CN117191781A - Nondestructive testing system and method for micro array hole through hole rate of composite wallboard - Google Patents
Nondestructive testing system and method for micro array hole through hole rate of composite wallboard Download PDFInfo
- Publication number
- CN117191781A CN117191781A CN202310425659.0A CN202310425659A CN117191781A CN 117191781 A CN117191781 A CN 117191781A CN 202310425659 A CN202310425659 A CN 202310425659A CN 117191781 A CN117191781 A CN 117191781A
- Authority
- CN
- China
- Prior art keywords
- data
- point cloud
- hole
- dimensional
- detection
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 80
- 239000002131 composite material Substances 0.000 title claims abstract description 32
- 238000002493 microarray Methods 0.000 title claims abstract description 30
- 238000009659 non-destructive testing Methods 0.000 title claims abstract description 21
- 238000001514 detection method Methods 0.000 claims abstract description 119
- 238000012545 processing Methods 0.000 claims abstract description 55
- 238000007405 data analysis Methods 0.000 claims abstract description 30
- 238000000605 extraction Methods 0.000 claims abstract description 22
- 238000001914 filtration Methods 0.000 claims abstract description 11
- 238000005259 measurement Methods 0.000 claims description 40
- 238000004422 calculation algorithm Methods 0.000 claims description 38
- 238000007781 pre-processing Methods 0.000 claims description 23
- 238000007726 management method Methods 0.000 claims description 20
- 238000004458 analytical method Methods 0.000 claims description 17
- 230000008676 import Effects 0.000 claims description 15
- 230000000007 visual effect Effects 0.000 claims description 15
- 238000004364 calculation method Methods 0.000 claims description 14
- 239000012636 effector Substances 0.000 claims description 13
- 238000012800 visualization Methods 0.000 claims description 13
- 230000008569 process Effects 0.000 claims description 11
- 238000013523 data management Methods 0.000 claims description 9
- 238000013079 data visualisation Methods 0.000 claims description 6
- 238000007689 inspection Methods 0.000 claims description 6
- 238000011068 loading method Methods 0.000 claims description 6
- 238000009877 rendering Methods 0.000 claims description 6
- 230000005540 biological transmission Effects 0.000 claims description 5
- 230000011218 segmentation Effects 0.000 claims description 5
- 230000000694 effects Effects 0.000 claims description 4
- 238000009434 installation Methods 0.000 claims description 4
- 230000001066 destructive effect Effects 0.000 claims 5
- 239000013598 vector Substances 0.000 description 20
- 239000000523 sample Substances 0.000 description 19
- 238000002474 experimental method Methods 0.000 description 14
- 230000006870 function Effects 0.000 description 12
- 238000010586 diagram Methods 0.000 description 10
- 239000011159 matrix material Substances 0.000 description 9
- 230000009466 transformation Effects 0.000 description 8
- 230000000903 blocking effect Effects 0.000 description 7
- 239000011148 porous material Substances 0.000 description 5
- 238000012360 testing method Methods 0.000 description 5
- 238000012795 verification Methods 0.000 description 5
- 230000003993 interaction Effects 0.000 description 4
- 238000009826 distribution Methods 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 230000001788 irregular Effects 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 238000004088 simulation Methods 0.000 description 3
- 238000013519 translation Methods 0.000 description 3
- 239000011157 advanced composite material Substances 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000000105 evaporative light scattering detection Methods 0.000 description 2
- 230000005484 gravity Effects 0.000 description 2
- 239000012535 impurity Substances 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000013450 outlier detection Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 229910000838 Al alloy Inorganic materials 0.000 description 1
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 239000008186 active pharmaceutical agent Substances 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000003754 machining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 238000013439 planning Methods 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000007794 visualization technique Methods 0.000 description 1
Landscapes
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention discloses a nondestructive testing system and method for the micro array hole through hole rate of a composite wallboard. And generating three-dimensional point cloud data, filtering to remove noise, finely extracting the three-dimensional point cloud data of the micro-holes on the detection plane, analyzing and judging whether the holes are plugged, and customizing a detection data analysis report template and outputting a data analysis report by the output report module. The method can perform denoising pretreatment on the imported point cloud data, so that the data is smooth, the scanning data can be subjected to tiny Kong Jingxi extraction and hole quality detection, and a unified mathematical expression model is formed by analyzing the mode of the micropores, so that an accurate quality detection result is obtained. Practical application proves that the method can remarkably improve the accuracy and the integrity of point cloud data processing, quicken the processing speed of the point cloud data, improve the degree of automation and humanization, and ensure that the through hole rate detection work is normally carried out.
Description
Technical Field
The invention belongs to the technical field of quality detection of composite wall boards, and particularly relates to a nondestructive detection system and method for micro array hole through hole rate of a composite wall board.
Background
The application of the advanced composite material in the aircraft structure is compatible with the aluminum alloy due to the excellent performances of high strength, high temperature resistance, corrosion resistance, light weight and the like, and the advanced composite material becomes the field of the most rapid development of the current material technology. The level of aeronautical composite material performance and its application in structures has become an important indicator of the structural advancement of aircraft. The processing precision of the composite material wallboard (composite material wallboard) is one of important indexes in the aircraft manufacturing process. The micro array hole through hole rate detection of the composite material wallboard is one step of quality detection, the traditional method generally adopts a probe to carry out contact measurement, the advantage is that deeper hole depth can be measured, the disadvantage is that the detection efficiency is low due to the fact that the array holes have relatively more small holes, time and labor are consumed when the traditional method is used, the overall speed of aircraft manufacturing is indirectly reduced, and in addition, unknown influence on hole making quality is possibly caused by contact measurement. Therefore, the detection of the through hole rate of the micro array holes of the composite wallboard is realized in an efficient, nondestructive and automatic mode, and the method is particularly important for improving the production efficiency.
Disclosure of Invention
The invention aims to provide a nondestructive testing system and method for the micro array hole through hole rate of a composite wallboard, and aims to solve the problem of the method.
The invention is realized mainly by the following technical scheme:
the nondestructive testing system for the micro array hole through rate of the composite wallboard comprises a data acquisition and processing subsystem and a measuring robot subsystem, wherein the data acquisition and processing subsystem comprises a data preprocessing module, a data analysis module and an output report module; the data preprocessing module is used for generating three-dimensional point cloud data and filtering and noise filtering preprocessing, the data analysis module is used for finely extracting micro-hole three-dimensional point cloud data on a detection plane and analyzing and judging whether the micro-hole three-dimensional point cloud data is a plugged hole or not, and the output report module is used for customizing a detection data analysis report template and outputting a data analysis report.
In order to better realize the invention, the data acquisition and processing subsystem further comprises a data visualization module, a data import and export module and a data management module; the data visualization module is used for managing and visualizing the data, loading the data, displaying the rendered point cloud data and displaying the imported data; the data import and export module is used for reading, processing and outputting data and reading various parameters of the detection and analysis system in real time according to a transmission protocol; the data management module is used for carrying out coordinate system management, data file management and data output management, and carrying out historical data analysis, viewing and editing on the original data and the calculation result stored in the database.
In order to better realize the invention, the measuring robot subsystem further comprises a measuring robot, wherein two oppositely arranged laser scanners are arranged on the measuring robot, the light emitting directions of the two laser scanners are parallel to each other, the light emitting positions are flush, and the laser scanners are used for acquiring point cloud data of a hole to be measured and transmitting the point cloud data to the data acquisition and processing subsystem.
The invention is realized mainly by the following technical scheme:
the nondestructive testing method for the micro array hole through hole rate of the composite wallboard is carried out by adopting the nondestructive testing system, and the detection analysis is carried out by adopting a data acquisition and processing subsystem, and comprises the following steps:
step S100: according to a scanning path planned by the measuring robot, acquiring a data time sequence by using a laser scanner, splicing two-dimensional point cloud outlines, and generating three-dimensional point cloud data;
step S200: filtering and denoising preprocessing is carried out on the obtained three-dimensional point cloud data;
step S300: sequentially carrying out boundary extraction and cluster segmentation on the preprocessed three-dimensional point cloud data to extract hole features in the point cloud data;
step S400: and detecting the quality of the holes, and judging whether the holes are plugged.
