Water turbine top cover inner cavity global measurement method based on mobile base station

Abstract

According to the water turbine top cover inner cavity global measurement method based on the mobile base station, accurate acquisition of section point cloud data is achieved by constructing a camera and a structured light model. A unified reference is provided for global point cloud splicing by building a mobile reference station, a coordinate conversion formula is deduced to unify point cloud data under all the references, and finally the three-dimensional shape of an inner cavity is reconstructed. The invention discloses a water turbine top cover inner cavity global measurement method based on a mobile base station. The method comprises the following steps: building a measurement system; camera calibration and structure cursor calibration; fixing a reference mobile camera, and obtaining point cloud poses of all positions under the same reference; fixing a camera moving reference, and obtaining a conversion relation between the two references through camera shooting and formula derivation; continuously replacing the mobile camera and the base station and measuring; and the plurality of local point cloud models form an inner cavity overall three-dimensional point cloud model. A coordinate conversion formula is deduced to achieve unification of point cloud coordinate reference, global measurement of the inner cavity is completed, the space efficiency and the detection precision are high, and the method can adapt to complex working conditions.

Description

Technical Field

[0001] The invention belongs to the technical field of machine vision measurement, and relates to a method for measuring the entire inner cavity of a turbine top cover based on a mobile reference station. Background Art

[0002] As a key device for converting water energy into mechanical energy, the operation stability and safety of the turbine are crucial to the entire hydropower station. However, the turbine cover has a narrow and long cavity structure, which is subjected to the impact of high-speed water flow, vibration, temperature changes and other harsh working conditions for a long time. The inner cavity is prone to defects such as cracks, wear, and corrosion. Since the defects are generated randomly and the positions are not fixed, it is necessary to conduct full-area cavity detection and find out the defect locations to provide a basis for maintenance. However, there are limitations in the traditional detection of the inner cavity of the turbine cover: on the one hand, the traditional detection method is limited by the large size of the turbine cover structure, the small closed space, the large area to be inspected, and the high difficulty of detection. On the other hand, the traditional detection method has problems such as low detection efficiency, insufficient accuracy, and susceptibility to environmental interference. Visual measurement technology, with its high precision, high efficiency, strong penetration and high visualization, shows great potential in the internal detection of complex and narrow spaces. Therefore, studying the full-area visual measurement method of the inner cavity of the turbine cover is of great significance to improving the operation and maintenance level of the hydropower industry.

[0003] In 2020, Wang Xinxin from Inner Mongolia University of Science and Technology published a paper titled "Design of Structured Light Vision Inspection System for Pipeline Inner Surface", which proposed a method to use one-dimensional guide rails, lasers, cameras and computer systems to capture light strip deformation through laser projection and camera, combined with pipeline movement control, laser and camera position adjustment and image processing technology to achieve three-dimensional morphology detection and reconstruction of the inner surface of short experimental pipelines. However, due to the material strength of the fixed bracket and the length limit of the pipeline to be measured, it is not suitable for long-distance measurement.

[0004] In 2023, Hu Chengcheng from Tianjin University published a paper titled "Research on the Three-Dimensional Morphology Measurement System and Key Technologies for the Inner Surface of Narrow and Long Pipelines". The paper designed an axial distance measurement and spatial attitude orientation device to measure the axial displacement and rotation angle of the pipeline crawler, and achieved the unification of local scanning data in the same coordinate system. Finally, the precise splicing of each segment of point cloud data was achieved through line structured light scanning, multi-step calibration and posture solution. However, since the paper can only achieve axial measurement for straight pipelines, it is unable to measure circular orbits.

[0005] In summary, existing lumen detection methods have significant limitations. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide a method for full-area measurement of the inner cavity of a turbine top cover based on a mobile reference station. By introducing a mobile reference station system, full-area, non-contact measurement of the inner cavity of the turbine top cover is achieved, the measurement accuracy and efficiency are improved, and reliable data support is provided for the maintenance and management of the turbine.

