Method and system for measuring spatial pose of pipeline end

Abstract

The invention provides a method and system for measuring the spatial pose of a pipeline end, and the method comprises the following steps: carrying out the scanning measurement of the pipeline end, so as to obtain the spatial pose information of the pipeline end, and the spatial pose information comprises three-dimensional point cloud data and / or two-dimensional image information; preprocessing the spatial pose information, and filtering and denoising the three-dimensional point cloud data and / or the two-dimensional image information; extracting feature information of the end of the pipeline from the preprocessed spatial pose information; and obtaining the spatial pose of the pipeline end according to the feature information. According to the measurement method provided by the invention, the measurement efficiency is improved while the measurement precision is improved, all modules are in close cooperation, so that the measurement process is efficient and smooth, and the precision is greatly improved; the method is high in adaptability, and can be suitable for targets of different materials and different measurement shapes.

Description

Technical Field

[0001] The present invention relates to the technical field of posture measurement, and in particular to a method and system for measuring the spatial posture of a pipeline end, and a display device. Background Art

[0002] In the process of vigorous development of modern industry, many key fields such as aerospace, automobile manufacturing, petrochemical and shipbuilding are highly dependent on the stable operation of pipeline systems. The precise installation of pipeline systems is the core element to ensure the performance, safety and reliability of the overall equipment, and its importance is self-evident. Taking the shipbuilding industry as an example, the pipelines of ship pipeline systems are generally long, with a wide variety of specifications, intricate spatial structures and extremely dense arrangements. At the same time, it also involves a variety of materials, and the parameters of its pipelines, including length, specifications, spatial structure and arrangement, operating pressure and medium, are extremely complex. This puts extremely high technical requirements on pipeline assembly. Therefore, the high-precision determination of the spatial posture of the pipeline end is particularly critical. Accurate spatial posture measurement can ensure that the median misalignment of the end face during pipeline assembly reaches a high accuracy, thereby effectively reducing the stress applied during assembly, significantly improving the tightness of the pipeline, and thus ensuring the performance, safety and reliability of the overall pipeline equipment.

[0003] However, when measuring the spatial position of the end of complex pipelines such as ships, a single sensor measurement device is currently used. This device is easily restricted by the environment and target characteristics, resulting in inaccurate or incomplete data and low measurement accuracy. The measurement system has poor adaptability, and different measurement systems need to be replaced for different pipe heads, which limits the measurement range. The data analysis and processing capabilities are also relatively weak.

[0004] Therefore, it is urgent to develop a method and system for measuring the spatial posture of a pipeline system to address the shortcomings of traditional technology and promote the development of related industries. Summary of the invention

[0005] In view of the problems existing in the measurement accuracy of pipeline ends in the prior art as described above, the present application provides a method and system for measuring the spatial posture of a pipeline end.

[0006] On the one hand, the present application provides a method for measuring the spatial position and posture of a pipeline end, comprising the following steps:

[0007] Scan and measure the pipeline end to obtain spatial position information of the pipeline end, wherein the spatial position information includes three-dimensional point cloud data and / or two-dimensional image information;

[0008] Preprocessing the spatial posture information, and filtering and denoising the three-dimensional point cloud data and / or the two-dimensional image information;

[0009] Extracting feature information of the pipeline end from the preprocessed spatial posture information;

[0010] The spatial position of the pipeline end is obtained according to the feature information.

[0011] Optionally, before scanning the pipeline end, the method further includes:

[0012] Build a measurement system;

[0013] The surface of the pipe end is sprayed with a coating, the thickness of which is between 0.01 mm and 0.05 mm.

[0014] Fixing the measuring system to the area to be measured at the end of the pipeline;

[0015] Set the measurement environment: the ambient temperature is between 19℃ and 21℃, and the ambient humidity is between 45% and 55%.

[0016] Optionally, if the spatial posture information includes the three-dimensional point cloud data and the two-dimensional image information, the three-dimensional point cloud data and the two-dimensional image information are synchronously acquired information, and the time synchronization error is controlled within ±0.003s.

[0017] Optionally, acquiring the spatial position and posture of the pipeline end according to the feature information further includes:

[0018] Clarify the constraint rules of the feature points, wherein the constraint rules include the distance or angle between the feature points of the pipeline end;

[0019] Determining the weight of the feature point;

[0020] Integrating the feature information according to the constraint rules and the feature point weights;

[0021] The spatial position of the pipeline end is determined based on the trust region strategy weighted least square method.

[0022] Optionally, preprocessing the spatial posture information further includes:

[0023] Performing noise reduction processing on the three-dimensional point cloud data based on a hybrid filtering algorithm of point cloud density and curvature features;

[0024] The two-dimensional image information is denoised by grayscale, multi-scale adaptive histogram equalization, texture direction enhancement and deep learning denoising.

[0025] Optionally, extracting feature information of the pipeline end from the preprocessed spatial posture information includes:

[0026] Extracting the feature information in the three-dimensional point cloud data based on an improved region growing algorithm combined with a shape prior knowledge base method;

[0027] The feature information in the two-dimensional image information is extracted based on a multi-scale fusion Canny edge detection algorithm combined with deep learning feature extraction.

[0028] On the other hand, the present application provides a pipeline end spatial posture measurement system, comprising:

[0029] A scanning module, used to scan the pipeline end to obtain spatial posture information of the pipeline end;

[0030] The data processing unit optimizes the spatial posture information to determine the spatial posture of the pipeline end.

[0031] The display and storage unit comprises a display screen and a memory, and the display and storage unit is used to display and store all data information.

[0032] Optionally, the scanning module includes:

[0033] A laser scanning device, mounted on a bracket with six degrees of freedom, the laser scanning device being used to obtain three-dimensional point cloud data containing position information;

[0034] and / or a visual measuring device, provided with a plurality of camera modules, the visual measuring device being used to obtain two-dimensional image information including appearance information of the pipe end;

[0035] and / or an inertial measurement unit, for acquiring two-dimensional image information containing relative position information of the pipeline end.

