Industrial model scanning method, system and equipment based on 3D point cloud data

Abstract

The invention relates to the technical field of data processing, in particular to an industrial model scanning method, system and equipment based on 3D point cloud data, and the method comprises the steps: fixing a to-be-scanned industrial model and a reference object on the same rotating platform, and starting the rotating platform and 3D scanning equipment based on a driving instruction, scanning the industrial model and the reference object at a first rotation speed according to a preset sequence, identifying and obtaining a first position based on the obtained first 3D point cloud data, determining a first scanning area, scanning the first scanning area to obtain second 3D point cloud data of the industrial model, and fusing the second 3D point cloud data into the first 3D point cloud data to obtain a second 3D point cloud data of the industrial model; and calculating an error between the first 3D point cloud data and preset reference data of the reference object, generating a correction vector based on an error calculation result, applying the correction vector to the originally obtained first 3D point cloud data, and recombining to generate a high-precision 3D model. According to the invention, the scanning precision and efficiency of the industrial model are improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and more specifically to an industrial model scanning method, system and equipment based on 3D point cloud data. Background Art

[0002] With the continuous advancement of industrial technology and the rapid development of intelligent manufacturing, three-dimensional (3D) scanning technology is increasingly being used in industrial production. In particular, in the fields of reverse engineering, quality inspection, and mold manufacturing of complex industrial models, 3D scanning technology has become an indispensable tool. However, traditional scanning methods often have problems such as insufficient accuracy, low efficiency, and high cost, which make it difficult to meet the needs of modern industry for high precision and high efficiency.

[0003] Similar prior art includes a Chinese patent application with publication number CN111402342B, which discloses a 3D point cloud processing method based on multiple feature points that can be used for industrial robot calibration. The invention includes: generating 3D point cloud data of the calibration plate; generating a two-dimensional image from the 3D point cloud data according to a preset mapping rule; processing the two-dimensional image to obtain target feature point information; comparing the target feature value of the calibration plate with the theoretical value, and when the error is less than the preset value, proceed to the next step; when it is greater than the preset value, reduce the image traversal interval, return to the previous step and solve again; perform inverse mapping calculation according to the preset mapping rule; and generate calibration plate target data. Although the hand-eye calibration test data can be obtained efficiently and accurately, the accuracy and efficiency of scanning are not considered.

[0004] Similar prior art also includes a Chinese patent application with publication number CN115619780B, which discloses a laser scanning image quality assessment method and system, the method comprising: determining reference object information from the target to be scanned; obtaining scanning point cloud data and determining the scanning point cloud coordinate system; adjusting the object data coordinates according to the scanning point cloud coordinate system; determining the object contour features, combining the object size information, and determining the size assessment information of each reference object; determining the position assessment information according to the object position information; and obtaining the quality assessment result of the scanned image according to the size assessment information and position assessment information of each reference object. Although the quality of the laser scanning image is improved, it cannot be used to scan various complex industrial models.

[0005] Therefore, the present invention provides an industrial model scanning method, system and device based on 3D point cloud data to improve the scanning accuracy and efficiency of industrial models, and is also suitable for fast and accurate scanning of various complex industrial models. Summary of the invention

[0006] The industrial model and the reference object to be scanned are fixed on the same rotating platform, the rotating platform and the 3D scanning device are started based on a driving instruction, the industrial model and the reference object are scanned at a first rotation speed in a preset order, and based on the acquired first 3D point cloud data, a first position is identified and acquired, a first scanning area is determined, and the first scanning area is scanned to acquire second 3D point cloud data of the industrial model, which is then merged into the first 3D point cloud data, and an error between the first 3D point cloud data and preset reference data of the reference object is calculated. Based on the result of the error calculation, a correction vector is generated, the correction vector is applied to the originally acquired first 3D point cloud data, and the adjusted first 3D point cloud data is recombined to generate a high-precision 3D model, thereby improving the scanning accuracy and efficiency of the industrial model.

[0007] To achieve the above object, the present invention provides an industrial model scanning method based on 3D point cloud data, which is implemented by executing the following steps: Step S1: fixing the industrial model to be scanned and the reference object on the same rotating platform, wherein the industrial model is located in the middle of the reference object, and the reference object is a plurality of identical objects; Step S2: based on the driving instruction issued by the control unit, the rotating platform and the 3D scanning device are started, the rotating platform rotates at a first rotation speed according to a preset sequence, and the 3D scanning device scans the industrial model and the reference object to obtain first 3D point cloud data of the industrial model; Step S3: Based on the first 3D point cloud data, a first position is identified and acquired, and a first scanning area is determined based on the first position. The 3D scanning device scans the first scanning area and acquires second 3D point cloud data of the industrial model. The second 3D point cloud data is fused into the first 3D point cloud data, and then input into the control unit; Step S4: The control unit receives the first 3D point cloud data, and calculates the error between the first 3D point cloud data and the preset reference data of the reference object, generates a correction vector based on the result of the error calculation, applies the correction vector to the originally acquired first 3D point cloud data, and obtains the adjusted first 3D point cloud data, and recombines the adjusted first 3D point cloud data to generate a 3D model of the industrial model.