[0007] In order to solve the above technical problems, the technical solution adopted by the present invention is: a method for measuring the entire inner cavity of a turbine top cover based on a mobile reference station, comprising the following steps:

[0008] S1, build the measurement system, using camera, laser projector, camera track trolley, mobile reference station, laser projector fixture, high-precision calibration plate and calibration plate track trolley to build the measurement system;

[0009] S2, calibration, combining camera calibration and structured light calibration to obtain the three-dimensional point cloud coordinates in the camera coordinate system;

[0010] S3, morphology acquisition, fix the base station, and move the measurement component multiple times to obtain point cloud data of different sections in the camera coordinate system, and use the coordinate system transformation relationship to obtain the pose of each point cloud under its own reference;

[0011] S4, fix the measurement component, move the reference station, capture the image, calculate the external parameters to determine the conversion relationship between the two references;

[0012] S5, replace the mobile measurement component and the mobile reference station to realize data acquisition and reference transfer, and complete the three-dimensional point cloud stitching process around the inner cavity.

[0013] In S1, the specific process of setting up the measurement system includes:

[0014] S1-1, the bottom of the camera track trolley is connected to the laser projector fixture by threaded connection to ensure the accuracy and stability of 3D imaging;

[0015] S1-2, connect the camera track trolley and the camera with screws to ensure that the camera remains fixed when the camera track trolley is running;

[0016] S1-3, a high-precision calibration plate with a micro high-brightness backlight, which can produce a clear and bright calibration plate image from the camera;

[0017] S1-4, fix the high-precision calibration plate on the calibration plate track trolley to ensure that the posture between the two vehicles is changed multiple times to obtain external parameters.

[0018] In S2, the center of the initial high-precision calibration plate is the origin O, the direction perpendicular to the calibration plate is the Z axis, and the directions parallel to the two sides of the calibration plate are established to form the world coordinate system O. w -X w Y w Zw ; Set the camera coordinate system to O c -X c Y c Z c , with the camera optical center O c As the origin of the coordinate system, Z c The X axis is consistent with the camera optical axis. c Axis and Y c The camera coordinate system is established by using the plane formed by the axes to be parallel to the imaging plane; the pixel coordinate system is O-UV, with the upper left corner of the image as the origin and the two edges of the image at right angles to each other as the UV axes.

[0019] In S2, during the detection process, the camera internal and external parameters are obtained according to the standard parameter information of the high-precision calibration plate, and the specific formula for building the camera imaging model is as follows:

[0020]

[0021]

[0022] Where: (u,v) T is the pixel coordinate of the object image; (u0,v0) is the coordinate of the principal point of the camera capturing the image, (f x ,f y ) is the scale factor in the two-axis direction of the image coordinate system, which constitutes the internal parameter matrix W; R is the rotation matrix, T is the translation matrix, which constitutes the external parameter M; (X w ,Y w ,Z w ) T is the object image (assuming the calibration plate is in the world coordinate system Z w = 0) in the plane); H is the perspective projection matrix; Z c is the scale factor;

[0023] After obtaining the camera intrinsic parameters through camera calibration, ensure that the camera and laser projector are rigidly connected through the fixture, turn on the laser projector to project the annular structured light; adjust the position of the high-precision calibration plate so that it is within the irradiation range of the annular structured light, and calculate the linear equation of any laser stripe in the camera coordinate system by analyzing the laser stripes projected on the high-precision calibration plate and using the camera calibration parameters. The complete equation of the structured light plane is obtained by combining multiple sets of linear equations as follows:

[0024] ax c +by c +cz c +d=0(3).

