A method for extracting duct ridges based on 3D point clouds

Abstract

This invention discloses a method for extracting duct ridges based on 3D point clouds, comprising the following steps: First, input the nominal radius of the duct and the original 3D point cloud of its surface, calculate the surface point normal vectors of the 3D point cloud, spatially rasterize the surface points, and extract the effective raster of the duct region; then, by comparing the number of times the effective raster is penetrated by the surface point normal, obtain the initial point for duct ridge growth and optimize it; finally, grow from the initial point to both sides, perform fitting optimization based on the corresponding surface point set, calculate all ridge points, and order the ridge points to complete the duct ridge extraction. This invention can adapt to incomplete 3D point cloud data of duct surfaces, is not affected by the 3D morphology of the duct, has high accuracy, good versatility, and good stability.

Description

Technical Field

[0001] This invention relates to the field of computer vision technology and discloses a method for extracting duct ridges based on three-dimensional point clouds. Background Technology

[0002] Complex curved conduits are crucial components in aerospace manufacturing, and their quality inspection has become an indispensable part of the manufacturing process. With advancements in manufacturing technology, the installation density of conduits in products and equipment is increasing, and their three-dimensional spatial shapes are becoming more complex. These conditions place higher demands on the surface quality measurement methods for conduits.

[0003] In recent years, computer vision-based methods for measuring catheter surface quality have been widely researched and applied in the industrial field. The core of computer vision-based catheter measurement methods is extracting the three-dimensional centerline of the catheter, i.e., the ridge line. Based on the three-dimensional discrete point coordinates of the ridge line, most of the catheter's key quality parameters can be calculated. Currently, the main existing ridge line extraction method is multi-view stereo vision reconstruction. This method uses the catheter's outline centerline in the image as the catheter axis and employs methods such as epipolar correction, feature point extraction, and matching reconstruction to fit and obtain a three-dimensional ridge line model of the catheter. This method is fast, but it has high requirements for the placement of the catheter, especially for catheters with complex shapes, which can easily lead to matching ambiguities, affecting the measurement results. Furthermore, this method requires that the catheter's edge contour be as complete and unobstructed as possible in the field of view of multiple cameras; otherwise, it will affect the accuracy of the catheter ridge line reconstruction.

[0004] Therefore, how to utilize incomplete catheter surface data, unaffected by the three-dimensional morphology of the catheter, and extract a complete and accurate central ridge line is an urgent problem to be solved in catheter measurement technology. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the technical problem to be solved by this invention is to provide a method for extracting catheter ridges based on three-dimensional point clouds. This method uses a tracking growth method to obtain continuous central ridge points, which can effectively improve the quality inspection level of the catheter surface and help improve the manufacturing quality of the catheter.

[0006] The technical solution adopted by this invention to achieve the above objectives is: a method for extracting duct ridges based on three-dimensional point clouds, comprising the following steps:

[0007] Step 1: Set the nominal radius R of the conduit and the original 3D point cloud P of its surface;

[0008] Step 2: Preprocess the 3D point cloud of the duct curved section to obtain surface points with normal vectors and an effective grid of the duct region;

[0009] Step 3: By comparing the number of times the effective grid is penetrated by the surface point normal, the initial point for ridge growth of the duct is obtained and optimized; the center of the grid with the highest number of penetrations is taken as the initial point for ridge growth.

[0010] Step 4: Track the growth from the initial point to both sides and calculate all the ridge points of the ducts;

[0011] Step 5: Based on the growth sequence of the ridge points on both sides of the initial ridge point, the ridge points are ordered to complete the ridge extraction of the duct.

[0012] The preprocessing of the 3D point cloud of the curved section of the duct to obtain surface points with normal vectors and an effective raster of the duct region includes the following steps:

[0013] a. Calculate the normal vector of each surface point pi in the 3D point cloud of the duct, with a search radius rnl = R / 8;

[0014] b. Spatial rasterization of catheter surface points: Calculate the minimum circumscribed cuboid of catheter surface points, such that each side of the cuboid is parallel to the x, y and z axes respectively. Cut the space containing the cuboid continuously into tiny cubes along the x, y and z axes according to the side length rgrid, i.e., the raster.

