A method for measuring the size of a fish hook tongue

Abstract

The application discloses a kind of measurement methods suitable for the measurement of fish-tail characteristic size, comprising the following steps: 1) multiple polyhedral targets are arranged around the fish-tail to be measured;Robot carries out multiple pose photographing with point cloud sensor;2) image features are extracted in each pose and three-dimensional point cloud is solved;3) with the data of mark point on polyhedral target as medium, the point cloud splicing between two adjacent point cloud sets is realized;4) ensure that the point cloud after splicing has covered the characteristic region to be measured, then obtain the three-dimensional point cloud slice and normal of test position according to the test requirement of characteristic to be measured;5) call the virtual gauge point cloud template constructed in advance, and realize accurate matching of three-dimensional point cloud slice and virtual gauge point cloud template by using the registration method of directed distance constraint;6) the measurement result output of characteristic to be measured is completed by using the point line constraint of gauge. The method can realize online measurement of characteristic size without pasting points on the surface of fish-tail, and complete the data output of fish-tail detection.

Description

Technical Field

[0001] This invention belongs to the field of industrial vision measurement, specifically relating to a method for measuring the feature dimensions of hook tongues. Background Technology

[0002] Currently, the common method for measuring hook tongue characteristic dimensions such as hook tongue bulge, hook tongue locking surface wear, and hook tongue S-surface wear is still mainly manual operation by workers using handheld gauges. This measurement mode is inefficient and can no longer meet the speed requirements of current maintenance operations.

[0003] Utilizing industrial vision measurement to measure the dimensional features of the hook tongue can effectively enhance the automation of this process. However, when inspecting workpieces using vision measurement equipment, precise workpiece positioning is a prerequisite for ensuring measurement accuracy. Before measurement begins, the workpiece's six degrees of freedom must be precisely constrained: three points for planar locking of translation and rotation, two points for linear orientation, and one point for final fixation. If the workpiece lacks planes or feature holes with guaranteed machining accuracy, it is difficult to achieve precise workpiece positioning through fixture engagement.

[0004] The hook tongue has a complex surface morphology and some missing planes and positioning holes, making it difficult to obtain complete and high-precision point cloud data that can be used to calculate the hook tongue's feature parameters. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method for measuring the characteristic dimensions of hook tongues, which enables online measurement of characteristic dimensions without the need for affixing points to the workpiece surface, and completes the data output for hook tongue detection.

[0006] Therefore, the technical solution of the present invention is as follows:

[0007] A method for measuring the characteristic dimensions of a hook tongue includes the following steps:

[0008] 1) The hook tongue is fixed to the test station, and multiple polyhedral targets are set around it; the robot carries a point cloud sensor to take pictures of it in multiple poses, covering the feature area to be tested.

[0009] When a point cloud sensor takes pictures in multiple poses, it can capture any feature surface of the same polyhedral target while acquiring the features of the target surface in two adjacent poses.

[0010] The polyhedral target has at least one feature face, and each feature face has at least three non-collinear marker points; let the marker points on each polyhedral target be the point set C. k The value of k is a positive integer less than or equal to the number of polyhedral targets; the relative positions of the markers on each polyhedral target are known.

[0011] 2) Under each photographic pose of the point cloud sensor, extract image features and solve the 3D point cloud to obtain the surface 3D point cloud data and the marker point 3D point cloud data under the current pose, which are recorded as the point cloud set of the current pose.

[0012] 3) Let any point cloud be the source point set A, and the point cloud set of its adjacent positions be the target set B. Use the marker point data on the polyhedral target as a medium to realize the point cloud stitching between the source point set A and the target set B.

[0013] 4) If the stitched point cloud obtained in step 3) completely covers the feature region to be tested, proceed directly to step A;

[0014] Conversely, the stitched point cloud is denoted as the source point set A, and the point cloud set at its adjacent position is denoted as the target set B. The point cloud stitching between the source point set A and the target set B is achieved again using the marker point data on the polyhedral target as a medium. This operation is repeated until the stitched point cloud completely covers the feature region to be tested, and then step A is executed.