[0025] In S3, turn on the laser projector and start the camera to capture an image with an annular structured light; let the projection point of the intersection of the annular structured light and the inner cavity in the image be P = (u, v, 1) T , the coordinates of the point in the camera coordinate system are P′=(X c ′,Y c ′,Z c ′) and the camera optical center O c =(0,0,0) and we get the straight line O c P′, we can get the straight line O c P′ passes through the structured light plane point coordinate P″=(X c ″,Y c ″,Z c "),satisfy:

[0026]

[0027] Where: r = (r x ,r y ,r z ) is a point on the structured light plane, and its normal vector is For the straight line O c P′ direction vector;

[0028] The coordinates of the intersections of different straight lines and the structured light plane in the camera coordinate system are obtained through repeated calculations. The coordinates of multiple intersections constitute a single cross-section point cloud. A single cross-section point cloud data acquisition method is adopted to continuously change the position of the mobile base station and use the camera to obtain each cross-section point cloud, which is convenient for the subsequent construction of the three-dimensional model of the inner cavity.

[0029] In S4, the calibration plate track trolley is moved to ensure that there is enough distance between the camera and the high-precision calibration plate, so that the calibration plate image is clearly within the camera field of view; first, the high-precision calibration plate is kept stationary, and different positions of the calibration plate are taken as the local coordinate system O wi -X wi Y wi Z wi (i=1,2,...,n) and the initial position of the calibration plate is taken as the global coordinate system O w1 -X w1 Y w1 Z w1 ; Secondly, turn on the laser projector and obtain the structured light point cloud image through camera shooting. The coordinates P of each point can be calculated by formula 4 ci =(X ci ,Y ci ,Z ci ) T (i=1,2,...,n); Then, taking the initial position as an example, analyze the image taken by the camera to obtain the external parameter M1, and realize the point cloud coordinate from O according to formula 5 c -Xc Y c Z c Convert to O w1 -X w1 Y w1 Z w1 ; Finally, the camera track car is moved n times to shoot n point cloud images. According to formula 5, n sets of point cloud poses in the local coordinate system can be obtained by multiple calculations. The specific formula for the coordinate system conversion is as follows:

[0030]

[0031] Where: (X wi ,Y wi ,Z wi ) T is the coordinate of the point cloud in the i-th local coordinate system; M i Convert the point cloud camera coordinate system to the extrinsic parameter of the i-th local coordinate system.

[0032] In S5, during the movement of the camera track trolley, the high-precision calibration plate will exceed the camera field of view and the camera will be out of focus, so the moving object is replaced; during this period, the local coordinate system of the high-precision calibration plate is changed from O to w1 -X w1 Y w1 Z w1 Move to O w2 -X w2 Y w2 Z w2 At this time, it is only necessary to realize the coordinate transformation between the world coordinate systems; after moving the position of the high-precision calibration plate, open the camera 1 to shoot and obtain O w2 -X w2 Y w2 Z w2 Convert to O c -X c Y c Z c The external parameter M1′, with the camera coordinate system as the intermediate quantity, can realize the transformation from the local coordinate system to the global coordinate system. The specific formula is as follows:

[0033]

[0034] Where: M n is the nth external parameter in the i-th local coordinate system; M i ′ is the external parameter of the i+1th local coordinate system converted to the camera coordinate system.

[0035] Repeat S4 and S5, keep the high-precision calibration plate still, and use the second local coordinate system O w2 -X w2 Y w2 Zw2 As the reference, move the camera track car to change the camera posture under the second reference. At this time, move the camera n times to shoot n calibration plate images, and get the point cloud coordinates in O c -X c Y c Z c Downconvert to O w2 -X w2 Y w2 Z w2 The external parameter M i According to formula 6, multiple local coordinate systems O w2 -X w2 Y w2 Z w2 The point cloud pose is transformed to the global coordinate system O w1 -X w1 Y w1 Z w1 , realize point cloud stitching; repeat steps S4 and S5, so that the camera track trolley and the calibration plate track trolley travel along the inner cavity for one week, during which time the point cloud data is continuously collected and processed, and the obtained three-dimensional point cloud coordinates are uniformly converted to the global coordinate system O w1 -X w1 Y w1 Z w1 Thus forming a complete three-dimensional point cloud model of the inner cavity.

