Laser measurement system suitable for large complex curved surface and measurement method thereof

Abstract

The invention discloses a laser measurement system suitable for a large complex curved surface and a measurement method thereof.The measurement system comprises a control module, a positioning module, a pre-scanning module and a high-precision scanning module, and the control module comprises a high-precision scanning planning unit and a scanning analysis unit; the high-precision scanning module comprises a three-dimensional scanning platform, a universal mechanical arm and a distance measuring laser; the measurement method comprises the steps of pre-scanning to generate an initial model, scanning path planning and scanning attitude control, high-precision scanning to collect point cloud data, preprocessing and marking of the high-precision scanning point cloud data, local optimization and supplementary scanning strategy generation and high-precision scanning model generation. According to the measurement system and method, the dynamic scanning path of the high-precision scanning module is intelligently planned based on the pre-scanning result, the pre-scanning and high-precision scanning strategies are combined, high-density coverage and uniform sampling of the large complex curved surface are achieved, measurement is accurate, the anti-interference capacity is high, and the model scanning precision is high; a dynamic supplementary scanning mechanism is automatically marked and triggered in the scanning process, and the quality of scanned data is ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of laser processing, and particularly relates to a laser measurement system suitable for large complex curved surfaces and a measurement method thereof. BACKGROUND

[0002] The laser measurement system is a high-precision measurement equipment with a laser sensor as a core and a multi-axis motion platform to realize non-contact three-dimensional data acquisition. In the system, laser path planning is one of the key links to ensure measurement coverage, data integrity and measurement efficiency. That is, under the premise of known geometric information of a target object, a group of spatial scanning trajectories and corresponding laser attitude parameters are generated to guide the laser head to move in a specific way and collect data. Reasonable path planning can avoid occlusion and blind area, improve scanning density, reduce redundant motion, and thus improve overall measurement accuracy and efficiency.

[0003] For large complex curved surfaces, the existing laser scanning paths are mostly preset straight lines or uniform sampling methods. Common strategies include: regular grid scanning method: the target is divided into regular grid areas, and linear scanning is performed in sequence, which is suitable for flat or slowly varying curved surfaces; equal interval slicing method: the three-dimensional model is sliced along the fixed axis direction, and a closed contour path is generated on each layer, which is used for laser measurement similar to CT slicing; equal distance surface unfolding method: the three-dimensional curved surface is unfolded into a two-dimensional plane and then the path is planned, and the measurement accuracy is easily affected by the curvature change; static attitude + uniform path strategy: in most actual applications, the incident direction of the laser head remains in a fixed attitude (i.e. not adjusted with the curvature or normal of the curved surface), and only the motion trajectory of the laser head in the plane (such as straight line, grid, etc. uniform path) is controlled to realize the scanning coverage of the curved surface.

[0004] The above strategies have many problems. For example, path planning relies on manual setting and is difficult to adapt to complex geometric changes. In high-curvature, occlusion, and concave areas, the point cloud is incomplete due to insufficient scanning density. When the laser incidence angle is not optimized, edge occlusion and reflection interference are easily caused. Dynamic error compensation is insufficient, so that the traditional measurement technology based on static model matching cannot real-time perceive and correct the measurement errors caused by dynamic factors such as mechanical system vibration, thermal deformation, and local springback of the workpiece. Data is easily disturbed by image noise and reflection characteristics of different material curved surfaces, and traditional filtering algorithms are difficult to effectively distinguish between real features and noise. In addition, traditional calibration relies on a single interpolation algorithm, which has insufficient accuracy when processing global error distribution, especially in curvature mutation areas.

[0005] For example, the galvanometer correction laser measurement method proposed in CN106815822A has a three-dimensional adaptability defect. The bilinear interpolation model only compensates for two-dimensional plane geometric distortion, and does not consider the dynamic deformation of the Z-axis of a three-dimensional curved surface. In the measurement of a double-curvature curved surface such as a ship plate, the laser beam cannot be perpendicular to the incident due to the change of the curved surface normal, the spot offset error accumulates to ±0.15mm, and there is a lack of curvature adaptive mechanism. The scanning path may overlap or have a gap in the curvature mutation area, which requires manual intervention, and the measurement efficiency is reduced by 40%. SUMMARY

[0006] The purpose of the present application is to overcome the defects in the prior art and provide a laser measurement system and method suitable for large and complex curved surfaces, which can overcome one or more problems caused by the limitations and defects of related technologies to some extent.

