Coordinate measuring machine and coordinate measuring method

Abstract

The invention discloses a coordinate measuring machine and a coordinate measuring method, the coordinate measuring machine comprises a rack, a workbench, a moving frame and a measuring assembly, the workbench is arranged on the rack and is used for placing a workpiece to be measured; the measuring assembly comprises a laser scanning measuring head and a contact type measuring head, the laser scanning measuring head and the contact type measuring head are both arranged on the moving frame, the laser scanning measuring head is used for conducting non-contact type three-dimensional scanning measurement on the workpiece to be measured, and the contact type measuring head is used for conducting contact type precision measurement on the workpiece to be measured; wherein the movable frame can move relative to the to-be-measured workpiece so as to drive the contact type measuring head to be in contact with the to-be-measured workpiece for measurement, or drive the laser scanning measuring head to carry out scanning measurement under the condition that the laser scanning measuring head is not in contact with the to-be-measured workpiece. According to the technical scheme, the laser scanning measuring head and the contact type measuring head are integrated on the same coordinate measuring machine, seamless switching between non-contact type rapid scanning and contact type precision measurement is achieved, and the flexibility and efficiency of measurement are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of workpiece measurement, and in particular to a coordinate measuring machine and a coordinate measuring method. Background Art

[0002] Coordinate measuring machine is a kind of precision three-dimensional measuring instrument, which is widely used in industrial manufacturing. It can measure the geometric size, shape and position of workpieces with high precision by precisely controlling the movement of the probe in three-dimensional space. This kind of measuring equipment can quickly and accurately obtain the dimensional data of complex parts, providing key information for quality control and production optimization.

[0003] In the existing coordinate measurement technology, measuring machines are mainly divided into two types. One is to use a contact measurement method to obtain dimensional data by directly contacting the workpiece surface with a probe. This method has high accuracy but low efficiency. The other is to use non-contact laser scanning technology, which can quickly collect a large number of data points on the workpiece surface, but the accuracy is slightly insufficient in some precision measurement scenarios. This single-function design leads to the problem of limited application range of measuring equipment. In actual production, engineers often need to choose between these two types of measuring machines according to different measurement requirements, which not only increases the equipment investment cost, but also reduces the flexibility and efficiency of the measurement process. Summary of the invention

[0004] The main purpose of the present invention is to provide a coordinate measuring machine, aiming to solve the technical problem that the existing coordinate measuring machine can only perform contact measurement or non-contact measurement and cannot flexibly switch the measurement mode.

[0005] To achieve the above-mentioned purpose, the coordinate measuring machine proposed in the present invention includes a frame, a workbench, a movable frame and a measuring component, wherein the workbench is arranged on the frame, and the workbench is used to place a workpiece to be measured; the movable frame is movably arranged on the frame; the measuring component includes a laser scanning probe and a contact measuring head, and the laser scanning probe and the contact measuring head are both arranged on the movable frame, the laser scanning probe is used to perform non-contact three-dimensional scanning measurement of the workpiece to be measured, and the contact measuring head is used to perform contact precision measurement of the workpiece to be measured; wherein the movable frame can move relative to the workpiece to be measured to drive the contact measuring head to contact the workpiece to be measured for measurement, or drive the laser scanning probe to perform scanning measurement without contacting the workpiece to be measured.

[0006] Optionally, the movable frame includes a Z1 axis and a Z2 axis disposed on the top of the workbench, the Z1 axis and the Z2 axis extend in parallel and are arranged at intervals, and the Z1 axis and the Z2 axis can both move in the height direction of the frame; The laser scanning probe is arranged on the Z1 axis, and the contact type measuring head is arranged on the Z2 axis.

[0007] Optionally, the workbench includes a table body and a turntable rotatably disposed on the table body, a rotating shaft of the turntable extends in a height direction of the frame, and the turntable is used to place a workpiece to be measured.

[0008] The present invention also proposes a coordinate measurement method, which acquires point cloud data of a workpiece based on a laser scanning probe, extracts features from the point cloud data, and obtains feature information of the workpiece; Generate a sampling strategy for contact measurement based on the feature information, measure the workpiece using a contact measurement head according to the sampling strategy, and obtain discrete point data; fusing the point cloud data and the discrete point data to obtain a prediction model of the workpiece surface; Identify the area to be optimized based on the prediction model, perform supplementary measurement in the area to be optimized and update the prediction model, and repeat the process until a preset condition is met; The final prediction model is used to reconstruct the surface model of the workpiece, extract the geometric feature parameters of the workpiece, and calculate the statistical confidence intervals of the geometric feature parameters.

[0009] Optionally, acquiring point cloud data of the workpiece based on a laser scanning probe, and performing feature extraction on the point cloud data to obtain feature information of the workpiece include: Use the laser scanning probe to scan the workpiece in all directions to obtain the original point cloud data of the workpiece; Performing noise reduction and filtering processing on the original point cloud data to obtain optimized point cloud data; Performing multi-scale decomposition on the optimized point cloud data to obtain point cloud data sets at multiple scale levels; Performing geometric feature analysis on the point cloud data sets at the multiple scale levels to extract edge contours, surface curvature, local flatness, and micro-texture features of the workpiece; The extracted geometric features at each scale level are comprehensively analyzed to generate feature information of the workpiece.

[0010] Optionally, generating a sampling strategy for contact measurement based on the feature information, measuring the workpiece using a contact measurement head according to the sampling strategy, and acquiring discrete point data includes: Analyzing the feature information to identify key geometric feature areas and complex curved surface areas on the workpiece surface; Generate the initial sampling point distribution based on the identified key geometric feature areas and complex surface areas, combined with the preset measurement accuracy requirements; The initial sampling point distribution is optimized by using an adaptive density algorithm, and according to a preset curvature threshold and a flatness threshold, the number of sampling points is increased in an area exceeding the curvature threshold, and the number of sampling points is reduced in an area below the flatness threshold, so as to obtain an optimized sampling point distribution; Based on the optimized sampling point distribution and in combination with the spatial position information of the workpiece, a motion path of the contact measurement head is planned to generate a sampling strategy for contact measurement; Controlling the contact measurement head to move to a predetermined position of each sampling point according to the sampling strategy to perform contact measurement; The spatial coordinates and normal vector information of each sampling point are recorded to form a discrete point data set of the workpiece.

[0011] Optionally, the step of fusing the point cloud data and the discrete point data to obtain a prediction model of the workpiece surface includes: Performing a rigid body transformation on the point cloud data to align it and the discrete point data to the same coordinate system; Constructing a NURBS parameterized surface function, performing least squares fitting on the point cloud data, and obtaining an initial NURBS surface model; Using the discrete point data to locally modify the initial NURBS surface model, and obtaining an optimized NURBS surface model by adjusting the positions and weights of the control points; Calculating the Euclidean distance between the optimized NURBS surface model and the point cloud data and the discrete point data to generate a spatial error distribution field; Based on the spatial error distribution field, adaptively subdividing the optimized NURBS surface model, inserting new nodes and control points in areas where the error value exceeds a predetermined threshold; A weighted least squares algorithm is applied to optimize the subdivided NURBS surface model as a whole according to the respective measurement uncertainties of the point cloud data and the discrete point data, so as to obtain a prediction model of the workpiece surface.

