is estimated on the pre-processed point cloud data to obtain the curvature point cloud feature set, and the boundary feature is estimated on the pre-processed point cloud data to obtain the boundary point cloud feature set;

[0008] The curvature point cloud feature set and the boundary point cloud feature set are fused to obtain the seam feature point cloud data;

[0009] The neighboring point search of the seam feature point cloud data is performed to obtain the neighboring point distance and the neighboring point direction vector, and the RANSAC algorithm is used to fit the plane where the seam feature point cloud data is located to obtain the plane normal vector;

[0010] Construct a projection coordinate system based on the direction vector of the neighboring points and the plane normal vector, and project the seam feature point cloud data onto the projection coordinate system;

[0011] Performing straight line fitting on the seam feature point cloud data projected on the projection coordinate system to obtain a fitting line, and calculating the distance from the seam feature point cloud data to the fitting line to obtain step difference data;

[0012] The gap data is calculated using the neighboring point distance and step difference data, and the step difference data and gap data are used as the measurement data of the target gap.

[0013] Optionally, preprocessing includes:

[0014] Point cloud data denoising and downsampling processing.

[0015] Optionally, local curvature mutation estimation is performed on the pre-processed point cloud data to obtain a curvature point cloud feature set, including:

[0016] Determine two neighboring points in the preprocessed point cloud data, and determine a first normal vector of a first point of the two neighboring points and a second normal vector of a second point of the two neighboring points;

[0017] Construct an osculating circle about the two neighboring points based on the first normal vector and the second normal vector, and determine the central angle of the two neighboring points and the angle between the line segment connecting the two neighboring points and the radius of the osculating circle in the osculating circle;

[0018] The curvature of the two neighboring points is calculated based on the central angle of the two neighboring points in the osculating circle and the angle between the line segment connecting the two neighboring points and the radius of the osculating circle. The calculation of the curvature satisfies the following relationship:

[0019]

[0020] Where k represents the curvature, β represents the central angle of two neighboring points in the osculating circle, α represents the angle between the line segment connecting the two neighboring points and the radius of the osculating circle, |pq i | represents the vector of neighboring points.

[0021] Determine the local curvature threshold, and when the curvature is greater than the local curvature threshold, use the two neighboring points as curvature point cloud features; when the curvature is less than or equal to the local curvature threshold, discard the two neighboring points;

[0022] After all the curvatures of the points in the preprocessed point cloud data are calculated, all the curvature point cloud features are used as a curvature point cloud feature set.

[0023] Optionally, boundary feature estimation is performed on the pre-processed point cloud data to obtain a boundary point cloud feature set, including:

[0024] Determine the neighborhood of the point in the preprocessed point cloud data, and project the point in the preprocessed point cloud data onto the tangent plane of its neighborhood to obtain the projection point;

[0025] Determine the projected neighboring points of the midpoint of the preprocessed point cloud data among the projected points, and calculate the angle between the lines connecting two consecutive projected neighboring points and the midpoint of the preprocessed point cloud data in a clockwise or counterclockwise direction with the points in the preprocessed point cloud data as corner points;

[0026] Determine a boundary angle threshold, and when the angle is greater than the boundary angle threshold, use the projected neighboring points corresponding to the angle between the point in the preprocessed point cloud data and the point in the preprocessed point cloud data as boundary point cloud features; when the angle is less than the boundary angle threshold, discard the projected neighboring points corresponding to the angle between the point in the preprocessed point cloud data and the point in the preprocessed point cloud data;

[0027] After all the angles between the points in the preprocessed point cloud data and their corresponding projected neighboring points are calculated, all the boundary point cloud features are used as the boundary point cloud feature set.

[0028] Optionally, the curvature point cloud feature set and the boundary point cloud feature set are fused to obtain seam feature point cloud data, including:

[0029] The curvature point cloud feature set and the boundary point cloud feature set are added together to obtain the total feature set, and the clipping area is determined based on the size of the target gap.

[0030] The total feature set is clipped through the clipping area to obtain the seam feature point cloud data.

[0031] Optionally, neighbor point search is performed on the seam feature point cloud data to obtain the neighbor point distance and neighbor point direction vector, including:

[0032] Taking the target gap as a reference, the pair of seam feature point cloud data is divided into a first pair of seam feature point cloud data located on one side of the target gap and a second pair of seam feature point cloud data located on the other side of the target gap;

[0033] Searching for neighboring points of the first pair of seam feature point cloud data in the second pair of seam feature point cloud data, and taking the distance between a point in the first pair of seam feature point cloud data and its corresponding neighboring point in the second pair of seam feature point cloud data as the neighboring point distance;

[0034] The direction vector between the point in the first pair of seam feature point cloud data and the corresponding neighboring point in the second pair of seam feature point cloud data is used as the neighboring point direction vector.