[0008] As a preferred technical solution of the present invention, in step S2, obtaining the first 3D point cloud data of the industrial model further includes: The 3D scanning device emits a laser beam to the industrial model and records the emission time point. The laser beam is reflected after being irradiated to the measuring point on the industrial model. The receiver on the 3D scanning device captures the reflected laser beam and records the time point of capturing the laser beam as the receiving time point. Based on the receiving time point and the transmitting time point, calculating the time difference between the receiving time point and the transmitting time point, and converting the time difference into the distance between the laser transmitting point and the industrial model measuring point; As the rotating platform and the 3D scanning device move, the distance between the laser emission point and each measuring point on the industrial model is continuously recorded, and the distance data is converted into points in three-dimensional space to form first 3D point cloud data.

[0009] As a preferred technical solution of the present invention, step S3 further includes: Preprocessing the first 3D point cloud data, and projecting the preprocessed first 3D point cloud data onto a 2D plane from multiple perspectives to generate multiple two-dimensional images corresponding to the first 3D point cloud data, and recording the distance from each of the first 3D point cloud data to the 2D plane as depth information, and creating depth maps of different perspectives based on the two-dimensional image and the depth information, wherein the 2D plane is a virtual plane set on the back of the first 3D point cloud data; Based on the depth map, a position where the missing data is identified as a first position, a size of the first position is measured, and multiple first positions whose sizes of the first position are greater than or equal to a first threshold are obtained. A first scanning area is determined based on the first position. The 3D scanning device first scans the first scanning area from an edge to a center of each of the first positions, and then scans from the center to an edge of each of the first positions to obtain second 3D point cloud data of the industrial model, and the second 3D point cloud data is merged into the first 3D point cloud data to supplement the data of the first position, and new first 3D point cloud data after the data of the first position is supplemented is obtained.

[0010] As a preferred technical solution of the present invention, after acquiring the new first 3D point cloud data after supplementing the data of the first position, the method further includes: The depth map is updated, and the depth maps of different perspectives are merged together to form a complete three-dimensional model, the depth maps of different perspectives and the three-dimensional model are compared, and the position of the missing data is identified as the second position again, an interpolation algorithm is applied to obtain interpolation data of the second position, and the interpolation data is converted into the form of 3D point cloud data, and the interpolation data is gradually filled in the second position along the shape of the industrial model, and merged into the first 3D point cloud data to obtain new first 3D point cloud data.

[0011] As a preferred technical solution of the present invention, obtaining the second 3D point cloud data of the industrial model also includes: Based on multiple first positions, a rotation path of the rotating platform is determined, and based on the number of the first positions, a second rotation speed of the rotating platform is determined. The first scanning area is scanned based on the rotation path and the second rotation speed of the rotating platform, and when scanning from the edge to the center of each first position, the rotation speed of the rotating platform is converted from the second rotation speed to the third rotation speed, and when scanning from the center to the edge of each first position, the rotation speed of the rotating platform is converted from the third rotation speed to the second rotation speed, to acquire second 3D point cloud data, wherein the first rotation speed is greater than the second rotation speed, and the second rotation speed is greater than the third rotation speed.

[0012] As a preferred technical solution of the present invention, identifying the position of missing data as the first position based on the depth map also includes: Based on the depth map, the distribution of the first 3D point cloud data at different depth levels is visualized, and blank or discontinuous areas on the depth map are defined as locations of missing data.

[0013] As a preferred technical solution of the present invention, the interpolation data is fused into the first 3D point cloud data, further comprising: The original scan data and the interpolated data in the first 3D point cloud data are marked, and overlapping data in the first 3D point cloud data is deleted based on the original scan data and the interpolated data.

[0014] As a preferred technical solution of the present invention, the step S4 further includes: Step S41: aligning the first 3D point cloud data of the industrial model with the preset reference data of the reference object, and determining the best spatial transformation mode so that the first 3D point cloud data and the preset reference data overlap in space, analyzing the difference between the first 3D point cloud data and the preset reference data, and extracting a first minimum deviation; Step S42: when the first minimum deviation is greater than or equal to a third threshold, calculating the deviation between each point cloud data in the first 3D point cloud data and the preset reference data, and determining a correction vector for each point cloud data based on the deviation, wherein the deviation includes a deviation direction and a deviation magnitude; Step S43: integrating the correction vector of each point cloud data into a correction matrix, and applying the correction vector to the first 3D point cloud data of the industrial model based on the correction matrix; Step S44: After the first 3D point cloud data of the industrial model is corrected, the difference between the first 3D point cloud data and the preset reference data is re-analyzed, and the second minimum deviation is extracted. If the second minimum deviation is still greater than or equal to the third threshold, the third position where the deviation occurs is obtained from the first 3D point cloud data, and the third position is re-scanned.