[0012] Optionally, the identifying the area to be optimized based on the prediction model, performing supplementary measurement in the area to be optimized and updating the prediction model, and repeating the process until a preset condition is met, includes: Performing curvature analysis on the prediction model to calculate the Gaussian curvature and the average curvature of each point on the model surface; Based on a pre-set curvature threshold vector, the area beyond the threshold range is marked as a geometrically complex area; Performing deviation analysis on the prediction model and the original point cloud data and the discrete point data, and calculating the Euclidean distance from each data point to the model surface; According to a predetermined deviation tolerance, the area where the deviation exceeds the tolerance is marked as a high error area; Performing a Boolean intersection operation on the geometrically complex region and the high error region to obtain a region to be optimized; In the area to be optimized, a spatial clustering algorithm is used to generate a distribution of supplementary sampling points; Controlling the contact measuring head to measure the supplementary sampling points to obtain supplementary discrete point data; The supplemented discrete point data is merged with the original data set, and the prediction model is updated using the NURBS interpolation method; Calculate the root mean square error between the updated prediction model and all measured data points; Determine whether the root mean square error is less than a preset threshold, or whether the number of iterations reaches a preset upper limit; if both conditions are not met, return to step "performing curvature analysis on the prediction model" to continue execution; if any condition is met, terminate the iteration process.

[0013] Optionally, the step of reconstructing the surface model of the workpiece by using the final prediction model, extracting geometric feature parameters of the workpiece and calculating statistical confidence intervals of the geometric feature parameters includes: Meshing the final prediction model to generate a high-density triangular mesh model; Performing geometric feature segmentation and recognition on the triangular mesh model to obtain a set of basic geometric elements; Performing parameterized fitting on each element in the basic geometric element set to obtain a geometric parameter set; Using statistical simulation methods, uncertainty analysis is performed on the geometric parameter set to obtain parameter probability distribution; Calculating a statistical confidence interval for each geometric parameter based on the parameter probability distribution; Performing shape deviation and surface roughness analysis on the triangular mesh model to extract relevant feature parameters; Using resampling techniques, evaluating the stability of the characteristic parameters and calculating the statistical confidence intervals of the characteristic parameters; Integrate geometric parameters, form deviations and surface roughness information to generate a comprehensive geometric feature report of the workpiece.

[0014] First, this solution cleverly solves the problem of limited application range of single-function measuring machines in the background technology by setting up both laser scanning probes and contact measuring heads on the same mobile frame. This design allows users to flexibly choose the measurement method according to measurement needs without switching between different types of measuring machines. For example, when measuring a workpiece with a complex shape, you can first use a laser scanning probe to quickly obtain the overall shape data, and then use a contact measuring head to accurately measure the key dimensions. This measurement strategy combines the advantages of the two measurement methods, ensuring both measurement efficiency and high precision of key parts. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0016] Figure 1 It is a structural schematic diagram of an embodiment of a coordinate measuring machine of the present invention; Figure 2 Schematic diagram of a flow chart of a coordinate measurement method according to an embodiment of the present invention.

[0017] Description of Figure Numbers: 1. Frame; 2. Workbench; 21. Table body; 22. Turntable; 3. Moving frame; 31. Z1 axis; 32. Z2 axis; 33. Gantry guide rail; 34. Column; 35. Crossbeam guide rail; 36. Slide; 4. Measuring component; 41. Laser scanning probe; 42. Contact measuring head.

[0018] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back...), the directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0021] In addition, the descriptions of "first", "second", etc. in the present invention are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, "and / or" in the full text includes three solutions. Taking A and / or B as an example, it includes technical solution A, technical solution B, and technical solution that satisfies both A and B. In addition, the technical solutions between the various embodiments can be combined with each other, which must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0022] The invention provides a coordinate measuring machine.

[0023] In the embodiment of the present invention, Figure 1 As shown, the coordinate measuring machine includes a frame 1, a workbench 2, a movable frame 3 and a measuring assembly 4, wherein the workbench 2 is arranged on the frame 1, and the workbench 2 is used to place the workpiece to be measured; the movable frame 3 is movably arranged on the frame 1; the measuring assembly 4 includes a laser scanning probe 41 and a contact measuring head 42, and both the laser scanning probe 41 and the contact measuring head 42 are arranged on the movable frame 3, the laser scanning probe 41 is used to perform non-contact three-dimensional scanning measurement on the workpiece to be measured, and the contact measuring head 42 is used to perform contact precision measurement on the workpiece to be measured; wherein the movable frame 3 can move relative to the workpiece to be measured to drive the contact measuring head 42 to contact with the workpiece to be measured for measurement, or drive the laser scanning probe 41 to perform scanning measurement without contacting the workpiece to be measured.

[0024] Specifically, the frame 1 is the basis of the entire system. In some embodiments, it can be made of high-rigidity materials such as granite or precision cast iron to ensure stability and accuracy during the measurement process. The workbench 2 is set on the frame 1 to provide a stable and accurate placement platform for the workpiece to be measured.

[0025] The movable frame 3 is a key part of the present invention, which is movably arranged on the frame 1 and can adopt high-precision linear guide rails and servo systems to achieve multi-axis precision movement.

[0026] The laser scanning probe 41 uses optical principles to perform non-contact three-dimensional scanning measurement. It quickly collects a large number of data points on the surface of the workpiece by emitting a laser beam and receiving reflected light. It is suitable for rapid measurement of complex shapes and can generate a complete three-dimensional model of the workpiece in a short time. The contact measuring head 42 adopts the mechanical trigger principle and can use a high-precision ruby ​​ball probe. When the probe contacts the surface of the workpiece, the precise trigger mechanism can capture extremely accurate single-point coordinate data. Although this method has a slow measurement speed, it can provide the highest level of measurement accuracy and is suitable for precise measurement of critical dimensions.

[0027] The integrated installation of the two probes is a major feature of the present invention. They are cleverly arranged on the mobile frame 3, which not only ensures their respective measurement performance, but also enables fast switching. This design allows users to flexibly choose the measurement method according to specific needs, greatly improving the measurement efficiency and flexibility. For example, when measuring a complex workpiece, the laser scanning probe 41 can be used to quickly obtain the overall shape, and then the contact probe can be used to accurately measure the key dimensions. This combination gives full play to the advantages of the two measurement technologies.

[0028] In the actual measurement process, the mobile frame 3 can accurately control the movement of the two probes. When performing laser scanning, the mobile frame 3 drives the scanning probe to move along a predetermined path to quickly collect surface data. When performing contact measurement, the mobile frame 3 accurately controls the contact between the probe and the workpiece surface to ensure the accuracy of the measurement. This flexible measurement method not only improves efficiency, but also can adapt to various complex measurement tasks.