[0015] c. Extracting valid grids in the conduit region: Calculate the distance between the center point of each grid and its nearest surface point. If the distance is less than the conduit radius R, the grid is considered a valid grid; otherwise, the grid is discarded. The center of the extracted valid grid is denoted as gn.

[0016] The process of obtaining and optimizing the initial point for ridge growth of the duct by comparing the number of times the effective grid is penetrated by the surface point normal includes the following steps:

[0017] a. Calculate the initial growth point of the ridge line of the duct:

[0018] Calculate the effective grid center g n To the surface point set P n Each point p in i normal vector ν i Distance d of the line ni If d ni If the value is less than ε, it is considered that a valid grid has been penetrated once;

[0019] The center of the effective grid that is penetrated the most times is taken as the initial growth point of the ridge line of the vessel, and denoted as S. k The set of grid penetration points corresponding to the initial growth point is denoted as SP. k ;

[0020] b. Optimize the initial growth point of the ridge line in the duct: Using a random sampling consensus algorithm, perform spatial cylindrical fitting on the surface point set SPk corresponding to the initial growth point Sk of the ridge line to obtain a point Ck through which the central axis of the cylinder passes and the unit direction vector. Projecting the initial growth point Sk onto the central axis of the cylinder yields the optimized initial growth point Sk', which can be calculated as follows:

[0021]

[0022] The calculation of the effective grid center g n To the surface point set P n Each point p in i normal vector ν i Distance d of the line ni If d ni <ε is considered a valid grid cell being penetrated once; including:

[0023] With effective grid center g n Centered on the catheter, define a region of radius R, with a distance p from the catheter surface point that is less than the nominal radius of the catheter. i , forming the surface point set P n , can be represented as:

[0024] P n ={p i |R>‖g n -p i ||,p i ∈P}

[0025] Through surface point set P n Each point p in i and the normal vector ν of that point i To form a straight line from the direction vectors, calculate the effective grid center g. n Distance d to the line ni , is represented as:

[0026] d ni =‖(p i -g n )-(p i -g n )·v i ·v i ||

[0027] Set a penetration threshold ε, ε = r grid / 2; if d ni <ε, is considered a valid grid cell being penetrated once; for each valid grid cell center g n Calculate its corresponding point set P n Number of times it penetrates.

[0028] The process of tracking growth from the initial point outwards to both sides and calculating all the ridge points of the duct includes the following steps:

[0029] a. Taking the initial growth point S of the ridge line as an example k The tangent direction is used to search for candidate points S for the next ridge point. k+1 Determine the point set CP on the catheter surface. k+1 In the set CP k+1 Select infinitesimal cylinders of length Δ to form the set of intersecting points SC. k+1 The micro-short cylinder is S k The tangent to ' is the central axis of the cylinder;

[0030] b. Point set SC k+1 Perform spatial cylindrical fitting and select ridge candidate points S k+1 Projecting onto the cylinder axis yields the optimized ridge point S'. k+1 Compare the tangent vectors at the ridge point. Tangent vector to the previous ridge point The direction of the ridge line is used to determine whether the tangent direction of that ridge point should be adjusted.

[0031] Set the ridge point S' k+1 and tangent vector Substitute 4(a)-4(b) and continue to grow the ridge point downwards;

[0032] c. If the intersection point set SC k+1 If the number of points in the ridge is less than a, update the ridge candidate points S. k+1 The calculation is as follows:

[0033]

[0034] l′ step =5·l step

[0035] Repeat step 5(a) to calculate the updated ridge candidate points S. k+1 Find the corresponding intersection point set and update the point set SC. k+1 If the updated point set SC k+1 If the number of points in the matrix is ​​greater than a, proceed to step 5(b) to calculate and update the ridge point S'. k+1 and tangent vector Otherwise, it is assumed that the ridge point grows to the end of the duct, and growth in that direction stops;

[0036] d. After the ridge point on one side of the initial ridge point has grown, proceed with the growth of the ridge point on the other side; adjust the ridge...

