A method for measuring long catheter splicing based on two-dimensional targets

Abstract

This invention discloses a method for measuring the splicing of long catheters based on two-dimensional targets, comprising the following steps: First, selecting measurement segment locations on the catheter and installing two-dimensional targets; acquiring images of the catheter at each measurement segment location using multiple cameras and reconstructing a three-dimensional model of that measurement segment; then, obtaining the coded and sorted center coordinates of the target dots through image processing, matching common target dots in adjacent measurement segments, and calculating the coordinate system transformation matrix between adjacent measurement segments; finally, transforming the three-dimensional models of all measurement segments of the catheter to the same coordinate system to obtain a complete three-dimensional model of the long catheter. This invention is adaptable to the three-dimensional attitude measurement of catheters with large length specifications, is not limited by the type of two-dimensional target, has good flexibility and versatility, and exhibits good stability.

Description

Technical Field

[0001] This invention relates to the field of computer vision technology and discloses a method for measuring the splicing of long ducts based on two-dimensional targets. Background Technology

[0002] As the primary carrier of system media, conduit components play a crucial role in aerospace manufacturing technology, directly impacting key indicators such as product functionality and lifespan. To meet the demands of miniaturization and lightweighting, the installation posture and spatial shape of conduits are becoming increasingly complex, making posture measurement at critical locations a vital step in the manufacturing process. Generally, a three-dimensional model of the conduit needs to be reconstructed through precise measurements, and its key features are compared with the theoretically designed CAD model to evaluate whether the manufacturing quality meets installation requirements, thereby ensuring high product quality and performance.

[0003] In recent years, with the rapid development of computer vision technology, vision-based catheter measurement methods have been widely used in the industrial field. Currently, the mainstream measurement method utilizes multi-view stereo vision technology to reconstruct and obtain the 3D model and pose information of the catheter. This method processes the 2D image of the catheter to extract its edge contour lines or central axis, and then performs feature matching on the catheter images captured by multiple cameras to reconstruct the actual 3D model of the catheter. This method offers high measurement accuracy and good reconstruction results, but it is limited by the field of view of multiple cameras, limiting the range of catheter lengths that can be measured. Especially for long catheters (>2m) with few bends, this method can only obtain a partial 3D model of the catheter and cannot evaluate the overall bending quality of the catheter.

[0004] In summary, current methods for measuring long catheters cannot obtain their overall three-dimensional model and three-dimensional orientation information. 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 long catheter splicing measurement method based on a two-dimensional target. This method uses two-dimensional target splicing measurement to obtain a complete three-dimensional model of the long catheter, which can effectively improve the quality inspection level of long catheter parts and help improve the manufacturing quality of the catheter.

[0006] The technical solution adopted by the present invention to achieve the above objectives is: a method for measuring long catheters based on a two-dimensional target, comprising the following steps:

[0007] A method for measuring long catheters based on a two-dimensional target involves capturing images of the catheter and a two-dimensional target, processing these images, importing three-dimensional models of the catheter at different locations, and then stitching together the three-dimensional model of the entire catheter according to coordinate system transformations to measure catheter pose. The method includes the following steps:

[0008] S1: Based on the distance L between the ends of the long tube and the single measurement range D×E of the vision device, the long tube is divided into [L / D] measurement segments, and a two-dimensional target is set at the end point between each measurement segment of the long tube;

[0009] S2: Place the long conduit with a two-dimensional target under the vision device, move the measurement segment in sequence, and simultaneously trigger multiple industrial cameras of the vision device to acquire multi-angle images of the current measurement segment of the conduit during each movement;

[0010] S3: Under the current measurement segment, reconstruct the three-dimensional model of the duct in the current measurement segment using multi-angle images;

[0011] S4: Under the current measurement segment, arbitrarily select two images and perform the following i-iiii processing respectively to obtain the three-dimensional coordinates of the target circle center in the world coordinate system after encoding and sorting;

[0012] i. Extract the pixel coordinates of the center of each dot on the target;

[0013] ii. Calculate the number of targets and obtain the pixel coordinates of the center of the large and small dots of each target;

[0014] iii. Sort the dots of each target and encode the target;

[0015] iiii. Calculate the three-dimensional coordinates of the center of the target's dot;

[0016] S5: Match the three-dimensional coordinates of the common target circle of adjacent measurement segments and calculate the transformation matrix between adjacent measurement segments;

[0017] S6: Transform the 3D models of all measured catheter segments into the same coordinate system, stitch them together to obtain a complete 3D model of the long catheter, and then perform measurements.