[0035] Optionally, a projected coordinate system is constructed based on the direction vectors of the neighboring points and the plane normal vector, including:

[0036] The projection normal vector of the projection plane is calculated based on the direction vector of the neighboring point and the plane normal vector. The calculation formula satisfies the following relationship:

[0037] f p =e i ×f n ;

[0038] Where, f p represents the projection normal vector, e i Represents the direction vector of the neighboring point, f n represents the plane normal vector;

[0039] Construct a projected coordinate system with the projection normal vector as the coordinate axis.

[0040] Optionally, gap data is calculated using neighbor point distances and step difference data, including:

[0041] According to the distance between the nearest points and the step difference data, a right triangle of the step difference data and the gap data is constructed, and the side length of the right triangle is calculated to obtain the gap data. The calculation formula satisfies the following relationship:

[0042]

[0043] Where d represents gap data, l represents the distance between neighboring points, and h represents the step difference data.

[0044] In a second aspect, an embodiment of the present application provides a gap measurement system based on point cloud data, characterized in that it includes a processor and a memory;

[0045] Memory for storing computer programs;

[0046] The processor is configured to implement any one of the method steps described in the first aspect when executing a program stored in the memory.

[0047] Beneficial effects:

[0048] The present invention provides a gap measurement method based on point cloud data, which obtains a curvature point cloud feature set by estimating the local curvature mutation of the acquired pre-processed point cloud data, and obtains a boundary point cloud feature set by estimating the boundary features of the pre-processed point cloud data; then fuses the curvature point cloud feature set with the boundary point cloud feature set to obtain the gap feature point cloud data; and projects the gap feature point cloud data onto a projection coordinate system; performs linear fitting on the gap feature point cloud data projected onto the projection coordinate system to obtain a fitting line, and calculates the distance from the gap feature point cloud data to the fitting line to obtain step difference data; and then can be used The gap data is calculated based on the distance between neighboring points and the step difference data, and the step difference data and the gap data are used as the measurement data of the target gap. Compared with other algorithms, this method obtains the curvature point cloud feature set by performing local curvature mutation estimation on the preprocessed point cloud data, and obtains the boundary point cloud feature set by performing boundary feature estimation on the preprocessed point cloud data. In the case of incomplete boundaries, the local curvature mutation estimation and boundary feature estimation can be used to extract more dimensional information from the point cloud data for feature extraction. In addition, this method is compatible with the measurement of both complete boundary data and incomplete boundary data, and has better robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 Flowchart of a gap measurement method based on point cloud data according to a preferred embodiment of the present invention;

[0050] Figure 2 Flowchart of a method for acquiring seam feature point cloud data according to a preferred embodiment of the present invention;

[0051] Figure 3 Flowchart of a gap step difference measurement method according to a preferred embodiment of the present invention;

[0052] Figure 4 Schematic diagram of a method for estimating a local curvature mutation according to a preferred embodiment of the present invention;

[0053] Figure 5 Schematic diagram of a boundary feature estimation method according to a preferred embodiment of the present invention;

[0054] Figure 6 A schematic projection plane diagram of a preferred embodiment of the present invention;

[0055] Figure 7 A schematic diagram of gap step difference measurement according to a preferred embodiment of the present invention;

[0056] Figure 8 This is a seam projection diagram of Pm150000 in a preferred embodiment of the present invention;

[0057] Figure 9 This is a seam projection diagram of Pm150002 in a preferred embodiment of the present invention;

[0058] Figure 10 This is the seam projection diagram of Pm151002 in the preferred embodiment of the present invention. DETAILED DESCRIPTION

[0059] The following is a clear and complete description of the technical solutions of the present invention. It should be understood that the embodiments described are only a portion of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are intended to fall within the scope of protection of the present invention.

[0060] Unless otherwise defined, the technical or scientific terms used in the present invention shall have the usual meanings understood by persons of ordinary skill in the field to which the present invention belongs. The words "first", "second" and similar terms used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, words such as "one" or "a" do not indicate a quantity limitation, but rather indicate the existence of at least one. Words such as "connected" or "connected" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the object being described changes, the relative positional relationship also changes accordingly.