[0036] Optionally, the data processing unit includes:

[0037] A data preprocessing subunit, which preprocesses the spatial posture information;

[0038] A feature extraction and matching subunit, used to extract feature information from the spatial pose information after preprocessing;

[0039] The spatial posture calculation and optimization subunit further optimizes the extracted feature information to determine the spatial posture of the pipeline end.

[0040] Optionally, an electromagnetic shielding device is also provided to reduce external electromagnetic interference to the measurement system.

[0041] As described above, the spatial posture measurement method and system provided by the present invention have at least the following beneficial technical effects:

[0042] The spatial posture measurement method and system proposed in the present invention have achieved remarkable technical effects in terms of measurement accuracy, data processing, measurement analysis, adaptability and stability, and application:

[0043] Improve measurement accuracy: Multi-source data acquisition and fusion, the laser scanning device, visual measurement device and inertial measurement unit work together, and ultimately the calculated pipe end position coordinate error is controlled within ±0.03mm, and the attitude angle error is controlled within ±0.05°, which is far higher than the accuracy of traditional measurement technology.

[0044] Intelligent data processing: Use hybrid filtering algorithm to process 3D point cloud data, and multi-scale adaptive histogram equalization algorithm to process 2D images, with data integrity exceeding 99%; use specific algorithms to extract point cloud and image feature information, with feature extraction accuracy exceeding 98%; after integration of constraint rules and feature point weights, matching accuracy is higher than 97%; spatial posture calculation uses relevant algorithms, which are intelligent and efficient, and effectively guarantee the overall measurement quality.

[0045] High measurement and analysis efficiency: The various modules of the measurement system work closely together and have excellent performance, which greatly shortens the measurement time and improves the measurement efficiency; the laser scanning module obtains 3D point cloud data, the visual measurement module and the inertial measurement unit obtain 2D image data, the large-scale parallel computing architecture and dynamic task allocation of the data processing unit, the convenient interaction and large-capacity and high-speed storage of the display and storage units make the entire measurement process run efficiently and smoothly and the measurement accuracy is greatly improved.

[0046] Enhanced adaptability and stability: It has good adaptability to pipe ends of different materials and shapes as well as complex and changeable measurement environments. The laser scanning head can adaptively adjust the power and wavelength, the visual measurement device camera has a variety of anti-interference functions, the inertial measurement unit can work stably in harsh environments and the self-diagnosis accuracy rate is more than 99%, and the intelligent cache and pre-fetch mechanism of the data processing unit ensures the stability and accuracy of the measurement results.

[0047] Widely used: not only can the spatial posture of real-time acquired data be accurately measured, but also the spatial posture can be obtained by optimizing the already taken photos, stored 3D data or 2D data. It greatly improves the data processing efficiency, and the intelligent data processing can adapt to the complex pipe end shape and measurement environment changes. The modules of the measurement system work closely together, with excellent performance, greatly shortening the measurement time, and improving the measurement accuracy and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 Shown is a flow chart of the spatial posture measurement method provided in Example 1.

[0049] Figure 2Shown is a schematic diagram of the composition of the spatial posture measurement system provided in Example 2.

[0050] Reference numerals

[0051] 10. Scanning module; 11. Laser scanning device; 12. Visual measurement device; 13. Inertial measurement unit; 20. Data processing unit; 21. Data preprocessing subunit; 22. Feature extraction and matching subunit; 23. Spatial posture calculation and optimization subunit; 24. Cache area; 25. Cooling system; 30. Display and storage unit; 31. Light-emitting diode display screen; 32. Data redundancy backup module; 33. Encrypted storage module; 34. Quick retrieval function module. DETAILED DESCRIPTION

[0052] The following describes the embodiments of the present invention through specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention.

[0053] It should be noted that the illustrations provided in this embodiment only illustrate the basic concept of the present invention in a schematic manner. Although the illustrations only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation, the form, quantity, positional relationship and proportion of each component in actual implementation can be changed at will under the premise of realizing the technical solution of this party, and the component layout form may also be more complicated.

[0054] Embodiment 1

[0055] This embodiment provides a method for measuring the spatial position of a pipeline end. Figure 1 As shown, it is a flow chart of the spatial posture measurement method provided by this embodiment, which specifically includes the following steps:

[0056] S1: Scan and measure the pipeline end to obtain spatial position information containing the pipeline end, wherein the spatial position information includes three-dimensional point cloud data and / or two-dimensional image information.

[0057] The scanning device scans the pipe end to obtain spatial posture information of the pipe end. Generally, the spatial posture information includes the spatial position information of the pipe end and the posture information of whether the pipe end has angle offset such as rotation. Optionally, the spatial posture information only includes three-dimensional point cloud data; optionally, the spatial posture information only includes two-dimensional image information; optionally, the spatial posture information includes a collection of three-dimensional point cloud data and two-dimensional image information. Specifically, if the spatial posture information includes three-dimensional point cloud data and two-dimensional image information, the three-dimensional point cloud data and the two-dimensional image information are information acquired synchronously. Specifically, the time synchronization error is controlled within ±0.003s.

[0058] Specifically, before scanning and measuring the pipeline end, the method further includes:

[0059] S11: Build the measurement system.

[0060] Specifically, both the three-dimensional point cloud data and the two-dimensional image information contain the spatial position information of the pipe end, but the results obtained by scanning with different scanning devices are different. Different measuring devices can be selected according to the actual situation of the pipe end and the measurement environment. Generally, the scanning device includes a three-dimensional scanning device and a two-dimensional scanning device. Specifically, the scanning device is installed on an electric bracket, which supports and drives the scanning device to move. Specifically, the initial positioning accuracy of the electric bracket is controlled within ±0.005mm, the repeated positioning accuracy is controlled within ±0.003mm, and the movement speed of the electric bracket is between 0.01 and 0.5m / s. Specifically, a timing device is also included to ensure that the three-dimensional point cloud data and the two-dimensional image information are collected at the same time.