[0015] The present invention also provides an industrial model scanning system based on 3D point cloud data as described above, comprising the following modules: A positioning unit, used to fix the industrial model to be scanned and a reference object on the same rotating platform, wherein the industrial model is located in the middle of the reference object, and the reference object is a plurality of identical objects; The scanning unit is configured to: start the rotating platform and the 3D scanning device, the rotating platform rotates at a first rotation speed according to a preset sequence, and the 3D scanning device scans the industrial model and the reference object to obtain first 3D point cloud data of the industrial model; Based on the first 3D point cloud data, a first position is identified and acquired, and a first scanning area is determined based on the first position, the 3D scanning device scans the first scanning area and acquires second 3D point cloud data of the industrial model, and the second 3D point cloud data is merged into the first 3D point cloud data, and then inputted into the control unit; A generating unit is used for the control unit to receive the first 3D point cloud data, calculate the error between the first 3D point cloud data and the preset reference data of the reference object, generate a correction vector based on the result of the error calculation, apply the correction vector to the originally acquired first 3D point cloud data, and obtain the adjusted first 3D point cloud data, recombine the adjusted first 3D point cloud data, and generate a 3D model of the industrial model.

[0016] The present invention also provides a computing device, the device comprising: Memory and processor; The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the above-mentioned industrial model scanning method based on 3D point cloud data is implemented.

[0017] Compared with the prior art, the beneficial effects of the present invention are at least as follows: 1. The technical solution of the present invention ensures that the relative positions of the industrial model and the reference object are stable during the scanning process by fixing the industrial model and the reference object to be scanned on the same rotating platform. Based on the driving instruction issued by the control unit, the rotating platform and the 3D scanning device are started, and the rotating platform rotates at a first rotation speed in a preset order. At the same time, the 3D scanning device scans the industrial model and the reference object to achieve an all-round scan of the industrial model and the reference object, obtain comprehensive and continuous 3D point cloud data, and improve the scanning efficiency. Based on the first 3D point cloud data obtained, the first position is identified and obtained, and the first scanning area is determined based on the first position. The 3D scanning device scans the first scanning area and obtains the second 3D point cloud data of the industrial model to improve the scanning accuracy of the complex area in the industrial model, enhance the integrity of the overall point cloud data, and fuse the second 3D point cloud data into the first 3D point cloud data to form a more complete point cloud data set.

[0018] 2. The technical solution of the present invention receives the first 3D point cloud data, calculates the error between the first 3D point cloud data and the preset reference data of the reference object, generates a correction vector based on the result of the error calculation, and applies the correction vector to the originally acquired first 3D point cloud data to eliminate or reduce the error in the scanned data and improve the accuracy of the data. The adjusted first 3D point cloud data is obtained, and the adjusted first 3D point cloud data is recombined to generate a high-precision 3D model of the industrial model. Through the coordination between the above steps, the present invention improves the scanning accuracy and efficiency of the industrial model, and the method is also suitable for fast and accurate scanning of various complex industrial models. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0020] Figure 1 It is a flowchart of the steps of the industrial model scanning method based on 3D point cloud data in the present invention; Figure 2 This is a structural diagram of the industrial model scanning system based on 3D point cloud data in the present invention. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0022] It is understood that the terms "first", "second", etc. used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are only used to distinguish a first element from another element. For example, without departing from the scope of this application, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script.

[0023] With the continuous advancement of industrial technology and the rapid development of intelligent manufacturing, three-dimensional (3D) scanning technology is increasingly being used in industrial production. In particular, in the fields of reverse engineering, quality inspection, and mold manufacturing of complex industrial models, 3D scanning technology has become an indispensable tool. However, traditional scanning methods often have problems such as insufficient accuracy, low efficiency, and high cost, which make it difficult to meet the needs of modern industry for high precision and high efficiency.

[0024] In view of the above technical problems, the present invention proposes Figure 1 The industrial model scanning method based on 3D point cloud data is implemented by executing the following steps: Step S1: The industrial model to be scanned and the reference object are fixed on the same rotating platform, the industrial model is located in the middle of the reference object, and the reference object is a plurality of identical objects.

[0025] Specifically, when scanning the industrial model, since the rotating platform can rotate arbitrarily, the industrial model and reference object placed on the rotating platform will become unstable, resulting in inaccurate scanned data. Therefore, it is necessary to fix the industrial model and the reference object to ensure that the relative position of the industrial model and the reference object is stable during the scanning process. At the same time, the industrial model is placed in the middle of the reference object. The known characteristics of the reference object (such as shape, size, etc.) can be used to assist the subsequent correction process to improve the accuracy and reliability of the scanned data and provide a benchmark for subsequent data correction. Regarding the fixing method of the industrial model and the reference object on the rotating platform, the industrial model and the reference object can be firmly fixed on the rotating platform using a clamp, screws or other fixing devices.

[0026] Step S2: Based on the driving instruction issued by the control unit, the rotating platform and the 3D scanning device are started, the rotating platform rotates at a first rotation speed in a preset order, and the 3D scanning device scans the industrial model and the reference object to obtain the first 3D point cloud data of the industrial model.