[0037] The growth direction of the line from its initial point can be calculated as follows:

[0038]

[0039] Repeat steps 5(a)-5(c) to complete the ridge point growth on the other side of the ridge initial point.

[0040] The initial growth point S of the ridge line k The tangent direction is used to search for candidate points S for the next ridge point. k+1 Determine the point set CP on the catheter surface. k+1 In the set CP k+1 Select infinitesimal cylinders of length Δ to form the set of intersecting points SC. k+1 ;include:

[0041] The initial growth point S of the ridge line k Its tangent vector is the unit direction vector of the central axis of the cylinder.

[0042] Set growth step size l step (l step =r grid / 2), then the candidate point S of the next ridge point k+1 It can be calculated as:

[0043]

[0044] Selection of candidate points S k+1 The set of CP points on the duct surface with a distance less than 1.1R k+1 , can be represented as:

[0045] CP k+1 ={p i |1.1R>‖S k+1 -p i ||,p i ∈P}

[0046] In the collection CP k+1 The inner length is selected as Δ (Δ = r) grid The short cylinders ( / 2) form the set of intersecting points SC. k+1 , can be represented as:

[0047]

[0048] The pair set SC k+1 Perform spatial cylindrical fitting and select ridge candidate points S k+1 Projecting onto the cylinder axis yields the optimized ridge point S'. k+1 Compare the tangent vectors at the ridge point. Tangent vector to the previous ridge point The direction of the ridge line is used to determine whether the tangent direction at that ridge point should be adjusted; including:

[0049] Set a threshold 'a' for the number of points fitted to the cylinder. If the set of intersecting points SC... k+1 The number of points in the set SC is greater than a, for point set SC k+1 Perform a spatial cylinder fitting to obtain a point C through which the cylinder's axis passes. k+1 and the unit vector of the central axis of the cylinder Candidate ridge point S k+1 Projecting this onto the cylinder axis yields the optimized ridge point S'. k+1 It can be calculated as:

[0050]

[0051] unit vector of the central axis of the cylinder S' is the ridge point k+1 The tangent vectors at the ridge point are compared. Tangent vector to the previous ridge point The direction of the ridge point's tangent vector; if the directions are different, the tangent vector of the ridge point... The calculation is as follows:

[0052]

[0053] The present invention has the following beneficial effects and advantages:

[0054] 1. The present invention provides a method for extracting duct ridges based on three-dimensional point clouds. The method obtains the initial point of ridge growth by statistically analyzing the number of effective grid penetrations, and calculates the ridge of the entire duct using a tracking growth method. All three-dimensional data of the duct surface are involved in the calculation, which improves the extraction accuracy of the ridges.

[0055] 2. The duct ridge extraction method based on three-dimensional point cloud described in this invention can automatically change the ridge growth stopping condition according to the duct diameter parameters, adapt to incomplete three-dimensional point cloud data of the duct surface, and has strong algorithm robustness and high computational efficiency.

[0056] 3. The method of the present invention is based on the three-dimensional point cloud data of the catheter surface, without considering the placement and three-dimensional shape of the catheter. It has good versatility and can be applied to the quality inspection unit of the catheter digital production system, which helps to improve the production efficiency and manufacturing quality of catheters. Attached Figure Description

[0057] Figure 1 This is a flowchart of the method of the present invention;

[0058] Figure 2 This is a schematic diagram of the effective grid principle for extracting the duct region in the method of the present invention;

[0059] Figure 3 This is a schematic diagram of the duct intersection contour in the method of the present invention;

[0060] Figure 4 This is a schematic diagram illustrating the principle of selecting candidate initial growth points for ducts in the method of this invention.

[0061] Figure 5 This is a flowchart illustrating the duct ridge growth process in the method of the present invention.