In order to better implement the present invention, further, the step S100 includes the steps of:
Step S110: reading the motion parameters, the motion path and the scanning frequency of a laser scanner of the measuring robot, calculating the acquisition time of each group of two-dimensional coordinate values from detection, and calculating the position of the end effector of the measuring robot in the track when each group of two-dimensional coordinate values are acquired;
step S120: calculating the position relation between two scanner data coordinate systems according to the installation position of the fixed tool, converting the data of the scanner B into the coordinate system of the scanner A, and combining the two groups of coordinate values into one group;
step S130: dividing the data by taking the position adjacent relation obtained by calculation in the step S110 as a reference and a certain time period, and dividing all the two-dimensional coordinate value components into a plurality of subgroups;
step S140: according to the uniform motion relation, under the condition that the measurement intervals are the same, the distance between each two-dimensional coordinate value is constant, so that multiple groups of two-dimensional measurement values of each group are arranged along the Y direction to form local three-dimensional data;
step S150: and (3) forming local three-dimensional data by using a method of the step (S140) from a plurality of groups of two-dimensional measured values in all subgroups, wherein the track position corresponding to the acquisition moment of the two-dimensional coordinate values of the first group in each subgroup is used as a positioning standard, and the generation of the three-dimensional point cloud data is completed.
In order to better implement the present invention, in step S200, a weight angle criterion is used to detect a noise point, and when the boundary attribute is greater than a set threshold, the noise point is determined as the noise point; or detecting a noise point by adopting a neighboring point criterion, and judging the point as the noise point when the point in the radius of the fixed neighborhood is smaller than a set threshold value; and processing the point cloud data containing the noise points and the outliers by using an FCM algorithm, and retaining the edge characteristics of the holes in the point cloud while removing the noise points.
In order to better implement the present invention, further, the step S400 includes the steps of:
step S410: inputting hole edge point cloud data to be detected, and searching neighborhood data within a certain radius range in the original point cloud;
step S420: according to the neighborhood data fitting plane, calculating the normal direction of the plane;
step S430: fitting by using the hole characteristic point cloud to obtain a fitting circle center, and calculating the three-dimensional coordinates of the circle center;
step S440: taking the center coordinates as a searching starting point, taking the plane normal direction as a searching direction, setting the searching distance as a certain fixed value, and setting the point cloud obtained by searching as hole bottom data;
step S450: if the hole bottom point cloud data does not exist or the calculated distance is larger than a set threshold value, the hole is considered to be a through hole; if the calculated distance is smaller than the set threshold value, performing secondary judgment;
Step S460: the projection of the hole bottom point cloud is simplified, the projection area of the plug adjacent to the fitting plane in the hole center is calculated, if the projection area of the plug is larger than a set threshold value, the hole is judged to be plugged, and otherwise, the hole is a through hole;
step S470: counting the number of all detected holes and the number of the blocked holes, calculating the hole passing rate, and recording the corresponding number of rows and columns of each blocked hole in the digital-analog, so that a detector can quickly position the blocked holes.
To better implement the present invention, further comprising visualizing the data:
step A1: constructing an octree point cloud space index structure of the OBB outer surrounding box, and reading point cloud data;
step A2: constructing an LOD model based on the octree point cloud space index structure;
step A3: selecting LOD (on-demand) levels to be visualized based on the distance of the view cone viewpoint, and selecting nodes according to the visibility judgment of view cone shielding rejection;
step A4: when the visual angle changes, according to the visual angle information of the current frame, preloading the nodes possibly used by the next frame in advance;
step A5: reading the node selected in the step S3, and performing visual rendering;
step A6: and when the visual angle is unchanged, performing point cloud self-adaptive filling, and improving the point cloud visualization effect.
To better implement the present invention, further, the method further includes importing and exporting data:
step B1: after the detection flow starts, a related data interface is called to read scanning data of the laser scanner and a motion track of an end effector of the measuring robot in real time;
step B2: inputting a corresponding three-dimensional point cloud data generation module;
step B3: invoking a related point cloud data format template aiming at the three-dimensional point cloud data type;
step B4: corresponding data import and export codes are written.
In order to better implement the invention, further, the detection is performed by using a measuring robot subsystem, comprising the following steps:
step S1: the method comprises the steps that a measured workpiece is mounted on a detection platform and is positioned and clamped through a tool, and meanwhile the measured workpiece is divided into a plurality of waiting areas;
step S2: adjusting the pose of the measuring robot to an initial detection position;
step S3: the measuring robot carries a laser scanner to move at a uniform speed, the light emitting position of a laser line is equal to the distance between the surface to be measured, the laser scanner scans and detects a first area to be measured on a workpiece, and point cloud data are collected and stored in real time to finish the condition detection of the through hole of the area to be measured;
Step S4: the measuring robot carries the laser scanner to the next station, and the step S2 and the step S3 are repeated until all areas to be measured are detected;
step S5: the data acquisition and processing subsystem processes and analyzes the point cloud data to obtain the through hole rate and marks the blocked hole;
step S6: outputting the detection result.
The beneficial effects of the invention are as follows:
(1) According to the invention, single measurement data of the laser scanner is realized through the data preprocessing module, two-dimensional contour data are spliced according to the motion trail of the end effector of the measuring robot planned by the measuring hardware system, and three-dimensional point cloud data are generated;
(2) According to the invention, operations such as micro Kong Jingxi extraction and hole quality detection on scanning data can be realized through the data analysis module, and in addition, the scale of the measured three-dimensional point cloud data is larger than that of the hole, so that a uniform mathematical expression model is formed by analyzing the mode of the micro holes, and an accurate quality detection result can be obtained. Practical application proves that the accuracy and the integrity of point cloud data processing can be obviously improved, the processing speed of the point cloud data is accelerated, the degree of automation and the degree of humanization are improved, and the through hole rate detection work is ensured to be normally carried out.
(3) The invention supports the visual management of mass point cloud data, the display of color and intensity information point clouds and the management organization function of multiple projects, and is convenient for user interaction.
(4) The data input of the invention is two-dimensional and three-dimensional point cloud data, and the data supported by the system for reading, writing and processing comprises but is not limited to PCD, STL, TXT and other common three-dimensional point cloud storage formats.
(5) The invention has perfect coordinate system management, measurement data file management and measurement data output management functions, and can realize historical data analysis, checking and editing.
Drawings
FIG. 1 is a schematic diagram of a micro array of holes on a surface of a composite wall panel;
FIG. 2 is a schematic diagram of a measurement robot workspace;
FIG. 3 is a schematic diagram of a measurement robot load capacity;
FIG. 4 is a diagram of the overall architecture of the nondestructive inspection system of the present invention;
fig. 5 is a sectional view of a region to be measured of a part to be measured according to an embodiment of the present invention.
FIG. 6 is a schematic flow chart of the nondestructive testing of the present invention;
FIG. 7 is a schematic diagram of a sample for detection experiments;
FIG. 8 is a schematic block diagram of a sample detection experiment calibration method;
FIG. 9 is a schematic diagram of sample detection experiment laser scanner acquisition data;
FIG. 10 is a schematic diagram of the result of sample detection experiment data stitching;
FIG. 11 is a schematic diagram of error detection of the splice result of sample detection experimental data;
FIG. 12 is a schematic block diagram of a sample detection experiment detection algorithm;
FIG. 13 is a schematic view of point cloud data and analysis of a sample detection experiment single well;
FIG. 14 is a photograph showing the actual sample test;
fig. 15 is a graph of the results of the algorithm in the sample detection experiment.
Detailed Description
Example 1:
a nondestructive testing system for the through hole rate of a micro array of holes of a composite wallboard, as shown in fig. 4, comprises a data acquisition and processing subsystem:
and a data preprocessing module: for splicing two-dimensional point cloud contours according to a scanning path planned by a measuring robot by utilizing a laser scanner to acquire data time sequences, generating three-dimensional point cloud data, and filtering and denoising the obtained three-dimensional point cloud data;
and a data analysis module: the method is used for identifying and finely extracting three-dimensional point cloud data of micro holes on a detection plane, forming a unified mathematical expression model by analyzing the modes of the micro holes, detecting the quality of the holes on the basis, and judging whether the holes are plugged or not;
further, the data acquisition and processing subsystem further comprises the following modules:
and a visualization module: the method is used for data management, data loading and visualization, point cloud data display after rendering and imported data display;
And a data import and export module: the system is used for reading, processing and outputting data and reading various parameters of the detection and analysis system in real time according to a transmission protocol;
and a data management module: the system is used for carrying out coordinate system management, data file management and data output management, and carrying out historical data analysis, checking and editing on the original data and calculation results stored in the database;
and a report outputting module: the method is used for customizing the detection data analysis report template and outputting a corresponding data analysis report according to the template.
Preferably, the data analysis report output by the output report module includes: product name, number of detection holes of detection elements, number of blocked holes, hole passing rate, position coordinates of the blocked holes in the point cloud data and related measurement parameters.