[0029] Optionally, the mobile frame 3 includes a Z1 axis 31 and a Z2 axis 32 disposed on the top of the workbench 2, the Z1 axis 31 and the Z2 axis 32 extend in parallel and are arranged at intervals, and the Z1 axis 31 and the Z2 axis 32 can both move in the height direction of the frame 1; The laser scanning probe 41 is disposed on the Z1 axis 31 , and the contact type measuring head 42 is disposed on the Z2 axis 32 .

[0030] Specifically, the mobile frame 3 also includes a gantry guide rail 33 located on one side of the workbench 2 and extending in the Y-axis direction, a column 34 slidably arranged on the gantry guide rail 33, a beam guide rail 35 extending in the X-axis direction is arranged on the column 34, and a slide 36 slidably arranged on the beam guide rail 35. The Z1 axis 31 and the Z2 axis 32 are both arranged on the slide 36, the plane formed by the X and Y axes is the horizontal plane, and the Z axis direction is the direction perpendicular to the plane, that is, the height direction of the frame 1. This design forms a complete three-dimensional coordinate system, so that the measuring head can move freely in three dimensions, thereby realizing comprehensive measurement of the workpiece.

[0031] When the column 34 moves in the Y-axis direction, it will drive the entire measurement system, including the laser scanning probe 41 and the contact measuring head 42, to move along the length direction of the workbench 2. When the slide 36 on the beam guide 35 slides in the X-axis, it will drive the Z1 axis 31 and the Z2 axis 32, as well as the probes they carry, to move along the width direction of the workbench 2. This design ensures that the measurement system can cover the entire plane range of the workpiece. The movement of the Z1 axis 31 and the Z2 axis 32 realizes the measurement in the height direction. The Z1 axis 31 drives the laser scanning probe 41 to move up and down, so that it can scan at different heights, and the Z2 axis 32 drives the contact measuring head 42 to move up and down, so that it can accurately contact different height positions on the surface of the workpiece.

[0032] The laser scanning probe 41 is arranged on the Z1 axis 31, and the contact measuring probe 42 is arranged on the Z2 axis 32. This design has significant advantages over the method of directly arranging the two probes on the column 34 or the slide 36. It increases the flexibility of measurement, allows two measurement methods to be performed simultaneously at different heights, and improves the measurement efficiency. Through independent Z-axis control, the height of each probe can be adjusted more accurately, reducing mutual interference and thus improving the measurement accuracy. In addition, the design of two Z axes allows the system to measure workpieces with a larger height range without being limited by the single Z-axis travel, thereby expanding the measurement range.

[0033] The movement of the column 34, the slide 36 and the Z1 axis 31 and the Z2 axis 32 can be achieved by a precise transmission system. Specifically, in some embodiments, the movement of the column 34 on the gantry guide rail 33 can adopt a gear rack transmission mechanism, wherein the rack is fixed on the gantry guide rail 33, and the gear meshing therewith is driven by a servo motor. The movement of the slide 36 on the beam guide rail 35 can use a ball screw transmission system, in which the ball screw is installed on the beam, the slide 36 is connected to the screw through a ball nut, and the servo motor drives the screw to rotate to achieve the translation of the slide 36. The vertical movement of the Z1 axis 31 and the Z2 axis 32 can also be driven by a ball screw, and each axis is equipped with an independent screw and servo motor to ensure accurate height control. These transmission systems are equipped with high-precision grating rulers or encoders for real-time monitoring and feedback of position information, thereby realizing closed-loop control and ensuring the accuracy and repeatability of movement.

[0034] Optionally, the workbench 2 includes a table body 21 and a turntable 22 rotatably disposed on the table body 21 , a rotating shaft of the turntable 22 extends in the height direction of the frame 1 , and the turntable 22 is used to place a workpiece to be measured.

[0035] Specifically, the turntable 22 can be rotated by manual operation or a driving member. The driving member can be a stepper motor or a servo motor, which can provide precise angle control. The motor is connected to the turntable 22 through a reducer to ensure smooth and precise rotation. The introduction of the turntable 22 achieves multiple beneficial effects: first, it allows the workpiece to rotate during the measurement process, so that the parts that are difficult to measure directly can be turned to a position that is easy to measure, greatly improving the comprehensiveness of the measurement. Secondly, through rotation, multi-angle scanning and measurement of the workpiece can be achieved, which is particularly useful for workpieces with complex shapes and can obtain more complete three-dimensional data. Furthermore, the use of the turntable 22 reduces the complex movement requirements of the measuring head, simplifies the measurement path planning, and improves the measurement efficiency. Finally, for workpieces with rotational symmetry, the turntable 22 can help achieve fast full-circle measurement and greatly shorten the measurement time. In general, the introduction of the turntable 22 significantly enhances the flexibility and measurement capabilities of the coordinate measuring machine, enabling it to better adapt to the measurement needs of various complex-shaped workpieces.

[0036] Another embodiment of the present application provides a coordinate measurement method. Figure 2 A flow chart of a coordinate measurement method provided in another embodiment of the present application. In this embodiment, the method includes: See also Figure 2 , acquiring point cloud data of the workpiece based on a laser scanning probe, performing feature extraction on the point cloud data, and obtaining feature information of the workpiece; In one embodiment of the present invention, the point cloud data of the workpiece is obtained based on a laser scanning probe, and features are extracted from the point cloud data to obtain feature information of the workpiece, including: using a laser scanning probe to perform an omnidirectional scan of the workpiece to obtain original point cloud data of the workpiece; performing noise reduction and filtering on the original point cloud data to obtain optimized point cloud data; performing multi-scale decomposition on the optimized point cloud data to obtain point cloud data sets of multiple scale levels; performing geometric feature analysis on the point cloud data sets of multiple scale levels to extract edge contours, surface curvature, local flatness and micro-texture features of the workpiece; and performing comprehensive analysis on the extracted geometric features of each scale level to generate feature information of the workpiece.

[0037] Specifically, first of all, point cloud data refers to a set of points in a three-dimensional coordinate system, each point has its specific X, Y, and Z coordinate values, which are used to represent the geometric information of the object surface. In this method, a laser scanning probe is used to scan the workpiece in all directions to obtain the original point cloud data of the workpiece. The laser scanning probe calculates the spatial position of the surface point of the object by emitting a laser beam and receiving the reflected light, thereby forming point cloud data. This step controls the moving frame to drive the laser scanning probe to perform a comprehensive scan around the workpiece to ensure that the complete geometric information of the workpiece is captured. For example, for a complex turbine blade, the laser scanning probe will move along a predetermined path and scan the blade from multiple angles, including the leading edge, trailing edge, pressure surface, and suction surface of the blade to obtain complete surface data.

[0038] After obtaining the original point cloud data, it needs to be subjected to noise reduction and filtering to obtain optimized point cloud data. Noise reduction refers to removing random errors and outliers in the data, while filtering is used to smooth the data and reduce high-frequency noise. This step aims to reduce the noise and outliers generated during the measurement process and improve data quality. Common methods include statistical outlier filtering and smoothing filtering. Statistical outlier filtering is based on the distance statistics between a point and its neighboring points to judge and delete outliers. Smoothing filtering reduces data fluctuations by weighted averaging the neighborhood of each point. Taking turbine blades as an example, there will be outliers caused by surface reflection or environmental interference. By setting appropriate thresholds, those points that are obviously deviated from the blade surface are deleted. At the same time, the remaining points are smoothed using methods such as moving average or Gaussian filtering to reduce surface noise and make the outline and details of the blade clearer.