[0027] Specifically, the control unit issues instructions manually or through control information to start the rotating platform and the 3D scanning device. The rotating platform rotates at a first rotation speed in a preset sequence, such as continuous rotation, segmented rotation, clockwise rotation, or counterclockwise rotation, and the 3D scanning device synchronously scans the rotating industrial model and reference object to obtain preliminary point cloud data. The rotating platform rotates at a first rotation speed in a preset sequence, which can achieve an all-round scan of the industrial model and reference object, obtain comprehensive and continuous 3D point cloud data, and improve scanning efficiency. The 3D scanning device can be a laser scanner.

[0028] Step S3: Based on the first 3D point cloud data, identify and obtain the first position, determine the first scanning area based on the first position, use a 3D scanning device to scan the first scanning area, and obtain the second 3D point cloud data of the industrial model, merge the second 3D point cloud data into the first 3D point cloud data, and then input it into the control unit.

[0029] Specifically, when the 3D scanning device scans the industrial model for the first time, it scans at the same first rotation speed in a preset order, and the basic 3D point cloud data obtained is relatively equal within a certain range, that is, the first 3D point cloud data. This is relatively comprehensive for scanning the simple-shaped area in the industrial model, but for the complex area in the industrial model, the data obtained will be lacking. Therefore, it is necessary to further scan the complex area in the industrial model. In the above scheme, based on the first 3D point cloud data of the preliminary scan, multiple first positions that need to be refined are identified, such as edges, curved surfaces, holes, etc. The first scanning area is determined, which surrounds the first position and covers the part that requires higher precision. The 3D scanning device performs a second scan on the first scanning area to obtain the second 3D point cloud data, and the second 3D point cloud data is merged into the first 3D point cloud data to form a more complete point cloud data set. In order to improve the scanning accuracy of complex areas in industrial models and enhance the integrity of the overall point cloud data. How to identify multiple first positions will be described in detail below.

[0030] Step S4: The control unit receives the first 3D point cloud data, and calculates the error between the first 3D point cloud data and the preset reference data of the reference object, generates a correction vector based on the result of the error calculation, applies the correction vector to the originally acquired first 3D point cloud data, and obtains the adjusted first 3D point cloud data, and recombine the adjusted first 3D point cloud data to generate a 3D model of the industrial model.

[0031] Specifically, equipment errors and environmental interference are inevitable during the scanning process, so it is also necessary to correct the scanned data based on the reference object to generate a high-precision 3D model of the industrial model. Specifically, the control unit receives the fused first 3D point cloud data, calculates the error between the first 3D point cloud data and the preset reference data of the reference object, the reference data is the precise CAD model or known size of the reference object, generates a correction vector based on the error calculation result, applies the correction vector to the originally acquired first 3D point cloud data, and performs error correction to eliminate or reduce the error in the scanned data and improve the accuracy of the data. The corrected first 3D point cloud data is recombined to generate a 3D model of the industrial model.

[0032] Through the coordination of the above steps, the present invention improves the scanning accuracy and efficiency of industrial models, and the method is also suitable for fast and accurate scanning of various complex industrial models.

[0033] Furthermore, in the above step S2, obtaining the first 3D point cloud data of the industrial model further includes: The 3D scanning device emits a laser beam to the industrial model and records the emission time. The laser beam is reflected after being irradiated to the measuring point on the industrial model. The receiver on the 3D scanning device captures the reflected laser beam and records the time point of capturing the laser beam as the receiving time point. Based on the receiving time point and the transmitting time point, the time difference between the receiving time point and the transmitting time point is calculated, and the time difference is converted into the distance between the laser transmitting point and the measuring point of the industrial model; As the rotating platform and the 3D scanning device move, the distance between the laser emission point and each measurement point on the industrial model is continuously recorded, and the distance data is converted into points in three-dimensional space to form the first 3D point cloud data.

[0034] Specifically, traditional methods may be limited by measurement accuracy or data acquisition methods, resulting in loss of model details. Laser scanning technology can provide high-precision three-dimensional coordinate data to ensure that every detail of the industrial model is accurately recorded. In addition, laser scanning does not require direct contact with the industrial model, avoiding model damage or measurement errors that may be caused by contact. Therefore, laser scanning is used to obtain data. The specific implementation method is that the laser scanner emits a laser beam to the industrial model and records the emission time point. After the laser is irradiated to the measurement point on the industrial model, it is reflected. The receiver on the device will capture the reflected laser and record the reception time point. The time difference is converted into the distance between the laser emission point and the industrial model measurement point through the propagation speed of the laser in the air and the pre-set parameters. As the rotating platform and the 3D scanning device move, multiple distances can be obtained, and these distance data are converted into points in three-dimensional space to form the first 3D point cloud data, realizing high-precision data acquisition, non-contact measurement and rapid scanning of the industrial model.