[0015] Step A: Obtain the 3D point cloud slices and normals of the test location according to the test requirements of the feature to be tested;

[0016] 5) Retrieve the pre-constructed virtual gauge point cloud template and use the directed distance constraint registration method to achieve accurate matching between the 3D point cloud slices obtained in step A and the virtual gauge point cloud template;

[0017] 6) Use the point and line constraints of the gauge to complete the measurement results of the feature to be measured.

[0018] Furthermore, the point cloud stitching between source point set A and target point set B includes the following steps:

[0019] ① Let A1 be the three-dimensional point cloud data of the hook tongue surface in the source point set A, and A2 be the three-dimensional point cloud data of the hook tongue surface;

[0020] Let B1 be the 3D point cloud data of the hook tongue surface in the target set B, and B2 be the 3D point cloud data of the marker point;

[0021] Construct an error function using corresponding points in the overlapping regions of A1 and B1. ;

[0022] ②Find the point set C k The markers corresponding to A2 and B2 are denoted as points in point set A2. Point set C k The corresponding point in Let point B2 be... Point set C k The corresponding point in Construct the error function:

[0023]

[0024] (R,T) is the position transformation matrix between the source point set A and the target point set B;

[0025] It is a point set C k The position transformation matrix to point set A, where m is the number of marker points in A2 and n is the number of marker points in B2;

[0026] ③ Construct the final objective function The objective function is solved using an optimization algorithm to obtain the position transformation matrix (R,T) and residual values ​​of the source point set A and the target point set B; M represents the penalty factor, which is a preset value;

[0027] ④ Using the position transformation matrix (R,T) of the source point set A and the target point set B obtained in step ③, transform them to the same coordinate system, align them, merge them, and realize point cloud stitching.

[0028] Furthermore, in step ①, an error function is constructed. The method is as follows:

[0029] ;

[0030] in: and Let R be the set of points in the overlapping regions A1 and B1, respectively, where R is the rotation matrix, T is the translation matrix, and n is the number of points in the overlapping region involved in the calculation.

[0031] Furthermore, the optimization method is as follows: Algorithm or Gaussian iteration method; preferably, M > 10 8 .

[0032] The method provided by this invention follows the People's Republic of China Railway Industry Standard No. TB / T 2048-2016, which is titled: Locomotive and Rolling Stock Coupler Buffer Device Measuring Instrument Type 13 Coupler Inspection Measuring Instrument.

[0033] The regions to be tested are the outer region of the hook tongue nose, the hook tongue locking surface, and / or the S-surface region.

[0034] Furthermore, the pre-constructed virtual gauge point cloud template is matched with the measurement results of the feature to be measured;

[0035] When the measurement result of the feature to be measured is wear of the hook and tongue lock surface, the virtual gauge point cloud template is a hook and tongue lock surface measuring ruler (i.e., standard TB / T 2048-2016, pages 7 and 8, Figure 13 Hook and Tongue Lock Surface Measuring Ruler).

[0036] When the measurement result of the feature to be measured is that the hook tongue is bulging, the virtual gauge point cloud template is the hook tongue bulging detection template (i.e., standard TB / T 2048-2016, page 7, Figure 12 Hook Tongue bulging detection template).

[0037] When the measurement result of the feature to be measured is the wear of the hook tongue S-surface, the virtual gauge point cloud template is the hook tongue S-surface wear measuring ruler (i.e., standard TB / T 2048-2016, pages 8 and 9, Figure 14 Hook tongue S-surface wear measuring ruler).