[0039] Next, the optimized point cloud data is decomposed at multiple scales to obtain point cloud data sets at multiple scale levels. Multiscale decomposition is a technique that represents data at different spatial resolutions, which can be understood as observing an object from different "distances". The purpose of this step is to analyze the features of the workpiece at different spatial resolutions, which helps to capture comprehensive information from large-scale structures to subtle details. For turbine blades, multiscale decomposition can be performed using methods such as wavelet transform or Laplace pyramid. Wavelet transform decomposes the signal into different frequency components through specific mathematical transformations, while Laplace pyramid achieves multiscale representation by gradually reducing the resolution. At the coarsest scale, the overall shape and main curved surface of the blade can be obtained; at the intermediate scale, the leading and trailing edge contours of the blade can be identified; at the finest scale, the tiny textures and processing marks on the surface can be captured.

[0040] Then, geometric feature analysis is performed on the point cloud data sets at multiple scale levels to extract the edge contour, surface curvature, local flatness and micro-texture features of the workpiece. Geometric feature analysis involves a variety of mathematical and computational methods to quantitatively describe the shape characteristics of the object. The edge contour refers to the contour line of the object surface, which can be identified by detecting discontinuities in the point cloud. The surface curvature describes the degree of curvature of the surface, which is usually obtained by fitting the local surface and calculating the principal curvature. The local flatness indicates the degree of deviation of the surface from the ideal plane, which can be evaluated by least squares plane fitting. The micro-texture features reflect the subtle structure of the surface, which is usually quantified by analyzing the local height changes. For turbine blades, the principal curvature analysis method can be used to extract the main surface features of the blade at the coarse scale; at the medium scale, the edge detection algorithm can be applied to identify the contour line of the blade; at the fine scale, the local curvature and roughness parameters can be calculated to characterize the surface quality. For example, by analyzing the local curvature distribution, the uneven areas on the blade surface can be identified, which is crucial for evaluating the aerodynamic performance of the blade.

[0041] Finally, the extracted geometric features at each scale level are comprehensively analyzed to generate the feature information of the workpiece. This step integrates the feature information extracted at different scales to form a comprehensive feature description of the workpiece. The comprehensive analysis involves technologies such as feature fusion, statistical analysis, and pattern recognition, with the goal of generating a data set that can fully represent the geometric characteristics of the workpiece. For turbine blades, this includes the overall dimensional parameters of the blades (such as chord length, torsion angle, etc.), surface curvature distribution map, accurate description of edge profile, and statistical data of surface roughness. These comprehensive feature information provides an important basis for subsequent precise measurement and quality assessment.

[0042] This multi-step, multi-scale feature extraction method has many advantages. First, the full-scale scanning ensures the integrity of the data and avoids local omissions. Second, noise reduction and filtering processing improve data quality and reduce errors in subsequent analysis. Multi-scale decomposition allows the workpiece to be analyzed at different spatial resolutions, capturing both the overall shape and identifying subtle features. Separate geometric feature analysis can select the most appropriate algorithm for features of different scales, improving the accuracy and efficiency of feature extraction. Finally, the comprehensive analysis step integrates multi-scale information to provide a comprehensive and detailed description of the workpiece features. This method is particularly suitable for the measurement of complex-shaped workpieces, such as precision parts in industries such as aerospace and automotive manufacturing, and can meet the needs of high-precision and high-efficiency measurement. Through this comprehensive and systematic analysis, detailed geometric features of the workpiece can be obtained, providing key data support for subsequent precision machining, quality control, and performance evaluation.

[0043] Please continue reading Figure 2, generating a sampling strategy for contact measurement based on the feature information, measuring the workpiece using a contact measurement head according to the sampling strategy, and acquiring discrete point data; In one embodiment of the present invention, the method of generating a sampling strategy for contact measurement based on the feature information, measuring the workpiece using a contact measurement head according to the sampling strategy, and acquiring discrete point data includes: analyzing the feature information to identify key geometric feature areas and complex curved surface areas on the workpiece surface; generating an initial sampling point distribution based on the identified key geometric feature areas and complex curved surface areas in combination with preset measurement accuracy requirements; optimizing the initial sampling point distribution using an adaptive density algorithm, increasing the number of sampling points in an area exceeding the curvature threshold and reducing the number of sampling points in an area below the flatness threshold according to preset curvature thresholds and flatness thresholds, to obtain an optimized sampling point distribution; planning a motion path of the contact measurement head based on the optimized sampling point distribution and in combination with the spatial position information of the workpiece, and generating a sampling strategy for contact measurement; controlling the contact measurement head to move to a predetermined position of each sampling point according to the sampling strategy to perform contact measurement; and recording the spatial coordinates and normal vector information of each sampling point to form a discrete point data set of the workpiece.

[0044] First, the feature information is analyzed to identify the key geometric feature areas and complex surface areas on the workpiece surface. Feature information refers to the geometric features of the workpiece extracted from the point cloud data in the previous step, including edge contours, surface curvature, local flatness, and microtexture. Key geometric feature areas refer to areas that have a significant impact on the function or performance of the workpiece, such as edges, corners, or specific functional surfaces. Complex surface areas refer to surface parts with large curvature changes or complex shapes. For example, for a turbine blade, the leading edge and trailing edge of the blade may be identified as key geometric feature areas, while the curved surface of the blade may be identified as a complex surface area.

[0045] Based on the identified key geometric feature areas and complex surface areas, combined with the preset measurement accuracy requirements, an initial sampling point distribution is generated. The preset measurement accuracy requirements are determined based on the design specifications and quality control standards of the workpiece. The initial sampling point distribution refers to the preliminary plan for selecting measurement points on the workpiece surface. In this step, different densities of sampling points are assigned according to the importance and complexity of different areas. For example, a higher density of sampling points are assigned in critical areas such as the leading and trailing edges of turbine blades, while a lower density of sampling points are assigned in relatively flat areas. This strategy ensures that more measurement data is obtained in important areas while avoiding over-measurement in less important areas.

[0046] Next, the adaptive density algorithm is used to optimize the initial sampling point distribution. The adaptive density algorithm is a method that can automatically adjust the sampling point density according to the surface characteristics. The preset curvature threshold and flatness threshold are introduced here as judgment criteria. The curvature threshold is used to identify the curvature of the surface, while the flatness threshold is used to evaluate the flatness of the surface. The algorithm will increase the number of sampling points in areas exceeding the curvature threshold to better capture the shape of complex surfaces; and reduce the number of sampling points in areas below the flatness threshold to improve measurement efficiency. For example, for turbine blades, the algorithm will automatically increase the sampling points in areas with high curvature on the blade surface (such as the curved part of the leading edge or trailing edge); while in the relatively flat area in the middle of the blade, the sampling points will be appropriately reduced. This optimization ensures that the measurement efficiency is maximized while maintaining the measurement accuracy.