[0007] In order to achieve the above-mentioned purpose, the technical scheme of the present application is as follows:

[0008] A laser measurement system suitable for large and complex curved surfaces, comprising a control module, a positioning module, a pre-scanning module and a high-precision scanning module, the control module comprising a high-precision scanning planning unit and a scanning analysis unit, the positioning module performing positioning and clamping of a workpiece, and aligning the scanning starting point of the workpiece with the coordinate origin of the CAD model stored therein; The pre-scanning module comprises a high-speed laser scanner and a two-dimensional driving mechanism. The high-speed laser scanner is arranged directly above the workpiece and is driven by the two-dimensional driving mechanism to move along a horizontal trajectory, and is used to collect an initial model of the workpiece and send it to the high-precision scanning planning unit for planning of path, posture and path point density; The high-precision scanning module comprises a three-dimensional scanning platform, a universal mechanical arm and a ranging laser. The ranging laser has full freedom, is arranged on the universal mechanical arm and can rotate universally, and is rotatably installed on the three-dimensional scanning platform, so that the laser of the ranging laser always vertically enters the workpiece surface to collect data. The universal mechanical arm and the three-dimensional scanning platform adjust the collection speed, orientation, position and posture of the ranging laser according to the control instructions generated by the high-precision scanning planning unit. The collection data of the ranging laser is sent to the scanning analysis unit. The scanning analysis unit processes the scanning data, analyzes the quality and error, generates a supplementary scanning instruction to the high-precision scanning module, and is used to supplement the area that does not meet the scanning quality requirements. The scanning analysis unit finally outputs a high-precision scanning model.

[0009] Further, the control module further comprises an echo monitoring module and a deviation correction unit, an electronic level is integrated on the three-dimensional scanning platform and the universal mechanical arm, the electronic level and the echo monitoring module respectively send dynamic monitoring data to the scanning analysis unit, the scanning analysis unit generates error compensation parameters and sends them to the deviation correction unit, and the deviation correction unit sends deviation correction instructions to the three-dimensional scanning platform or the universal mechanical arm, for dynamically correcting high-precision scanning data and performing deviation correction operations.

[0010] A laser measurement method for a large and complex curved surface, comprising the following steps:

[0011] S1, generating an initial model by pre-scanning; aligning the starting point of the pre-scanning module with the coordinate origin of the CAD model, starting the pre-scanning module, collecting three-dimensional point cloud data of the workpiece, and generating an initial model of a rough grid structure;

[0012] S2, scanning path planning and scanning posture control; extracting a path seed point set of the initial model according to a uniform sampling strategy, calculating the principal curvatures of the path seed points, and planning the path step length and the density distribution of the scanning sub-path according to the principal curvatures and the curvature gradient of the path seed points to generate a complete scanning path; obtaining the surface normal vector of each scanning path point on the complete scanning path to guide and control the scanning posture of the ranging laser, and coupling the complete scanning path, the path step length, and the surface normal vector to output a dynamic scanning path;

[0013] S3, collecting high-precision scanning point cloud data; driving and controlling the ranging laser to perform high-precision dynamic scanning by scanning patches divided according to the initial model according to the dynamic scanning path, adjusting the scanning posture in real time according to the surface normal vector when the ranging laser reaches each scanning path point, so that the measurement laser beam is incident and reflected along the normal of the curved surface point, and collecting reflected data to generate high-precision scanning point cloud data;

[0014] S4, preprocessing and marking of high-precision scanning point cloud data; the preprocessing further includes real-time preprocessing and centralized preprocessing, the real-time preprocessing is performed after each scanning patch is completed, and the centralized preprocessing is performed after the scanning task is completed; the high-precision scanning point cloud data after real-time preprocessing and centralized preprocessing is respectively subjected to point cloud data quality evaluation, and scanning errors are monitored, and the area with low point cloud data quality and scanning error exceeding the threshold is marked as a "low-quality area to be supplemented";

[0015] S5, local optimization and supplement scanning strategy; optimizing or reconstructing the dynamic scanning path in the "low-quality area to be supplemented", starting the high-precision scanning module for local supplement scanning until the high-precision scanning point cloud data meets the requirements;

[0016] S6, updating the high-precision scanning point cloud data and fitting to generate a final high-precision scanning model.

[0017] Further, in the scanning path planning, the path seed point of the curved surface is expressed in a parameterized form, and then the first fundamental form and the second fundamental form of each path seed point on the initial model are calculated, the principal curvatures k1 and k2 of the path seed point are determined, and the density function is constructed :

[0018] ;

[0019] wherein, , and are self-defined weight parameters, + + , is the curvature gradient;

[0020] The density field distribution is calculated according to the density function, the path planning is preferentially performed on a high-density area, a plurality of curve trajectories are generated along the direction of the principal curvature k1, the initial model is divided into a plurality of scanning strips, and different path step lengths are set according to the curvatures on the curve trajectories.