[0061] Example 1:

[0062] See Figure 1-7 , an embodiment of the present application provides a gap measurement method based on point cloud data, comprising:

[0063] Acquire original point cloud data of components on both sides of the target gap, and preprocess the original point cloud data of components on both sides of the target gap to obtain preprocessed point cloud data of components on both sides of the target gap;

[0064] The local curvature mutation is estimated on the pre-processed point cloud data to obtain the curvature point cloud feature set, and the boundary feature is estimated on the pre-processed point cloud data to obtain the boundary point cloud feature set;

[0065] The curvature point cloud feature set and the boundary point cloud feature set are fused to obtain the seam feature point cloud data;

[0066] The neighboring point search of the seam feature point cloud data is performed to obtain the neighboring point distance and the neighboring point direction vector, and the RANSAC algorithm is used to fit the plane where the seam feature point cloud data is located to obtain the plane normal vector.

[0067] Construct a projection coordinate system based on the direction vector of the neighboring points and the plane normal vector, and project the seam feature point cloud data onto the projection coordinate system;

[0068] Performing straight line fitting on the seam feature point cloud data projected on the projection coordinate system to obtain a fitting line, and calculating the distance from the seam feature point cloud data to the fitting line to obtain step difference data;

[0069] The gap data is calculated using the neighboring point distance and step difference data, and the step difference data and gap data are used as the measurement data of the target gap.

[0070] In the above embodiment, a curvature point cloud feature set is obtained by performing local curvature mutation estimation on the acquired pre-processed point cloud data, and a boundary feature estimation is performed on the pre-processed point cloud data to obtain a boundary point cloud feature set; the curvature point cloud feature set and the boundary point cloud feature set are then fused to obtain the seam feature point cloud data; and the seam feature point cloud data is projected on the projection coordinate system; a straight line fitting is performed on the seam feature point cloud data projected on the projection coordinate system to obtain a fitting straight line, and the distance from the seam feature point cloud data to the fitting straight line is calculated to obtain step difference data; the gap data can be calculated using the neighboring point distance and the step difference data, and the step difference data and the gap data are used as the measurement data of the target gap; compared with other algorithms, this method can extract more dimensional information from the point cloud data for feature extraction by using local curvature mutation estimation and boundary feature estimation when the boundary is incomplete, and this method is compatible with the measurement of complete boundary data and incomplete boundary data, and has better robustness.

[0071] Optionally, preprocessing includes:

[0072] Point cloud data denoising and downsampling processing.

[0073] Optionally, local curvature mutation estimation is performed on the pre-processed point cloud data to obtain a curvature point cloud feature set, including:

[0074] Determine two neighboring points in the preprocessed point cloud data, and determine a first normal vector of a first point of the two neighboring points and a second normal vector of a second point of the two neighboring points;

[0075] Construct an osculating circle about the two neighboring points based on the first normal vector and the second normal vector, and determine the central angle of the two neighboring points and the angle between the line segment connecting the two neighboring points and the radius of the osculating circle in the osculating circle;

[0076] The curvature of the two neighboring points is calculated based on the central angle of the two neighboring points in the osculating circle and the angle between the line segment connecting the two neighboring points and the radius of the osculating circle. The calculation of the curvature satisfies the following relationship:

[0077]

[0078] Where k represents the curvature, β represents the central angle of two neighboring points in the osculating circle, α represents the angle between the line segment connecting the two neighboring points and the radius of the osculating circle, |pq i | represents the vector of neighboring points.

[0079] Determine the local curvature threshold, and when the curvature is greater than the local curvature threshold, use the two neighboring points as curvature point cloud features; when the curvature is less than or equal to the local curvature threshold, discard the two neighboring points;

[0080] After all the curvatures of the points in the preprocessed point cloud data are calculated, all the curvature point cloud features are used as a curvature point cloud feature set.

[0081] In the above embodiment, if Figure 4 As shown, the point in the preprocessed point cloud data is regarded as point p, and there are M neighboring points for p, q i is the i-th neighbor point, point p has a normal vector N, point q i The normal vector is M i , we can calculate the points p and q i , establish an osculating circle to estimate the curvature k of point p

[0082]

[0083] Where k represents the curvature, β represents the central angle of two neighboring points in the osculating circle, α represents the angle between the line segment connecting the two neighboring points and the radius of the osculating circle, |pq i | represents the vector of the nearest neighbor points;

[0084] By setting the local curvature threshold K, if the curvature k of the calculated point is greater than the threshold, it is considered to be a seam feature point and added to the seam feature point set.