[0047] Based on the optimized sampling point distribution and combined with the spatial position information of the workpiece, the motion path of the contact measuring head is planned and the sampling strategy for contact measurement is generated. The spatial position information includes the position and orientation of the workpiece in the CMM workspace. Motion path planning needs to consider the motion efficiency of the measuring head, collision avoidance, and optimization of the measurement sequence. The sampling strategy includes not only the location of the sampling points, but also the approach direction of the measuring head and the measurement sequence. For example, for a turbine blade, the overall profile of the blade is measured first, and then the key areas are measured in detail, while ensuring that the motion path of the measuring head is smooth and efficient.

[0048] The contact measuring head is then controlled to move to the predetermined position of each sampling point according to the sampling strategy to perform contact measurement. This step involves precise motion control of the measuring machine and the trigger mechanism of the measuring head. The contact measuring head usually contains a high-precision trigger sensor, which accurately records the coordinates of the trigger point when the probe touches the workpiece surface. For the measurement of turbine blades, the measuring head moves precisely to each predetermined sampling point, touches the surface and records the data, and then moves to the next point. This process requires precise motion control to ensure the accuracy and repeatability of the measurement.

[0049] Finally, the spatial coordinates and normal vector information of each sampling point are recorded to form a discrete point data set of the workpiece. Spatial coordinates refer to the precise position of the sampling point in three-dimensional space, usually expressed as X, Y, and Z coordinates. The normal vector refers to the vertical direction of the surface where the point is located, which provides the orientation information of the surface at that point. The normal vector is very important for subsequent surface reconstruction and feature analysis. A discrete point data set is a set of data containing these spatial coordinates and normal vector information, which represents the precise geometric information of the workpiece surface. For example, for a turbine blade, this data set contains hundreds or thousands of measurement points, each of which has its precise position and surface orientation information.

[0050] This adaptive sampling method based on feature information has many advantages. First, it can adjust the measurement strategy according to the specific characteristics of the workpiece to ensure that sufficient measurement data is obtained in key areas. Secondly, through the adaptive density algorithm, the measurement efficiency can be optimized while ensuring the measurement accuracy and reducing unnecessary measurement points. Furthermore, path planning combined with the spatial position information of the workpiece can improve the efficiency and safety of the measurement process. Finally, by recording the coordinates and normal vector information of each point, comprehensive and accurate basic data is provided for subsequent data analysis and surface reconstruction. This method is particularly suitable for high-precision measurement of workpieces with complex shapes, such as precision parts in industries such as aerospace and automotive manufacturing. It can significantly improve measurement efficiency while ensuring measurement accuracy, providing key support for product quality control and production optimization.

[0051] Please continue reading Figure 2 , fusing the point cloud data and the discrete point data to obtain a prediction model of the workpiece surface; In one embodiment of the present invention, the point cloud data and the discrete point data are fused to obtain a prediction model of the workpiece surface, including: performing a rigid body transformation on the point cloud data to align it with the discrete point data to the same coordinate system; constructing a NURBS parameterized surface function, performing least squares fitting on the point cloud data, and obtaining an initial NURBS surface model; locally correcting the initial NURBS surface model using the discrete point data, and obtaining an optimized NURBS surface model by adjusting the positions and weights of control points; calculating the Euclidean distance between the optimized NURBS surface model and the point cloud data and the discrete point data to generate a spatial error distribution field; based on the spatial error distribution field, performing adaptive subdivision on the optimized NURBS surface model, and inserting new nodes and control points in areas where the error value exceeds a predetermined threshold; applying a weighted least squares algorithm, and optimizing the subdivided NURBS surface model as a whole according to the respective measurement uncertainties of the point cloud data and the discrete point data, to obtain a prediction model of the workpiece surface.

[0052] Specifically, first, a rigid body transformation is performed on the point cloud data to align it with the discrete point data to the same coordinate system. Rigid body transformation refers to the operation of translating and rotating an object without changing the shape and size of the object. The purpose of this step is to ensure that the data from two different sources are consistent in spatial position. For example, for a turbine blade, the point cloud data obtained by laser scanning will have a slight deviation in spatial position from the discrete point data obtained by contact measurement. Through rigid body transformation, the two sets of data can be accurately aligned, laying the foundation for subsequent fusion processing.

[0053] Next, a NURBS (Non-Uniform Rational B-Splines) parametric surface function is constructed, and the point cloud data is fitted with the least squares method to obtain the initial NURBS surface model. NURBS is a mathematical model used to represent complex surfaces. It has the characteristics of high flexibility and high precision and is widely used in computer-aided design and manufacturing. Least squares fitting is a mathematical optimization technique used to find the best function match for a set of data. In this step, a preliminary mathematical model of the workpiece surface can be obtained by NURBS fitting a large amount of point cloud data. Taking the turbine blade as an example, this initial NURBS model can roughly reflect the overall shape and main surface features of the blade.

[0054] Then, the initial NURBS surface model is locally modified using the discrete point data, and the optimized NURBS surface model is obtained by adjusting the control point positions and weights. Discrete point data usually comes from high-precision contact measurements and has high local accuracy. Control points are key points that define the shape of the NURBS surface, and the weights determine the degree of influence of the control points on the surface shape. By adjusting these parameters, the NURBS model can be made to better conform to the high-precision discrete point data. For example, in key areas of turbine blades (such as the leading and trailing edges), the control points are adjusted to ensure that the model is highly consistent with the actual measurement data in these areas.

[0055] Next, the Euclidean distance between the optimized NURBS surface model and the point cloud data and discrete point data is calculated to generate a spatial error distribution field. Euclidean distance is the straight-line distance between two points in three-dimensional space. By calculating the distance between a point on the model surface and the nearest measured point, an error distribution graph representing the accuracy of the model can be obtained. The spatial error distribution field intuitively shows the fit quality of the model in different areas. For turbine blades, this error distribution field will show that the error is larger in complex surface areas (such as the curved part of the blade) and smaller in flat areas.

[0056] Based on the spatial error distribution field, the optimized NURBS surface model is adaptively subdivided, and new nodes and control points are inserted in areas where the error value exceeds a predetermined threshold. Adaptive subdivision is a technique that dynamically adjusts the complexity of the model based on local features. The predetermined threshold is an acceptable error range set according to the measurement accuracy requirements. By adding nodes and control points in high error areas, the fitting accuracy of the model in these areas can be improved. For example, if a certain surface area of ​​a turbine blade is found to have a large error, more control points will be added in this area so that the model can more accurately describe the complex local geometry.

[0057] Finally, the weighted least squares algorithm is applied to optimize the subdivided NURBS surface model as a whole according to the measurement uncertainties of the point cloud data and discrete point data, and obtain the prediction model of the workpiece surface. The weighted least squares algorithm takes into account the reliability of different data sources and gives higher weights to more reliable data. The measurement uncertainty reflects the credibility of the measurement results. Usually, the uncertainty of contact measurement is lower, while the uncertainty of laser scanning is higher. Through this weighted optimization, the advantages of the two measurement methods can be fully utilized to obtain an optimal surface model that combines high-density point cloud data and high-precision discrete point data.