[0062] Figure 6 This is a schematic diagram illustrating the principle of duct ridge line tracking growth in the method of this invention;

[0063] Figure 7 (a) is the original three-dimensional point cloud image of the catheter surface in the method of the present invention;

[0064] Figure 7 (b) is a diagram showing the duct ridge line extraction result in the method of the present invention; Detailed Implementation

[0065] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0066] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0067] like Figures 1-6 As shown, the method for extracting duct ridges based on three-dimensional point clouds according to the present invention includes the following steps:

[0068] Step 1: Set the nominal radius R of the conduit and the original 3D point cloud P of its surface;

[0069] Step 2: Perform the following preprocessing steps 2.1 to 2.3 on the surface 3D point cloud of the duct bend section to obtain surface points with normal vectors and an effective raster of the duct region:

[0070] Step 2.1: Calculate the surface point normal vector of the duct's 3D point cloud; including: for each point p in the 3D point cloud i Define a radius r with it as the center. nl Region (r) nl =R / 8), to obtain other points within the region. Since these points all belong to the duct surface, when the region radius r nlWhen the values ​​are small, these points can be considered to approximately form a plane. A plane fitting is then performed on these points, and the unit normal vector ν of the plane is calculated. i The unit normal vector ν of the plane i That is, the surface normal vector of that point.

[0071] Step 2.2: Spatial rasterization of the above catheter surface points; including: calculating the minimum circumscribed cuboid of the catheter surface points, such that each side of the cuboid is parallel to the x, y, and z axes respectively, and rasterizing the space containing the cuboid along the x, y, and z axes according to the side length r. grid The grid is continuously cut into tiny cubes. All grids are numbered, and the center of each grid is denoted as g. m This is used to simplify subsequent calculations;

[0072] Step 2.3: As Figure 2 As shown, the effective grid for extracting the duct region is shown; this includes: calculating the center point g of each grid. m Its nearest surface point p i The distance d between mi , can be represented as:

[0073] d mi =‖g m -p i ‖ (1)

[0074] If d mi If the radius of the raster is less than the conduit radius R, the raster is considered a valid raster; otherwise, the raster is discarded. The center of the extracted valid raster is denoted as g. n ;

[0075] Step 3: Process the surface points with normal vectors and the effective grid of the duct region as described in steps 3.1 to 3.2 to obtain the initial growth points of the duct ridges:

[0076] Step 3.1: As Figures 3-4 As shown, the duct can be approximated as being formed by splicing together countless micro-cylinders along a ridge. The intersection contour line between the micro-cylinders can be approximated as a perfect circle, and the surface point normal vectors on the intersection contour line converge inward at the center of the circle, which is the duct ridge point. Therefore, the number of times the effective grid is penetrated by the surface point normal of the duct is counted, and the center of the grid with the highest number of penetrations is taken as the initial growth point.

[0077] Due to the large number of points on the duct surface, to improve computational efficiency, we search for duct surface points near the effective grid to reduce the search range. The effective grid center g is used as the reference point. n Centered on the catheter, define a region of radius R, with a distance p from the catheter surface point that is less than the nominal radius of the catheter. i , forming the surface point set P n , can be represented as

[0078] Pn ={p i |R>‖g n -p i ||,p i ∈P} (2)

[0079] Through surface point set P n Each point p in i and the normal vector ν of that point i To form a straight line from the direction vectors, calculate the effective grid center g. n Distance d to the line ni , is represented as:

[0080] d ni =‖(p i -g n )-(p i -g n )·v i ·v i ‖ (3)

[0081] Set a penetration threshold ε, ε = r grid / 2; if d ni <ε, is considered a valid grid cell being penetrated once. For each valid grid cell center g n Calculate its surface point set P n The number of penetrations is used to determine the initial growth point of the ridge line of the vessel, with the effective grid center having the most penetrations, denoted as S. k The set of surface points corresponding to the initial growth point is denoted as SP. k ;

[0082] Step 3.2: To improve accuracy, the initial ridge growth point obtained in Step 3.1 is optimized. A random sampling consensus algorithm is used to optimize the initial ridge growth point S. k The corresponding surface point set SP k By performing spatial cylinder fitting, we can obtain a point C through which the central axis of the cylinder passes. k and unit direction vector The initial growth point S k Projecting onto the central axis of the cylinder yields the optimized initial growth point S. k ', can be calculated as:

[0083]