[0061] Specifically, in this embodiment, a laser scanning device is used to obtain three-dimensional point cloud data, a visual scanning device and an inertial measurement unit are used to obtain two-dimensional image information, and atomic clock information is used to synchronously collect multi-source data. The ranging range of the laser scanning device is between 0.1 and 100 meters; the measurement accuracy is controlled within ±0.005 meters, and the power is between 10 4 ~10 5μW; the resolution of the visual scanning device is greater than or equal to 25 million pixels, and it can capture color images with a color depth of 16 bits, and contains optical and electronic image stabilization and angle compensation of ±2°; the inertial measurement unit obtains two-dimensional data of the motion state and spatial posture of the pipeline end. By changing the measurement angle of the inertial measurement unit, the three-dimensional data of the pipeline end can be presented, and the time form and posture of the pipeline end in space can be understood from a visual level. Specifically, in this embodiment, the inertial measurement unit adopts a high-precision fiber optic gyro inertial measurement unit (IMU), the gyro zero bias stability is better than 0.01° / h, and the accelerometer zero bias stability is better than 10-4m / s2, so that the attitude angle error is controlled within ±0.05°, and the measurement accuracy reaches more than 99%.

[0062] S12: spraying a coating on the surface of the pipe end, wherein the thickness of the coating is between 0.01 mm and 0.05 mm.

[0063] Specifically, before scanning the pipe end, a layer of coating needs to be sprayed on the end surface to better reflect light and improve the recognition of image features, which is beneficial for subsequent data processing. Generally, the coating thickness cannot affect the accuracy of the pipe end itself. Specifically, the coating thickness is between 0.01mm and 0.05mm.

[0064] S13: Fixing the measuring system to the area to be measured at the end of the pipeline.

[0065] S14: Set the measurement environment: the ambient temperature is between 19°C and 21°C, and the ambient humidity is between 45% and 55%.

[0066] Generally, specific environmental conditions are required for measuring the pipe end to ensure that the electronic components and sensors of the measuring equipment are in a relatively stable working state, to avoid equipment performance fluctuations and measurement data deviations caused by excessively high or low temperatures, excessive or low humidity, etc., which helps to ensure the accuracy and stability of the measurement. Specifically, the ambient temperature is between 20°C ± 1°C, and the ambient humidity is between 45% and 55%. Specifically, in this embodiment, the temperature is 20°C and the humidity is 50%.

[0067] S2: Preprocess the spatial posture information, and perform filtering and denoising on the three-dimensional point cloud data and / or the two-dimensional image information.

[0068] Generally, the three-dimensional point cloud data and two-dimensional image information are processed to lay the foundation for subsequent feature extraction, precise matching and spatial pose calculation. Generally, methods for processing three-dimensional point cloud data include: point cloud density algorithm, curvature feature hybrid filtering algorithm, statistical filtering algorithm, radius filtering algorithm, voxel grid filtering algorithm and other algorithms that effectively remove isolated noise points; methods for processing two-dimensional image information include: grayscale, histogram equalization, Gaussian filtering algorithm, median filtering algorithm and other algorithms that retain image edges and details. The present invention is not limited to the above algorithms, and any one of the methods or a combination of several methods can be used to process three-dimensional point cloud data and two-dimensional image information. Specifically, the steps for processing three-dimensional point cloud data and two-dimensional images in this embodiment are as follows:

[0069] For the processing of three-dimensional point cloud data, a hybrid filtering algorithm based on point cloud density and curvature features is used to identify and remove low-density noise points according to the local density distribution of the point cloud, effectively improving the data quality. The specific steps are as follows: Point cloud density calculation, point cloud density calculation uses an adaptive neighborhood radius method, and the neighborhood radius range is 0.5-5mm. Since the density of different areas in the point cloud data may be different, for example, the point cloud at the edge of the pipe end may be sparse, while other parts are relatively dense. The adaptive neighborhood radius can be dynamically adjusted according to the local sparsity of the point cloud, setting a larger neighborhood radius (such as 2mm) for sparse areas and a smaller neighborhood radius (such as 0.8mm) for dense areas. This adaptive neighborhood radius method can more accurately identify the local density distribution of the point cloud and remove low-density noise points that are significantly deviated from the normal density range. Curvature feature calculation: The curvature feature is calculated based on a local quadratic surface fitting method. The curvature is calculated by fitting the local point cloud data with a quadratic surface. During the fitting process, the fitting error is required to be controlled within ±0.01mm. Strict error control can ensure that the calculated curvature feature is closer to the actual situation, making the subsequent edges and key feature points screened out based on the curvature feature more accurate. After this processing, the retained data integrity is higher than 99%.