[0058] This fusion processing method has many advantages. First, it combines the high efficiency of laser scanning and the high precision of contact measurement, making full use of the advantages of the two measurement methods. Secondly, the NURBS model provides a continuous and smooth surface representation, which is conducive to subsequent analysis and processing. Furthermore, adaptive subdivision and weighted optimization ensure that the model has sufficient accuracy in key areas while avoiding overfitting. Finally, the prediction model generated by this method not only accurately reflects the actual geometry of the workpiece, but also contains measurement uncertainty information, providing a reliable basis for quality control and tolerance analysis. This fusion method is particularly suitable for high-precision modeling of complex-shaped workpieces, such as precision parts in industries such as aerospace and automotive manufacturing, and can provide key support for product design, manufacturing and quality control.

[0059] Please continue reading Figure 2 , identifying the area to be optimized based on the prediction model, performing supplementary measurements in the area to be optimized and updating the prediction model, and repeating the process until a preset condition is met; In one embodiment of the present invention, the method of identifying the area to be optimized based on the prediction model, performing supplementary measurements in the area to be optimized and updating the prediction model, and repeating the process until a preset condition is met, includes: performing curvature analysis on the prediction model, calculating the Gaussian curvature and the average curvature of each point on the model surface; based on a preset curvature threshold vector, marking the area beyond the threshold range as a geometrically complex area; performing deviation analysis on the prediction model and the original point cloud data and discrete point data, and calculating the Euclidean distance from each data point to the model surface; based on a predetermined deviation tolerance, marking the area where the deviation exceeds the tolerance as a high error area; and separating the geometrically complex area from the high error area. The method comprises the following steps: performing a Boolean intersection operation on the region to obtain the region to be optimized; generating the distribution of supplementary sampling points by using a spatial clustering algorithm in the region to be optimized; controlling the contact measuring head to measure the supplementary sampling points to obtain supplementary discrete point data; merging the supplementary discrete point data with the original data set, and updating the prediction model by using the NURBS interpolation method; calculating the root mean square error between the updated prediction model and all the measured data points; judging whether the root mean square error is less than a preset threshold value, or whether the number of iterations reaches a preset upper limit; if both conditions are not met, returning to the step of "performing a curvature analysis on the prediction model" to continue the execution; if any condition is met, terminating the iteration process.

[0060] Specifically, first, the prediction model is subjected to curvature analysis, and the Gaussian curvature and mean curvature of each point on the model surface are calculated. Curvature is an important geometric feature that describes the degree of curvature of a surface. Gaussian curvature refers to the product of the two principal curvatures of a point on the surface, while mean curvature is the arithmetic mean of the two principal curvatures. These two types of curvature provide different information about the local shape of the surface. For example, for a turbine blade, different parts of the blade surface will have different curvature values. Highly curved areas such as the leading edge and the trailing edge have larger curvature values, while relatively flat areas in the middle have smaller curvatures.

[0061] Next, based on the pre-set curvature threshold vector, the areas beyond the threshold range are marked as geometrically complex areas. The curvature threshold vector is a set of predefined curvature values ​​used to determine the complexity of the surface. This threshold is set based on the design specifications or manufacturing requirements of the workpiece. Marking areas beyond these thresholds as geometrically complex areas means that these areas require more intensive measurements or more sophisticated modeling. Taking turbine blades as an example, the leading edge, trailing edge, and some special functional surfaces of the blades will be marked as geometrically complex areas.

[0062] The predicted model is then analyzed for deviations from the original point cloud data and the discrete point data, and the Euclidean distance from each data point to the model surface is calculated. The Euclidean distance is the straight-line distance between two points in three-dimensional space. The purpose of this step is to evaluate how well the predicted model matches the actual measured data. For turbine blades, this means calculating the shortest distance from each measured point (whether it is point cloud data from laser scanning or discrete point data from contact measurement) to the predicted model surface.

[0063] Based on the pre-determined deviation tolerance, areas where the deviation exceeds the tolerance are marked as high error areas. The deviation tolerance is the maximum acceptable difference between the model and the actual data. This tolerance value is usually set based on the accuracy requirements of the workpiece. Areas that exceed this tolerance are considered to be areas where the model fit is not good enough and need further optimization. For example, if certain areas of a turbine blade, such as complex cooling channel outlets, show large deviations, these areas will be marked as high error areas.

[0064] The Boolean intersection of the geometrically complex areas and the high error areas is performed to obtain the areas to be optimized. The Boolean intersection operation refers to finding areas that meet both conditions at the same time. The purpose of this step is to identify those areas that are both geometrically complex and have large fitting errors, which are most in need of optimization and additional measurements. For turbine blades, this includes certain parts of the leading and trailing edges, or special functional areas on the blade surface.

[0065] In the area to be optimized, a spatial clustering algorithm is used to generate the distribution of additional sampling points. The spatial clustering algorithm is a method for reasonably distributing points in three-dimensional space. It ensures that the additional sampling points are evenly distributed in the area to be optimized while avoiding being too close to existing measurement points. This maximizes the amount of information from the additional measurement points. For example, in the area to be optimized of a turbine blade, the algorithm generates a series of new sampling points that are distributed in the gaps between the original measurement points.

[0066] The contact probe is controlled to measure the additional sampling points to obtain additional discrete point data. This step involves precise control of the coordinate measuring machine. The contact probe is guided to each newly generated sampling point position for high-precision point measurement. These new measurement points provide additional geometric information of the area to be optimized, which helps to improve the overall accuracy of the model.

[0067] The supplementary discrete point data is merged with the original data set, and the prediction model is updated using the NURBS interpolation method. NURBS (Non-Uniform Rational B-Spline) interpolation is an advanced surface modeling technique that can create smooth and continuous surfaces based on discrete data points. By incorporating newly measured points into the model, NURBS interpolation can locally adjust the model shape to better conform to the actual geometry. This process involves adjusting the position and weight of the control points to optimize the fit of the surface.

[0068] Calculate the root mean square error between the updated prediction model and all measured data points. The root mean square error is an important indicator for evaluating the overall fit quality of the model. It is obtained by calculating the square average of the distances from all measured points to the model surface and then taking the square root. The smaller this value is, the better the fit between the model and the actual measured data.

[0069] Finally, determine whether the root mean square error is less than the preset threshold or whether the number of iterations reaches the preset upper limit. This is a key step in deciding whether to continue the optimization process. The preset threshold represents the maximum acceptable error level, and the preset upper limit is to prevent infinite loops. If the root mean square error is still greater than the preset threshold and the number of iterations has not reached the upper limit, return to the curvature analysis step and start a new round of optimization. If any termination condition is met, the iteration process ends.