[0084] Step 4: As Figure 6As shown, since the ridges of the duct are continuous and first-order differentiable, the tangents at adjacent points on the ridge are approximately parallel. After determining a point on the ridge, a growth method can be used to obtain its neighboring points, i.e., candidate points for the next ridge point. Steps 4.1 to 4.4 are performed on the duct surface points and the initial growth points of the ridges to obtain disordered, discrete ridge points:

[0085] Step 4.1: Initial growth point S of the ridge line k Its tangent vector is the unit direction vector of the cylinder's central axis obtained in step 3.2. Set growth step size l step Then the candidate point S of the next ridge point k+1 It can be calculated as:

[0086]

[0087] l step =r grid / 2 (6)

[0088] The above ridge point S k The tangent vector of ' is used as the normal vector, and it passes through the candidate point S of the ridge. k+1 Take the cross section Plane k+1 Cross-section Plane k+1 The intersection circle obtained by the intersection with the duct surface is the ridge point S. k+1 The corresponding intersection circle. Due to the cross-section Plane k+1 An infinitely large value may result in multiple intersections with the duct, affecting ridge line calculations. Therefore, a candidate point S is selected. k+1 The set of CP points on the duct surface with a distance less than 1.1R k+1 , can be represented as:

[0089] CP k+1 ={p i |1.1R>‖S k+1 -p i ||,p i ∈P} (7)

[0090] Ridge candidate point S k+1 The corresponding intersecting circle can be approximated as a short, infinitesimal cylinder of length Δ, in the set CP. k+1 Inside and in Plane k+1 The points on the intersection point set SC k+1 , can be represented as:

[0091]

[0092] Δ=r grid / 2 (9)

[0093] Step 4.2: Set the threshold 'a' for the number of fitting points on the cylinder, and determine the intersection point set SC. k+1 The relationship between the number of points and the threshold 'a':

[0094] If the intersection point set SC k+1 The number of points in the set SC is greater than a, for point set SC k+1 Perform a spatial cylinder fitting to obtain a point C through which the cylinder's axis passes. k+1 and the unit vector of the central axis of the cylinder Candidate ridge point S k+1 Projecting this onto the cylinder axis yields the optimized ridge point S'. k+1 It can be calculated as:

[0095]

[0096] unit vector of the central axis of the cylinder S' is the ridge point k+1 The tangent vectors at the ridge point are compared. Tangent vector to the previous ridge point The direction of the ridge point's tangent vector; if the directions are different, the tangent vector of the ridge point... The calculation is as follows:

[0097]

[0098] Set the ridge point S' k+1 and tangent vector Substitute steps 4.1 and 4.2 into the equation and continue to grow the ridge points downwards.

[0099] Step 4.3: If the intersection set of points SC is SC k+1 If the number of points in the ridge is less than a, update the ridge candidate points S. k+1 The calculation is as follows:

[0100]

[0101] l′ step =5·l step (13)

[0102] Repeat step 4.1 to calculate the updated ridge candidate points S. k+1 Find the corresponding intersection point set and update the point set SC. k+1 If the updated point set SC k+1 If the number of midpoints is greater than a, repeat step 4.2 to calculate and update the ridge point S'. k+1 and tangent vector Otherwise, it is assumed that the ridge point grows to the end of the duct, and growth in that direction is stopped.

[0103] Step 4.4: After the ridge point growth on one side of the initial ridge point is completed, perform ridge point growth on the other side. Adjust the growth direction of the initial ridge point, which can be calculated as follows:

[0104]

[0105] Repeat steps 4.1-4.3 to complete the ridge point growth on the other side of the ridge initial point.

[0106] Step 5: Based on the growth order of the ridge points on both sides of the initial ridge point, sort and organize them to obtain the ordered ridge points from one end of the duct to the other end, thus completing the ridge extraction of the duct.

[0107] Example:

[0108] like Figure 7 (a) is the original three-dimensional point cloud image of the catheter surface input in the method of the present invention, such as Figure 7 (b) is the result of the original three-dimensional point cloud ridge line extraction and processing by the method of the present invention. The nominal diameter of the conduit is 20 mm, and the grid side length is r. grid With a diameter of 0.8 mm and a cylinder fitting point threshold of 157, the extraction results show that the average radius of all ridge points corresponding to the cylinder fitting is 19.93 mm, which is 0.07 mm different from the nominal diameter value, proving that the ridge extraction accuracy of the catheter is high.