[0018] The two-dimensional target is used for assembling the three-dimensional model of the segmented catheter. The two-dimensional target is a dot matrix composed of large and small dots, and the center distance between each adjacent dot in the matrix is ​​equal. All the large dots are connected in series to form a polygon, and the distance between each large dot and the centroid of the polygon is not equal.

[0019] The two-dimensional target is a 4*4 dot matrix, and the polygon is composed of three large circles (M1, M2, M3). M1 is a fixed point, always located in the first row and first column; M2 and M3 are variable points, located in any position in the third or fourth row and are adjacent points.

[0020] Extracting the pixel coordinates of the center of each dot on the target includes the following steps:

[0021] S4.1: Use the Canny operator to extract subpixel edges of the region of interest to obtain an edge image containing target dots, conduits, and connecting fixtures;

[0022] S4.2: Use roundness to remove interfering edges outside the target dots in the edge image;

[0023] S4.3: Perform least-squares fitting on the contours of all target dots on the current target to calculate the center coordinates p of all target dots. i (i = 1, 2, ..., m).

[0024] The calculation of the number of targets and the acquisition of the center coordinates of the large and small dots corresponding to each target include the following steps:

[0025] S4.4: Calculate the average contour length of all target dots. Dots with a length greater than the average are marked as large dots, and dots with a length less than the average are marked as small dots. The number of large dots is recorded as n, and the number of targets is n / 3.

[0026] S4.5: If n = 3, proceed to S4.6; if n > 3, divide the dots corresponding to different targets according to the distance clustering algorithm, as follows:

[0027] (4) Randomly select n / 3 points from the large circle as the initial class centers, denoted as μ. j (j = 1, 2, ..., n / 3);

[0028] (5) Calculate each point p i With class center μ j distance d ij And divide the circle into d ij Minimum class center μ j Corresponding category C j (j = 1, 2, ..., n / 3);

[0029] (6) The target error function can be calculated as follows:

[0030]

[0031] If E < 300, the classification stops, indicating that the dots corresponding to different targets have been divided. The dots in each category are denoted as p. jk (j=1,2,…,n / 3)(k=1,2,…,3m / n), where p jk (j=1,2,…,n / 3)(k=1,2, 3) is the center of the great circle, p jk (j=1,2,…,n / 3)(k=4,5,…,3m / n) is the center of the small circle; otherwise, return to step (1) and continue to update μ. j and Cj .

[0032] The process of sorting the dots of each target and encoding the targets includes the following steps:

[0033] S4.6: Category C for each target j Calculate the centroid q of its three great spheres. j (j=1,2,…,n / 3), calculate the distances between the three great circles and the centroid. The great circle with the largest distance is M1, denoted as p. j1 The other two are M2 and M3, denoted as p. j2 and p j3 ;

[0034] S4.7: Determine the x-axis direction of the target: p j2 Point to p j3 The vector is denoted as ν 23 Passing point p j1 、with ν 23 Draw a straight line l1 with the direction vector as the vector; extract the small circle p. jk In the sequence (k = 4, 5, ..., 3m / n), the point whose distance to line l1 is less than the threshold is denoted as p. j4 p j5 and p j6 Calculate p j1 Pointing to p respectively j4 p j5 and p j6 The vector, denoted as ν 14 ν 15 and ν 16 ; calculate ν respectively 23 and ν 14 ν 15 ν 16 The included angle θ i (i = 1, 2, 3), if all are less than 90°, then the x-axis direction is defined as vector ν. 23 The direction of the x-axis; otherwise, the direction of the x-axis is defined as vector ν. 23 The opposite direction. The unit vector in the x-axis direction is denoted as ν. x ;