[0021] Further, the high-precision scanning point cloud data in step S3 includes coordinate data, echo intensity, laser flight time and curvature value, and the marking of an insufficient intensity area is further included in step S4. A signal intensity threshold is set, an area with echo signal intensity lower than the signal intensity threshold is marked, and the attitude of the ranging laser is adjusted for local rescan.

[0022] Further, the real-time preprocessing further includes dynamic correction; after each scanning slice area is completed, the state error of the scanning point is calculated according to the dynamic monitoring data, an outlier distance threshold for screening state error data is set, high-precision scanning point cloud data is randomly selected and assumed to be an inner point, the inner points and the outer points are distinguished according to the outlier distance threshold, the inner points are selected multiple times until a maximum inner point set is reached, the least square method is used to fit and analyze the maximum inner point set after screening, an accurate error model is constructed, error compensation parameters are generated, the high-precision scanning point cloud data of the last scanning slice area is fine-compensated, and the control parameters of the high-precision scanning module are fine-adjusted.

[0023] Further, the centralized preprocessing method includes: performing format conversion, denoising, key point extraction and splicing on the high-precision scanning point cloud data collected in each scanning slice area to generate a complete curved surface point cloud, using a surface normal guided reconstruction algorithm to fill in the missing area, completing scanning model fitting, and registering and denoising the scanning model based on the initial model.

[0024] Further, the error monitoring includes normal error monitoring and distance error monitoring. The normal error is the included angle between the normal of the scanning point and the normal of the adjacent point on the initial model , and the distance error is the distance d between the coordinates of the scanning point and the adjacent point on the initial model.

[0025] ;

[0026] ;

[0027] wherein, is the normal vector of the adjacent point on the initial model, is the normal vector of the adjacent point on the initial model; is the coordinate of the scanning point on the scanning model, is the coordinate of the adjacent point on the initial model;

[0028] A two-dimensional error matrix is constructed based on the normal error and the distance error, and according to a preset error threshold, in real-time preprocessing and centralized preprocessing, a 'low-quality to-be-supplemented scanning area' is marked respectively.

[0029] Further, the method for evaluating the quality of point cloud data is: different principal curvatures of the scanning model surface are divided by using different scale analysis grids, the quality of point cloud in different analysis grids is analyzed, and according to the preset point cloud density requirement and error tolerance of different principal curvatures, the quality of point cloud data in the analysis grid is evaluated in real time.

[0030] Further, the step S3 further includes automatic adjustment of the scanning parameter of the laser ranging module in the high-precision scanning process, a signal intensity limit value is set, when the echo intensity exceeds the signal intensity limit value, and combined with the curvature analysis, the scanning point is located in the convex reflection area or the concave shadow area, the scanning parameter of the laser ranging module is automatically adjusted.

[0031] Compared with the prior art, the laser measurement system suitable for large complex curved surfaces and the measurement method thereof have the following beneficial effects:

[0032] 1. The initial model generated based on pre-scanning intelligently plans a dynamic scanning path of a high-precision scanning module, the scanning path is automatically encrypted in a high-curvature area, the details of the complex curved surface are ensured to be captured completely, the scanning frequency is reduced in a low-curvature area, the measurement efficiency is improved, the pre-scanning and the high-precision scanning strategy are combined, high-density coverage and uniform sampling of the large complex curved surface are realized, and high-quality data is provided for three-dimensional modeling;

[0033] 2. The control module monitors and analyzes scanning data quality in real time, marks an abnormal area with low scanning quality, and triggers a dynamic supplement scanning mechanism to re-optimize the path and execute supplement scanning, so that the data quality and reliability are ensured;

[0034] 3. The curvature-guided path encryption and real-time attitude control make the ranging laser head and the normal of the scanning curved surface keep approximately vertical, cooperate with a data post-processing algorithm, effectively solve the problems of sparse point cloud and attitude mismatch in a high-curvature area, improve the accuracy of complex curved surface measurement, and ensure the reliability of the measurement result.

[0035] 4. Strong anti-interference ability, through real-time preprocessing and centralized preprocessing algorithm, removing noise of scanning process data and result data, optimizing measurement parameters, effectively suppressing interference, and being beneficial to output high-precision point cloud data; meanwhile, the high-precision scanning module has self-monitoring and self-adaptive adjustment functions, realizes closed-loop adjustment through a feedback mechanism, and reduces interference in the data acquisition process. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 Fig. 1 is a structural schematic diagram of a laser measurement system of the present application;

[0037] Figure 2 Fig. 2 is a measurement flowchart of a laser measurement method of the present application;

[0038] Figure 3 Fig. 3 is a planning schematic diagram of a dynamic scanning path;

[0039] Figure 4 Fig. 4 is an adjustment schematic diagram of a laser head incident direction, and the adjustment angle is a. DETAILED DESCRIPTION

[0040] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings of the embodiments of the present application. Obviously, the described embodiments are only the best embodiments of the present application, rather than all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the present application.