[0038] Furthermore, step A: the method for obtaining 3D point cloud slices and normals is as follows:

[0039] When the measurement result of the feature to be measured is the wear of the hook tongue lock surface, determine the side where the hook tongue pin hole is located, offset it along its normal direction to the hook tongue area by a specific length, extract the point cloud slice of the hook tongue lock surface area, and calculate the normal of each point; the point cloud slice meets the measurement requirements of the hook tongue lock surface measuring ruler (i.e., standard TB / T 2048-2016, page 32, A.13 hook tongue lock surface wear measurement requirements);

[0040] When the measurement result of the feature to be tested is that the hook tongue is bulging outward, determine the side where the hook tongue pin hole is located, offset it along its normal direction to the hook tongue area by a specific length, cut out the point cloud slice of the outer region of the hook tongue nose, and calculate the normal of each point; the point cloud slice meets the measurement requirements of the hook tongue bulging detection template (i.e., standard TB / T 2048-2016, page 32, A.12 hook tongue bulging detection requirements);

[0041] When the measurement result of the feature to be measured is the wear of the S-surface of the hook tongue, determine the position of the edge surface of the hook tongue nose, and offset it to the hook tongue area by multiple different distances along the normal direction of the edge surface. Extract point cloud slices of multiple S-surfaces in the hook tongue nose area, and calculate the normal of each point. The point cloud slices meet the measurement requirements of the hook tongue S-surface wear measuring ruler (i.e., standard TB / T2048-2016, page 33, A.14 hook tongue S-surface wear detection requirements).

[0042] Furthermore, the directed distance-constrained registration method includes the following steps:

[0043] a) Using the virtual gauge point cloud template as a reference, transform the 3D point cloud slice data to the reference using the initial transformation matrix (R0,T0);

[0044] The initial transformation matrix (R0, T0) was confirmed manually, using a reference frame, and a reference target.

[0045] b) Using the corresponding position that needs to be matched with the object to be measured during the gauge inspection process as the target point, the nearest neighbor point in the 3D point cloud slice data is searched using the K-nearest neighbor algorithm to form the nearest neighbor-nearest neighbor point pair;

[0046] c) Construct an objective function E with the constraint of minimizing the directed distance between all nearest neighbor pairs, and calculate the residuals using optimization methods:

[0047] d) If the residual meets the threshold or exceeds the required number of iterations, stop the iteration and output the matching RT; otherwise, update the initial transformation matrix in step a) with this RT and repeat steps a) to c).

[0048] Furthermore, step c) constructs the objective function E using the following formula:

[0049]

[0050]

[0051]

[0052] : Three-dimensional rotation matrix;

[0053] Translation matrix;

[0054] The k-th iteration The three-dimensional coordinates of the source point;

[0055] The k-th iteration The three-dimensional coordinates of the target point corresponding to each source point;

[0056] The transpose of the unit normal vector at the target point;

[0057] Total number of iterations;

[0058] The number of points involved in the calculation during the k-th iteration;

[0059] Penalty factor, with a value of 10. 8 .

[0060] Furthermore, step 6) involves using the point and line constraints of the gauge to complete the measurement results of the feature to be measured as follows:

[0061] A rectangular coordinate system is constructed based on the test position of the gauge. The reading of the feature to be measured is obtained in the rectangular coordinate system according to the gauge's usage instructions. The difference between this reading and the reference value is taken as the measurement result. The output of the test results shall conform to the requirements of standard TB / T 2048-2016.

[0062] This invention first provides a method for measuring the dimensions of hook tongue features. Based on the rigid invariance between marker points on a polyhedron, it utilizes point cloud feature similarity as supplementary information to participate in point cloud stitching. This not only retains the high precision advantage of the marker point stitching method but also eliminates the need to attach points to the workpiece. Then, based on the overall three-dimensional point cloud of the workpiece, three-dimensional point cloud slices are generated according to the workpiece feature measurement requirements. Based on a point cloud matching method with directed distance constraints, the precise matching of the gauge point cloud template and the slices is achieved, thereby obtaining the dimensional results of the feature to be determined. Detailed Implementation

[0063] The method provided by this invention follows the People's Republic of China Railway Industry Standard TB / T 2048-2016, entitled "Measuring Instruments for Locomotive and Rolling Stock Coupler Buffer Devices - Type 13 Coupler Inspection Measuring Tool." It is mainly used to test properties that cannot be detected by conventional visual inspection methods: wear on the coupler tongue locking surface, outer expansion of the coupler tongue, and wear on the S-surface of the coupler tongue.