[0035] S4.8: Determine the y-axis direction of the target: passing through point p j2 Draw a straight line l2 along the x-axis; extract the small circle p. jk (except p) j2 and p j3 The point in the circle whose distance to line l is less than the threshold is denoted as p. j7 and p j8 ; Calculate p j2 Pointing to p respectively j7 and p j8 The vector, denoted as ν27 and ν 28 The vector product G can be expressed as:

[0036] G = v 2h ·v x (h=3,7,8) (3)

[0037] Wherein, the vector v corresponding to the minimum G value 2h In (h=3,7,8), p j2 The dot it points to is denoted as p. jh (h = 3, 7, 8), large dot p j1 Point to p jh The direction is defined as the y-axis direction;

[0038] S4.9: Sort the coordinates of the dots of each target according to the determined x and y directions, and update p after sorting. jk Based on the row and column positions of the large dots M2 and M3, different targets are encoded.

[0039] The process of matching the three-dimensional coordinates of the common target point of adjacent measurement segments before and after measurement yields the transformation matrix between adjacent measurement segments, including the following steps:

[0040] Based on the target codes in the measurement locations, the common targets in the current measurement segment and the previous measurement segment are matched. Assume there are z common target dots in the matching areas of the previous and current measurement segments, and the three-dimensional coordinates of these common target dots are p... t and p t For the sequence '(t=1,2,…,z), by solving the least squares equation, the transformation matrix H between the two measurement segments is obtained, and the objective function can be expressed as:

[0041]

[0042] The process of transforming the 3D models of the catheter at all measurement locations to the same coordinate system and stitching them together to obtain a complete 3D model of the long catheter includes the following steps:

[0043] Assume the three-dimensional model of the catheter at the a-th measurement segment is S. a (a = 1, 2, ..., [L / D]), the coordinate system transformation matrix between the b-th and (b+1)-th measurement segment positions is H. b (b=1,2,…,[L / D]-1), transforming the three-dimensional model of the catheter at all measurement segments to the same coordinate system yields the complete three-dimensional model S of the long catheter, which can be represented as:

[0044]

[0045] The present invention has the following beneficial effects and advantages:

[0046] 1. The present invention provides a method for measuring the splicing of long catheters based on two-dimensional targets. By identifying and encoding two-dimensional targets, the catheters are measured and spliced ​​in segments to obtain a three-dimensional model of the entire long catheter.

[0047] 2. The method of this invention designs a two-dimensional target based on circular marker points, proposes a two-dimensional target recognition method, achieves high-precision visual positioning, has good versatility, and can be applied to other related measurement scenarios in the industrial field.

[0048] 3. The long catheter measurement method based on two-dimensional targets described in this invention is not limited by the type of two-dimensional targets, has good flexibility, and helps to improve the production efficiency and manufacturing quality of catheters. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the long catheter measuring device in the method of the present invention;

[0050] Figure 2 This is a flowchart of the method of the present invention;

[0051] Figure 3 This is a schematic diagram illustrating the principle of two-dimensional target recognition in the method of this invention.

[0052] Figure 4 This is a schematic diagram illustrating the principle of long catheter splicing measurement in the method of the present invention;

[0053] Figure 5 This is a diagram showing the two-dimensional target recognition result in the method of the present invention;

[0054] Figure 6 This is a diagram showing the measurement results of long catheter splicing in the method of the present invention;

[0055] Among them, 1 is the device body, 101 is the industrial camera, 102 is the platform support rod, 103 is the conduit support platform, 104 is the backlight source, 105 is the metal frame, and 2 is the host computer. Detailed Implementation

[0056] 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.

[0057] 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.

[0058] like Figure 2 As shown, the hardware device for measuring long conduits includes a device body 1 and a host computer 2. The device body 1 includes an industrial camera 101, a support platform connecting rod 102, a support platform 103, a backlight source 104, and a metal frame 105. Eight industrial cameras are fixed to the top of the metal frame, and their arrangement ensures that the measurement range of each camera can cover the entire backlight source. The conduit support platform and the backlight source are located on the same plane, and the size of the support platform can be changed according to the conduit dimensions.