[0109] In summary, the method of the present invention provides a catheter ridge extraction method based on three-dimensional point cloud, which has high accuracy, good versatility, and strong robustness, and helps to improve the production efficiency and manufacturing quality of catheters.

[0110] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for extracting duct ridges based on three-dimensional point clouds, characterized in that, Includes the following steps: Step 1: Set the nominal radius R of the conduit and the original 3D point cloud P of its surface; Step 2: Preprocess the 3D point cloud of the duct bend to obtain surface points with normal vectors and an effective grid of the duct region; the preprocessing of the 3D point cloud of the duct bend to obtain surface points with normal vectors and an effective grid of the duct region includes the following steps: 2(a). For each point p in the 3D point cloud i Define a radius r with it as the center. nl Region, r nl =R / 8, obtain other points within the region; perform plane fitting on these points, and calculate the unit normal vector ν of the plane. i ; 2(b). Spatial rasterization of catheter surface points: Calculate the minimum circumscribed cuboid of the catheter surface points, such that each side of the cuboid is parallel to the x, y, and z axes respectively. Then, rasterize the space containing the cuboid along the x, y, and z axes according to the side length r. grid The grid is obtained by continuously cutting it into tiny cubes; 2(c). Extracting valid grids in the conduit region: Calculate the distance between the center point of each grid and its nearest surface point. If the distance is less than the nominal radius R of the conduit, the grid is considered a valid grid; otherwise, the grid is discarded. The center of the extracted valid grid is denoted as g. n ; Step 3: By comparing the number of times the effective grid is penetrated by the surface point normal, the initial point for ridge growth of the duct is obtained and optimized; the center of the grid with the highest number of penetrations is taken as the initial point for ridge growth. Step 4: Track the growth from the initial point outwards to both sides and calculate all the ridge points of the duct; the process of tracking the growth from the initial point outwards to both sides and calculating all the ridge points of the duct includes the following steps: 4(a). The initial growth point S of the ridge line k The tangent direction is used to search for candidate points S for the next ridge point. k+1 Determine the point set CP on the catheter surface. k+1 In the set CP k+1 Select infinitesimal cylinders of length Δ to form the set of intersecting points SC. k+1 The micro-short cylinder is S k The tangent to ' is the central axis of the cylinder; 4(b). For the point set SC k+1 Perform spatial cylindrical fitting and select ridge candidate points S k+1 Projecting onto the cylinder axis yields the optimized ridge point S'. k+1 Compare the tangent vectors at the ridge point. Tangent vector to the previous ridge point The direction of the ridge line is used to determine whether the tangent direction of that ridge point should be adjusted. Set the ridge point S' k+1 and tangent vector Substitute 4(a)-4(b) and continue to grow the ridge point downwards; 4(c). If the intersection set of points SC k+1 If the number of points in the ridge is less than a, update the ridge candidate points S. k+1 The calculation is as follows: l′ step ＝5·l step Among them l step This refers to the growth step size; Repeat step 4(a) to calculate the updated ridge candidate points S. k+1 Find the corresponding intersection point set and update the point set SC. k+1 ; If the updated point set SC k+1 If the number of points in the matrix is ​​greater than a, proceed to step 4(b) to calculate and update the ridge point S'. k+1 and tangent vector Otherwise, it is assumed that the ridge point grows to the end of the duct, and growth in that direction stops; 4(d). After the ridge point on one side of the ridge initial point has grown, perform ridge point growth on the other side; adjust the growth direction of the ridge initial point, calculated as follows: Repeat steps 4(a)-4(c) to complete the ridge point growth on the other side of the ridge initial point; Step 5: Based on the growth sequence of the ridge points on both sides of the initial ridge point, the ridge points are ordered to complete the ridge extraction of the duct.