Further, still include measurement robot subsystem, measurement robot subsystem includes measurement robot, switch board, demonstrator, precision compensation module and off-line programming software, wherein:
the precision compensation module is used for calibrating the measuring robot, establishing a model between the space point of the measuring robot and the error of the measuring robot, compensating the space coordinate of the measuring robot in real time in the operation process of the measuring robot, and correcting the absolute positioning error so as to ensure the absolute positioning precision of the measuring robot.
Further, the data acquisition and processing subsystem comprises two oppositely arranged laser scanners, the light emergent directions of the two laser scanners are parallel to each other, and the light emergent positions are flush.
Further, the measuring robot subsystem can be used for detecting the surface of an irregular product and guaranteeing that the laser line emitting direction of the laser scanner is perpendicular to the surface to be measured and the hole to be measured is in the scanning range of the laser scanner.
Further, the two laser scanners respectively acquire point cloud data of the hole to be detected, the data acquisition and processing subsystem splices the two groups of data, and the splicing result is used as basic data for judging the hole blocking, so that detection blind areas caused by a triangular reflection principle of a single sensor are avoided.
The method can perform denoising pretreatment on the imported point cloud data, so that the data is smooth, the scanning data can be subjected to tiny Kong Jingxi extraction and hole quality detection, and a unified mathematical expression model is formed by analyzing the mode of the micropores, so that an accurate quality detection result is obtained. Practical application proves that the method can remarkably improve the accuracy and the integrity of point cloud data processing, quicken the processing speed of the point cloud data, improve the degree of automation and humanization, and ensure that the through hole rate detection work is normally carried out.
Example 2:
a nondestructive testing system for the through hole rate of micro array holes of a composite wall plate is shown in figure 1 and is used for detecting the through hole rate of the micro array holes of the composite wall plate. As shown in fig. 4, the system includes a data visualization module, a data import and export module, a data preprocessing module, a data analysis module, a data management module and an output report module.
Preferably, the UI interface is mainly divided into a menu bar, a toolbar, a data tree, a measurement window, a visualization window, a shortcut tool and sub-windows of each module, wherein:
menu bar: the system comprises files, data processing, data management, display, plug-in units, views and help tabs, and has the main functions of some point cloud data processing basic operations such as reading, writing, loading and displaying of point clouds, and provides access to various modules such as query of a database, report output and the like.
Tool bar: the integration of common functions mainly comprises the functions of opening, storing and deleting point cloud, operating a database, outputting a report of a measurement result and the like.
Data tree: and displaying the loaded point cloud data structure information, selecting one or more point clouds for operation, such as hole blocking detection and the like, and controlling the point clouds displayed in the visualization area.
Visualization area (window): and displaying and interacting all the visualized information of the point cloud data and the three-dimensional digital-analog in the window area, performing operations such as zooming the point cloud, interactive inquiry on a measurement result and the like, and displaying a color difference model for plugging the hole according to the measurement result.
A shortcut tool: mainly provides operations for fast control of the visualization window, such as selecting laser scanner position, setting rotation center, observing model data from various directions (up and down, left and right, front and back) according to coordinate data, etc.
Measurement result area (window): the method mainly provides the result display aiming at the system customization function, such as the display of three-dimensional coordinates in local data of the plugged holes, the display of the through hole ratio and the like.
Preferably, (a) a data visualization module: the method is used for managing mass data, loading and visualizing large-scale data, displaying point cloud data after rendering and displaying imported data. Preferably, the main algorithm of the loading and visualization method of the three-dimensional point cloud data comprises the following steps:
s1, constructing an octree point cloud space index structure of an OBB outer surrounding box, and reading point cloud data;
s2, constructing an LOD model based on an octree point cloud space index structure;
s3, selecting LOD levels to be visualized based on the view point distance of the view cone, and selecting nodes according to the visibility judgment of view cone shielding rejection;
S4, when the view angle changes, preloading nodes possibly used by the next frame in advance according to the view angle information of the current frame;
s5, reading the node selected in the S3, and performing visual rendering;
and S6, when the visual angle is unchanged, performing point cloud self-adaptive filling, and improving the point cloud visualization effect.
UI and interaction: clicking to open the file, selecting the file to be processed, or calling the module in the stage of displaying the detection result and opening the point cloud data with the blocked holes; the input point cloud data is visualized, and rotation, translation, scaling, clipping and deleting operations of the large-scale point cloud data can be supported.
Preferably, the invention supports rotation, translation, scaling, cropping, deletion of large-scale point cloud data. The mouse clicks the point cloud, and the left mouse button is set in the full selection state: select region, ctrl+left key: deselect region, delete: deleting the selected area, the mouse middle key: rotating, shift+mouse right button: scaling, alt+ middle bond: translation, S: a rotation center is set. The information is displayed below the operation interface, so that user interaction is facilitated.
Preferably, (two) a data import/export module: the system is used for reading, processing and outputting data, and reading various parameters of the detection and analysis system in real time according to a transmission protocol, wherein the read, processed and output data comprises, but is not limited to, point cloud files in a format of PCD, PLY, TXT and the like; the data import and export module supports reading parameters such as a motion track of the measuring robot and data acquisition frequency of the line laser, and controls hardware through a data transmission protocol to finish a measuring task. When the module supports connection with a measuring robot and a laser scanner, measuring data and parameters are imported to a computer through an API interface of the module, and control parameters are exported to a hardware system through the interface so as to realize real-time control according to a planned measuring path.
Preferably, the algorithm of the data import and export module includes:
s1, after a detection flow starts, calling a related data interface to read scanning data of a laser scanner and a motion track of an end effector of a measuring robot in real time;
s2, inputting a corresponding three-dimensional point cloud data generation module;
s3, aiming at the three-dimensional point cloud data type, calling a related point cloud data format template;
s4, writing corresponding data import and export codes.
UI and interaction: in the point cloud data generation stage, calling the module to acquire parameters of a laser scanner and a measuring robot; when the system starts to process data, the module is called to read point cloud data; after the detection flow is finished, the system outputs a detection result and visualizes the detection result, and a user can select to export a certain segment of point cloud data to a designated path. Clicking a 'file' button in a menu bar to import point cloud data to be processed. Clicking a 'start detection' button in a menu bar, enabling a data acquisition system to start detection, enabling a measuring robot end effector to clamp a laser scanner to acquire data, and finally segmenting to form local point cloud data.
Preferably, (iii) a data preprocessing module: the main functions of the data preprocessing module are generation of three-dimensional point cloud data and denoising of original point cloud data. Preferably, the data preprocessing module is used for acquiring data time sequence by using a laser scanner according to a scanning path planned by the measuring robot, splicing a plurality of two-dimensional point cloud contours to generate three-dimensional point cloud data, analyzing and diagnosing the data quality by using a spatial distribution rule of the three-dimensional point cloud data of the detection line laser, filtering and noise-removing preprocessing the three-dimensional point cloud data on the basis, filtering part of noise point cloud, and improving the data quality.
1. Three-dimensional point cloud data generation
According to the detection target, a laser scanner (2D laser profiler) is adopted to collect data of the object to be detected. The laser scanner, also called laser profiler, adopts the principle of laser triangle reflection, and can collect the two-dimensional profile information of different material surfaces. The laser beam is amplified to form a static laser line to be projected onto the surface of the object to be measured to form diffuse reflection through a special lens group, and the reflected light is projected onto a sensitive photosensitive matrix through a high-quality optical system. In addition to the distance information (Z-axis) of the laser scanner to the surface being measured, the controller can also calculate position information (X-axis) along the laser line from the image information. In a two-dimensional coordinate system with a scanner as an origin, the scanner measures and outputs a set of two-dimensional coordinate values. And the three-dimensional measurement can be realized by moving the object to be measured or the scanner probe.
According to the principle, three-dimensional point cloud data are required to be generated according to the measurement process for subsequent data analysis. And calculating the adjacent relation of the two-dimensional coordinate values collected by the laser scanner each time according to the motion parameters, the motion path and the scanning frequency of the scanner of the measuring robot. And for a plurality of groups of two-dimensional coordinate values, approximately converting the two-dimensional coordinate values into local plane data according to the adjacent relation on the track, and processing and integrating the plurality of groups of two-dimensional coordinate values in batches for subsequent data analysis.
2. Preprocessing of three-dimensional point cloud data
The local feature estimation operation of the point cloud is complex, the existence of the outlier directly affects subsequent operations such as curvature calculation, registration, feature extraction, curved surface reconstruction, visualization and the like, and even causes an erroneous calculation result, so that the basic pretreatment mainly is the removal of sparse outliers in the point cloud data, the subsequent processing work of the point cloud is greatly affected by the operation, and the analysis result of the point cloud data plays a very key role. Due to the fact that outliers are unorganized, disordered, sparse and inconsistent in density, outlier detection is generally complex, and traditional outlier detection methods mainly comprise a statistical distribution method, a distance method, a clustering method, a density method and the like.