[0041] As shown in Fig. 1, the present embodiment provides a laser measurement system suitable for large complex surfaces, which comprises a control module, a positioning module, a pre-scanning module and a high-precision scanning module; wherein, Figure 1 The workpiece is positioned and clamped by the clamping terminal of the positioning module, the CAD model of the workpiece is stored in the positioning module, and a vision auxiliary unit is configured, and the vision auxiliary unit is used to align the scanning starting point of the pre-scanning module with the CAD model coordinate origin corresponding to the workpiece;

[0042] The pre-scanning module integrates an image processing unit and a high-speed laser scanner, the high-speed laser scanner is arranged directly above the workpiece and is driven to move along a horizontal trajectory by a two-dimensional driving mechanism, is used for fast and rough scanning of the surface of the large workpiece and modeling, and generates an initial model of a rough grid to provide basic data for subsequent measurement;

[0043] The control module comprises a high-precision scanning planning unit and a scanning analysis unit; the high-precision scanning planning unit plans a high-precision scanning path, a scanning posture and a path point density according to the curvature calculation and analysis of the initial model by the pre-scanning module;

[0044] ​

[0045] The high-precision scanning module performs dynamic scanning according to the dynamic scanning path planned by the high-precision scanning planning unit, and includes a three-dimensional scanning platform, a universal mechanical arm, and a ranging laser. The ranging laser is arranged on the universal mechanical arm and can rotate universally. The universal mechanical arm is rotatably installed on the three-dimensional scanning platform. Therefore, the ranging laser has full freedom and can be flexibly adapted to complex curved surfaces with different curvature characteristics for posture adjustment, so that the laser of the ranging laser always vertically enters the workpiece surface for data collection. The universal mechanical arm and the three-dimensional scanning platform adjust the collection speed, position, orientation, and posture of the ranging laser according to the control instructions of the high-precision scanning planning unit. The high-precision scanning point cloud data collected by the ranging laser is sent to the scanning analysis unit. The scanning analysis unit processes, quality evaluates, and error analyzes the high-precision scanning point cloud data, generates a supplementary scanning instruction, and feeds back to the high-precision scanning module to drive the ranging laser to perform local supplementary scanning on the area that does not meet the scanning quality requirement. The scanning analysis unit finally outputs a high-precision scanning model.

[0046] In the embodiment, the control module further includes an echo monitoring module and a deviation correction unit, and electronic levels are integrated at key positions of the three-dimensional scanning platform and the universal mechanical arm. The electronic levels can adopt inclination sensors to monitor the level state of each key position of the platform in real time. The deviation correction unit corrects local deviations caused by dynamic factors such as deviation of the three-dimensional scanning platform, workpiece springback, and system jitter in real time. The scanning analysis unit is configured with a high-performance embedded processor, which receives dynamic monitoring data collected by the electronic levels and the echo monitoring module in real time. The scanning analysis unit analyzes the dynamic monitoring data and generates error compensation parameters to automatically and dynamically compensate the collected high-precision scanning point cloud data. At the same time, the error compensation parameters are sent to the deviation correction unit. The deviation correction unit generates deviation correction instructions to fine-tune the adjusting mechanism of the mechanical arm support foot or the guide rail platform, quickly correct the levelness, eliminate the deviation caused by jitter and uneven foundation during the measurement movement, and provide a stable reference for subsequent measurement. The ranging laser is internally provided with an adjusting motor and an angle sensor. The adjusting motor can quickly and accurately adjust the angle of the laser head of the ranging laser according to the deviation correction instructions, so as to ensure that the laser beam is perpendicular to the measurement surface.