[0070] For the processing of two-dimensional image information, the edge clarity is improved by graying, multi-scale adaptive histogram equalization, combining texture direction enhancement and deep learning noise reduction. The specific steps are as follows: 1) First, the two-dimensional image information is grayed and multi-scale adaptive histogram equalization is performed: graying can simplify the image data structure while retaining the brightness information of the image, reducing the amount of calculation for subsequent processing. Multi-scale adaptive histogram equalization is to perform histogram equalization on the local area of ​​the image at different scales, disperse the concentrated grayscale, make the grayscale distribution of the image more uniform, and adaptively adjust the histogram equalization parameters of each local area according to the characteristics of different areas of the image, so as to better retain the local detail information of the image while enhancing the overall contrast of the image and improve the visual effect of the image. 2) Then combine the texture direction enhancement processing: for the texture features that may exist on the surface of the pipe end, the local contrast is adjusted along the texture direction. For example, for the area with longitudinal texture on the surface of the pipe end, the contrast is adjusted along the longitudinal texture direction to make the texture details clearer. This helps to highlight the texture information in the image and provide more detailed basis for subsequent feature extraction. 3) Finally, the two-dimensional image is denoised by combining deep learning denoising: the deep learning model has a strong learning ability and can automatically extract and remove noise features in the image by learning a large number of noisy and noise-free image pairs. While removing the noise, the edge and detail information of the image are retained to make the image clearer. After the above steps, the edge clarity of the image can be improved by more than 40%, which is conducive to subsequent edge detection and feature extraction tasks. 4) After completing the above processing, the image can be sharpened: a gradient-based edge sharpening algorithm is used to adjust the pixel value according to the gradient information of the image. The edge part of the image usually has a large gradient change. By enhancing these gradient changes, the edge can be made clearer and more obvious. However, during the sharpening process, care should be taken to avoid over-sharpening to prevent artifacts from appearing on the edge, which will affect the subsequent analysis and understanding of the image.

[0071] S3: Extracting feature information of the pipe end from the preprocessed spatial posture information.

[0072] Generally, methods for extracting geometric information from three-dimensional point cloud data include: region growing algorithm, shape prior knowledge base method, principal component analysis algorithm, curvature estimation and key point extraction algorithm, feature histogram algorithm, etc. Methods for extracting feature information from two-dimensional image information include: multi-scale Canny edge detection algorithm, deep learning feature extraction network, scale-invariant feature transform (SIFT) algorithm, accelerated robust feature algorithm, histogram of oriented gradients (HOG) algorithm, etc. The algorithm for extracting feature information from three-dimensional point cloud data and two-dimensional image information of the present invention is not limited to the above algorithms, and any one or a combination of several methods can be used. Specifically, in this embodiment, the method for extracting feature information is as follows:

[0073] The geometric features of three-dimensional point cloud data are extracted based on the improved region growing algorithm combined with the shape prior knowledge base method. First, based on the preliminary analysis results of the point cloud data, the shape model that may match is selected from the shape prior knowledge base as a reference. Then, the rules of region growth are dynamically adjusted according to the geometric feature parameters of the selected model. For example, if the pipe end is judged to be circular, the growth rule is customized around the characteristics of the circle: the distance deviation threshold between the growth point and the seed point is set according to the radius range of the circle, such as 0.1 times the radius of the circle; the angle threshold between the growth direction and the tangent direction of the circle is precisely controlled, such as 5°. At the same time, in order to ensure the accuracy and reliability of the growth, the threshold is determined by statistical learning. By performing statistical analysis and machine learning on a large amount of historical pipe end point cloud data, a threshold system for different shape models is established. The threshold is not fixed, but will be continuously optimized with the accumulation of new data and the update of the model, so that the algorithm can adapt to pipe ends of different specifications, materials and processes. Ultimately, the geometric shape characteristics of the pipe end can be accurately extracted from the complex point cloud data, and the error between the extracted edge contour and the actual one can be successfully controlled within ±0.05mm, ensuring that the extracted geometric features are highly consistent with the actual pipe end.

[0074] Generally, the traditional region growing algorithm starts with a seed point and gradually merges qualified points in the neighborhood into the growing region according to the preset similarity criterion until the stop condition is met. The improved region growing algorithm here fully considers the complex geometric characteristics of the pipe end and the diversity of actual measurement scenarios.

[0075] For two-dimensional image information, the multi-scale fusion Canny edge detection algorithm is combined with a deep learning feature extraction network to extract features. The multi-scale fusion Canny edge detection algorithm is an upgrade and expansion of the classic Canny algorithm. Multi-scale fusion processes images at different scales, and each scale can capture edge information at different levels. Small scales can detect subtle edge changes in images, while large scales are better at obtaining macro and stable edge contours. It can overcome the limitations of single-scale detection. The deep learning feature extraction network can mine deep and abstract feature information from two-dimensional images, which complements the edge features extracted by the multi-scale fusion Canny edge detection algorithm. The combination of these two technologies has enabled the accuracy of extracting pipe end features from two-dimensional image information to reach over 98%.

[0076] S4: Acquire the spatial position of the pipeline end according to the feature information.

[0077] Specifically, obtaining the spatial position and posture of the pipe end according to the feature information includes integrating the feature information extracted from the three-dimensional data point cloud data and the two-dimensional information, and extracting the spatial position and posture of the pipe end from the integrated information. Specifically, the following steps are included:

[0078] The present invention is not limited to the matching method. This embodiment provides a matching method based on spatial geometric constraints and feature similarity weighting.

[0079] S41: clarify the constraint rules of the feature points, the constraint rules include the distance or angle between the feature points of the pipe end: that is, clarify the constraint rules of the geometric features, and construct the spatial geometric constraint conditions according to the spatial structural characteristics of the pipe end. For example, for the feature points of the pipe end port, the geometric relationship between them, such as the distance and angle, is fixed. In addition, there are also the shape of the pipe port, the relative position relationship of the feature points, etc., which can be used as constraint rules.

[0080] S42 then determines the weight of the feature points: different weights (0.2 to 1.8) are assigned to the extracted features according to their importance in the characteristics of the pipe end. For key features that can more accurately and representatively reflect the characteristics of the pipe end (such as the center of the port, the vertex of the bend, etc.), higher weights (such as 2.5, 1.8) are assigned; and relatively minor features are given lower weights (such as 0.2).

[0081] S43 integrates the feature information according to the constraint rules and feature weights. For example, for a feature point on the edge of the pipe end detected in both the three-dimensional point cloud data and the two-dimensional image, calculate their spatial position relationship (such as whether the distance deviation and angle deviation between the two points meet the geometric constraints) and visual feature similarity (such as the grayscale change similarity of the edge, the texture feature similarity, etc.), and calculate the comprehensive match in combination with their respective weights.