[0070] This iterative optimization method has many advantages. First, it can automatically identify areas that need to be optimized and focus resources where improvements are most needed. Second, by combining geometric complexity and fitting errors, high-precision measurement results are ensured in critical areas. Third, the iterative process allows the model to be gradually improved until the predetermined accuracy requirements are achieved. Finally, this method can effectively balance measurement accuracy and efficiency, avoiding over-measurement or under-measurement. This method is particularly effective for workpieces with complex shapes, such as aircraft engine components or precision mechanical parts, and can optimize the use of measurement resources while ensuring high accuracy.

[0071] Please continue reading Figure 2 , using the final prediction model to reconstruct the surface model of the workpiece, extract the geometric feature parameters of the workpiece and calculate the statistical confidence intervals of the geometric feature parameters.

[0072] In one embodiment of the present invention, the method of reconstructing the surface model of the workpiece using the final prediction model, extracting the geometric feature parameters of the workpiece and calculating the statistical confidence intervals of the geometric feature parameters includes: meshing the final prediction model to generate a high-density triangular mesh model; performing geometric feature segmentation and identification on the triangular mesh model to obtain a set of basic geometric elements; performing parameter fitting on each element in the set of basic geometric elements to obtain a set of geometric parameters; performing uncertainty analysis on the set of geometric parameters using a statistical simulation method to obtain a parameter probability distribution; calculating the statistical confidence interval of each geometric parameter based on the parameter probability distribution; performing shape deviation and surface roughness analysis on the triangular mesh model to extract relevant feature parameters; using resampling technology to evaluate the stability of the feature parameters and calculate the statistical confidence intervals of the feature parameters; integrating geometric parameters, shape deviation and surface roughness information to generate a comprehensive geometric feature report of the workpiece.

[0073] Specifically, first, the final prediction model is meshed to generate a high-density triangular mesh model. Meshing is the process of converting a continuous surface model into a discrete triangular mesh. A high-density triangular mesh model can accurately approximate a complex surface with a large number of small triangles. For example, for a turbine blade, meshing will generate hundreds of thousands or even millions of small triangles to accurately describe every subtle surface change of the blade. This high-density mesh model provides a detailed and accurate digital representation for subsequent geometric analysis.

[0074] Next, the triangular mesh model is segmented and identified for geometric features to obtain a set of basic geometric elements. Geometric feature segmentation refers to the process of decomposing a complex surface model into multiple basic geometric shapes. Basic geometric elements include planes, cylinders, spheres, cones, etc. For turbine blades, this step will identify the main curved surface of the blade, the cylindrical shapes of the leading and trailing edges, and the existing flat parts (such as the blade root). This segmentation and identification helps to simplify the analysis of complex shapes, making subsequent parameterization and measurement more accurate and efficient.

[0075] Then, parametric fitting is performed on each element in the set of basic geometric elements to obtain a set of geometric parameters. Parametric fitting refers to the use of mathematical equations to describe the shape and size of each basic geometric element. For example, for the cylindrical surface (such as the leading edge) identified in the turbine blade, its radius and axial direction are extracted as parameters; for the plane part, its normal vector and position information are extracted. These parameters accurately define the geometric features of the workpiece and provide a basis for subsequent analysis and quality assessment.

[0076] The uncertainty analysis of the set of geometric parameters is performed using statistical simulation methods to obtain the parameter probability distribution. Statistical simulation methods, such as Monte Carlo simulation, evaluate the variability of the parameters through multiple random sampling and calculations. This step takes into account various uncertainties in the measurement process, such as the accuracy of the measuring instrument, the influence of environmental conditions, etc. For each geometric parameter of the turbine blade, such as the radius of the leading edge, this analysis will produce a probability distribution that represents the range of values ​​of this parameter and its probability.

[0077] Based on the parameter probability distribution, a statistical confidence interval is calculated for each geometric parameter. A statistical confidence interval provides a range within which the true value of the parameter may fall, usually accompanied by a confidence level (such as 95% confidence). For example, for the radius of the leading edge of a turbine blade, a conclusion might be drawn such as "with 95% confidence, the true radius is between 4.98 mm and 5.02 mm." This statement not only gives an estimate of the parameter, but also quantifies the uncertainty of the estimate.

[0078] The triangular mesh model is subjected to shape deviation and surface roughness analysis to extract relevant feature parameters. Shape deviation analysis compares the difference between the actual shape and the ideal shape, while surface roughness analysis evaluates the microstructure of the surface. For turbine blades, shape deviation analysis evaluates the degree of deviation of the blade profile from the design model; surface roughness analysis measures the microscopic undulations on the blade surface. These analyses provide important indicators of the geometric quality of the workpiece.

[0079] Resampling techniques are used to evaluate the stability of characteristic parameters and calculate statistical confidence intervals of characteristic parameters. Resampling techniques, such as the bootstrap method, simulate the process of repeated measurements by randomly sampling multiple times from the original data. This method can evaluate the variability of characteristic parameters (such as shape deviation and surface roughness indicators) and calculate their statistical confidence intervals. For example, for the surface roughness parameters of turbine blades, it may be concluded that "with 95% confidence, the average roughness Ra is between 0.8μm and 1.2μm."

[0080] Finally, the geometric parameters, shape deviation and surface roughness information are integrated to generate a comprehensive geometric feature report of the workpiece. This report summarizes all the key geometric features of the workpiece, including basic dimensions, shape deviation, surface quality, etc., as well as the statistical confidence intervals of these features. For turbine blades, the report includes key dimensions of the blades (such as chord length, torsion angle, etc.), profile deviation, surface roughness, etc., and each parameter is accompanied by its statistical confidence interval.

[0081] This comprehensive method of geometric feature analysis and statistical evaluation has many advantages. First, it provides a comprehensive description of the workpiece's geometric features, including macroscopic shape and microscopic surface characteristics. Second, through statistical analysis and calculation of confidence intervals, the uncertainty of the measurement results is quantified, providing a reliable basis for quality control and decision-making. Furthermore, this method takes into account various uncertainties in the measurement process and improves the reliability and interpretability of the results. Finally, the generation of a comprehensive report provides comprehensive and detailed information support for product design, manufacturing process optimization and quality control. This method is particularly suitable for complex workpieces with high precision requirements, such as key components in the fields of aerospace, precision machinery, etc., and can significantly improve the accuracy and reliability of product quality control.

[0082] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. All equivalent structural changes made by using the contents of the present invention specification and drawings under the inventive concept of the present invention, or directly / indirectly applied in other related technical fields are included in the patent protection scope of the present invention.

Claims

1. A coordinate measuring machine, characterized in that: include: Rack (1); A workbench (2) is arranged on the frame (1), and the workbench (2) is used to place a workpiece to be measured; A movable frame (3) movably arranged on the frame (1); A measuring assembly (4), comprising a laser scanning probe (41) and a contact measuring head (42), wherein the laser scanning probe (41) and the contact measuring head (42) are both arranged on the moving frame (3), the laser scanning probe (41) is used for performing non-contact three-dimensional scanning measurement on the workpiece to be measured, and the contact measuring head (42) is used for performing contact precision measurement on the workpiece to be measured; The movable frame (3) is capable of moving relative to the workpiece to be measured, so as to drive the contact measuring head (42) to contact the workpiece to be measured for measurement, or to drive the laser scanning measuring head (41) to perform scanning measurement without contacting the workpiece to be measured.