Preferably, (four) a data analysis module: the method is used for identifying and finely extracting three-dimensional point cloud data of micro holes on a detection plane, forming a unified mathematical expression model by analyzing the modes of the micro holes, detecting the quality of the holes on the basis, and judging whether the holes are plugged. Preferably, the data analysis module detects the main plane where the round hole is located through a local sampling principle, provides a basis for micro-hole extraction, identifies the micro-hole based on a template matching and quick searching method, extracts relevant parameters, classifies detection elements into a mathematical expression model according to the requirement of a hole quality detection project, and self-adaptively adjusts the detection parameters based on the micro-hole extraction result to obtain an accurate quality detection result.
Preferably, (fifth) the data management module: the method is used for carrying out coordinate system management, data file management and data output management, and carrying out historical data analysis, viewing and editing on the original data and the calculation result stored in the database. The system needs to configure the source data of the result analysis report, including the system integrates the data of each business system to form the basic source data needed by the intelligent analysis report, and meanwhile, the system needs to configure part of three-dimensional point cloud data with the plugged holes and store the three-dimensional coordinate information of the part of three-dimensional point cloud data on the space track.
1. Coordinate system management
The method is used for managing coordinate system information of each group of two-dimensional coordinate values, recording the calculated coordinate transformation relation to a base coordinate system and providing data support for three-dimensional point cloud data generation.
2. Measurement data file management
And storing the measurement original data and the calculation result in a database, and analyzing, checking and editing historical data to provide data support for the generation of subsequent measurement reports.
3. Measurement data output management
The measurement results are subjected to statistical analysis visualization, the measurement results of the plugged holes and the coordinates of the measurement results under a reference coordinate system can be visually displayed, and an interactive whole machine measurement result display interface is provided for visual guidance of subsequent manual plugging detection.
Preferably, (six) output reporting module: the method is used for customizing the detection data analysis report template and outputting a corresponding data analysis report according to the template, wherein the output data analysis report comprises, but is not limited to, product names, the number of detection holes of detection elements, the number of plugging holes, the rate of holes, the position coordinates of the plugging holes in the point cloud data and the detailed information of the measurement parameters.
Example 3:
a nondestructive testing method for the through hole rate of a micro array hole of a composite wallboard is carried out by adopting the system and comprises the following steps:
(1) According to a scanning path planned by the measuring robot, acquiring a data time sequence by using a laser scanner, splicing two-dimensional point cloud outlines, and generating three-dimensional point cloud data;
(2) Filtering and denoising preprocessing is carried out on the obtained three-dimensional point cloud data;
(3) Sequentially carrying out boundary extraction and cluster segmentation on the preprocessed three-dimensional point cloud data to extract hole features in the point cloud data;
(4) And detecting the quality of the holes, and judging whether the holes are plugged.
Preferably, the three-dimensional point cloud data generation algorithm:
s1, reading motion parameters, a motion path and scanning frequency of a laser scanner of a measuring robot, calculating acquisition time of each group of two-dimensional coordinate values from detection, and calculating the position of an end effector of the measuring robot in a track when each group of two-dimensional coordinate values are acquired;
S2, calculating the position relation between the two scanner data coordinate systems according to the installation position of the fixed tool, converting the data of the scanner B into the coordinate system of the scanner A, and combining the two groups of coordinate values into one group;
s3, dividing the data by taking the position adjacent relation obtained by calculation in the S1 as a reference and a certain time period, and dividing all the two-dimensional coordinate value components into a plurality of subgroups;
s4, according to a uniform motion relationship, under the condition that the measurement intervals are the same, the distance between each two-dimensional coordinate value is constant, so that multiple groups of two-dimensional measurement values of each group are arranged along the Y direction to form local three-dimensional data;
s5, forming local three-dimensional data by using a method of S4 on a plurality of groups of two-dimensional measured values in all groups, wherein the track position corresponding to the acquisition moment of the first group of two-dimensional coordinate values in each group is used as a positioning standard, and generating three-dimensional point cloud data is completed.
Preferably, the preprocessing algorithm of the three-dimensional point cloud data:
detecting a noise point by adopting a weight angle criterion, and judging the point as the noise point when the boundary attribute is larger than a set threshold value; or detecting a noise point by adopting a neighboring point criterion, and judging the point as the noise point when the point in the radius of the fixed neighborhood is smaller than a set threshold value; and processing the point cloud data containing the noise points and the outliers by using an FCM algorithm, and retaining the edge characteristics of the holes in the point cloud while removing the noise points.
Preferably, the fine Kong Jingxi extraction algorithm:
and sequentially carrying out boundary extraction and cluster segmentation on the preprocessed three-dimensional point cloud data to extract hole features in the point cloud data.
Preferably, the pore mass detection algorithm:
s1, inputting hole edge point cloud data to be detected, and searching neighborhood data within a certain radius range in original point cloud;
s2, fitting a plane according to the neighborhood data, and calculating the normal direction of the plane;
s3, fitting by using the hole characteristic point cloud to obtain a fitting circle center, and calculating the three-dimensional coordinates of the circle center;
s4, taking the center coordinates as a searching starting point, taking a plane normal direction as a searching direction, setting a searching distance as a certain fixed value, and setting point clouds obtained by searching as hole bottom data;
s5, if no hole bottom point cloud data exists or the calculated distance is larger than a set threshold value, the hole is considered to be a through hole; if the calculated distance is smaller than the set threshold value, performing secondary judgment;
s6, the projection of the hole bottom point cloud is simplified, the projection area of the plug adjacent to the fitting plane in the hole center is calculated, if the projection area of the plug is larger than a set threshold value, the hole is judged to be blocked, and otherwise, the hole is a through hole;
s7, counting the number of all detected holes and the number of the plugged holes, calculating the hole passing rate, and recording the corresponding row number and column number of each plugged hole in the digital-analog, so that a detector can quickly position the plugged holes.
Further, the algorithm of the visualization module includes:
s1, constructing an octree point cloud space index structure of an OBB outer surrounding box, and reading point cloud data;
s2, constructing an LOD model based on an octree point cloud space index structure;
s3, selecting LOD levels to be visualized based on the view point distance of the view cone, and selecting nodes according to the visibility judgment of view cone shielding rejection;
s4, when the view angle changes, preloading nodes possibly used by the next frame in advance according to the view angle information of the current frame;
s5, reading the node selected in the S3, and performing visual rendering;
and S6, when the visual angle is unchanged, performing point cloud self-adaptive filling, and improving the point cloud visualization effect.
Further, the algorithm of the data import and export module comprises:
s1, after a detection flow starts, calling a related data interface to read scanning data of a laser scanner and a motion track of an end effector of a measuring robot in real time;
s2, inputting a corresponding three-dimensional point cloud data generation module;
s3, aiming at the three-dimensional point cloud data type, calling a related point cloud data format template;
s4, writing corresponding data import and export codes.
Further, the test is performed using the measurement robot subsystem, as shown in fig. 6, comprising the steps of:
(1) The method comprises the steps that a measured workpiece is mounted on a detection platform and is positioned and clamped through a special tool, and meanwhile the measured workpiece is divided into a plurality of waiting areas;
(2) Adjusting the pose of the measuring robot to an initial detection position;
(3) The measuring robot carries a laser scanner to move at a uniform speed, the light emitting position of a laser line is equal to the distance between the surface to be measured, the laser scanner scans and detects a first area to be measured on a workpiece, and point cloud data are collected and stored in real time to finish the condition detection of the through hole of the area to be measured;
(4) The measuring robot carries the laser scanner to the next station, and the step (2) and the step (3) are repeated until all areas to be measured are detected;
(5) The data acquisition and processing subsystem processes and analyzes the point cloud data to obtain the through hole rate and marks the blocked hole;
(6) And outputting the comprehensive detection result.
The method can realize operations such as micro Kong Jingxi extraction and hole quality detection on the scanning data, and the actual application proves that the method can obviously improve the accuracy and the integrity of point cloud data processing, accelerate the processing speed of the point cloud data, improve the degree of automation and the degree of humanization, and ensure that the detection work of the through hole rate is normally carried out.