[0085] Optionally, boundary feature estimation is performed on the pre-processed point cloud data to obtain a boundary point cloud feature set, including:

[0086] Determine the neighborhood of the point in the preprocessed point cloud data, and project the point in the preprocessed point cloud data onto the tangent plane of its neighborhood to obtain the projection point;

[0087] Determine the projected neighboring points of the midpoint of the preprocessed point cloud data among the projected points, and calculate the angle between the lines connecting two consecutive projected neighboring points and the midpoint of the preprocessed point cloud data in a clockwise or counterclockwise direction with the points in the preprocessed point cloud data as corner points;

[0088] Determine a boundary angle threshold, and when the angle is greater than the boundary angle threshold, use the projected neighboring points corresponding to the angle between the point in the preprocessed point cloud data and the point in the preprocessed point cloud data as boundary point cloud features; when the angle is less than the boundary angle threshold, discard the projected neighboring points corresponding to the angle between the point in the preprocessed point cloud data and the point in the preprocessed point cloud data;

[0089] After all the angles between the points in the preprocessed point cloud data and their corresponding projected neighboring points are calculated, all the boundary point cloud features are used as the boundary point cloud feature set.

[0090] In the above embodiment, if Figure 5 As shown in the figure, the boundary point features of point cloud are crucial features in understanding and processing 3D data. In this paper, the boundary feature points of the seam are extracted by the angle criterion method.

[0091] First, for point p, select its neighborhood Np Project onto its tangent plane, take point p as the corner point, calculate the angle between the line connecting two consecutive adjacent points and p in a clockwise direction (counterclockwise is also possible), and select the largest angle θ max , it is obvious that the angle of the boundary point is larger than that of the internal point, so the boundary point is extracted.

[0092] Optionally, the curvature point cloud feature set and the boundary point cloud feature set are fused to obtain seam feature point cloud data, including:

[0093] The curvature point cloud feature set and the boundary point cloud feature set are added together to obtain the total feature set, and the clipping area is determined based on the size of the target gap.

[0094] The total feature set is clipped through the clipping area to obtain the seam feature point cloud data.

[0095] Optionally, neighbor point search is performed on the seam feature point cloud data to obtain the neighbor point distance and neighbor point direction vector, including:

[0096] Taking the target gap as a reference, the pair of seam feature point cloud data is divided into a first pair of seam feature point cloud data located on one side of the target gap and a second pair of seam feature point cloud data located on the other side of the target gap;

[0097] Searching for neighboring points of the first pair of seam feature point cloud data in the second pair of seam feature point cloud data, and taking the distance between a point in the first pair of seam feature point cloud data and its corresponding neighboring point in the second pair of seam feature point cloud data as the neighboring point distance;

[0098] The direction vector between the point in the first pair of seam feature point cloud data and the corresponding neighboring point in the second pair of seam feature point cloud data is used as the neighboring point direction vector.

[0099] Optionally, a projected coordinate system is constructed based on the direction vectors of the neighboring points and the plane normal vector, including:

[0100] The projection normal vector of the projection plane is calculated based on the direction vector of the neighboring point and the plane normal vector. The calculation formula satisfies the following relationship:

[0101] f p =e i ×f n ; (2)

[0102] Where, f p represents the projection normal vector, e i Represents the direction vector of the neighboring point, f n represents the plane normal vector;

[0103] Construct a projected coordinate system with the projection normal vector as the coordinate axis.

[0104] In the above embodiment, if Figure 6 As shown, the neighbor point search is performed on the two areas contained in the subinterval, and the distance l between the neighbor point pairs and the direction vector e between the neighbor points are recorded. i ; Use the RANSAC algorithm to fit the two side planes to obtain the plane normal vector f n , and according to the direction vector e between the neighboring points i and the plane normal vector f n , get the normal vector f of the projection plane p =e i ×f n , project the point cloud data of the subinterval onto the projection coordinate system.

[0105] Optionally, gap data is calculated using neighbor point distances and step difference data, including:

[0106] According to the distance between the nearest points and the step difference data, a right triangle of the step difference data and the gap data is constructed, and the side length of the right triangle is calculated to obtain the gap data. The calculation formula satisfies the following relationship:

[0107]

[0108] Where d represents gap data, l represents the distance between neighboring points, and h represents the step difference data.