2. The coordinate measuring machine according to claim 1, characterized in that The movable frame (3) comprises a Z1 axis (31) and a Z2 axis (32) arranged on the top of the workbench (2); the Z1 axis (31) and the Z2 axis (32) extend in parallel and are arranged at intervals; and the Z1 axis (31) and the Z2 axis (32) can both move in the height direction of the frame (1); The laser scanning probe (41) is arranged on the Z1 axis (31), and the contact type measuring head (42) is arranged on the Z2 axis (32).

3. The coordinate measuring machine according to claim 1, characterized in that The workbench (2) comprises a table body (21) and a turntable (22) rotatably arranged on the table body (21); the rotation axis of the turntable (22) extends in the height direction of the frame (1); and the turntable (22) is used for placing a workpiece to be measured.

4. A coordinate measurement method, characterized in that: The coordinate measurement method is applied to the coordinate measuring machine according to any one of claims 1 to 3, and the coordinate measurement method comprises: Acquire point cloud data of the workpiece based on a laser scanning probe, perform feature extraction on the point cloud data, and obtain feature information of the workpiece; Generate a sampling strategy for contact measurement based on the feature information, measure the workpiece using a contact measurement head according to the sampling strategy, and obtain discrete point data; fusing the point cloud data and the discrete point data to obtain a prediction model of the workpiece surface; Identify the area to be optimized based on the prediction model, perform supplementary measurement in the area to be optimized and update the prediction model, and repeat the process until a preset condition is met; The final prediction model is used to reconstruct the surface model of the workpiece, extract the geometric feature parameters of the workpiece, and calculate the statistical confidence intervals of the geometric feature parameters.

5. The coordinate measurement method according to claim 4, characterized in that: The step of acquiring point cloud data of a workpiece based on a laser scanning probe and extracting features from the point cloud data to obtain feature information of the workpiece includes: Use the laser scanning probe to scan the workpiece in all directions to obtain the original point cloud data of the workpiece; Performing noise reduction and filtering processing on the original point cloud data to obtain optimized point cloud data; Performing multi-scale decomposition on the optimized point cloud data to obtain point cloud data sets at multiple scale levels; Performing geometric feature analysis on the point cloud data sets at the multiple scale levels to extract edge contours, surface curvature, local flatness, and micro-texture features of the workpiece; The extracted geometric features at each scale level are comprehensively analyzed to generate feature information of the workpiece.

6. The coordinate measurement method according to claim 4, characterized in that: The step of generating a sampling strategy for contact measurement based on the feature information, measuring the workpiece using a contact measurement head according to the sampling strategy, and acquiring discrete point data includes: Analyzing the feature information to identify key geometric feature areas and complex curved surface areas on the workpiece surface; Generate the initial sampling point distribution based on the identified key geometric feature areas and complex surface areas, combined with the preset measurement accuracy requirements; The initial sampling point distribution is optimized by using an adaptive density algorithm, and according to a preset curvature threshold and a flatness threshold, the number of sampling points is increased in an area exceeding the curvature threshold, and the number of sampling points is reduced in an area below the flatness threshold, so as to obtain an optimized sampling point distribution; Based on the optimized sampling point distribution and in combination with the spatial position information of the workpiece, a motion path of the contact measurement head is planned to generate a sampling strategy for contact measurement; Controlling the contact measurement head to move to a predetermined position of each sampling point according to the sampling strategy to perform contact measurement; The spatial coordinates and normal vector information of each sampling point are recorded to form a discrete point data set of the workpiece.

7. The coordinate measurement method according to claim 4, characterized in that: The step of fusing the point cloud data and the discrete point data to obtain a prediction model of the workpiece surface includes: Performing a rigid body transformation on the point cloud data to align it and the discrete point data to the same coordinate system; Constructing a NURBS parameterized surface function, performing least squares fitting on the point cloud data, and obtaining an initial NURBS surface model; Using the discrete point data to locally modify the initial NURBS surface model, and obtaining an optimized NURBS surface model by adjusting the positions and weights of the control points; Calculating the Euclidean distance between the optimized NURBS surface model and the point cloud data and the discrete point data to generate a spatial error distribution field; Based on the spatial error distribution field, adaptively subdividing the optimized NURBS surface model, inserting new nodes and control points in areas where the error value exceeds a predetermined threshold; A weighted least squares algorithm is applied to optimize the subdivided NURBS surface model as a whole according to the respective measurement uncertainties of the point cloud data and the discrete point data, so as to obtain a prediction model of the workpiece surface.

8. The coordinate measurement method according to claim 4, characterized in that: The step of identifying the area to be optimized based on the prediction model, performing supplementary measurement in the area to be optimized and updating the prediction model, and repeating the steps until a preset condition is met, includes: Performing curvature analysis on the prediction model to calculate the Gaussian curvature and the average curvature of each point on the model surface; Based on a pre-set curvature threshold vector, the area beyond the threshold range is marked as a geometrically complex area; Performing deviation analysis on the prediction model and the original point cloud data and the discrete point data, and calculating the Euclidean distance from each data point to the model surface; According to a predetermined deviation tolerance, the area where the deviation exceeds the tolerance is marked as a high error area; Performing a Boolean intersection operation on the geometrically complex region and the high error region to obtain a region to be optimized; In the area to be optimized, a spatial clustering algorithm is used to generate a distribution of supplementary sampling points; Controlling the contact measuring head to measure the supplementary sampling points to obtain supplementary discrete point data; The supplemented discrete point data is merged with the original data set, and the prediction model is updated using the NURBS interpolation method; Calculate the root mean square error between the updated prediction model and all measured data points; Determine whether the root mean square error is less than a preset threshold, or whether the number of iterations reaches a preset upper limit; if both conditions are not met, return to step "performing curvature analysis on the prediction model" to continue execution; if any condition is met, terminate the iteration process.

9. The coordinate measurement method according to claim 4, characterized in that: The method of reconstructing the surface model of the workpiece by using the final prediction model, extracting the geometric feature parameters of the workpiece and calculating the statistical confidence interval of the geometric feature parameters includes: Meshing the final prediction model to generate a high-density triangular mesh model; Performing geometric feature segmentation and recognition on the triangular mesh model to obtain a set of basic geometric elements; Performing parameterized fitting on each element in the basic geometric element set to obtain a geometric parameter set; Using statistical simulation methods, uncertainty analysis is performed on the geometric parameter set to obtain parameter probability distribution; Calculating a statistical confidence interval for each geometric parameter based on the parameter probability distribution; Performing shape deviation and surface roughness analysis on the triangular mesh model to extract relevant feature parameters; Using resampling techniques, evaluating the stability of the characteristic parameters and calculating the statistical confidence intervals of the characteristic parameters; Integrate geometric parameters, form deviations and surface roughness information to generate a comprehensive geometric feature report of the workpiece.

Images