[0047] Based on the above laser measurement system, the embodiment further provides a laser measurement method for a large complex curved surface. Before measurement, the six-degree-of-freedom movement capability of the three-dimensional scanning platform and the universal mechanical arm is calibrated by the electronic levels to ensure stable measurement reference. The workpiece to be measured is fixed on the guide rail platform using a special clamp or a magnetic attraction device to avoid displacement or shaking during measurement. The scanning starting point of the pre-scanning module is aligned with the origin of the workpiece coordinate system by using a visual auxiliary system. The measurement process includes the following contents and steps:

[0048] S1, an initial model is generated by pre-scanning; a pre-scanning module is started, a high-speed laser scanner is horizontally moved, a workpiece surface is quickly and fully scanned, three-dimensional point cloud data is generated by an image processor at the same time, after the three-dimensional point cloud data is preliminarily preprocessed to remove outliers, an initial model of a rough grid is generated by a triangulation algorithm, and the initial model contains overall contour and low-precision curvature distribution information of the workpiece, thereby providing basic geometric data for subsequent scanning path planning.

[0049] S2, scanning path planning; according to a uniform sampling strategy, a path seed point set of the initial model is extracted, a scanning strategy is determined based on differential geometry theory, principal curvatures k1 and k2 of each path seed point P i on the initial model are calculated, the path seed point P i is expressed in a parameterized form , and then:

[0050] (1)

[0051] In formula (1), e= , is a first-order partial derivative of p with respect to a parameter u, , is a first-order partial derivative of p with respect to a parameter v, , , is a second-order partial derivative of p with respect to the parameter u, , is a mixed partial derivative, , is a second-order partial derivative of p with respect to the parameter v, , and a calculation result of formula (1) is distributed according to curvatures of , and .

[0052] A curvature gradient is calculated, a density function is constructed according to the principal curvature and the curvature gradient, and a density requirement of scanning sampling is represented:

[0053] (2)

[0054] In formula (2), a, b and c are self-defined weight parameters, , and , + + , The higher the value of the area (such as high-curvature corners, recesses), the greater the demand for scanning sampling density. The system presets the D value of the priority planning area according to the density field distribution, prioritizes high-density areas for path planning, uses the curvature gradient direction as a pointer, and generates a curve trajectory along the principal curvature direction, as shown in Figure 3 According to the density field distribution of different curvatures, different path step lengths and different scanning path widths are configured, the initial model is divided into several scanning strips, and the incident direction of the laser head at each scanning path point is determined by its normal vector to ensure that the angle between the laser beam direction and the normal vector of the scanning path point is less than 5°, as shown in Figure 2 The path point spacing in the high-curvature area is set to 0.1 mm, and the path point spacing in the low-curvature area is set to 0.5 mm, forming a variable-step scanning strategy; the high-precision scanning planning unit couples the scanning path, the path step length, and the surface normal vector to output a dynamic scanning path.

[0055] S3, performing high-precision scanning to collect point cloud data; the three-dimensional scanning platform and the universal mechanical arm drive and control the ranging laser according to the high-precision scanning path to perform dynamic scanning, and the ranging laser performs dynamic scanning piece by piece according to the scanning piece area divided by the initial model. When each scanning path point is reached, the universal mechanical arm adjusts the posture of the laser head of the ranging laser in real time according to the surface normal vector, so that the laser beam is incident and reflected along the normal of the scanning path point. The ranging laser has an angle sensor built-in to realize closed-loop control of the posture adjustment without manual intervention. The laser head automatically collects the reflected data of the workpiece surface with 650 nm wavelength laser, generates about 1000 points per second, and forms preliminary high-precision scanning point cloud data; the high-precision scanning point cloud data includes scanning path point three-dimensional coordinates, echo intensity, laser flight time, curvature value, etc.

[0056] The echo monitoring module monitors the echo signal intensity, and according to the preset signal intensity limit value, when the echo signal intensity exceeds the range, the scanning analysis unit combines the curvature analysis, and when the scanning point is located in the convex reflective area or the concave shadow area, the ranging laser is automatically adjusted in real time to collect data by adjusting the laser power, exposure time and other parameters, avoiding data loss in high-reflective or concave areas.

[0057] S4, preprocessing and marking of high-precision scanning point cloud data; the preprocessing further includes real-time preprocessing and centralized preprocessing. The real-time preprocessing takes the initial model self-adaptive divided rough grid unit as the scanning piece area, and each scanning piece area is executed for real-time preprocessing. The real-time preprocessing not only evaluates the quality of the point cloud data in real time and monitors the scanning error to mark the "low-quality area to be compensated" in real time, but also can compensate and correct local deviations in real time.