[0109] In the above embodiment, if Figure 7 As shown in the figure, a straight line is fitted to the projected point cloud to obtain a fitted straight line, the distance from the feature point to the straight line is calculated to obtain the step difference h, and finally the gap is obtained by the neighbor point distance l and the step difference h.

[0110] Example 2:

[0111] See Figure 8-10 In this embodiment, three workpieces with different gaps and step differences are first collected through real acquisition for data testing. For the convenience of recording, in the real acquisition situation, the three workpieces with different gaps and step differences are named as Region 1, Region 2, and Region 3 respectively. Based on the above Region 1, Region 2, and Region 3, data testing under real acquisition is carried out. The results of the real acquisition data test are shown in Table 1.

[0112] Table 1 Real data collection test

[0113]

[0114]

[0115]

[0116] As a comparison with the actual acquisition, three workpiece gap data with different gaps and step differences are simulated by simulation. For the convenience of recording, the first workpiece gap data is named Pm150000. Figure 8 As shown, the gap data of the second workpiece is named Pm150002, and its gap data is as follows Figure 9 As shown, the gap data of the third workpiece is named Pm151002, and its gap data is as follows Figure 10 As shown, a simulation data test was performed based on the above simulated gap data, and the results of the simulation data test are shown in Table 2.

[0117] Table 2 Simulation data test

[0118]

[0119]

[0120] It can be seen from the data in Tables 1 and 2 that this method has a good measurement effect for both the real data of workpiece gaps obtained through real camera sampling and the simulated data constructed by simulating real workpiece gaps. Considering that the real collected data may have incomplete boundary data in point cloud data collection compared to the simulated data, it is not difficult to see from the data in Table 1 that this method can extract more dimensional information from the point cloud data for feature extraction by using local curvature mutation estimation and boundary feature estimation in the case of incomplete boundaries, thereby reducing the measurement error of the data. In addition, this method is compatible with the measurement of both complete boundary data and incomplete boundary data, and has better robustness.

[0121] The embodiment of the present application also provides a gap measurement system based on point cloud data, including a processor and a memory;

[0122] Memory for storing computer programs;

[0123] The processor is configured to implement any one of the method steps described in the gap measurement method based on point cloud data when executing the program stored in the memory.

[0124] The above-mentioned gap measurement system based on point cloud data can implement various embodiments of the above-mentioned gap measurement method based on point cloud data and achieve the same beneficial effects. Here, no further details are given.

[0125] The above describes in detail the preferred embodiments of the present invention. It should be understood that those skilled in the art can make numerous modifications and variations based on the concepts of the present invention without inventive effort. Therefore, any technical solutions that can be derived by those skilled in the art through logical analysis, reasoning, or limited experimentation based on the concepts of the present invention and the prior art should be within the scope of protection defined by the claims.

Claims

1. A gap measurement method based on point cloud data, characterized in that: include: Acquire original point cloud data of components on both sides of the target gap, and preprocess the original point cloud data of components on both sides of the target gap to obtain preprocessed point cloud data of components on both sides of the target gap; Performing local curvature mutation estimation on the pre-processed point cloud data to obtain a curvature point cloud feature set, and performing boundary feature estimation on the pre-processed point cloud data to obtain a boundary point cloud feature set; Fusing the curvature point cloud feature set with the boundary point cloud feature set to obtain seam feature point cloud data; Performing a neighbor point search on the seam feature point cloud data to obtain a neighbor point distance and a neighbor point direction vector, and using a RANSAC algorithm to fit the plane where the seam feature point cloud data is located to obtain a plane normal vector; Constructing a projection coordinate system according to the neighboring point direction vector and the plane normal vector, and projecting the seam feature point cloud data onto the projection coordinate system; Performing straight line fitting on the seam feature point cloud data projected on the projection coordinate system to obtain a fitting straight line, and calculating the distance from the seam feature point cloud data to the fitting straight line to obtain step difference data; Gap data is calculated using the neighbor point distance and the step difference data, and the step difference data and the gap data are used as measurement data of the target gap.