[0059] like Figures 1-4 As shown, the method for measuring long catheters based on a two-dimensional target according to the present invention includes the following steps:

[0060] Step 1: Let the distance between the ends of the long conduit be L, and the measurement area of ​​the multi-camera system be D×E (D is the long side, E is the short side). Select [L / D] ([] is the rounding up sign) measurement segment positions on the long conduit and set one two-dimensional target at each end of the long conduit and at the endpoints between each measurement segment, for a total of ([L / D]-1) two-dimensional targets, which are used for stitching together the three-dimensional models of each measurement segment of the conduit. Figure 3 As shown, the two-dimensional target consists of 3 large circles (M1, M2, M3) and 13 small circles, with the center distance between any two adjacent circles being equal. Among them, M1 is a fixed point (always located in the 1st row and 1st column), while M2 and M3 are variable points (located in any position in the 3rd or 4th row, ensuring they are adjacent points).

[0061] Step 2: Set the position of a measurement segment containing the end of the catheter as the initial measurement segment position, and manually place the measurement segment catheter on the support platform 103 and the backlight source 104.

[0062] Step 3: Simultaneously trigger 8 industrial cameras to acquire images of the duct from different perspectives. Using steps 3-5 of the patent entitled "Measuring a Duct under Robot Clamping" with authorization announcement number CN112729150B and authorization announcement date of October 22, 2021, the method for removing the two-dimensional target and clamping components in this invention is the same as the method for removing the robot clamping components in the cited patent, thereby obtaining the three-dimensional model of the duct measured at the measurement segment position.

[0063] Step 4: Select two cameras and process the images they capture using steps 4.1 to 4.9 respectively to identify and encode the two-dimensional target:

[0064] i. Extract the center coordinates of all target dots (steps 4.1-4.3):

[0065] Step 4.1: Use the Canny operator to extract subpixel edges of the region of interest to obtain an edge image containing the target dots, conduits, and connecting fixtures.

[0066] Step 4.2: Filter out interfering edges in the edge image other than the target dots, i.e., remove edges with a contour roundness less than the threshold. In the experiment, the threshold was set to 0.7.

[0067] Step 4.3: Perform least-squares fitting on the contours of all target dots to calculate the center coordinates p of all target dots. i (i = 1, 2, ..., m).

[0068] ii. Calculate the number of targets and obtain the center coordinates of the large and small dots corresponding to each target (steps 4.4-4.5):

[0069] Step 4.4: Calculate the average contour length of all target dots. Dots with a contour length greater than the average are called large dots, and those with a contour length less than the average are called small dots. The number of large dots is recorded as n, and the number of targets is n / 3.

[0070] Step 4.5: If the number of targets is 1, it means that the position is the initial measurement position or the termination measurement position, and proceed directly to step 4.6; if the number of targets is greater than 1, it means that the position is the intermediate measurement position, and there are more than 2 targets.

[0071] An improved distance clustering algorithm is used to distinguish the dots corresponding to different targets. The steps are as follows:

[0072] (1) Randomly select n / 3 points from the large circle as the initial class centers, denoted as μ. j (j = 1, 2, ..., n / 3);

[0073] (2) Calculate each point p i With class center μ j distance d ij And divide the circle into d ij Minimum class center μ j Corresponding category C j (j = 1, 2, ..., n / 3);

[0074] (3) The target error function can be calculated as follows:

[0075]

[0076] If E < 300, the classification stops, indicating that the dots corresponding to different targets have been divided. The dots in each category are denoted as p. jk (j=1,2,…,n / 3)(k=1,2,…,3m / n), where p jk (j=2,3,…,n / 3)(k=1,2, 3) are large dots, p jk (j=1,2,…,n / 3)(k=4,5,…,3m / n) are small dots; otherwise, return to step (1) and continue updating μ. j and C j .

[0077] iii. Sort the dots of each target and encode the target (steps 4.4-4.7):

[0078] Step 4.6: Class C for each target j Calculate the centroid q of its three great spheres. j (j=1,2,…,n / 3) can be represented as:

[0079]

[0080] Calculate the distances from the three great circles to the centroid. The great circle with the largest distance is M1, denoted as p. j1 The other two are M2 and M3, denoted as p. j2 and p j3 .