[0036] The main beneficial effects of the present invention are:

[0037] For closed and narrow cavities, a mobile reference station is built to complete the precise transformation of point cloud data through camera calibration and structural optical calibration. The point cloud data sets under each benchmark are obtained with the high-precision calibration plate as the benchmark, and the coordinate conversion formula is derived to achieve the unification of the point cloud coordinate benchmark, completing the full-domain measurement of the inner cavity, with high spatial efficiency and detection accuracy, and can adapt to complex working conditions. In addition, the camera and the high-precision calibration plate are equipped with a track car, which improves the degree of automation of the measurement and is highly feasible. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0039] Figure 1 This is the structural diagram of the intracavity measurement system based on the mobile reference station.

[0040] Figure 2 This is a flow chart of the full-area measurement of the inner cavity of the top cover according to the method of the present invention.

[0041] Figure 3 This is a schematic diagram of the point cloud splicing principle under the reference of the present invention.

[0042] Figure 4Schematic diagram of the principle of realizing reference transfer of point cloud coordinates in the present invention.

[0043] Figure 5 Schematic diagram comparing the theoretical model of the inner cavity of the top cover of the present invention and the reconstruction result.

[0044] Explanation of the reference numerals: 1-camera, 2-laser projector, 3-high-precision calibration plate, 4-camera track trolley, 5-mobile base station, 6-calibration plate track trolley, 7-laser projector fixture. DETAILED DESCRIPTION

[0045] like Figure 1 to Figure 5 A method for measuring the entire inner cavity of a turbine top cover based on a mobile reference station comprises the following steps:

[0046] S1, build the measurement system, using camera, laser projector, camera track trolley, mobile reference station, laser projector fixture, high-precision calibration plate and calibration plate track trolley to build the measurement system;

[0047] S2, calibration, combining camera calibration and structured light calibration to obtain the three-dimensional point cloud coordinates in the camera coordinate system;

[0048] S3, morphology acquisition, fix the base station, and move the measurement component multiple times to obtain point cloud data of different sections in the camera coordinate system, and use the coordinate system transformation relationship to obtain the pose of each point cloud under its own reference;

[0049] S4, fix the measurement component, move the reference station, capture the image, calculate the external parameters to determine the conversion relationship between the two references;

[0050] S5, replace the mobile measurement component and the mobile reference station to realize data acquisition and reference transfer, and complete the three-dimensional point cloud stitching process around the inner cavity.

[0051] Embodiment 1,

[0052] like Figure 1 In the embodiment, the intracavity measurement system includes 1- camera, 2- laser projector, 3- high-precision calibration plate, 4- camera track trolley, 5- mobile base station, 6- calibration plate track trolley, 7- laser projector fixture.

[0053] Figure 3 This is a schematic diagram of the principle of point cloud stitching under the reference of the present invention, wherein the dotted line position is the initial position of the camera 1, and the solid line position is the actual position of the camera 1 after movement.

[0054] Attached Figure 4 Schematic diagram of the principle of realizing reference transfer of point cloud coordinates of the present invention, wherein the dotted line position is the initial position of the high-precision calibration plate 3, and the solid line position is the actual position of the high-precision calibration plate 3 after movement.

[0055] Figure 5 It is a schematic diagram comparing the theoretical model of the inner cavity of the top cover of the present invention and the reconstruction result, wherein the reconstruction result is a local three-dimensional model constructed by splicing multiple cross-section point clouds.

[0056] A camera 1 with a resolution of 1440×1080 pixels, a signal-to-noise ratio of 41dB, and a pixel size of 3.45μm×3.45μm is selected, and a laser projector 2 with a minimum line width of 50μm and a divergence angle of 45° is selected, and the two are rigidly connected to each other. The high-precision calibration plate 3 is a solid circle calibration plate made of ceramic material, with a size of 10mm×10mm, a number of solid circles of 7×7, a solid circle diameter of 0.625mm, and a solid circle spacing of 1.25mm. The inner cavity structure is a circular ring, and its cavity inner diameter is 760mm and the thickness is 20mm.