[0058] The dynamic correction method is: after each scanning area is completed, the state error between the scanning points and the curved surface points is calculated according to the dynamic monitoring data of the electronic level and the echo monitoring module, including the error of the inclination angle in each direction and the echo intensity error, the RANSAC algorithm is used to screen inliers and outliers, specifically, the distance threshold for screening state error data is set, the high-precision scanning point cloud data is randomly selected as inliers, the inliers and outliers are distinguished according to the distance threshold, the model is fitted to remove outliers, and the above process is iterated until the maximum inlier set is obtained, finally, the least square method is used to fit and analyze the maximum inlier set, an accurate local error model is constructed, the error compensation parameters are calculated, the high-precision scanning point cloud data of the last scanning area is compensated, and the levelness of the high-precision scanning module and the angle of the laser head are corrected in real time combined with the monitoring feedback of the electronic level and the echo monitoring module, as shown in Figure 4 The error between the laser head angle and the surface normal vector monitored by the electronic level is calculated to obtain the posture adjustment angle a.

[0059] After the full coverage scanning task of the workpiece is completed, centralized preprocessing is performed; the centralized processing is used for point cloud denoising, key point extraction, cropping and splicing of high-precision scanning point cloud data of each scanning area, complete scanning model fitting based on bilinear interpolation and RANSAC algorithm, missing area is filled by using surface normal guided reconstruction algorithm, quality assessment is performed on complete curved surface point cloud data, global registration is performed on multi-view point cloud and initial model based on ICP algorithm, error≤0.05mm, model denoising processing: median filter is used to remove salt and pepper noise, and morphological opening operation is used to eliminate small burrs, finally, error monitoring and calculation are performed on the processed complete curved surface point cloud data for supplementary marking;

[0060] The quality evaluation method of the point cloud data is as follows: firstly, the point cloud data collected by the above method is subjected to point cloud denoising, key point extraction, and cutting, and then subjected to local scanning model fitting, filling, local registration, and model denoising; the local scanning model and the complete scanning model are subjected to grid element division, for a large and complex curved surface, different grid sizes are used to divide the model curved surface with different principal curvatures, different sampling density requirements and error tolerances are set for different principal curvatures, and the quality of the point cloud data of each grid element is independently judged, for example: in a low-curvature or approximately planar area, a large-size grid of 10 mm to 15 mm is set to quickly screen the sparse sampling; in a high-curvature area (such as an area with a curvature radius less than 10 mm), the grid size is reduced to 2 to 3 mm to enhance the detection capability of missing scanning or error mutation, and the point density lower than 80% of the theoretical value (such as a low-curvature area < 2 points / mm2 and a high-curvature area < 10 points / mm2) is marked; according to the preset error tolerance: a high-curvature area ≤ 0.05 mm and a low-curvature area ≤ 0.2 mm, the minimum distance between the sampling points and the fitted curved surface in different curvature areas is analyzed; accordingly, the low-quality point cloud data with a sampling density lower than the sampling density requirement and sampling data not meeting the error tolerance is marked as a “low-quality area to be supplemented”;

[0061] The error monitoring includes normal error monitoring and distance error monitoring, the normal error is the included angle between the normal of the scanning point and the normal of the adjacent point on the initial model , and the distance error is the distance d between the coordinates of the scanning point and the adjacent point on the initial model:

[0062] (3)

[0063] (4)

[0064] In formula (3) and formula (4), is the normal vector of the scanning point on the scanning model, is the normal vector of the adjacent point on the initial model; is the coordinate of the scanning point on the scanning model, is the coordinate of the adjacent point on the initial model;

[0065] A two-dimensional error matrix is constructed based on the normal error and the distance error, in order to facilitate marking, the preset error threshold and color gradient color number are used to map the error value to the corresponding color gradient, and in real-time preprocessing and centralized preprocessing, the color gradient is used to mark the “low-quality area to be supplemented” with scanning errors exceeding the threshold, in this example, the scanning area with a distance error exceeding ±0.03 mm and a normal deviation > 8° is marked.

[0066] In addition, the marks of real-time preprocessing and centralized preprocessing also include the marks of insufficient intensity areas, and the areas with insufficient signal intensity after the adjustment of the acquisition parameters of the ranging laser and the adjustment of the laser head angle after preprocessing are adjusted and rescanned again, and the signal intensity threshold is set to mark the areas with insufficient intensity.

[0067] S5, local optimization and rescan strategy; according to the curvature characteristics and marking reasons of each marked rescan area, for the rescan area that does not meet the sampling density requirement, the high-precision scanning planning unit optimizes and readjusts the density function weight of the area, proportionally reduces the path step and path spacing, regenerates the encrypted path, and drives the high-precision scanning module to perform local rescan; for the area with error exceeding the threshold or insufficient intensity, the reasons are analyzed in detail, such as system deviation, workpiece rebound or shielding factors, and after corresponding correction, adjustment, optimization or reconstruction, the high-precision scanning module is driven to perform local rescan until the high-precision scanning point cloud data meets the requirements.