[0082] S44 Optimize matching results: Use random sampling consistency algorithm to optimize the preliminary matching results. Data points that meet the model are considered as inliers, and those that do not meet the model are considered as outliers, thereby removing incorrectly matched feature point pairs, further improving the accuracy and reliability of matching, and making the final matching accuracy reach more than 97%.

[0083] S45: Determine the spatial position and posture of the pipe end. Generally, algorithms for determining the position and posture of an object in space include classical algorithms (including quaternion algorithms, Kalman filter algorithms, etc.) and algorithms based on machine learning (including neural network algorithms, random forest algorithms, etc.). Specifically, in this embodiment, the spatial position and posture of the pipe end is calculated based on the weighted least squares method based on the trust region strategy. The specific steps are as follows:

[0084] First, feature data with different degrees of importance are obtained in the feature extraction and information integration steps, and these data are assigned corresponding weights (0.2 to 1.8). Then, an objective function is constructed, the goal is to minimize the sum of square errors between the spatial pose calculated based on these weighted features and the actual measured data. The mathematical optimization method solves this objective function to determine the spatial pose of the pipe end. At the same time, the trust region radius is set, and its initial value is between 0.1 and 1 mm, which is equivalent to defining a relatively reliable range for the calculation process. Within this range, searching and optimizing the spatial pose based on the weighted least squares method is considered to be more reasonable and more likely to be close to the real result. Specifically, the trust region radius is dynamically adjusted according to the changes in the objective function.

[0085] The spatial position of the pipeline end can be preliminarily determined through the above method.

[0086] Optionally, in order to make the spatial posture of the pipeline port more accurate, further optimization can be performed based on the iterative nearest point algorithm. Specifically, in the point sets corresponding to different data sources (three-dimensional point cloud data and three-dimensional spatial information corresponding to the matched two-dimensional image, etc.), the closest point pairs are continuously searched. Based on these closest point pairs found, the transformation relationship (including rotation and translation) between the two sets of point sets is calculated, and then this transformation relationship is applied to the current spatial posture estimation, and it is updated and adjusted. After multiple such iterative processes, the estimation of the spatial posture becomes closer and closer to the actual situation, and the accuracy of the calculation is continuously improved.

[0087] Optionally, in order to avoid data anomalies caused by factors such as data points with large errors, environmental interference, and complex models, the LGS~RANSAC strategy is introduced. The LGS~RANSAC strategy can screen out reliable data for calculation. It will randomly select a part of the data points, fit a spatial pose model based on these points, and then use this model to determine whether all other data points conform to the model. Those that conform are regarded as inliers (that is, reliable data points), and those that do not conform are excluded as outliers (possibly abnormal data points). Through such a screening mechanism, the interference of bad data on spatial pose calculations is reduced, and the quality of the data based on which the calculation is based is improved, so that the final calculation results are more accurate.

[0088] Optionally, in order to make the calculation more accurate and precise and obtain a spatial pose close to the real one, an adaptive dynamic adjustment step size is introduced. The step size is set between 0.05 and 0.5 mm. This step size refers to the magnitude of each adjustment to the spatial pose during the iterative calculation process. During the preliminary calculation, if it is found that the iteration converges quickly, the step size can be appropriately increased to speed up the overall calculation process; if it is found that the convergence speed slows down or instability occurs during the iteration process, the step size can be reduced accordingly to make the calculation more precise to ensure the final calculation accuracy.

[0089] In short, through the above method, a very good effect was finally achieved, which enabled the iterative convergence speed to be increased by more than 40%. At the same time, the position coordinate error was controlled within ±0.03mm, and the attitude angle error was controlled within ±0.05°, realizing high-precision calculation of the spatial posture of the pipeline end.

[0090] In addition, the method for measuring the spatial posture of the pipe end provided in this embodiment also has a relatively powerful optimization function, and its application scenarios are very wide. It can not only accurately measure the spatial posture of data acquired in real time, but also optimize the photos that have been taken and the stored three-dimensional data to obtain the spatial posture.

[0091] Embodiment 2

[0092] This embodiment also provides a spatial posture measurement system. This embodiment also takes the spatial posture measurement system of the pipeline end as an example. Figure 2As shown, it is a schematic diagram of the composition of the spatial posture measurement system provided in this embodiment. The spatial posture measurement system of the pipeline end provided in this embodiment includes a scanning module 10, a data processing unit 20 and a display and storage unit 30. The scanning module 10 is used to scan the pipeline end to obtain the spatial posture information of the pipeline end; the data processing unit 20 is used to optimize the spatial posture information obtained by the scanning module 10 to determine the spatial posture of the pipeline end; the display and storage unit 30 is used to display all data information and properly store these data information.

[0093] Generally, the scanning module 10 includes a laser scanning device, a structured light three-dimensional scanner, a visual measurement device, a linear array camera, a panoramic camera, an inertial measurement unit, and other devices that can obtain the position information of the pipe end. Optionally, the scanning module 10 can be a scanning device or a combination of multiple scanning devices. Specifically, in this embodiment, the scanning module 10 includes a laser scanning device 11, a visual measurement device 12, and an inertial measurement unit 13. The laser scanning device 11 is used to obtain three-dimensional point cloud data of the scanned target, and the visual measurement device 12 and the inertial measurement unit 13 are used to obtain two-dimensional image information. In actual applications, the scanning device can be selected for scanning according to the actual situation of the pipe end, and a combination of one or more of them can be selected.

[0094] Specifically, in the actual application operation process, due to the many differences in the environment, physical characteristics and precision requirements of different pipeline ends, a suitable scanning device can be selected for scanning according to the specific conditions of the actual pipeline ends. Optionally, the operator can manually select the corresponding scanning device based on his or her own professional knowledge and experience; or, in order to further improve the automation and accuracy of the scanning work, a computer control module can be added. According to the pre-set algorithm and the various conditions of the actual pipeline ends obtained in real time, the most suitable scanning device or scanning device combination is independently selected to carry out the scanning work.