Example 4:
a nondestructive testing method for the through hole rate of a micro array hole of a composite wallboard is carried out by adopting the system:
data preprocessing:
(1) Three-dimensional point cloud data generation algorithm
S1, reading motion parameters, a motion path and scanning frequency of a laser scanner of a measuring robot, calculating acquisition time of each group of two-dimensional coordinate values from detection, and calculating the position of an end effector of the measuring robot in a track when each group of two-dimensional coordinate values are acquired;
s2, calculating the position relation between the two scanner data coordinate systems according to the installation position of the fixed tool, converting the data of the scanner B into the coordinate system of the scanner A, and combining the two groups of coordinate values into one group;
s3, dividing the data by taking the position adjacent relation obtained by calculation in the S1 as a reference and a certain time period, and dividing all the two-dimensional coordinate value components into a plurality of subgroups;
s4, according to a uniform motion relationship, under the condition that the measurement intervals are the same, the distance between each two-dimensional coordinate value is constant, so that multiple groups of two-dimensional measurement values of each group are arranged along the Y direction to form local three-dimensional data;
s5, forming local three-dimensional data by using a method of S4 from a plurality of groups of two-dimensional measured values in all groups, wherein the track position corresponding to the acquisition moment of the two-dimensional coordinate values of the first group in each group is used as a positioning standard, and generating three-dimensional point cloud data is completed;
(2) Preprocessing algorithm for three-dimensional point cloud data
And processing the point cloud data containing the noise points and the outliers by using an FCM algorithm, and retaining the edge characteristics of the holes in the point cloud while removing the noise points.
And detecting a noise point by adopting a weight angle criterion, and judging the point as the noise point when the boundary attribute is larger than a set threshold value.
Boundary properties:
wherein prpjectFor point x i The center of gravity of the projection points of all neighborhood points on the local tangential plane, r is the radius of the local semicircular disc area, and the calculation formula is as follows:
wherein project (y j ) Origin x i Projection point, x on local tangential plane i Is the gravity center point thereof.
And detecting a noise point by adopting a neighboring point criterion, and judging the point as the noise point when the point in the radius of the fixed neighborhood is smaller than a set threshold value.
The FCM algorithm clusters the remaining points and updates the data points to a cluster center.
The objective function of the clustering algorithm is:
parameter mu jk Sample point x j Relative clustering center O k Wherein μ is the membership of jk Representing point x j Probability belonging to the k-th class.
Wherein:
d 2 jk (x j ，o k )＝||(1/W jk )(x j -o k )|| 2 ；
d 2 jk for sample point x j Relative clustering center O k Square distance of W k The sum of membership of all points to cluster k. The FCM algorithm can smooth small-scale noise points, and is beneficial to the subsequent analysis algorithm of micropore point cloud extraction.
And (II) data analysis:
1. fine extraction of microperforated three-dimensional point cloud data
And detecting micro-hole features in the three-dimensional point cloud data, and determining the three-dimensional coordinates of each hole feature so as to carry out hole quality detection subsequently.
2. Hole quality detection function
And analyzing point clouds in the corresponding data neighborhood according to the positions of the hole features, searching hole bottom data in the hole axis direction, and identifying the acquired hole bottom point clouds based on a template matching and quick searching method so as to distinguish a honeycomb structure from a blockage and judge the blocked holes.
Preferably, the fine Kong Jingxi extraction algorithm:
and sequentially carrying out boundary extraction and cluster segmentation on the preprocessed three-dimensional point cloud data to extract hole features in the point cloud data.
The normal vector is one of the differential geometric characteristics, and the normal vector of each data point on a discrete surface model is calculated in the fields of computer imaging, surface feature analysis, surface reconstruction, computer vision and the like, and is one of the basic steps. Assuming that the local quadric is S (u, v) =s (u, v, h (u, v)), the first order partial derivatives and normal of the surface at this point are as follows:
wherein: h (u, v) =au 2 +buv+cv 2 +du+ev+f。
The normal vector calculation formula for the surface is as follows:
Where a, b, c, d, e, f is the surface equation coefficient.
However, when the curved surface data is not available or the curved surface is reconstructed from the point cloud data to have low curved surface efficiency, the normal vector problem of a certain point is calculated by fitting a local plane by using a least square method in the embodiment.
And taking the normal vector of the fitted curved surface of the neighborhood point of a point as the true normal vector of the point. This approach considers the normal vector information of all neighbors, thus improving the accuracy and robustness of the normal vector estimation of one irregularly sampled data point. The main advantage of this approach is that the normal vector estimation at a point depends on the nearest neighbor point of that point, and that the normal vector can be direction-adjusted during the estimation process. Compared with the prior method, the method is based on the neighborhood information, so that the neighborhood size of the method can be adjusted, and the unorganized point cloud data normal vector estimation is more accurate and robust.
The normal vector of the point cloud data obtained by the method points to different directions, so that the normal vector directions of all the data are unified by direction adjustment, and the method is convenient for comparing different normal vectors and ensuring the accuracy of the feature extraction result. When the normal vector n of the test point i Normal vector n to neighborhood point k If the dot product is greater than zero, no adjustment is needed, and if it is less than zero, the latter is multiplied by-1.
The boundary extraction algorithm determines whether a point is a boundary point based on the normal vector change condition of a point neighborhood. We observe that there is a significant change in the direction of the normal vector between the hole wall and the hole surface, so when the angle between the normal vector of a point and the normal vector of its neighborhood point is greater than the set threshold, then the point is considered as the boundary point. Similarly, using the neighborhood link angle criterion, we can determine if a point is a boundary point.
Preferably, the pore mass detection algorithm:
s1, inputting hole edge point cloud data to be detected, and searching neighborhood data within a certain radius range in original point cloud;
s2, fitting a plane according to the neighborhood data, and calculating the normal direction of the plane;
s3, fitting by using the hole characteristic point cloud to obtain a fitting circle center, and calculating the three-dimensional coordinates of the circle center;
s4, taking the center coordinates as a searching starting point, taking a plane normal direction as a searching direction, setting a searching distance as a certain fixed value, and setting point clouds obtained by searching as hole bottom data;
s5, if no hole bottom point cloud data exists or the calculated distance is larger than a set threshold value, the hole is considered to be a through hole; if the calculated distance is smaller than the set threshold value, performing secondary judgment;
s6, the projection of the hole bottom point cloud is simplified, the projection area of the plug adjacent to the fitting plane in the hole center is calculated, if the projection area of the plug is larger than a set threshold value, the hole is judged to be blocked, and otherwise, the hole is a through hole;
S7, counting the number of all detected holes and the number of the plugged holes, calculating the hole passing rate, and recording the corresponding row number and column number of each plugged hole in the digital-analog, so that a detector can quickly position the plugged holes.
The method can realize operations such as micro Kong Jingxi extraction and hole quality detection on the scanning data, and the actual application proves that the method can obviously improve the accuracy and the integrity of point cloud data processing, accelerate the processing speed of the point cloud data, improve the degree of automation and the degree of humanization, and ensure that the detection work of the through hole rate is normally carried out.
Example 5:
a nondestructive testing system for the through hole rate of a micro array hole of a composite wallboard consists of a measuring robot subsystem and a data acquisition and processing subsystem.
(1) Measurement robot subsystem: the device comprises a measuring robot, a control cabinet, a demonstrator, a precision compensation module and off-line programming software.
Measuring robot: according to the weight of the end effector clamp and the laser scanner, the safety coefficient and the effective working space are comprehensively considered, and the KUKA KR 16R 1610-2 industrial six-axis serial robot is selected in the embodiment, and the main technical indexes are shown in the following table 1.
Table 1 main technical indexes of industrial six-axis serial robot
| Sequence number | Project | Parameters (parameters) |
| 1 | Maximum range of motion | 1612mm |
| 2 | Maximum load capacity | 20kg |
| 3 | Rated load | 16KG |
| 4 | Pose repeatability (ISO 9283) | ±0.04mm |
| 5 | Number of axes | 6 |
| 6 | Floor area | 430.5mm x 370mm |
| 7 | Weight of (E) | About 255kg |
| 8 | Protection system | IP65 |
| 9 | Operating ambient temperature | 5 ℃ to 55 DEG C |
The measuring robot meets the weight and mounting position requirements of the end effector, the available working space schematic of which is shown in fig. 2, and the load capacity of which is shown in fig. 3.
The end effector fixture is used to enable mounting and connection between the flange and the laser scanner, assisting in TCP measurements.
The measuring robot is provided with a control cabinet and a demonstrator, wherein an additional axle box is integrated on the control cabinet.
An industrial six-axis robot can simultaneously set 8 Cartesian and 8 working spaces related to the axes, the working spaces allow superposition, the measuring robot only moves within the set and activated working space, the actual position is continuously calculated and monitored according to the set safety parameters, and if the monitoring limit range or the safety parameters are exceeded, the monitoring is responded, and the measuring robot automatically stops, so that safety guarantee is provided for equipment and personnel.