[0057] Embodiment 2,

[0058] The following is a detailed description of the specific measurement implementation:

[0059] (1) Measurement system construction

[0060] First, use screws and bolts to fix the camera 1 and the laser projector 2 to the trolley and the fixture respectively. The camera track trolley 4 and the laser projector fixture 7 are connected by bolts to ensure the tightness of the connection. Secondly, the camera 1 is connected to the camera track trolley 4 by screws, and the three constitute the measurement part. Finally, the high-precision calibration plate 3 is fixed to the back plate of the calibration plate track trolley 6 by screws, and the micro high-brightness backlight source is turned on while moving the calibration plate track trolley 6. When a clear and bright calibration plate image can be obtained with the help of the camera 1, stop moving the calibration plate track trolley 6 to facilitate the calibration of the camera model.

[0061] (2) Camera and structured light positioning

[0062] The present invention adopts Zhang's calibration method to solve the internal and external parameters of the camera, uses the camera 1 to shoot 12 images of the high-precision calibration plate 3 at different angles, and analyzes and calculates the corner points of the chessboard to obtain the internal parameter matrix The structured light calibration of the present invention adopts the laser stripe fitting light plane method. The annular structured light is projected onto the high-precision calibration plate 3 with the help of the laser projector 2. By changing the position of the high-precision calibration plate 3, the equations of the two laser stripe fitting lines in the camera coordinate system can be obtained, and then the coefficients of the structured light plane equation are obtained as a=-0.018030, b=-0.015171, c=0.190667, and d=0.074542.

[0064] (3) Acquisition of the three-dimensional morphology of the inner cavity surface

[0065] 1) Point cloud data acquisition

[0066] After obtaining the camera internal parameters and determining the structured light plane, move the calibration board track trolley 6 to the depth of field range in front of the camera 1, and keep the laser projector 2 turned on. Turn on the camera 1 to shoot the annular structured light point cloud image, and the point cloud data set can be obtained as Ω=(p1,p2,…,p n ), where p i =(u,v,1) T , the point cloud dataset Ω′=(p1′,p2′,…,p n ′). Each point is related to the camera optical center O c =(0,0,0) and we get the straight line O c P i ′(i=1,2,…,n) Line equation, the line O can be obtained by calculating the formula 4 c P i ′ passes through the structured light plane in the camera coordinate system point cloud data set Ω″=(p1″,p2″,…,p n ″), the poses of the point cloud datasets are combined to obtain the shape of a single cross-section point cloud.

[0067] 2) The calibration board is stationary, and the camera is moving

[0068] Keep the calibration plate track trolley 6 stationary within the depth of field of the camera 1, and take the position of the high-precision calibration plate 3 as the global coordinate system O w1 -X w1 Y w1 Z w1 The camera track trolley 4 moves 8 times to obtain 8 sets of images containing the high-precision calibration plate 3, ensuring that the images are clear and complete and that one image is taken each time it moves. The 8 external parameters obtained through software analysis and mathematical calculations are:

[0069]

[0070]

[0071]

[0072]

[0073] Through step 1), the point cloud of camera 1 at 8 positions can be obtained as Ω i ″=(p1″,p2″,…,p n ″)(i=1,2,...,8), calculate according to formula 5 to obtain w1 -X w1 Y w1 Z w1 The coordinates of the next point cloud Ω wi ″=(p w1 ″,pw2 ″,…,p wn ″)(i=1,2,...,8), a total of 8 coordinate system transformations can be combined into a three-dimensional point cloud image under a single reference.