[0068] S6, after the supplementary point cloud data after local rescan is denoised and feature acquisition, the high-precision scanning point cloud data is updated, the final high-precision scanning model is generated by bilinear interpolation and RANSAC algorithm fitting, finally, the final high-precision scanning model is registered with the CAD theoretical model, the key error indicators such as overall mean square error (MSE) and maximum deviation are calculated, and the detection report is output, and the surface modeling accuracy and overall deviation of measurement and processing are intuitively presented.

[0069] Through the description of the above embodiments, those skilled in the art can clearly understand that the embodiments of the present application can be realized by means of software or software combined with necessary general hardware platform, and of course can be realized by hardware function, based on such understanding, the technical solutions of the present application can be embodied in the form of software product, which is stored in a storage medium, including a plurality of instructions for making a computer device, such as but not limited to personal computer, server, or network device, to execute all or part of the steps of the method described in any embodiment of the present application.

[0070] Although the embodiments of the present application have been shown and described, it can be understood by those skilled in the art that various changes, modifications, replacements and variations can be made to the embodiments without departing from the principles and spirits of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A laser measuring system suitable for large complex curved surfaces, characterized in that, The application relates to a high-precision laser scanning device and a high-precision laser scanning method. The device comprises a control module, a positioning module, a pre-scanning module and a high-precision scanning module, wherein the control module comprises a high-precision scanning planning unit and a scanning analysis unit; the positioning module performs positioning and clamping of a workpiece, aligns a scanning starting point of the workpiece with a coordinate origin of a CAD model stored in the workpiece, The pre-scanning module comprises a high-speed laser scanner and a two-dimensional driving mechanism, the high-speed laser scanner is arranged directly above the workpiece and is driven to move along a horizontal track by the two-dimensional driving mechanism, is used for collecting an initial model of the workpiece and is sent to the high-precision scanning planning unit to plan a path, a posture and a path point density; 2. The laser measurement system suitable for large complex curved surfaces according to claim 1, characterized in that: The high-precision scanning module comprises a three-dimensional scanning platform, a universal mechanical arm and a ranging laser, the ranging laser has full freedom, is arranged on the universal mechanical arm and can rotate in all directions, can be rotatably installed on the three-dimensional scanning platform, so that laser of the ranging laser always vertically enters a workpiece surface to collect data, the universal mechanical arm and the three-dimensional scanning platform adjust a collection speed, a position, a location and a posture of the ranging laser according to control instructions generated by the high-precision scanning planning unit, the collection data of the ranging laser is sent to the scanning analysis unit, the scanning analysis unit processes scanning data, analyzes quality and errors, generates a supplementary scanning instruction to the high-precision scanning module, is used for supplementary scanning of a region that does not meet a scanning quality requirement, and finally outputs a high-precision scanning model.

3. A method for measuring a large complex curved surface by laser, characterized in that, The control module further comprises a echo monitoring module and a deviation correction unit, an electronic level is integrated on the three-dimensional scanning platform and the universal mechanical arm, the electronic level and the echo monitoring module respectively send dynamic monitoring data to the scanning analysis unit, the scanning analysis unit generates error compensation parameters and sends the error compensation parameters to the deviation correction unit, the deviation correction unit sends a deviation correction instruction to the three-dimensional scanning platform or the universal mechanical arm, is used for dynamically correcting high-precision scanning data and performing a deviation correction operation. The method comprises the following steps: S1, an initial model is generated by pre-scanning; a scanning starting point of a pre-scanning module is aligned with a coordinate origin of a CAD model, the pre-scanning module is started, three-dimensional point cloud data of a workpiece is collected, and an initial model of a rough grid structure is generated; S2, scanning path planning and scanning posture control; path seed point sets of the initial model are extracted according to a uniform sampling strategy, principal curvatures of the path seed points are calculated, path step lengths and density distributions of scanning sub-paths are planned according to the principal curvatures and curvature gradients of the path seed points, a complete scanning path is generated, surface normal vectors of scanning path points on the complete scanning path are acquired, scanning postures of a ranging laser are guided and controlled, and a dynamic scanning path is output by coupling the complete scanning path, the path step lengths and the surface normal vectors; S3, high-precision scanning and collecting point cloud data; performing high-precision dynamic scanning on each scanning slice area divided according to the initial model under the driving and control of the dynamic scanning path, and adjusting the scanning posture in real time according to the surface normal vector when the ranging laser reaches each scanning path point, so that the measurement laser beam is incident and reflected along the normal of the curved surface point, and high-precision scanning point cloud data is collected; S4, preprocessing and marking of the high-precision scanning point cloud data; The preprocessing further includes real-time preprocessing and centralized preprocessing, the real-time preprocessing is performed after each scanning slice area is completed, and the centralized preprocessing is performed after the scanning task is completed; The high-precision scanning point cloud data after the real-time preprocessing and the centralized preprocessing is respectively subjected to point cloud data quality evaluation, and scanning errors are monitored, and regions with low point cloud data quality and scanning errors exceeding a threshold value are marked as "low-quality regions to be supplemented by scanning"; S5, local optimization and supplementing scanning strategy; optimizing or reconstructing the dynamic scanning path in the "low-quality region to be supplemented by scanning", starting the high-precision scanning module to perform local supplementing scanning, and stopping until the high-precision scanning point cloud data meets the requirements; S6, updating the high-precision scanning point cloud data and fitting to generate a final high-precision scanning model.