In the working space, the repeated positioning accuracy of the measuring robot is high, but the absolute positioning accuracy is poor.
And the precision compensation module is used for: the precision compensation module (comprising measuring robot calibration software and positioning precision compensation software) is used for calibrating the measuring robot, establishing a model between the space points of the measuring robot and errors of the measuring robot, compensating the space coordinates of the measuring robot in real time in the operation process of the measuring robot, and correcting absolute positioning errors, so that the absolute positioning precision of the measuring robot is ensured.
Offline programming software: in this embodiment, the offline programming software adopts mature general offline programming software robottmaster, which is the offline programming and simulation software with the strongest current function, and supports multiple robot brands such as KUKA. The Robotmaster consists of a robot ginseng interface, a robot simulation simulator and an off-line program post-processor for generating executable by the robot, and can complete the functions of reading three-dimensional models of workpieces, extracting and converting model data, converting a coordinate system, automatically generating robot program codes, planning tracks, checking interference, simulating and optimizing processes, generating executable codes and the like. The Robotmaster software has the function of extending the external axis of the robot, and can carry out linkage control and off-line simulation on the rotation of the six axes of the measuring robot and the workpiece by extending the seventh axis turntable so as to ensure that the measuring robot system works in the optimal position and posture.
(2) And the data acquisition and processing subsystem: the system comprises a data acquisition sensor, a detection computer and the micro array hole through hole rate automatic nondestructive testing and analysis system of the composite wallboard in the embodiment 1.
Data acquisition sensor: in the embodiment, a laser scanner is adopted as the data acquisition sensor, and a deep vision intelligent SR6000 series laser scanner is selected.
The data acquisition and processing subsystem comprises two oppositely arranged laser scanners, the light emergent directions of the two laser scanners are parallel to each other, the light emergent positions are flush, at the moment, the measuring robot subsystem can be used for detecting the surface of an irregular product, the laser line emission direction of the laser scanners is perpendicular to the surface to be detected, and the hole to be detected is in the scanning range of the laser scanners. The two laser scanners respectively acquire point cloud data of a hole to be measured, the data acquisition and processing subsystem is used for splicing the two groups of data, and the splicing result is used as basic data for judging the hole blocking, so that detection dead zones caused by a triangular reflection principle of a single sensor are avoided, for example, for an irregularly-shaped cylindrical part, micro holes with the diameter of 1.2mm are arranged in an inner wall array, the two laser scanners are required to be carried by a measuring robot to move at the moment perpendicular to the irregular inner wall, the light emitting direction is always guaranteed to be parallel to the axis of the hole, and then the two groups of data are spliced (the single sensor cannot acquire complete geometric profile).
And (3) detecting a computer: the detection computer is used for data acquisition, processing and storage, and in this embodiment, an associative graphic workstation is selected, and the workstation adopts an Intel (r) borui processor I7-9700, an nvidia professional display card, a 16G memory, and a 1t+512G hard disk.
Example 6:
a nondestructive testing method for the through hole rate of a micro array hole of a composite wallboard is carried out by adopting the system and comprises the following steps:
(1) The measured workpiece is arranged on a detection platform and is positioned and clamped through a special tool, and meanwhile, the measured workpiece is divided into a plurality of equal-width areas to be measured, and the detection area division mode and the light emitting direction of a line laser scanner are shown in figure 5;
(2) Adjusting the pose of the measuring robot to an initial detection position;
(3) The measuring robot carries a laser scanner to move at a uniform speed, the light emitting position of a laser line is equal to the distance between the surface to be measured, the laser scanner scans and detects a first area to be measured on a workpiece, and point cloud data are collected and stored in real time to finish the condition detection of the through hole of the area to be measured;
(4) The measuring robot carries the laser scanner to the next station, and the step (2) and the step (3) are repeated until all areas to be measured are detected;
(5) The data acquisition and processing subsystem processes and analyzes the point cloud data to obtain the through hole rate and marks the blocked hole;
(6) And outputting the comprehensive detection result.
Sample detection experiments were performed: a front view of the experimental sample is shown in figure 7,
data calibration experiment
1. Calibration method frame
As shown in fig. 8, the present calibration method is divided into two steps: and respectively performing transformation matrix calculation based on the checkerboard data, and applying the transformation matrix and the data to be spliced. The calibration method extracts angular points from original data, averages errors according to corresponding relations, and calculates a transformation matrix of laser scanner B data to laser scanner A data. Then, based on the matrix, the workpiece data collected by the laser scanner B are subjected to coordinate transformation and combined with the data of the laser scanner A, so that the calibration and data splicing of the double laser scanners are realized.
2. Calibration method
The experimental platform for verification consists of a KUKA KR 16R1610-2 robot, 2 laser scanners, a clamping device and a computer. The measuring robot clamps the laser scanner through the clamping device, adjusts the emergent laser surface to be basically vertical to the surface of the measured object, and then moves along a straight line at a uniform speed and a straight line, and the laser scanner synchronously scans the checkerboard. The computer collects and stores the data in real time, and analyzes and marks the data.
3. Detection result
Calculating a transformation matrix from the laser scanner B data to the laser scanner A data in a fixed state by a calibration method; and (3) applying the transformation matrix to splice the data acquired by the two laser scanners on the same track. As shown in fig. 9, the left data is data collected by the laser scanner a, the right data is data collected by the laser scanner B, and the splicing result is shown in fig. 10.
Meanwhile, two adjacent points at the same position are respectively taken from the two data to carry out error analysis. In the experiment, 6 positions are taken for analysis, the error range of the X direction is 0.0032 to-0.006mm, the error range of the Y direction is 0.02 to-0.0032 mm, the error range of the Z direction is +/-0.04 mm, and the detection result is shown in figure 11.
4. Calibration conclusion
By the existing calibration method, data splicing of two laser scanners can be achieved.
After data acquisition is carried out on actual workpieces, a calibration method is applied to calculate a transformation matrix, the matrix is applied to B data of a laser scanner, the error range of spliced data in the X direction is 0.0033mm to-0.006mm, the error range in the Y direction is 0.02mm to 0.032mm, and the error range in the Z direction is +/-0.04 mm. The spliced data can meet the judgment standard of the hole bottom data in the aspect of height and area.
(II) analysis Module verification
1. Algorithm framework
The detection algorithm is divided into three modules: the system is a line laser three-dimensional data diagnosis and preprocessing module, a micro Kong Jingxi extraction module and a hole quality detection module respectively, as shown in fig. 12. The data preprocessing module analyzes and diagnoses the data quality through the space distribution rule of the three-dimensional data of the detection line laser, filters partial noise point clouds and improves the data quality. Meanwhile, the main plane where the round hole is detected by the local sampling principle, so that a foundation is provided for extracting the fine hole. The hole extraction module forms a unified mathematical expression model by analyzing the mode of the micro holes, identifies the micro holes based on a template matching and quick searching method, and extracts related parameters. And the pore quality detection module classifies the detection elements into a mathematical expression model according to project requirements, and adaptively adjusts detection parameters based on the micro-pore extraction result to obtain an accurate quality detection result.
2. Experimental method
The experimental platform for verification consists of a KUKA KR 16R1610-2 robot, 2 laser scanners, a clamping device and a computer. The measuring robot clamps the laser scanner through the clamping device, adjusts the emergent laser surface to be basically vertical to the surface of the measured object, and then makes uniform linear motion along the linear track, and the laser scanner synchronously scans the data of the surface and the hole bottom of the measured object. And the computer collects and stores the point cloud data in real time, reconstructs the surface of the measured object through software, and analyzes and marks the surface.
In the verification experiment, the plane fitting is respectively carried out on the hole periphery plane data and the hole bottom data through processing and analyzing the hole bottom data of the small holes, so that noise point interference is eliminated. Specifically, assuming that the hole position and the aperture of a hole have been detected, the upper plane is fitted with surrounding surface point cloud data, hole bottom data is obtained by selecting the normal direction of the upper plane of the hole to search data points in the direction, then fitting the data points searched in the aperture range, and the statistical analysis method can exclude noise points and outer points, and only selecting inner points with small errors to fit the plane (the hole bottom data is a plane according to the hole bottom condition).
As shown in fig. 13, the measurement results are within the actual machining size range, and it can be explained that the laser scanner has the capability of measuring the hole bottom data of the actual workpiece to be measured.