[0064] The technical solution of the present invention will be described in detail below with reference to specific embodiments.

[0065] A method for measuring the characteristic dimensions of a hook tongue includes the following steps:

[0066] 1) The hook tongue is fixed at the test station, and multiple polyhedral targets are set around it; the robot with point cloud sensors takes multiple pose photos of it to cover the feature area to be tested; the feature area to be tested is the outer area of ​​the hook tongue nose, the hook tongue lock surface and / or the S-surface area.

[0067] When a point cloud sensor takes pictures in multiple poses, it can capture any feature surface of the same polyhedral target while acquiring the features of the target surface in two adjacent poses.

[0068] A polyhedral target has at least one feature face, and each feature face has at least three non-collinear marker points; let the marker points on each polyhedral target be the point set C. k The value of k is a positive integer less than or equal to the number of polyhedral targets; the relative positions of the markers on each polyhedral target are known.

[0069] 2) Under each photographic pose of the point cloud sensor, extract image features and solve the 3D point cloud to obtain the surface 3D point cloud data and the marker point 3D point cloud data under the current pose, which are recorded as the point cloud set of the current pose.

[0070] 3) Let any point cloud be the source point set A, and the point cloud of its adjacent positions be the target set B. Using the marker point data on the polyhedral target as a medium, the point cloud stitching between the source point set A and the target set B is achieved through the following steps:

[0071] ① Let A1 be the three-dimensional point cloud data of the hook tongue surface in the source point set A, and A2 be the three-dimensional point cloud data of the hook tongue surface;

[0072] Let B1 be the 3D point cloud data of the hook tongue surface in the target set B, and B2 be the 3D point cloud data of the marker point;

[0073] Using the corresponding points in the overlapping regions of A1 and B1, the error function is constructed using the following formula. ;

[0074] ;

[0075] in: and Let R be the set of points in the overlapping regions A1 and B1, respectively, where R is the rotation matrix, T is the translation matrix, and n is the number of points in the overlapping region involved in the calculation.

[0076] ②Find the point set C k The markers corresponding to A2 and B2 are denoted as points in point set A2. Point set C k The corresponding point in Let point B2 be... Point set C k The corresponding point in Construct the error function:

[0077]

[0078] (R,T) is the position transformation matrix between the source point set A and the target point set B;

[0079] It is a point set C k The position transformation matrix to point set A, where m is the number of marker points in A2 and n is the number of marker points in B2;

[0080] ③ Construct the final objective function The objective function is solved using an optimization algorithm to obtain the position transformation matrix (R,T) and residual values ​​between the source point set A and the target point set B; M represents the penalty factor, which is a preset value; in practical applications, the optimization method can be selected from... Algorithm or Gaussian iteration method; preferably, M > 10 8 .

[0081] ④ Using the position transformation matrix (R,T) of the source point set A and the target point set B obtained in step ③, transform them to the same coordinate system, align them, merge them, and realize point cloud stitching.

[0082] 4) If the stitched point cloud obtained in step 3) completely covers the feature region to be tested, proceed directly to step A;

[0083] Conversely, the stitched point cloud is denoted as the source point set A, and the point cloud set at its adjacent position is denoted as the target set B. The point cloud stitching between the source point set A and the target set B is achieved again using the marker point data on the polyhedral target as a medium. This operation is repeated until the stitched point cloud completely covers the feature region to be tested, and then step A is executed.