[0095] Specifically, if two or more scanning devices are used to obtain posture information, the timing device is turned on to ensure that the posture information obtained is obtained at the same time. Generally, the timing device includes a high-precision atomic clock, a microcontroller timing module, and an electronic timer. Specifically, in this embodiment, a high-precision atomic clock is used. The laser scanning device 11, the visual measurement device 12 and the inertial measurement unit 13 are connected through a precise clock synchronization interface, so that the three can perform data synchronization based on the atomic clock signal as the reference clock, and the synchronization accuracy is ±0.001s and has a clock drift automatic compensation function. The time synchronization error between the laser scanning device and the visual measurement device is controlled within ±0.003s. The inertial measurement unit records the posture change information during the measurement process at a sampling frequency of not less than 200Hz, and its measurement data is accurately synchronized and fused with the laser scanning and visual measurement data through a high-precision time synchronization module based on the atomic clock signal.

[0096] Specifically, the three-dimensional point cloud data acquired by the laser scanning device 11 includes the position coordinates of each point of the measurement target and the laser reflection intensity of each coordinate point, which further reflects the spatial contour and shape of the measurement target, as well as whether there is wear and stains on the surface of the object; the two-dimensional image information acquired by the visual measurement device 12 includes various visual features of the surface of the object, such as color distribution, texture details, shape contours, etc., and further extracts information such as edge features, feature points, shape descriptions, etc. from these two-dimensional images; the data output by the inertial measurement unit 13 are mainly linear acceleration values ​​(along three axes) and angular velocity values ​​(around three axes). These data are continuously changing values ​​that reflect the dynamic characteristics of the object during movement. Based on these basic data, after subsequent calculation and processing, relatively abstract data such as the attitude angle and position change trajectory of the object, which are closely related to the object's motion state and spatial attitude, can be obtained.

[0097] Specifically, the laser scanning device 11 is configured with: a high-precision laser scanning head and a six-degree-of-freedom electrically adjustable bracket. The laser scanning head is installed on a six-degree-of-freedom electrically adjustable bracket, and it can be flexibly adjusted in six degrees of freedom (including translation along the three coordinate axes of X, Y, and Z and rotation around these three coordinate axes) to meet the scanning needs of different angles and positions. In addition, the bracket has a very high positioning accuracy, the error can be controlled within ±0.005mm, and the repeatability can reach ±0.003mm, ensuring that the laser scanning head can accurately locate the required spatial position at each measurement, ensuring the accuracy and consistency of the scanning data. Moreover, its movement speed is between 0.01 and 0.5m / s, and the acceleration is between 0.1 and 2m / s 2The scanning head position can be adjusted quickly and smoothly according to the requirements of the scanning task to adapt to different scanning scenarios and rhythms. Specifically, a high-precision angle sensor is installed at each joint of the six-degree-of-freedom electrically adjustable bracket to monitor and feedback the bracket's motion posture in real time, so as to compensate and calibrate the position and posture of the laser scanning head more accurately during data processing.

[0098] Specifically, the laser scanning device 11 is also equipped with a frequency modulated continuous wave (FMCW) laser ranging system, a power adaptive adjustment and wavelength control module, and a data acquisition unit. The system uses advanced frequency modulated continuous wave laser ranging technology to accurately measure the distance between the laser scanning head and the target object. The ranging range can reach 0.1 to 100m, and the accuracy can reach ±0.005m, which can meet the scanning ranging requirements under various distance scales. The power adaptive adjustment and wavelength control module is responsible for adaptively adjusting and controlling the power and wavelength of the laser. The laser power can be adjusted within 10 4 ~10 5 The power can be adjusted between μW and μW, and the power can be changed dynamically according to factors such as the distance, material, and reflection characteristics of the scanning target. At the same time, the wavelength range can be adaptively adjusted. Different wavelengths have different advantages in different scanning environments and for different target object materials. By reasonably adjusting the wavelength, the effect of laser scanning can be further optimized and the quality and accuracy of the scanning data can be improved.

[0099] Specifically, the laser scanning device 11 is also equipped with an intelligent anti-obstruction module, which monitors the scanning path in real time. If an obstacle is encountered (such as part of the structure of the fixture obstructing certain areas of the pipe end), the scanning path is automatically adjusted to avoid the obstacle, ensuring that the three-dimensional point cloud data of the pipe end can be fully acquired.

[0100] Specifically, the laser scanning device automatically adjusts the scanning resolution and scanning frequency based on the preliminary shape estimation of the pipe end. For example, for areas with complex shapes and large curvature changes, the scanning resolution is increased to 0.1mm and the scanning frequency is increased to 1000Hz; for relatively smooth areas, the scanning resolution and frequency are appropriately reduced to improve the overall scanning efficiency while ensuring that sufficiently accurate data is obtained. Multiple high-resolution cameras of the visual measurement device capture images of the pipe end from different perspectives, and automatically adjust the shooting parameters according to changes in ambient light and target features. For example, when the ambient light is dark, the aperture is automatically increased and the exposure time is extended; when special textures or reflective features are detected on the surface of the pipe end, the focus and image enhancement parameters are adjusted.

[0101] The visual measuring device 12 is equipped with a high-resolution camera, an anti-shake and angle compensation system, a back-illuminated sensor, an ultra-low dispersion large aperture optical lens, an intelligent control system, and a high-speed optical fiber network connection system. The visual measuring device 12 is installed at a suitable position on the pipe end to ensure that the pipe end can be photographed from different perspectives from a suitable angle. A uniform annular light source is arranged around the camera, and the brightness, color, and illumination angle of the light source can be adjusted according to the measurement requirements to highlight the features of the pipe end, which is convenient for image acquisition and subsequent feature extraction.