3. Detection result
By observing the new sample, it is found that the holes in the local area are totally blocked, and data is collected for this area and the result is calculated using an algorithm. The detection situation is shown in fig. 14 and 15, wherein fig. 14 is an actual photograph, and the impurity is marked with an 'x' to block the holes; fig. 15 shows the result of the algorithm detection, and the impurity detected by the algorithm is marked "x" to block the hole. According to analysis of the detection result, the detection algorithm can effectively detect the blocking of the hole.
Meanwhile, the actual detection result shows that the data quantity at the edge close to the line laser measurement range is small, and the detection algorithm identifies less hole bottom data as scattered noise and filters out the scattered noise. When the scheme is actually implemented, the scanning station is planned, partial data are selected as effective input, and therefore complete and correct detection is carried out on the to-be-detected piece.
Meanwhile, the algorithm can output the three-dimensional coordinates of the detected blocked holes in the data, and provide guidance for sample post-processing.
4. Conclusion(s)
Under the condition of better input data quality, the existing algorithm can accurately detect the full-plugging hole; the total blocking hole bottom has about 600 points of data volume, is used for a verification algorithm, and can detect the total blocking hole in the middle section of the line outgoing laser visual field. The algorithm test is operated by single thread in the whole process, and the practical scheme implementation can select multi-threads to analyze multi-segment data at the same time, so that the detection efficiency meets the project requirement. The algorithm is embodied as shown in table 2 below.
Table 2 algorithm performance
| Experimental data | Technical index | Solution scheme | Optimized results | |
| Data size | 100mm×60mm | |||
| Number of holes contained | 159 | |||
| Misjudgment rate | 0 | ≤0.05％ | Multi-station multidirectional scanning | 0 |
| Operating efficiency (Single thread) | 159/30.493 = 5.214 pores/sec | Not less than 10 holes/second | Multithreaded data processing | 43 holes/second |
The relevant index is described with respect to the technical requirements.
a) Detection efficiency
The experiment is used for collecting data of 159 micro array holes on the surface of the simulation piece, the statistical time is 14.72s, and the detection efficiency requirement of 10 holes/s is met after conversion of about 10.8 holes/s.
b) Misjudgment rate
After data are collected, preprocessing and analyzing the original data to a certain extent, and detecting the position of each hole by an algorithm and analyzing the hole bottom data; meanwhile, through manual detection, the actual blocking condition of each small hole is detected by using a probe, and the false judgment rate of the detection method is proved to be 0 by comparing the detection results of the two times.
c) Data processing time
And after collecting and generating point cloud data, carrying out data analysis processing. The data processing part is independently completed by each functional part, and then the running time of the programs is added to obtain a final result. At present, the following computer configurations are used for practical testing: intel cool i58500 processor, integrated graphics card, 4G memory, 500GB. The experimental collection sample surface was 159 micro array wells, the processing time was about 30.49s using a single-threaded running algorithm, and the data processing efficiency was about 5.2 wells/s. Through memory optimization and multithreading parallel experiments, when 5 sections of data with similar scale are processed simultaneously, the data processing efficiency can reach about 43 holes/s.
Experimental results show that the scheme for detecting the micro-hole through-hole rate by using a laser scanner is feasible, and related indexes can be realized.
The foregoing description is only a preferred embodiment of the present invention, and is not intended to limit the present invention in any way, and any simple modification, equivalent variation, etc. of the above embodiment according to the technical matter of the present invention fall within the scope of the present invention.
Claims (10)
1. The nondestructive testing system for the micro array hole through rate of the composite wallboard is characterized by comprising a data acquisition and processing subsystem and a measuring robot subsystem, wherein the data acquisition and processing subsystem comprises a data preprocessing module, a data analysis module and an output report module; the data preprocessing module is used for generating three-dimensional point cloud data and filtering and noise filtering preprocessing, the data analysis module is used for finely extracting micro-hole three-dimensional point cloud data on a detection plane and analyzing and judging whether the micro-hole three-dimensional point cloud data is a plugged hole or not, and the output report module is used for customizing a detection data analysis report template and outputting a data analysis report.
2. The nondestructive testing system for the micro array hole through rate of the composite wallboard according to claim 1, wherein the data acquisition and processing subsystem further comprises a data visualization module, a data import and export module and a data management module; the data visualization module is used for managing and visualizing the data, loading the data, displaying the rendered point cloud data and displaying the imported data; the data import and export module is used for reading, processing and outputting data and reading various parameters of the detection and analysis system in real time according to a transmission protocol; the data management module is used for carrying out coordinate system management, data file management and data output management, and carrying out historical data analysis, viewing and editing on the original data and the calculation result stored in the database.
3. The nondestructive testing system for the hole through rate of the micro array of the composite wallboard according to claim 1 or 2, wherein the measuring robot subsystem comprises a measuring robot, two oppositely placed laser scanners are arranged on the measuring robot, the light emitting directions of the two laser scanners are parallel to each other, the light emitting positions of the two laser scanners are level, and the laser scanners are used for acquiring point cloud data of a hole to be tested and transmitting the point cloud data to the data acquisition and processing subsystem.
4. A nondestructive testing method for the through hole rate of a micro array hole of a composite wallboard by adopting the nondestructive testing system as claimed in any one of claims 1 to 3, wherein the detection analysis is carried out by adopting a data acquisition and processing subsystem, and the method comprises the following steps:
step S100: according to a scanning path planned by the measuring robot, acquiring a data time sequence by using a laser scanner, splicing two-dimensional point cloud outlines, and generating three-dimensional point cloud data;
step S200: filtering and denoising preprocessing is carried out on the obtained three-dimensional point cloud data;
step S300: sequentially carrying out boundary extraction and cluster segmentation on the preprocessed three-dimensional point cloud data to extract hole features in the point cloud data;
step S400: and detecting the quality of the holes, and judging whether the holes are plugged.
5. The method for non-destructive inspection of hole-through rates of a micro array of composite wall panels according to claim 4, wherein said step S100 comprises the steps of:
step S110: reading the motion parameters, the motion path and the scanning frequency of a laser scanner of the measuring robot, calculating the acquisition time of each group of two-dimensional coordinate values from detection, and calculating the position of the end effector of the measuring robot in the track when each group of two-dimensional coordinate values are acquired;
step S120: calculating the position relation between two scanner data coordinate systems according to the installation position of the fixed tool, converting the data of the scanner B into the coordinate system of the scanner A, and combining the two groups of coordinate values into one group;
step S130: dividing the data by taking the position adjacent relation obtained by calculation in the step S110 as a reference and a certain time period, and dividing all the two-dimensional coordinate value components into a plurality of subgroups;
step S140: according to the uniform motion relation, under the condition that the measurement intervals are the same, the distance between each two-dimensional coordinate value is constant, so that multiple groups of two-dimensional measurement values of each group are arranged along the Y direction to form local three-dimensional data;
step S150: and (3) forming local three-dimensional data by using a method of the step (S140) from a plurality of groups of two-dimensional measured values in all subgroups, wherein the track position corresponding to the acquisition moment of the two-dimensional coordinate values of the first group in each subgroup is used as a positioning standard, and the generation of the three-dimensional point cloud data is completed.
6. The nondestructive testing method of micro array hole through rate of composite wall board according to claim 5, wherein in the step S200, a weight angle criterion is adopted to detect a noise point, and when the boundary attribute is greater than a set threshold value, the noise point is determined; or detecting a noise point by adopting a neighboring point criterion, and judging the point as the noise point when the point in the radius of the fixed neighborhood is smaller than a set threshold value; and processing the point cloud data containing the noise points and the outliers by using an FCM algorithm, and retaining the edge characteristics of the holes in the point cloud while removing the noise points.
7. The method for non-destructive inspection of hole-through rates of a micro array of composite wall panels according to claim 4, wherein said step S400 comprises the steps of:
step S410: inputting hole edge point cloud data to be detected, and searching neighborhood data within a certain radius range in the original point cloud;
step S420: according to the neighborhood data fitting plane, calculating the normal direction of the plane;
step S430: fitting by using the hole characteristic point cloud to obtain a fitting circle center, and calculating the three-dimensional coordinates of the circle center;
step S440: taking the center coordinates as a searching starting point, taking the plane normal direction as a searching direction, setting the searching distance as a certain fixed value, and setting the point cloud obtained by searching as hole bottom data;
Step S450: if the hole bottom point cloud data does not exist or the calculated distance is larger than a set threshold value, the hole is considered to be a through hole; if the calculated distance is smaller than the set threshold value, performing secondary judgment;
step S460: the projection of the hole bottom point cloud is simplified, the projection area of the plug adjacent to the fitting plane in the hole center is calculated, if the projection area of the plug is larger than a set threshold value, the hole is judged to be plugged, and otherwise, the hole is a through hole;
step S470: counting the number of all detected holes and the number of the blocked holes, calculating the hole passing rate, and recording the corresponding number of rows and columns of each blocked hole in the digital-analog, so that a detector can quickly position the blocked holes.