[0035] Furthermore, the above step S3 also includes: preprocessing the first 3D point cloud data, and projecting the preprocessed first 3D point cloud data onto a 2D plane from multiple viewing angles to generate multiple two-dimensional images corresponding to the first 3D point cloud data, and further recording the distance from each first 3D point cloud data to the 2D plane as depth information, and creating depth maps of different viewing angles based on the two-dimensional image and the depth information, wherein the 2D plane is a virtual plane set on the back of the first 3D point cloud data; Based on the depth map, the position of the missing data is identified as the first position, the size of the first position is measured, and multiple first positions whose sizes are greater than or equal to the first threshold are obtained. The first scanning area is determined based on the first position. The 3D scanning device first scans the first scanning area from the edge to the center of each first position, and then scans from the center to the edge of each first position to obtain second 3D point cloud data of the industrial model, and the second 3D point cloud data is merged into the first 3D point cloud data to supplement the data of the first position, and obtain new first 3D point cloud data after the data of the first position is supplemented.

[0036] Specifically, in order to further scan the complex areas in the industrial model, multiple first positions need to be identified. The specific identification method is to pre-process the acquired first 3D point cloud data, including denoising, filtering, etc., to improve the data quality. A virtual 2D plane is set on the back of the first 3D point cloud data, and the pre-processed point cloud data is projected onto this plane from multiple perspectives to generate a corresponding two-dimensional image, simplify the data analysis and processing process, and improve the data processing efficiency. The distance from each point to the 2D plane is also recorded as depth information, and depth maps of different perspectives are created based on the two-dimensional image and depth information. The depth map is analyzed to identify the area representing missing data (i.e., the area where the depth information is abnormal or does not exist) as the first position. The sizes of these first positions are measured, and multiple first positions whose sizes are greater than or equal to the first threshold are screened out. The first position is where the integrity and accuracy of the industrial model have a greater impact. Start a local refinement scan, and determine the first scanning area based on the identified first position. The first scanning area covers important missing data positions and is appropriately expanded to ensure the integrity of the scan. By performing local refinement scanning on the first scanning area through the above scheme, a higher density second 3D point cloud data can be obtained, which improves the scanning accuracy of complex areas in the industrial model. The second 3D point cloud data is fused into the original first 3D point cloud data to supplement the missing data. During the fusion process, the alignment and accuracy matching of the data need to be considered. By identifying and supplementing the missing data, the final generated 3D model is more complete and accurate.

[0037] Further, after acquiring the new first 3D point cloud data after supplementing the data of the first position, the method further includes: Update the depth map and fuse the depth maps from different perspectives to form a complete three-dimensional model. Compare the depth maps and three-dimensional models from different perspectives, identify the location of the missing data as the second location again, apply the interpolation algorithm to obtain the interpolation data of the second location, and convert the interpolation data into the form of 3D point cloud data. Gradually fill in the interpolation data at the second location along the shape of the industrial model, and fuse it into the first 3D point cloud data to obtain new first 3D point cloud data.

[0038] Specifically, through the above steps, complex areas in the industrial model can be identified. However, for smaller areas in the complex area, further refinement is required to obtain a more accurate 3D model of the industrial model. The specific method is that after supplementing the data of the first position, the depth map of each perspective is regenerated and updated, and the depth maps of different perspectives are fused. The overlapping areas and conflicting data are processed by the algorithm to construct a complete three-dimensional model, and the depth maps of different perspectives and the fused three-dimensional model are compared. The location of the missing data that may still exist is identified through algorithm analysis, that is, the second position. The second position may be a smaller scanning blind area or a more detailed hole. For such places that are not easy to scan, an appropriate interpolation algorithm, such as linear interpolation, bilinear interpolation, cubic interpolation, etc., is used to generate interpolation data. The specific interpolation algorithm belongs to the prior art and will not be repeated here. The interpolation data is converted into the form of 3D point cloud data to ensure that it is consistent with the original point cloud data in format. The interpolation data is gradually filled in the second position along the shape of the industrial model and merged into the first 3D point cloud data. In the fusion process, it is necessary to ensure a smooth transition and seamless connection between the interpolation data and the original data. The completeness and accuracy of 3D point cloud data are further improved through updating and fusion of depth maps and identification and filling of secondary missing data. The data processing process is simplified and processing efficiency is improved through automated data identification, interpolation and filling processes. After multiple supplements and improvements, the 3D point cloud data further improves the completeness of the data and the accuracy of the model.

[0039] Furthermore, obtaining the second 3D point cloud data of the industrial model also includes: Based on multiple first positions, a rotation path of the rotating platform is determined, and based on the number of first positions, a second rotation speed of the rotating platform is determined. The first scanning area is scanned based on the rotation path and the second rotation speed of the rotating platform, and when scanning from the edge to the center of each first position, the rotation speed of the rotating platform is converted from the second rotation speed to the third rotation speed, and when scanning from the center to the edge of each first position, the rotation speed of the rotating platform is converted from the third rotation speed to the second rotation speed, and second 3D point cloud data is acquired, wherein the first rotation speed is greater than the second rotation speed, and the second rotation speed is greater than the third rotation speed.