[0074] 3) The camera is stationary and the calibration plate is moving

[0075] After completing the combination of point cloud graphics under the same reference, the camera track trolley 4 remains stationary, and the camera 1 has moved to its eighth position. Since the high-precision calibration plate 3 will exceed the field of view of the camera 1 and the camera 1 will be defocused, the calibration plate track trolley 6 is moved again to the depth of field range of the camera 1, and the second reference coordinate system of the calibration plate here is taken as O w2 -X w2 Y w2 Z w2 , turn on camera 1 to shoot and obtain external parameters through mathematical analysis According to formula 6, the point cloud coordinate Ω under the second reference can be realized wi ″=(p w1 ″,p w2 ″,…,p wn ″)(i=9,10,...,16) is converted to the global coordinate system O w1 -X w1 Y w1 Z w1 Corresponding coordinates Ω i ″′=(p1″′,p2″′,…,p n ″′)(i=9,10,...,16), complete the local point cloud stitching.

[0076] 4) Repeated measurement of point cloud pose information

[0077] After building the local point cloud model, first keep the calibration plate track trolley 6 stationary in the second reference coordinate system. Then, move the camera track trolley 4 8 times again and obtain a high-precision calibration plate 1 image through camera 1 to ensure that the image is clear and complete. Taking the point cloud at the ninth position of camera 1 as an example, the external parameter is obtained by combining the mathematical analysis of the calibration plate image. The camera coordinate point cloud of camera 1 at the ninth position obtained by step 1 is Ω9″=(p1″,p2″,…,p n ″), the second reference coordinate system O can be calculated according to formula 5 w2 -X w2 Y w2 Z w2 Point cloud is Ω w9 ″=(p w1 ″,p w2 ″,…,p wn According to formula 6, the point cloud can be transformed into the global coordinate system O w1 -X w1Y w1 Z w1 Corresponding point cloud Ω9″′=(p1″′,p2″′,…,p n ″′). Repeat the measurement, move the camera track trolley 4 and the calibration plate track trolley 6 around the inner cavity for one cycle, and during this period, continuously repeat steps 2) and 3) to build the overall point cloud model.

[0078] After building the point cloud stitching model, according to Figure 5 The stitching model of the point cloud is selected from multiple representative test data sets. By processing the test data set, the fitting inner diameter of the inner cavity is obtained to be 759.88mm. The fitting data is compared with the standard value (inner diameter 760mm), and the deviation is found to be 0.12mm. The model after point cloud stitching is highly consistent with the real structure, with a small deviation, which fully meets the needs of high-precision detection.

[0079] In the above implementation process, based on the construction of the mobile reference station 5, the accurate capture of the annular point cloud pose information is achieved. Faced with the closed and space-constrained inner cavity environment, the present invention establishes a mobile reference station 5, uses the high-precision calibration plate 3 as a reference benchmark, and uses the camera 1 to effectively collect the point cloud data set under each benchmark, which significantly improves the accuracy and reliability of the detection. The present invention builds a camera model and a ring structured light model to accurately obtain the point cloud pose. Based on this, the mobile measurement component and the reference benchmark are replaced, and the coordinate conversion formula is derived, which not only realizes the benchmark unification of the point cloud coordinates, but also breaks through the spatial limitations of traditional measurement technology.

[0080] The above-mentioned embodiments are only preferred technical solutions of the present invention and should not be regarded as limitations of the present invention. The embodiments and features in the embodiments of the present application can be arbitrarily combined with each other without conflict. The protection scope of the present invention shall be the technical solutions recorded in the claims, including the equivalent replacement solutions of the technical features in the technical solutions recorded in the claims. That is, equivalent replacement improvements within this scope are also within the protection scope of the present invention.

Claims

1. A method for measuring the entire inner cavity of a turbine top cover based on a mobile reference station, characterized in that: The steps include: S1, build the measurement system, using camera, laser projector, camera track trolley, mobile reference station, laser projector fixture, high-precision calibration plate and calibration plate track trolley to build the measurement system; S2, calibration, combining camera calibration and structured light calibration to obtain the three-dimensional point cloud coordinates in the camera coordinate system; S3, morphology acquisition, fix the base station, and move the measurement component multiple times to obtain point cloud data of different sections in the camera coordinate system, and use the coordinate system transformation relationship to obtain the pose of each point cloud under its own reference; S4, fix the measurement component, move the reference station, capture the image, calculate the external parameters to determine the conversion relationship between the two references; S5, replace the mobile measurement component and the mobile reference station to realize data acquisition and reference transfer, and complete the three-dimensional point cloud stitching process around the inner cavity.