[0084] Step A: Obtain the 3D point cloud slice and normal of the test location according to the test requirements of the feature to be tested; the specific method is as follows: when the measurement result of the feature to be tested is the wear of the hook tongue lock surface, determine the side where the hook tongue pin hole is located, offset it along its normal direction to the hook tongue area by a specific length, cut out the point cloud slice of the hook tongue lock surface area, and calculate and obtain the normal of each point; the point cloud slice meets the measurement requirements of the hook tongue lock surface measuring ruler (i.e., standard TB / T 2048-2016, page 32, A.13 hook tongue lock surface wear measurement requirements);

[0085] When the measurement result of the feature to be tested is that the hook tongue is bulging outward, determine the side where the hook tongue pin hole is located, offset it along its normal direction to the hook tongue area by a specific length, cut out the point cloud slice of the outer region of the hook tongue nose, and calculate the normal of each point; the point cloud slice meets the measurement requirements of the hook tongue bulging detection template (i.e., standard TB / T 2048-2016, page 32, A.12 hook tongue bulging detection requirements);

[0086] When the measurement result of the feature to be measured is the wear of the S-surface of the hook tongue, determine the position of the edge surface of the hook tongue nose, and offset it by 50mm, 150mm and 250mm respectively along the normal direction of the edge surface to the hook tongue area. Extract point cloud slices of the three S-surfaces in the hook tongue nose area, and calculate the normal of each point. The point cloud slices meet the measurement requirements of the hook tongue S-surface wear measuring ruler (i.e., standard TB / T2048-2016, page 33, A.14 hook tongue S-surface wear detection requirements).

[0087] 5) Retrieve the pre-constructed virtual gauge point cloud template and use the directed distance constraint registration method to achieve accurate matching between the slices of the 3D point cloud obtained in step A and the virtual gauge point cloud template.

[0088] When the measurement result of the feature to be measured is wear of the hook and tongue lock surface, the virtual gauge point cloud template is the hook and tongue lock surface measuring ruler (i.e., standard TB / T 2048-2016, pages 7 and 8, Figure 13 Hook and Tongue Lock Surface Measuring Ruler).

[0089] When the measurement result of the feature to be measured is that the hook tongue is bulging, the virtual gauge point cloud template is the hook tongue bulging detection template (i.e., standard TB / T 2048-2016, page 7, Figure 12 Hook Tongue bulging detection template).

[0090] When the measurement result of the feature to be measured is the wear of the hook tongue S-surface, the virtual gauge point cloud template is the hook tongue S-surface wear measuring ruler (i.e., standard TB / T 2048-2016, pages 8 and 9, Figure 14 Hook tongue S-surface wear measuring ruler).

[0091] The registration method with directed distance constraints includes the following steps:

[0092] a) Using the virtual gauge point cloud template as a reference, transform the 3D point cloud slice data to the reference using the initial transformation matrix (R0,T0);

[0093] The initial transformation matrix (R0, T0) was confirmed manually, using a reference frame, and a reference target.

[0094] b) Using the corresponding position that needs to be matched with the object to be measured during the gauge inspection process as the target point, the nearest neighbor point in the 3D point cloud slice data is searched using the K-nearest neighbor algorithm to form the nearest neighbor-nearest neighbor point pair;

[0095] c) Construct an objective function E with the constraint of minimizing the directed distance between all nearest neighbor pairs, and calculate the residuals using optimization methods:

[0096] d) If the residual meets the threshold or exceeds the required number of iterations, stop the iteration and output the matching RT; otherwise, update the initial transformation matrix in step a) with this RT and repeat steps a) to c).

[0097] Furthermore, step c) constructs the objective function E using the following formula:

[0098]

[0099]

[0100]

[0101] : Three-dimensional rotation matrix;

[0102] Translation matrix;

[0103] The k-th iteration The three-dimensional coordinates of the source point;

[0104] The k-th iteration The three-dimensional coordinates of the target point corresponding to each source point;

[0105] The transpose of the unit normal vector at the target point;

[0106] Total number of iterations;

[0107] The number of points involved in the calculation during the k-th iteration;

[0108] Penalty factor, with a value of 10. 8 .

[0109] 6) Utilize the point-line constraints of the gauge to output the measurement results of the feature to be measured. Specifically, construct a rectangular coordinate system based on the test position of the gauge, obtain the reading of the feature to be measured in the rectangular coordinate system according to the gauge usage method, and take the difference between the reading and the reference value as the measurement result. The output of the test results shall conform to the requirements of standard TB / T 2048-2016.