[0102] Specifically, the high-resolution camera has a resolution of more than 25 million pixels and can capture color images with a color depth of 16 bits, which can present richer and more delicate color levels. The anti-shake and angle compensation system includes optical and electronic anti-shake functions and an angle compensation (±2°) mechanism, which can effectively reduce the image blur problem caused by shaking and ensure that clear and sharp images are obtained. It can also correct the deviation of the camera shooting angle to a certain extent. The back-illuminated sensor has higher photosensitivity efficiency and lower noise level. It places the photodiode behind the circuit layer, which can receive light more fully, so that clear images can be captured under low light conditions, and can reduce noise in the image and improve the overall quality of the image. The ultra-low dispersion large aperture optical lens, on the one hand, helps to improve the image brightness in a dark environment, and on the other hand, can achieve a shallow depth of field effect, highlight the main object, and facilitate focusing and measuring the characteristics of the main part. The response time of the intelligent control system is less than 0.005s, which can quickly respond to various shooting needs and achieve efficient shooting operations. The high-speed optical fiber network connection system has a signal transmission rate of not less than 10Gbps. Can be quickly and accurately transmitted to subsequent processing devices or storage devices.

[0103] The inertial measurement unit 13 is configured with: a high-precision fiber optic gyroscope, a high-precision accelerometer, a temperature self-compensation module, an anti-vibration interference module, a self-diagnosis function module, an attitude solution and data processing module, and a precise clock synchronization interface. The inertial measurement unit 13 is installed at a stable center position of the measurement system and is thermally isolated to reduce the impact of ambient temperature changes on its measurement accuracy.

[0104] Specifically, the high-precision fiber optic gyroscope is mainly used to measure the angular velocity information of an object around each coordinate axis, and the gyroscope zero bias stability is better than 0.01° / h; the high-precision accelerometer is mainly responsible for measuring the linear acceleration of an object in the three coordinate axis directions. Its zero bias stability is better than 10 -4 m / s 2The anti-vibration interference module adopts a variety of means such as shock-absorbing structure design and filtering algorithm to effectively filter and isolate the impact of external vibration on internal sensors, prevent false signals caused by vibration from interfering with normal measurement data, and ensure the accuracy and reliability of measurement data; the fault monitoring and early warning part has a self-diagnosis function, which can monitor and evaluate the working status of each component (including fiber optic gyroscope, accelerometer, related circuits, etc.) in real time; information integration and calculation center: this module is responsible for receiving the original measurement data from high-precision fiber optic gyroscope and accelerometer, and using corresponding mathematical algorithms (such as attitude solution algorithms based on Newtonian mechanics and kinematics principles, etc.) to process and integrate these data to ensure that the attitude angle error of the final output is controlled within ±0.05°, ensuring the high accuracy of the measurement results, so that the measurement accuracy of the entire inertial measurement unit reaches more than 99%; in addition, it also includes a precise clock synchronization interface, through which the inertial measurement unit can achieve precise time synchronization with other modules (such as laser scanning devices, visual measurement devices, etc.).

[0105] Generally, the data processing unit 20 includes: a data preprocessing subunit 21 , a feature extraction and matching subunit 22 , a spatial posture calculation and optimization subunit 23 , a cache area 24 and a heat dissipation system 25 .

[0106] Specifically, the data preprocessing subunit 21 relies on a large-scale parallel computing architecture, combined with a field programmable gate array (FPGA) and a graphics processing unit (GPU) to realize its functions. The FPGA performs filtering operations on the acquired three-dimensional point cloud data to remove interference data such as noise points and abnormal points, and improve data quality. The GPU enhances the two-dimensional image by adjusting the image contrast, brightness, sharpening edges and other operations to make the features of the two-dimensional image clearer and more obvious; the feature extraction and matching subunit 22 uses a deep learning neural network dedicated chip (NPU) to accelerate the deep learning algorithm and combines traditional algorithms to optimize accuracy and speed to collaboratively complete feature extraction and matching; the spatial posture calculation and optimization subunit 23, with the help of a multi-core central processing unit (CPU) combined with distributed computing technology, can perform parallel processing on complex spatial posture calculation tasks to improve the calculation speed. In terms of specific algorithms, optimized spatial geometric transformation and iterative closest point algorithm are used to calculate the spatial position and posture of objects, and an intelligent data caching and pre-fetching mechanism is set up. For example, a cache area of ​​5GB is set to reduce data reading waiting time and speed up data access during the calculation process by reasonably caching and pre-fetching data, ultimately achieving an effect of improving data processing efficiency by more than 50%, ensuring that the spatial position and posture information of objects can be obtained accurately and quickly.

[0107] Specifically, the cache area 24 has a capacity between 1GB and 10GB, which plays a role in buffering and temporarily storing data. At the same time, it also has a dynamic task allocation function. According to the real-time load and processing capacity of each sub-unit, the data is reasonably allocated to the corresponding sub-unit for processing, so that each sub-process in the entire data processing process can be closely connected; the system is also equipped with a heat dissipation system 25, which adopts a combination of liquid cooling and air cooling to cope with the large amount of heat generated by the hardware equipment during long-term high-load operation. Liquid cooling circulates the coolant in the heat dissipation pipe to take away the heat generated by the equipment, and air cooling uses a fan to promote air flow to enhance the heat dissipation effect on the surface of the equipment.

[0108] Generally, the display and storage unit 30 has the important functions of both a display window and a data warehouse, and is used to fully display all data information and properly store these data. Generally, it is configured with an intuitive display device to facilitate users to observe the spatial posture of the target to be measured from different angles; it is also configured with a large-capacity storage device for long-term and stable storage. Specifically, the display and storage unit 30 provided in this embodiment is configured with: a light-emitting diode display screen 31, a data redundancy backup module 32, an encryption storage module 33, and a fast retrieval function module 34.