[0040] Specifically, by determining the rotation path of the rotating platform, it is possible to ensure that the scanning process is more efficient, reduce unnecessary scanning time and repetitive work, and improve scanning efficiency. The second rotation speed of the rotating platform is determined according to the number of the first position. When it is necessary to scan the edge of the first position, the second rotation speed can be used for scanning. The closer the scan is to the center of the first position, the finer the model. Therefore, in order to obtain more accurate data, it is necessary to appropriately reduce the rotation speed and convert it to the third rotation speed for scanning to ensure that the range of the first position can be fully scanned and more detailed data can be collected. Conversely, the third rotation speed is converted to the second rotation speed to improve scanning efficiency. Optimizing the scanning path and dynamically adjusting the scanning speed can reduce data loss and improve the quality and accuracy of the 3D point cloud data finally acquired.

[0041] Further, identifying the position of the missing data as the first position based on the depth map further includes: Based on the depth map, the distribution of the first 3D point cloud data at different depth levels is visualized, and the blank or discontinuous areas on the depth map are defined as the locations of missing data.

[0042] Specifically, first, the depth map is visualized. The depth map is essentially a two-dimensional image, in which the value of each pixel represents the distance (i.e., depth) from the corresponding point in the 3D point cloud data to the virtual 2D plane. Through appropriate color mapping or grayscale representation, these depth information can be displayed in an intuitive way. On the visualized depth map, the algorithm identifies blank or discontinuous areas, which usually correspond to the missing data part in the first 3D point cloud data, i.e., the first position. The location of the missing data is identified, confirmed and relevant information is recorded for subsequent local refinement scanning or data interpolation processing. The implementation method is to record the coordinates of each missing data position, the pixel coordinates on the depth map, and evaluate the size and shape of the missing area to determine whether further refinement scanning is required. Through the visualization of the depth map, the missing part in the first 3D point cloud data can be intuitively identified. This method not only improves the efficiency of data processing, but also improves the accuracy of recognition.

[0043] Furthermore, the interpolation data is fused into the first 3D point cloud data, further comprising: The original scan data and the interpolated data in the first 3D point cloud data are marked, and overlapping data in the first 3D point cloud data are deleted based on the original scan data and the interpolated data.

[0044] Specifically, in order to distinguish the original scan data in the first 3D point cloud data from the interpolation data generated later by the interpolation algorithm, they need to be marked. When fusing the interpolation data into the first 3D point cloud data, the original scan data and the interpolation data may overlap in spatial position. These overlapping data not only increase the complexity of data processing, but also may affect the accuracy and precision of the final 3D model. Therefore, these overlapping data need to be deleted to ensure that the fused first 3D point cloud data is both complete and accurate.

[0045] Furthermore, the above step S4 further includes: Step S41: aligning the first 3D point cloud data of the industrial model with the preset reference data of the reference object, and determining the best spatial transformation mode so that the first 3D point cloud data and the preset reference data overlap in space, analyzing the difference between the first 3D point cloud data and the preset reference data, and extracting the first minimum deviation; Step S42: when the first minimum deviation is greater than or equal to a third threshold, calculating the deviation between each point cloud data in the first 3D point cloud data and the preset reference data, and determining a correction vector for each point cloud data based on the deviation, wherein the deviation includes a deviation direction and a deviation magnitude; Step S43: integrating the correction vector of each point cloud data into the correction matrix, and applying the correction vector to the first 3D point cloud data of the industrial model based on the correction matrix; Step S44: After the first 3D point cloud data of the industrial model is corrected, the difference between the first 3D point cloud data and the preset reference data is re-analyzed, and the second minimum deviation is extracted. If the second minimum deviation is still greater than or equal to the third threshold, the third position where the deviation occurs is obtained from the first 3D point cloud data, and the third position is re-scanned.

[0046] Specifically, in the industrial model scanning method based on 3D point cloud data, it involves accurately aligning and correcting the first 3D point cloud data of the industrial model to ensure its high consistency with the preset reference data. The implementation method is to spatially align the first 3D point cloud data of the industrial model with the preset reference data of the reference object to eliminate the spatial deviation caused by the scanning position, angle or equipment error. The deviation amount between each point cloud data in the first 3D point cloud data and the preset reference data is calculated, including the deviation direction and deviation size. When the first minimum deviation is greater than or equal to the preset third threshold, it means that there is still a certain deviation in the aligned point cloud data and correction is required. Based on the deviation amount, a correction vector is determined for each point cloud data, which indicates the direction and distance to be moved in order to reduce the deviation. A correction matrix can be constructed, which contains the correction vector information of all point cloud data, and the correction matrix is ​​applied to transform the first 3D point cloud data to eliminate or reduce the deviation from the preset reference data. After each iteration, the difference between the corrected first 3D point cloud data and the preset reference data is reanalyzed, and the second minimum deviation is extracted. If the second minimum deviation is still greater than or equal to the third threshold, it means that the scan data of this position is inaccurate and the position needs to be rescanned. The specific scanning method can be the same as the scanning method of the first position in the above scheme. Through a series of precise alignment and correction operations, the high consistency between the first 3D point cloud data of the industrial model and the preset reference data is ensured.