2. The method for extracting duct ridges based on three-dimensional point clouds according to claim 1, characterized in that, The process of obtaining and optimizing the initial point for ridge growth of the duct by comparing the number of times the effective grid is penetrated by the surface point normal includes the following steps: a. Calculate the initial growth point of the ridge line of the duct: Calculate the effective grid center g n To the surface point set P n Each point p in i normal vector ν i Distance d of the line ni If d ni If the value is less than ε, it is considered that a valid grid cell has been penetrated once; ε is the penetration threshold. The center of the effective grid that is penetrated the most times is taken as the initial growth point of the ridge line of the vessel, and denoted as S. k The set of grid penetration points corresponding to the initial growth point is denoted as SP. k ; b. Optimize the initial growth point of the duct ridge: Using a random sampling consensus algorithm, perform spatial cylinder fitting on the surface point set SPk corresponding to the initial growth point Sk of the ridge to obtain the point Ck through which the central axis of the cylinder passes and the unit direction vector. Projecting the initial growth point Sk onto the central axis of the cylinder yields the optimized initial growth point Sk', calculated as follows: 。 3. The method for extracting duct ridges based on three-dimensional point clouds according to claim 2, characterized in that, The calculation of the effective grid center g n To the surface point set P n Each point p in i normal vector ν i Distance d of the line ni If d ni <ε is considered a valid grid being penetrated once; including: With effective grid center g n Centered on the catheter, define a region of radius R, with a distance p from the catheter surface point that is less than the nominal radius of the catheter. i , forming the surface point set P n , represented as: P n ={p i |R>||g n -p i ||,p i ∈P} Through surface point set P n Each point p in i and the normal vector ν of that point i To form a straight line from the direction vectors, calculate the effective grid center g. n Distance d to the line ni , represented as: d ni =||(p i -g n )-(p i -g n )·v i ·v i || Set a penetration threshold ε, ε = r grid / 2; if d ni <ε, is considered a valid grid cell being penetrated once; for each valid grid cell center g n Calculate its corresponding point set P n Number of times it penetrates.

4. The method for extracting duct ridges based on three-dimensional point clouds according to claim 1, characterized in that, The initial growth point S of the ridge line k The tangent direction is used to search for candidate points S for the next ridge point. k+1 Determine the point set CP on the catheter surface. k+1 In the set CP k+1 Select infinitesimal cylinders of length Δ to form the set of intersecting points SC. k+1 ; include: The initial growth point S of the ridge line k Its tangent vector is the unit direction vector of the central axis of the cylinder. Set growth step size l step , l step =r grid / 2, then the candidate point S of the next ridge point k+1 The calculation is as follows: Selection of candidate points S k+1 The set of CP points on the duct surface with a distance less than 1.1R k+1 , represented as: CP k+1 ={p i |1.1R>||S k+1 -p i ||,p i ∈P} In the collection CP k+1 The inner length is selected as Δ (Δ = r) grid The short cylinders ( / 2) form the set of intersecting points SC. k+1 , represented as:

5. The method for extracting duct ridges based on three-dimensional point clouds according to claim 1, characterized in that, The pair set SC k+1 Perform spatial cylindrical fitting and select ridge candidate points S k+1 Projecting onto the cylinder axis yields the optimized ridge point S'. k+1 Compare the tangent vectors at the ridge point. Tangent vector to the previous ridge point The direction of the ridge line is used to determine whether the tangent direction at that ridge point should be adjusted; including: Set a threshold 'a' for the number of points fitted to the cylinder. If the set of intersecting points SC... k+1 The number of points in the set SC is greater than a, for point set SC k+1 Perform spatial cylinder fitting to obtain point C through which the cylinder axis passes. k+1 and the unit vector of the central axis of the cylinder Candidate ridge point S k+1 Projecting onto the cylinder axis yields the optimized ridge point S'. k+1 The calculation is as follows: unit vector of the central axis of the cylinder S' is the ridge point k+1 The tangent vectors at the ridge point are compared. Tangent vector to the previous ridge point The direction of the ridge point's tangent vector; if the directions are different, the tangent vector of the ridge point... The calculation is as follows:

Images