[0109] Specifically, the light-emitting diode display screen 31 adopts an ultra-high-definition, high-refresh rate organic light-emitting diode display screen (OLED), which can clearly display three-dimensional posture information and various data details. At the same time, it also has a naked-eye 3D display function, and can see stereoscopic images without the aid of additional auxiliary equipment. It can more realistically present objects and related information in the three-dimensional space, and supports multi-touch interaction functions. Users can conveniently operate the displayed content by touching the screen, such as zooming and rotating the three-dimensional model, viewing data information from different angles, etc., which greatly improves the convenience and flexibility of human-computer interaction.

[0110] The storage module selects a high-speed storage array based on non-volatile memory (NVM), with a capacity of more than 10TB and a read and write speed of no less than 1GB / s. A data redundancy backup module 32, an encryption storage module 33, and a fast retrieval function module 34 are provided. The data redundancy backup module 32 backs up important data stored in the high-speed storage array according to certain strategies and algorithms, creates multiple copies and stores them in different physical locations or storage media; the encryption storage module 33 uses advanced encryption algorithms to encrypt the data stored in the high-speed storage array, converts the original data into ciphertext for storage, and can only restore the data to a readable plaintext form with authorization and the use of the correct decryption key; the fast retrieval function module 34 can quickly locate the target data in the large-capacity storage array based on the search conditions such as keywords, time range, data type, etc. entered by the user.

[0111] Optionally, the spatial posture measurement system provided in this embodiment is equipped with an electromagnetic shielding device to prevent external electromagnetic interference from affecting electronic components and sensors in the measurement system, further ensuring the accuracy and stability of the measurement.

[0112] In summary, this embodiment provides a spatial posture measurement system, in which these components cooperate and work together to enable them to process the acquired data efficiently and stably, meeting the high-precision and high-efficiency requirements of the entire measurement system for data processing.

[0113] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A method for measuring the spatial position of a pipeline end, characterized in that: include: Scan and measure the pipeline end to obtain spatial position information of the pipeline end, wherein the spatial position information includes three-dimensional point cloud data and / or two-dimensional image information; Preprocessing the spatial posture information, and filtering and denoising the three-dimensional point cloud data and / or the two-dimensional image information; Extracting feature information of the pipeline end from the preprocessed spatial posture information; The spatial position of the pipeline end is obtained according to the feature information.

2. The method for measuring the spatial position and posture of a pipeline end according to claim 1, characterized in that: Before scanning the pipeline end, the method further includes: Build a measurement system; Spraying a coating on the surface of the pipe end, wherein the thickness of the coating is between 0.01 mm and 0.05 mm; Fixing the measuring system to the area to be measured at the end of the pipeline; Set the measurement environment: the ambient temperature is between 19℃ and 21℃, and the ambient humidity is between 45% and 55%.

3. The method for measuring the spatial position and posture of a pipeline end according to claim 1, characterized in that: If the spatial posture information includes the three-dimensional point cloud data and the two-dimensional image information, the three-dimensional point cloud data and the two-dimensional image information are synchronously acquired information, and the time synchronization error is controlled within ±0.003s.

4. The method for measuring the spatial position and posture of a pipeline end according to claim 1, characterized in that: Acquiring the spatial position of the pipeline end according to the feature information also includes: Clarify the constraint rules of the feature points, wherein the constraint rules include the distance or angle between the feature points of the pipeline end; Determining the weight of the feature point; Integrating the feature information according to the constraint rules and the feature point weights; The spatial position of the pipeline end is determined based on the trust region strategy weighted least square method.

5. The method for measuring the spatial position and posture of a pipeline end according to claim 1, characterized in that: Preprocessing the spatial posture information further includes: Performing noise reduction processing on the three-dimensional point cloud data based on a hybrid filtering algorithm of point cloud density and curvature features; The two-dimensional image information is denoised by grayscale, multi-scale adaptive histogram equalization, texture direction enhancement and deep learning denoising.

6. The method for measuring the spatial position and posture of a pipeline end according to claim 1, characterized in that: Extracting feature information of the pipeline end from the preprocessed spatial posture information includes: Extracting the feature information in the three-dimensional point cloud data based on an improved region growing algorithm combined with a shape prior knowledge base method; The feature information in the two-dimensional image information is extracted based on a multi-scale fusion Canny edge detection algorithm combined with deep learning feature extraction.

7. A pipeline end spatial posture measurement system, characterized in that: include: A scanning module, used to scan and measure the pipeline end to obtain spatial position information of the pipeline end; A data processing unit optimizes the spatial posture information to determine the spatial posture of the pipeline end; The display and storage unit comprises a display screen and a memory, and the display and storage unit is used to display and store all data information.

8. The pipeline end space posture measurement system according to claim 7, characterized in that: The scanning module comprises: A laser scanning device, mounted on a bracket with six degrees of freedom, the laser scanning device being used to obtain three-dimensional point cloud data containing position information; and / or a visual measuring device, provided with a plurality of camera modules, the visual measuring device being used to obtain two-dimensional image information including appearance information of the pipe end; and / or an inertial measurement unit, for acquiring two-dimensional image information containing relative position information of the pipeline end.

9. The pipeline end space posture measurement system according to claim 7, characterized in that: The data processing unit comprises: A data preprocessing subunit, which preprocesses the spatial posture information; A feature extraction and matching subunit, used to extract feature information from the spatial pose information after preprocessing; The spatial posture calculation and optimization subunit further optimizes the extracted feature information to determine the spatial posture of the pipeline end.

10. The pipeline end space posture measurement system according to claim 7, characterized in that: An electromagnetic shielding device is also provided to reduce external electromagnetic interference to the measurement system.

Images