[0110] The table below shows the data on hook tongue lock surface wear, hook tongue expansion detection, and hook tongue S-surface wear obtained using the method provided by this invention, as well as the results of manual measurement. It can be seen that this method has high accuracy and can completely replace manual measurement.

[0111]

[0112] The foregoing description of specific exemplary embodiments of the present invention is for illustrative and descriptive purposes. It is not intended to be exhaustive, nor to limit the invention to the precise forms disclosed; obviously, many changes and variations are possible in accordance with the foregoing teachings. The exemplary embodiments were chosen and described to explain the specific principles of the invention and its practical application, thereby enabling others skilled in the art to implement and utilize various exemplary embodiments of the invention, as well as their different alternatives and modifications. The scope of the invention is intended to be defined by the appended claims and their equivalents.

Claims

1. A method for measuring the characteristic dimensions of a hook tongue, characterized in that... The steps include the following: 1) The hook tongue is fixed to the test station, and multiple polyhedral targets are set around it; the robot carries a point cloud sensor to take pictures of the target targets in multiple poses, covering the feature area to be tested. When a point cloud sensor takes pictures in multiple poses, it can capture any feature surface of the same polyhedral target while acquiring the features of the target surface in two adjacent poses. The polyhedral target has at least one feature face, and each feature face has at least three non-collinear marker points; let the marker points on each polyhedral target be the point set C. k The value of k is a positive integer less than or equal to the number of polyhedral targets; the relative positions of the markers on each polyhedral target are known. 2) Under each photographic pose of the point cloud sensor, extract image features and solve the 3D point cloud to obtain the surface 3D point cloud data and the marker point 3D point cloud data under the current pose, which are recorded as the point cloud set of the current pose. 3) Let any point cloud be the source point set A, and the point cloud set of its adjacent positions be the target set B. Use the marker point data on the polyhedral target as a medium to realize the point cloud stitching between the source point set A and the target set B. 4) If the stitched point cloud obtained in step 3) completely covers the feature region to be tested, proceed directly to step A; Conversely, the stitched point cloud is denoted as the source point set A, and the point cloud set at its adjacent position is denoted as the target set B. The point cloud stitching between the source point set A and the target set B is achieved again using the marker point data on the polyhedral target as a medium. This operation is repeated until the stitched point cloud completely covers the feature region to be tested, and then step A is executed. Step A: Obtain the 3D point cloud slices and normals of the test location according to the test requirements of the feature to be tested; When the measurement result of the feature to be measured is wear of the hook tongue lock surface, determine the side where the hook tongue pin hole is located, offset it along its normal direction to the hook tongue area by a specific length, extract the point cloud slice of the hook tongue lock surface area, and calculate the normal of each point; the point cloud slice meets the measurement requirements of the hook tongue lock surface measuring ruler. When the measurement result of the feature to be measured is that the hook tongue is bulging outward, determine the side where the hook tongue pin hole is located, offset it along its normal direction to the hook tongue area by a specific length, cut out the point cloud slice of the outer region of the hook tongue nose, and calculate the normal of each point; the point cloud slice meets the measurement requirements of the hook tongue bulging detection template. When the measurement result of the feature to be measured is the wear of the S-surface of the hook tongue, determine the position of the edge surface of the hook tongue nose, and offset it to the hook tongue area by multiple different distances along the normal direction of the edge surface. Extract point cloud slices of multiple S-surfaces in the hook tongue nose area, and calculate the normal of each point. The point cloud slices meet the measurement requirements of the hook tongue S-surface wear measuring ruler. 5) Retrieve the pre-constructed virtual gauge point cloud template and use the directed distance constraint registration method to achieve accurate matching between the 3D point cloud slices obtained in step A and the virtual gauge point cloud template; When the measurement result of the feature to be measured is wear of the hook tongue lock surface, the virtual gauge point cloud template is a hook tongue lock surface measuring ruler; When the measurement result of the feature to be measured is that the hook tongue is bulging outward, the virtual gauge point cloud template is the hook tongue bulging detection template; When the measurement result of the feature to be measured is wear on the S-surface of the hook tongue, the virtual gauge point cloud template is a wear measuring ruler for the S-surface of the hook tongue. 6) Use the point and line constraints of the gauge to complete the measurement results of the feature to be measured.