[0047] The present invention also provides a Figure 2 The industrial model scanning system based on 3D point cloud data shown includes the following modules: A positioning unit, used to fix the industrial model to be scanned and the reference object on the same rotating platform, the industrial model is located in the middle of the reference object, and the reference object is a plurality of identical objects; The scanning unit is configured to: start the rotating platform and the 3D scanning device, the rotating platform rotates at a first rotation speed according to a preset sequence, and the 3D scanning device scans the industrial model and the reference object to obtain first 3D point cloud data of the industrial model; Based on the first 3D point cloud data, a first position is identified and obtained, a first scanning area is determined based on the first position, a 3D scanning device scans the first scanning area, and obtains second 3D point cloud data of the industrial model, and the second 3D point cloud data is merged into the first 3D point cloud data, and then inputted into the control unit; A generating unit is used to control the control unit to receive the first 3D point cloud data, calculate the error between the first 3D point cloud data and the preset reference data of the reference object, generate a correction vector based on the result of the error calculation, apply the correction vector to the originally acquired first 3D point cloud data, and obtain the adjusted first 3D point cloud data, and recombine the adjusted first 3D point cloud data to generate a 3D model of the industrial model.

[0048] A computing device, comprising: Memory and processor; The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the above-mentioned industrial model scanning method based on 3D point cloud data is implemented.

[0049] In summary, the present invention fixes the industrial model and the reference object to be scanned on the same rotating platform to ensure that the relative position of the industrial model and the reference object is stable during the scanning process. Based on the driving instruction issued by the control unit, the rotating platform and the 3D scanning device are started, and the rotating platform rotates at a first rotation speed in a preset order. At the same time, the 3D scanning device scans the industrial model and the reference object to achieve an all-round scan of the industrial model and the reference object, and obtains comprehensive and continuous 3D point cloud data, while also improving the scanning efficiency. Based on the first 3D point cloud data obtained, the first position is identified and obtained, and the first scanning area is determined based on the first position. The 3D scanning device scans the first scanning area and obtains the second 3D point cloud data of the industrial model to improve the scanning accuracy of the complex area in the industrial model, enhance the integrity of the overall point cloud data, and fuse the second 3D point cloud data into the first 3D point cloud data to form a more complete point cloud data set.

[0050] The control unit receives the first 3D point cloud data, calculates the error between the first 3D point cloud data and the preset reference data of the reference object, generates a correction vector based on the result of the error calculation, and applies the correction vector to the originally acquired first 3D point cloud data to eliminate or reduce the error in the scanned data and improve the accuracy of the data. The adjusted first 3D point cloud data is obtained, and the adjusted first 3D point cloud data is recombined to generate a high-precision 3D model of the industrial model. Through the coordination between the above steps, the present invention improves the scanning accuracy and efficiency of the industrial model, and the method is also suitable for fast and accurate scanning of various complex industrial models.

[0051] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0052] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the above-mentioned program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0053] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0054] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

[0055] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An industrial model scanning method based on 3D point cloud data, characterized in that: The method comprises the following steps: Step S1: fixing the industrial model to be scanned and the reference object on the same rotating platform, wherein the industrial model is located in the middle of the reference object, and the reference object is a plurality of identical objects; Step S2: based on the driving instruction issued by the control unit, the rotating platform and the 3D scanning device are started, the rotating platform rotates at a first rotation speed according to a preset sequence, and the 3D scanning device scans the industrial model and the reference object to obtain first 3D point cloud data of the industrial model; Step S3: Based on the first 3D point cloud data, a first position is identified and acquired, and a first scanning area is determined based on the first position. The 3D scanning device scans the first scanning area and acquires second 3D point cloud data of the industrial model. The second 3D point cloud data is fused into the first 3D point cloud data, and then input into the control unit; Step S4: The control unit receives the first 3D point cloud data, and calculates the error between the first 3D point cloud data and the preset reference data of the reference object, generates a correction vector based on the result of the error calculation, applies the correction vector to the originally acquired first 3D point cloud data, and obtains the adjusted first 3D point cloud data, and recombines the adjusted first 3D point cloud data to generate a 3D model of the industrial model.

2. The method according to claim 1, characterized in that In the step S2, obtaining the first 3D point cloud data of the industrial model further includes: The 3D scanning device emits a laser beam to the industrial model and records the emission time point. The laser beam is reflected after being irradiated to the measuring point on the industrial model. The receiver on the 3D scanning device captures the reflected laser beam and records the time point of capturing the laser beam as the receiving time point. Based on the receiving time point and the transmitting time point, calculating the time difference between the receiving time point and the transmitting time point, and converting the time difference into the distance between the laser transmitting point and the industrial model measuring point; As the rotating platform and the 3D scanning device move, the distance between the laser emission point and each measuring point on the industrial model is continuously recorded, and the distance data is converted into points in three-dimensional space to form first 3D point cloud data.