4. The method for laser measurement suitable for large complex curved surfaces according to claim 3, characterized in that: In the scan path planning, the path seed point of the curved surface is expressed in a parameterized form, and then the first and second fundamental forms of each path seed point on the initial model are calculated, the principal curvatures k1 and k2 of the path seed point are determined, and a density function is constructed : ； wherein, , and are custom weight parameters, + + , is a curvature gradient; The density field distribution is obtained according to the density function, the path planning is preferentially performed on the high-density region, a plurality of curve trajectories are generated along the direction of the main curvature k1, the initial model is divided into a plurality of scanning strips, and different path steps are set according to the curvatures on the curve trajectories.

5. The method for laser measurement suitable for large complex curved surfaces according to claim 3, characterized in that: The high-precision scanning point cloud data in step S3 includes coordinate data, echo intensity, laser flight time and curvature value, and the marking of the region with insufficient intensity is further included in step S4, a signal intensity threshold value is set, the region with echo signal intensity lower than the signal intensity threshold value is marked, and the posture of the ranging laser is adjusted to perform local supplementing scanning.

6. The method for laser measurement suitable for large complex curved surfaces according to claim 5, characterized in that: The real-time preprocessing further includes dynamic deviation correction; after each scanning slice area is completed, the state error of the scanning point is calculated according to the dynamic monitoring data, an outlier distance threshold value for screening state error data is set, the high-precision scanning point cloud data is randomly selected to be assumed as an inner point for fitting, the inner point and the outer point are distinguished according to the outlier distance threshold value, the inner point set is selected multiple times until the maximum inner point set is reached, the least square method is used to fit and analyze the maximum inner point set after screening, an accurate error model is constructed, error compensation parameters are generated, the high-precision scanning point cloud data of the previous scanning slice area is finely compensated, and the control parameters of the high-precision scanning module are finely adjusted.

7. The method for laser measurement suitable for large complex curved surfaces according to any one of claims 3 to 6, characterized in that: The centralized preprocessing method includes: performing format conversion, denoising, key point extraction and splicing on the high-precision scanning point cloud data collected by each scanning slice area to generate complete curved surface point cloud, filling missing areas by using a surface normal guided reconstruction algorithm, completing scanning model fitting, and registering and denoising the scanning model based on the initial model.

8. The method for laser measurement of large complex surfaces according to claim 7, characterized in that: The monitoring of the scanning errors includes normal error monitoring and distance error monitoring, the normal error being an included angle between a normal of a scanning point and a normal of a neighboring point on the initial model , and the distance error being a distance d between a coordinate of the scanning point and a coordinate of the neighboring point on the initial model ； ； wherein, is a normal vector of a neighboring point on the initial model, is a normal vector of a neighboring point on the initial model; is a coordinate of a scan point on the scan model, is a coordinate of a neighboring point on the initial model; A two-dimensional error matrix is constructed based on the normal error and the distance error, and the "low-quality region to be supplemented by scanning" is marked in the real-time preprocessing and the centralized preprocessing according to a preset error threshold value.

9. The method for laser measurement of large complex surfaces according to claim 8, characterized in that: The point cloud data quality evaluation method is: using different scale analysis grid to divide the surface of the scanning model with different principal curvatures, analyzing the point cloud quality in different analysis grids, and evaluating the point cloud data quality in the analysis grid in real time according to the preset point cloud density requirement and error tolerance of different principal curvatures.

10. The method for laser measurement suitable for large complex curved surfaces according to claim 7, characterized in that: The step S3 also includes automatic adjustment of the scanning parameters of the laser ranging module in the high-precision scanning process, setting a signal intensity limit value, when the echo intensity exceeds the signal intensity limit value, combining the curvature analysis, and when the scanning point is located in the convex reflection area or the concave shadow area, the scanning parameters of the laser ranging module are automatically adjusted.

Images