2. The gap measurement method based on point cloud data according to claim 1, characterized in that: The step of performing local curvature mutation estimation on the pre-processed point cloud data to obtain a curvature point cloud feature set includes: Determining two neighboring points in the preprocessed point cloud data, and determining a first normal vector of a first point of the two neighboring points and a second normal vector of a second point of the two neighboring points; Constructing an osculating circle about two neighboring points based on the first normal vector and the second normal vector, and determining, in the osculating circle, a central angle of the two neighboring points and an angle between a line segment connecting the two neighboring points and a radius of the osculating circle; The curvature of the two neighboring points is calculated based on the central angle of the two neighboring points in the osculating circle and the angle between the line segment connecting the two neighboring points and the radius of the osculating circle. The calculation of the curvature satisfies the following relationship: Where k represents the curvature, β represents the central angle of two neighboring points in the osculating circle, α represents the angle between the line segment connecting the two neighboring points and the radius of the osculating circle, |pq i | represents the vector of the nearest neighbor points; Determine a local curvature threshold, and when the curvature is greater than the local curvature threshold, use two neighboring points as curvature point cloud features; when the curvature is less than or equal to the local curvature threshold, discard the two neighboring points; After all curvatures of points in the pre-processed point cloud data are calculated, all curvature point cloud features are taken as a curvature point cloud feature set.

3. The gap measurement method based on point cloud data according to claim 1, characterized in that: The step of estimating boundary features of the pre-processed point cloud data to obtain a boundary point cloud feature set includes: Determine a neighborhood of a point in the pre-processed point cloud data, and project the point in the pre-processed point cloud data onto a tangent plane of the neighborhood to obtain a projection point; Determine the projected neighboring points of the midpoint of the pre-processed point cloud data among the projected points, and calculate the angle between the line connecting two consecutive projected neighboring points and the midpoint of the pre-processed point cloud data in a clockwise or counterclockwise direction with the point in the pre-processed point cloud data as a corner point; Determine a boundary angle threshold, and when the angle is greater than the boundary angle threshold, use the projected neighboring points corresponding to the angle between the point in the preprocessed point cloud data and the point in the preprocessed point cloud data as boundary point cloud features; when the angle is less than the boundary angle threshold, discard the projected neighboring points corresponding to the angle between the point in the preprocessed point cloud data and the point in the preprocessed point cloud data; After all angles between points in the pre-processed point cloud data and their corresponding projected neighboring points are calculated, all boundary point cloud features are used as a boundary point cloud feature set.

4. The gap measurement method based on point cloud data according to claim 1, characterized in that: The step of fusing the curvature point cloud feature set with the boundary point cloud feature set to obtain seam feature point cloud data includes: Adding the curvature point cloud feature set and the boundary point cloud feature set to obtain a total feature set, and determining a clipping area based on the size of the target gap; The total feature set is clipped through the clipping area to obtain the seam feature point cloud data.

5. The gap measurement method based on point cloud data according to claim 1, characterized in that: The performing of neighbor point search on the seam feature point cloud data to obtain neighbor point distances and neighbor point direction vectors includes: Dividing the pair of seam feature point cloud data into first pair of seam feature point cloud data located on one side of the target seam and second pair of seam feature point cloud data located on the other side of the target seam with reference to the target seam; Searching for neighboring points of the first pair of seam feature point cloud data in the second pair of seam feature point cloud data, and taking the distance between a point in the first pair of seam feature point cloud data and its corresponding neighboring point in the second pair of seam feature point cloud data as a neighboring point distance; A direction vector between a point in the first pair of seam feature point cloud data and its corresponding neighboring point in the second pair of seam feature point cloud data is used as a neighboring point direction vector.

6. The gap measurement method based on point cloud data according to claim 1, characterized in that: The constructing of a projected coordinate system according to the neighboring point direction vector and the plane normal vector includes: The projection normal vector of the projection plane is calculated based on the direction vector of the neighboring point and the plane normal vector, and the calculation formula satisfies the following relationship: f p =e i ×f n ; Where, f p represents the projection normal vector, e i Represents the direction vector of the neighboring point, f n represents the plane normal vector; Construct a projected coordinate system with the projection normal vector as the coordinate axis.

7. The gap measurement method based on point cloud data according to claim 1, characterized in that: The calculating gap data by using the neighbor point distance and the step difference data includes: A right triangle of the step difference data and the gap data is constructed based on the neighbor point distance and the step difference data, and the gap data is obtained by calculating the side length of the right triangle. The calculation formula satisfies the following relationship: Where d represents gap data, l represents the distance between neighboring points, and h represents the step difference data.

8. The gap measurement method based on point cloud data according to claim 1, characterized in that: The pretreatment includes: Point cloud data denoising and downsampling processing.

9. A gap measurement system based on point cloud data, characterized in that: Including processor and memory; Memory for storing computer programs; A processor, configured to implement the method steps described in any one of claims 1 to 8 when executing a program stored in a memory.

Images