[0081] Step 4.7: Determine the x-axis direction of the target: p j2 Point to p j3 The vector is denoted as ν 23 Passing point p j1 、with ν 23 Draw a straight line l1 with the direction vector as the vector. Extract the small circle p. jk In the sequence (k = 4, 5, ..., 3m / n), the point whose distance to line l1 is less than the threshold (set to 6 in the experiment) is denoted as p. j4 p j5 and p j6 Calculate p j1 Pointing to p respectively j4 p j5 and p j6 The vector, denoted as ν 14 ν 15 and ν 16 Calculate ν separately. 23 and ν 14 ν 15 ν 16 The included angle θ i (i = 1, 2, 3), if all are less than 90°, then the x-axis direction is defined as vector ν.23 The direction of the x-axis; otherwise, the direction of the x-axis is defined as vector ν. 23 The opposite direction. The unit vector in the x-axis direction is denoted as ν. x .

[0082] Step 4.8: Determine the y-axis direction of the target: through point p j2 Draw a straight line l2 along the x-axis and extract the small circle p. jk (except p) j2 and p j3 The point in the circle whose distance to line l is less than the threshold (set to 6 in the experiment) is denoted as p. j7 and p j8 Calculate p j2 Pointing to p respectively j7 and p j8 The vector, denoted as ν 27 and ν 28 The vector product G can be expressed as:

[0083] G = v 2h ·v x (h=3,7,8) (3)

[0084] Wherein, the vector v corresponding to the minimum G value 2h In (h=3,7,8), p j2 The dot it points to is denoted as p. jh (h = 3, 7, 8), large dot p j1 Point to p jh The direction is defined as the y-axis direction.

[0085] Step 4.9: Sort the coordinates of the dots of each target according to the determined x and y directions, and update p after sorting. jk Different targets are encoded based on the row and column positions of the large dots M2 and M3.

[0086] iv. Calculate the three-dimensional coordinates of the target's dot (step 4.10):

[0087] Step 4.10: Based on the target dots identified in the two camera images, and using the intrinsic and extrinsic parameters of the industrial camera obtained from global calibration, the three-dimensional coordinates of the target dots in the world coordinate system are obtained through binocular reconstruction.

[0088] If the current measurement segment position is the initial position, move the conduit to the next measurement segment position and return to step 3 to continue execution, obtaining the three-dimensional coordinates of the target dots in the world coordinate system for each measurement segment.

[0089] Step 5: Process the target's dot coordinates as described in steps 5.1 and 5.2, match the corresponding target, and establish a transformation relationship:

[0090] Step 5.1: Match the common targets in the current measurement segment position with the previous measurement segment position based on the code of each target in the measurement segment position.

[0091] Step 5.2: Since the coordinate system relationship between different measurement segment positions undergoes a rigid transformation during catheter movement, assume there are z common target points in the matching area between the previous and current measurement segment positions, and the three-dimensional coordinates of these common target points are p... t and p t For the sequence '(t=1,2,…,z), by solving the least squares equation, the transformation matrix H between the positions of the two measurement segments is obtained, and the objective function can be expressed as:

[0092]

[0093] If the current measurement position is not the final measurement segment position, move the conduit to the next measurement segment position and return to step 3 to continue execution.

[0094] Step 6: After the entire catheter measurement is completed, assume that the three-dimensional model of the catheter at the position of the a-th measurement segment is S. a (a = 1, 2, ..., [L / D]), the coordinate system transformation matrix between the b-th and (b+1)-th measurement segment positions is H. b (j=1,2,…,[L / D]-1), transforming the three-dimensional model of the catheter at all measurement positions to the same coordinate system yields the complete three-dimensional model S of the long catheter, which can be represented as:

[0095]

[0096] Example:

[0097] like Figure 6 The result of two-dimensional target identification at a certain measurement location in the method of the present invention is shown. The nominal diameter of the catheter is 15 mm, the length is 3.7 m, and the measurement accuracy reaches 0.1 mm.