3. The method according to claim 1, characterized in that The step S3 further comprises: Preprocessing the first 3D point cloud data, and projecting the preprocessed first 3D point cloud data onto a 2D plane from multiple perspectives to generate multiple two-dimensional images corresponding to the first 3D point cloud data, and recording the distance from each of the first 3D point cloud data to the 2D plane as depth information, and creating depth maps of different perspectives based on the two-dimensional image and the depth information, wherein the 2D plane is a virtual plane set on the back of the first 3D point cloud data; Based on the depth map, a position where the missing data is identified as a first position, a size of the first position is measured, and multiple first positions whose sizes of the first position are greater than or equal to a first threshold are obtained. A first scanning area is determined based on the first position. The 3D scanning device first scans the first scanning area from an edge to a center of each of the first positions, and then scans from the center to an edge of each of the first positions to obtain second 3D point cloud data of the industrial model, and the second 3D point cloud data is merged into the first 3D point cloud data to supplement the data of the first position, and new first 3D point cloud data after the data of the first position is supplemented is obtained.

4. The method according to claim 3, characterized in that After acquiring the new first 3D point cloud data after supplementing the data of the first position, the method further includes: The depth map is updated, and the depth maps of different perspectives are merged together to form a complete three-dimensional model, the depth maps of different perspectives and the three-dimensional model are compared, and the position of the missing data is identified as the second position again, an interpolation algorithm is applied to obtain interpolation data of the second position, and the interpolation data is converted into the form of 3D point cloud data, and the interpolation data is gradually filled in the second position along the shape of the industrial model, and merged into the first 3D point cloud data to obtain new first 3D point cloud data.

5. The method according to claim 3, characterized in that: Acquiring second 3D point cloud data of the industrial model further includes: Based on multiple first positions, a rotation path of the rotating platform is determined, and based on the number of the first positions, a second rotation speed of the rotating platform is determined. The first scanning area is scanned based on the rotation path and the second rotation speed of the rotating platform, and when scanning from the edge to the center of each first position, the rotation speed of the rotating platform is converted from the second rotation speed to the third rotation speed, and when scanning from the center to the edge of each first position, the rotation speed of the rotating platform is converted from the third rotation speed to the second rotation speed, to acquire second 3D point cloud data, wherein the first rotation speed is greater than the second rotation speed, and the second rotation speed is greater than the third rotation speed.

6. The method according to claim 3, characterized in that Identifying the position of missing data as a first position based on the depth map further includes: Based on the depth map, the distribution of the first 3D point cloud data at different depth levels is visualized, and blank or discontinuous areas on the depth map are defined as locations of missing data.

7. The method according to claim 4, characterized in that The interpolation data is fused into the first 3D point cloud data, further comprising: The original scan data and the interpolated data in the first 3D point cloud data are marked, and overlapping data in the first 3D point cloud data are deleted based on the original scan data and the interpolated data.

8. The method according to claim 1, characterized in that The step S4 further includes: Step S41: aligning the first 3D point cloud data of the industrial model with the preset reference data of the reference object, and determining the best spatial transformation mode so that the first 3D point cloud data and the preset reference data overlap in space, analyzing the difference between the first 3D point cloud data and the preset reference data, and extracting a first minimum deviation; Step S42: when the first minimum deviation is greater than or equal to a third threshold, calculating the deviation between each point cloud data in the first 3D point cloud data and the preset reference data, and determining a correction vector for each point cloud data based on the deviation, wherein the deviation includes a deviation direction and a deviation magnitude; Step S43: integrating the correction vector of each point cloud data into a correction matrix, and applying the correction vector to the first 3D point cloud data of the industrial model based on the correction matrix; Step S44: After the first 3D point cloud data of the industrial model is corrected, the difference between the first 3D point cloud data and the preset reference data is re-analyzed, and the second minimum deviation is extracted. If the second minimum deviation is still greater than or equal to the third threshold, the third position where the deviation occurs is obtained from the first 3D point cloud data, and the third position is re-scanned.

9. An industrial model scanning system based on 3D point cloud data, used to implement the method according to any one of claims 1 to 8, characterized in that: The system includes the following modules: A positioning unit, used to fix the industrial model to be scanned and a reference object on the same rotating platform, wherein the industrial model is located in the middle of the reference object, and the reference object is a plurality of identical objects; The scanning unit is configured to: start the rotating platform and the 3D scanning device, the rotating platform rotates at a first rotation speed according to a preset sequence, and the 3D scanning device scans the industrial model and the reference object to obtain first 3D point cloud data of the industrial model; Based on the first 3D point cloud data, a first position is identified and acquired, and a first scanning area is determined based on the first position, the 3D scanning device scans the first scanning area and acquires second 3D point cloud data of the industrial model, and the second 3D point cloud data is merged into the first 3D point cloud data, and then inputted into the control unit; A generating unit is used for the control unit to receive the first 3D point cloud data, calculate the error between the first 3D point cloud data and the preset reference data of the reference object, generate a correction vector based on the result of the error calculation, apply the correction vector to the originally acquired first 3D point cloud data, and obtain the adjusted first 3D point cloud data, recombine the adjusted first 3D point cloud data, and generate a 3D model of the industrial model.

10. A computing device, comprising: Memory and processor; The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the industrial model scanning method based on 3D point cloud data described in any one of claims 1 to 8 is implemented.

Images