2. The method for measuring the inner cavity of a turbine top cover based on a mobile reference station according to claim 1 is characterized in that: In S1, the specific process of setting up the measurement system includes: S1-1, the bottom of the camera track trolley is connected to the laser projector fixture by threaded connection to ensure the accuracy and stability of 3D imaging; S1-2, connect the camera track trolley and the camera with screws to ensure that the camera remains fixed when the camera track trolley is running; S1-3, a high-precision calibration plate with a micro high-brightness backlight, which can produce a clear and bright calibration plate image from the camera; S1-4, fix the high-precision calibration plate on the calibration plate track trolley to ensure that the posture between the two vehicles is changed multiple times to obtain external parameters.

3. The method for measuring the entire inner cavity of a turbine top cover based on a mobile reference station according to claim 1 is characterized in that: In S2, the center of the initial high-precision calibration plate is the origin O, the direction perpendicular to the calibration plate is the Z axis, and the directions parallel to the two sides of the calibration plate are established to form the world coordinate system O. w -X w Y w Z w ; Set the camera coordinate system to O c -X c Y c Z c , with the camera optical center O c As the origin of the coordinate system, Z c The X axis is consistent with the camera optical axis. c Axis and Y c The camera coordinate system is established by using the plane formed by the axes to be parallel to the imaging plane; the pixel coordinate system is O-UV, with the upper left corner of the image as the origin and the two edges of the image at right angles to each other as the UV axes.

4. The method for measuring the entire inner cavity of a turbine top cover based on a mobile reference station according to claim 3 is characterized in that: In S2, during the detection process, the camera internal and external parameters are obtained according to the standard parameter information of the high-precision calibration plate, and the camera imaging model is built. The specific formula is as follows: Where: (u,v) T is the pixel coordinate of the object image; (u0,v0) is the coordinate of the principal point of the camera capturing the image, (f x ,f y ) is the scale factor in the two-axis direction of the image coordinate system, which constitutes the internal parameter matrix W; R is the rotation matrix, T is the translation matrix, which constitutes the external parameter M; (X w ,Y w ,Z w ) T is the object image (assuming the calibration plate is in the world coordinate system Z w = 0) in the plane); H is the perspective projection matrix; Z c is the scale factor; After obtaining the camera intrinsic parameters through camera calibration, ensure that the camera and laser projector are rigidly connected through the fixture, turn on the laser projector to project the annular structured light; adjust the position of the high-precision calibration plate so that it is within the irradiation range of the annular structured light, and calculate the linear equation of any laser stripe in the camera coordinate system by analyzing the laser stripes projected on the high-precision calibration plate and using the camera calibration parameters. The complete equation of the structured light plane is obtained by combining multiple sets of linear equations as follows: ax c +by c +cz c +d=0 (3)。 5. The method for measuring the entire inner cavity of a turbine top cover based on a mobile reference station according to claim 1 is characterized in that: In S3, turn on the laser projector and start the camera to capture an image with an annular structured light; let the projection point of the intersection of the annular structured light and the inner cavity in the image be P = (u, v, 1) T , the coordinates of the point in the camera coordinate system are P′=(X c ′,Y c ′,Z c ′) and the camera optical center O c =(0,0,0) and we get the straight line O c P′, we can get the straight line O c P′ passes through the structured light plane point coordinate P″=(X c ″,Y c ″,Z c "),satisfy: Where: r = (r x ,r y ,r z ) is a point on the structured light plane, and its normal vector is For the straight line O c P′ direction vector; Accumulate calculations to obtain the coordinates of the intersections of different straight lines and the structured light plane in the camera coordinate system. Multiple intersection coordinates form a single cross-section point cloud. A single cross-section point cloud data acquisition method is adopted, the position of the mobile base station is continuously changed, and the camera is used to obtain the point cloud of each cross section, which is convenient for the subsequent construction of the inner cavity three-dimensional model.