[0098] In summary, the method of the present invention provides a long catheter splicing measurement method based on a two-dimensional target, which has high accuracy, good versatility and flexibility, and helps to improve the production efficiency and manufacturing quality of catheters.

[0099] The above embodiments are only for illustrating the technical concept and features of the present invention and should not be construed as limiting the scope of protection of the present invention. All equivalent variations and modifications made within the scope of the claims of this invention should be considered within the scope of this invention.

Claims

1. A long catheter splicing measurement method based on a two-dimensional target, characterized by, By capturing and processing images of the catheter and a 2D target, importing 3D models of the catheter at different locations, and stitching them together according to coordinate system transformations to calculate the entire 3D model of the catheter for measuring catheter pose, the following steps are included: S1: Based on the end distance L of the long catheter and the single measurement range D×E of the visual device, the long catheter is divided into [ L / D] measurement segments, with one two-dimensional target set at the endpoint between each measurement segment of the long catheter; S2: Place the long conduit with a two-dimensional target under the vision device, move the measurement segment in sequence, and simultaneously trigger multiple industrial cameras of the vision device to acquire multi-angle images of the current measurement segment of the conduit during each movement; S3: Under the current measurement segment, reconstruct the three-dimensional model of the duct in the current measurement segment using multi-angle images; S4: Under the current measurement segment, arbitrarily select two images and perform the following i-iiii processing respectively to obtain the three-dimensional coordinates of the target circle center in the world coordinate system after encoding and sorting; i. Extract the pixel coordinates of the center of each dot on the target; ii. Calculate the number of targets and obtain the pixel coordinates of the center of the large and small dots of each target; iii. Sort the dots of each target and encode the target; iiii. Calculate the three-dimensional coordinates of the center of the target's dot; S5: Match the three-dimensional coordinates of the common target circle of adjacent measurement segments and calculate the transformation matrix between adjacent measurement segments; S6: Transform the 3D models of all measured catheter segments into the same coordinate system, stitch them together to obtain a complete 3D model of the long catheter, and then perform measurements.

2. The method for measuring long catheter splicing based on a two-dimensional target according to claim 1, characterized in that, The two-dimensional target is used for assembling the three-dimensional model of the segmented catheter. The two-dimensional target is a dot matrix composed of large and small dots, and the center distance between each adjacent dot in the matrix is ​​equal. All the large dots are connected in series to form a polygon, and the distance between each large dot and the centroid of the polygon is not equal.

3. The method for measuring the splicing of long catheters based on a two-dimensional target according to claim 2, characterized in that, The two-dimensional target is 4 A 4-dot matrix, wherein the polygon is composed of three large circles (M1, M2, M3); wherein M1 is a fixed point, always located in the first row and first column; M2 and M3 are variable points, located in any position in the third or fourth row and are adjacent points.

4. The method for measuring the splicing of long catheters based on a two-dimensional target according to claim 1, characterized in that, Extracting the pixel coordinates of the center of each dot on the target includes the following steps: S4.1: Use the Canny operator to extract subpixel edges of the region of interest to obtain an edge image containing target dots, conduits, and connecting fixtures; S4.2: Use roundness to remove interfering edges outside the target dots in the edge image; S4.3: Perform least-squares fitting on the contours of all target dots on the current target to calculate the center coordinates of all target dots. p i ( i =1, 2, …, m ).

5. The method for measuring long catheter splicing based on a two-dimensional target according to claim 1, characterized in that, The calculation of the number of targets and the acquisition of the center coordinates of the large and small dots corresponding to each target include the following steps: S4.4: Calculate the average contour length of all target dots. Dots longer than the average are designated as large dots, and those shorter than the average are designated as small dots. Count the number of large dots. n The number of targets is n / 3; S4.5: If n =3, execute S4.6; if n >3. Divide the dots corresponding to different targets using a distance clustering algorithm. The steps are as follows: (1) Randomly select from the large dots n Three points are used as the initial class centers, denoted as... μ j ( j =1, 2, …, n / 3); (2) Calculate each dot p i With class center μ j distance d ij And divide the dot into d ij Minimum class center μ j Corresponding category C j ( j =1, 2, …, n / 3); (3) The target error function can be calculated as follows: ； If E < 300, the classification stops, indicating that the dots corresponding to different targets have been divided. The dots in each category are denoted as follows: p jk ( j =1, 2, …, n / 3) ( k =1, 2, …, 3m / n ),in p jk ( j =1, 2, …, n / 3) ( k =1, 2, 3) is the center of the large circle. p jk ( j =1, 2, …, n / 3) ( k =4, 5, …, 3m / n If the center of the dot is 0, then return to step (1) and continue updating. μ j and C j .