2. The method for measuring the characteristic dimensions of a hook tongue as described in claim 1, characterized in that: The process of stitching point clouds between source point set A and target point set B includes the following steps: ① Let A1 be the three-dimensional point cloud data of the hook tongue surface in the source point set A, and A2 be the three-dimensional point cloud data of the hook tongue surface; Let B1 be the 3D point cloud data of the hook tongue surface in the target set B, and B2 be the 3D point cloud data of the marker point; Construct an error function using corresponding points in the overlapping regions of A1 and B1. ; ②Find the point set C k The markers corresponding to A2 and B2 are denoted as points in point set A2. Point set C k The corresponding point in Let point B2 be... Point set C k The corresponding point in Construct the error function: (R,T) is the position transformation matrix between the source point set A and the target point set B; It is a point set C k The position transformation matrix to point set A, where m is the number of marker points in A2 and n is the number of marker points in B2; ③ Construct the final objective function The objective function is solved using an optimization algorithm to obtain the position transformation matrix (R,T) and residual values ​​of the source point set A and the target point set B; M represents the penalty factor, which is a preset value; ④ Using the position transformation matrix (R,T) of the source point set A and the target point set B obtained in step ③, transform them to the same coordinate system, align them, merge them, and realize point cloud stitching.

3. The method for measuring the characteristic dimensions of a hook tongue as described in claim 2, characterized in that: In step ①, construct the error function. The method is as follows: ； in: and Let R be the set of points in the overlapping regions A1 and B1, respectively, where R is the rotation matrix, T is the translation matrix, and n is the number of points in the overlapping region involved in the calculation.

4. The method for measuring the characteristic dimensions of a hook tongue as described in claim 2, characterized in that: The optimization method is either the Levenberg-Marquardt algorithm or the Gaussian iteration method, where M > 10. 8 .

5. The method for measuring the characteristic dimensions of a hook tongue as described in claim 1, characterized in that: The regions to be tested are the outer region of the hook tongue nose, the hook tongue locking surface, and / or the S-surface region.

6. The method for measuring the characteristic dimensions of a hook tongue as described in claim 1, characterized in that: The registration method with directed distance constraints includes the following steps: a) Using the virtual gauge point cloud template as a reference, transform the 3D point cloud slice data to the reference using the initial transformation matrix (R0,T0); The initial transformation matrix (R0, T0) was confirmed manually, using a reference frame, and a reference target. b) Using the corresponding position that needs to be matched with the object to be measured during the gauge inspection process as the target point, the nearest neighbor point in the 3D point cloud slice data is searched using the K-nearest neighbor algorithm to form the nearest neighbor-nearest neighbor point pair; c) Construct an objective function E with the constraint of minimizing the directed distance between all nearest neighbor pairs, and calculate the residuals using optimization methods: d) If the residual meets the threshold or exceeds the required number of iterations, stop the iteration and output the matching RT; otherwise, update the initial transformation matrix in step a) with this RT and repeat steps a) to c). Step c) constructs the objective function E using the following formula: : Three-dimensional rotation matrix; Translation matrix; The k-th iteration The three-dimensional coordinates of the source point; The k-th iteration The three-dimensional coordinates of the target point corresponding to each source point; The transpose of the unit normal vector at the target point; Total number of iterations; The number of points involved in the calculation during the k-th iteration; Penalty factor, with a value of 10. 8 .

7. The method for measuring the characteristic dimensions of a hook tongue as described in claim 1, characterized in that: Step 6) The method for obtaining the measurement results of the feature to be measured using the point and line constraints of the gauge is as follows: Construct a rectangular coordinate system based on the test position of the gauge, and obtain the reading of the feature to be measured in the rectangular coordinate system according to the usage method of the gauge. The difference between the reading and the reference value is the measurement result.

Images