6. The method for measuring the entire inner cavity of a turbine top cover based on a mobile reference station according to claim 1 is characterized in that: In S4, the calibration plate track trolley is moved to ensure that there is enough distance between the camera and the high-precision calibration plate, so that the calibration plate image is clearly within the camera field of view; first, the high-precision calibration plate is kept stationary, and different positions of the calibration plate are taken as the local coordinate system O wi -X wi Y wi Z wi (i=1,2,...,n) and the initial position of the calibration plate is taken as the global coordinate system O w1 -X w1 Y w1 Z w1 ; Secondly, turn on the laser projector and obtain the structured light point cloud image through camera shooting. The coordinates P of each point can be calculated by formula 4 ci =(X ci ,Y ci ,Z ci ) T (i=1,2,...,n); Then, taking the initial position as an example, analyze the image taken by the camera to obtain the external parameter M1, and realize the point cloud coordinate from O according to formula 5 c -X c Y c Z c Convert to O w1 -X w1 Y w1 Z w1 ; Finally, the camera track car is moved n times to shoot n point cloud images. According to formula 5, n sets of point cloud poses in the local coordinate system can be obtained by multiple calculations. The specific formula for the coordinate system conversion is as follows: Where: (Xwi, Ywi, Zwi) T is the coordinate of the point cloud in the i-th local coordinate system; M i Convert the point cloud camera coordinate system to the extrinsic parameter of the i-th local coordinate system.

7. The method for measuring the entire inner cavity of a turbine top cover based on a mobile reference station according to claim 1 is characterized in that: In S5, during the movement of the camera track trolley, the high-precision calibration plate will exceed the camera field of view and the camera will be out of focus, so the moving object is replaced; during this period, the local coordinate system of the high-precision calibration plate is changed from O to w1 -X w1 Y w1 Z w1 Move to O w2 -X w2 Y w2 Z w2 At this time, it is only necessary to realize the coordinate transformation between the world coordinate systems; after moving the position of the high-precision calibration plate, open the camera 1 to shoot and obtain O w2 -X w2 Y w2 Z w2 Convert to O c -X c Y c Z c The external parameter M1′, with the camera coordinate system as the intermediate quantity, can realize the transformation from the local coordinate system to the global coordinate system. The specific formula is as follows: Where: M n is the nth external parameter in the i-th local coordinate system; M i ′ is the external parameter of the i+1th local coordinate system converted to the camera coordinate system.

8. The method for measuring the entire inner cavity of a turbine top cover based on a mobile reference station according to claim 1 is characterized in that: Repeat S4 and S5, keep the high-precision calibration plate still, and use the second local coordinate system O w2 -X w2 Y w2 Z w2 As the reference, move the camera track car to change the camera posture under the second reference. At this time, move the camera n times to shoot n calibration plate images, and get the point cloud coordinates in O c -X c Y c Z c Downconvert to O w2 -X w2 Y w2 Z w2 The external parameter M i According to formula 6, multiple local coordinate systems O w2 -X w2 Y w2 Z w2 The point cloud pose is transformed to the global coordinate system O w1 -X w1 Y w1 Z w1 , realize point cloud stitching; repeat steps S4 and S5, so that the camera track trolley and the calibration plate track trolley travel along the inner cavity for one week, during which time the point cloud data is continuously collected and processed, and the obtained three-dimensional point cloud coordinates are uniformly converted to the global coordinate system O w1 -X w1 Y w1 Z w1 Thus forming a complete three-dimensional point cloud model of the inner cavity.

Images