6. The method for measuring the splicing of long catheters based on a two-dimensional target according to claim 1, characterized in that, The process of sorting the dots of each target and encoding the targets includes the following steps: S4.6: Category C for each target j Calculate the centroid of its three great circles. q j ( j =1, 2, …, n / 3), calculate the distances between the three great circles and the centroid. The great circle with the largest distance is M1, denoted as M1. p j1 The other two are M2 and M3, denoted as p j2 and p j3 ; S4.7: Determine the x-axis direction of the target: p j2 Pointing to p j3 The vector is denoted as ν 23 past the point p j1 、with ν 23 Draw a straight line with direction vector. l 1 Extract small dots p jk ( k =4, 5, …, 3m / n (middle to straight line) l 1 Points whose distance is less than the threshold are denoted as . p j4 , p j5 and p j6 ,calculate p j1 Pointing to respectively p j4 , p j5 and p j6 The vector, denoted as ν 14 ν 15 and ν 16 ; calculate ν respectively 23 and ν 14 ν 15 ν 16 The included angle θ i ( i =1, 2, 3), if all are less than 90°, then the x-axis direction is defined as vector ν. 23 The direction of the x-axis; otherwise, the direction of the x-axis is defined as vector ν. 23 The unit vector in the opposite direction of the x-axis is denoted as ν. x ; S4.8: Determine the y-axis direction of the target: passing through the point p j2 Draw a straight line along the x-axis l 2 Extract small dots p jk (remove p j2 and p j3 (middle to straight line) l Points whose distance is less than the threshold are denoted as . p j7 and p j8 ;calculate p j2 Pointing to respectively p j7 and p j8 The vector, denoted as ν 27 and ν 28 Vector product G It can be represented as: （3） Among them, the smallest G The vector corresponding to the value ( h In the case of =3, 7, 8), the p j2 The dot it points to is denoted as p jh ( h =3, 7, 8), large dots p j1 point to p jh The direction is defined as the y-axis direction; S4.9: Sort the coordinates of the dots of each target according to the determined x and y directions, and update them after sorting. p jk Based on the row and column positions of the large dots M2 and M3, different targets are encoded.

7. The method for measuring the splicing of long catheters based on a two-dimensional target according to claim 1, characterized in that, The process of matching the three-dimensional coordinates of the common target point of adjacent measurement segments before and after measurement yields the transformation matrix between adjacent measurement segments, including the following steps: Based on the target codes at the measurement locations, the common targets in the current measurement segment and the previous measurement segment are matched. Assume that the common target dots in the matching areas of the previous and current measurement segments are... z The three-dimensional coordinates of the common target circle are as follows: p t and p t ' ( t =1, 2, …, z By solving the least squares equation, the transformation matrix between the two measurement segments is obtained. H The objective function can be expressed as: 。 8. The method for measuring the splicing of long catheters based on a two-dimensional target according to claim 1, characterized in that, The process of transforming the 3D models of the catheter at all measurement locations to the same coordinate system and stitching them together to obtain a complete 3D model of the long catheter includes the following steps: Assuming the catheter is in the first a The three-dimensional model of the location of each measurement segment is as follows S a ( a =1, 2, …,[ L / D]), the first b The and the ( b The coordinate system transformation matrix between the +1) measurement segment positions is: H b ( b =1, 2, …,[ L / D]-1), transform the 3D model of the catheter at all measurement segments to the same coordinate system to obtain the complete 3D model of the long catheter. S , can be represented as: 。

Images