8. The method for non-destructive inspection of the hole-through rate of a micro array of composite wall panels according to claim 4, further comprising visualizing the data:
step A1: constructing an octree point cloud space index structure of the OBB outer surrounding box, and reading point cloud data;
step A2: constructing an LOD model based on the octree point cloud space index structure;
step A3: selecting LOD (on-demand) levels to be visualized based on the distance of the view cone viewpoint, and selecting nodes according to the visibility judgment of view cone shielding rejection;
Step A4: when the visual angle changes, according to the visual angle information of the current frame, preloading the nodes possibly used by the next frame in advance;
step A5: reading the node selected in the step S3, and performing visual rendering;
step A6: and when the visual angle is unchanged, performing point cloud self-adaptive filling, and improving the point cloud visualization effect.
9. The method for non-destructive inspection of hole-through rates in a micro array of composite wall panels according to claim 4, further comprising the steps of:
step B1: after the detection flow starts, a related data interface is called to read scanning data of the laser scanner and a motion track of an end effector of the measuring robot in real time;
step B2: inputting a corresponding three-dimensional point cloud data generation module;
step B3: invoking a related point cloud data format template aiming at the three-dimensional point cloud data type;
step B4: corresponding data import and export codes are written.
10. A method for non-destructive inspection of the hole-through rate of a micro array of composite wall panels according to any one of claims 4 to 9, characterized in that it is performed by a measuring robot subsystem comprising the steps of:
step S1: the method comprises the steps that a measured workpiece is mounted on a detection platform and is positioned and clamped through a tool, and meanwhile the measured workpiece is divided into a plurality of waiting areas;
Step S2: adjusting the pose of the measuring robot to an initial detection position;
step S3: the measuring robot carries a laser scanner to move at a uniform speed, the light emitting position of a laser line is equal to the distance between the surface to be measured, the laser scanner scans and detects a first area to be measured on a workpiece, and point cloud data are collected and stored in real time to finish the condition detection of the through hole of the area to be measured;
step S4: the measuring robot carries the laser scanner to the next station, and the step S2 and the step S3 are repeated until all areas to be measured are detected;
step S5: the data acquisition and processing subsystem processes and analyzes the point cloud data to obtain the through hole rate and marks the blocked hole;
step S6: outputting the detection result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310425659.0A CN117191781A (en) | 2023-04-20 | 2023-04-20 | Nondestructive testing system and method for micro array hole through hole rate of composite wallboard |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310425659.0A CN117191781A (en) | 2023-04-20 | 2023-04-20 | Nondestructive testing system and method for micro array hole through hole rate of composite wallboard |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117191781A true CN117191781A (en) | 2023-12-08 |
Family
ID=88998569
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310425659.0A Pending CN117191781A (en) | 2023-04-20 | 2023-04-20 | Nondestructive testing system and method for micro array hole through hole rate of composite wallboard |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117191781A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117782944A (en) * | 2024-02-27 | 2024-03-29 | 山东宇洋汽车尾气净化装置有限公司 | Method for detecting porosity of automobile exhaust purification filter |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109117602A (en) * | 2018-10-17 | 2019-01-01 | 江西洪都航空工业集团有限责任公司 | Large scale covering digitizing detection method based on laser tracker |
| CN109655009A (en) * | 2017-10-10 | 2019-04-19 | 南京金正奇交通设备有限责任公司 | A kind of high-speed rail component rapid detection system |
| CN111102926A (en) * | 2020-01-02 | 2020-05-05 | 青岛国信海天中心建设有限公司 | Engineering quality detection method and system based on BIM |
| CN111986115A (en) * | 2020-08-22 | 2020-11-24 | 王程 | Accurate elimination method for laser point cloud noise and redundant data |
| CN112308974A (en) * | 2020-10-30 | 2021-02-02 | 南京航空航天大学 | Large-scale point cloud visualization method for improving octree and adaptive reading |
| CN112509118A (en) * | 2020-12-02 | 2021-03-16 | 南京航空航天大学 | Large-scale point cloud visualization method capable of preloading nodes and self-adaptive filling |
| CN112528525A (en) * | 2020-12-31 | 2021-03-19 | 河钢数字技术股份有限公司 | Visual industrial process management and control platform based on modeling technology |
| CN112763486A (en) * | 2020-11-30 | 2021-05-07 | 成都飞机工业（集团）有限责任公司 | Composite material wall plate array hole detection method based on line laser scanning |
| CN113865508A (en) * | 2021-09-28 | 2021-12-31 | 南京航空航天大学 | Automatic detection device and method for through hole rate of acoustic lining of honeycomb sandwich composite material |
| CN114486889A (en) * | 2022-01-14 | 2022-05-13 | 成都飞机工业（集团）有限责任公司 | Detection system and method for through hole rate of micro array holes of composite wallboard |
| CN114708400A (en) * | 2022-04-08 | 2022-07-05 | 广州高新兴机器人有限公司 | Three-dimensional laser noise reduction method and device, medium and robot |
| CN115615407A (en) * | 2022-10-11 | 2023-01-17 | 中铁十一局集团有限公司 | Solid structure substance quality detection method based on three-dimensional laser scanning |
-
2023
- 2023-04-20 CN CN202310425659.0A patent/CN117191781A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109655009A (en) * | 2017-10-10 | 2019-04-19 | 南京金正奇交通设备有限责任公司 | A kind of high-speed rail component rapid detection system |
| CN109117602A (en) * | 2018-10-17 | 2019-01-01 | 江西洪都航空工业集团有限责任公司 | Large scale covering digitizing detection method based on laser tracker |
| CN111102926A (en) * | 2020-01-02 | 2020-05-05 | 青岛国信海天中心建设有限公司 | Engineering quality detection method and system based on BIM |
| CN111986115A (en) * | 2020-08-22 | 2020-11-24 | 王程 | Accurate elimination method for laser point cloud noise and redundant data |
| CN112308974A (en) * | 2020-10-30 | 2021-02-02 | 南京航空航天大学 | Large-scale point cloud visualization method for improving octree and adaptive reading |
| CN112763486A (en) * | 2020-11-30 | 2021-05-07 | 成都飞机工业（集团）有限责任公司 | Composite material wall plate array hole detection method based on line laser scanning |
| CN112509118A (en) * | 2020-12-02 | 2021-03-16 | 南京航空航天大学 | Large-scale point cloud visualization method capable of preloading nodes and self-adaptive filling |
| CN112528525A (en) * | 2020-12-31 | 2021-03-19 | 河钢数字技术股份有限公司 | Visual industrial process management and control platform based on modeling technology |
| CN113865508A (en) * | 2021-09-28 | 2021-12-31 | 南京航空航天大学 | Automatic detection device and method for through hole rate of acoustic lining of honeycomb sandwich composite material |
| CN114486889A (en) * | 2022-01-14 | 2022-05-13 | 成都飞机工业（集团）有限责任公司 | Detection system and method for through hole rate of micro array holes of composite wallboard |
| CN114708400A (en) * | 2022-04-08 | 2022-07-05 | 广州高新兴机器人有限公司 | Three-dimensional laser noise reduction method and device, medium and robot |
| CN115615407A (en) * | 2022-10-11 | 2023-01-17 | 中铁十一局集团有限公司 | Solid structure substance quality detection method based on three-dimensional laser scanning |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117782944A (en) * | 2024-02-27 | 2024-03-29 | 山东宇洋汽车尾气净化装置有限公司 | Method for detecting porosity of automobile exhaust purification filter |
| CN117782944B (en) * | 2024-02-27 | 2024-05-24 | 山东宇洋汽车尾气净化装置有限公司 | Method for detecting porosity of automobile exhaust purification filter |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7024032B2 (en) | Method for assessing fit and alignment of a manufactured part | |
| JP4933008B2 (en) | Method and medium for computer-aided manufacturing measurement analysis | |
| CN101196389B (en) | Image measuring system and method | |
| JP2014527663A (en) | Coordinate measurement system data organization | |
| US7683900B2 (en) | System and method for processing a point cloud in a simulated three-dimensional space | |
| CN107088892A (en) | A kind of industrial robot motion accuracy checking method based on binocular vision | |
| CN117191781A (en) | Nondestructive testing system and method for micro array hole through hole rate of composite wallboard | |
| CN111540001A (en) | Method for detecting axial direction of air film hole of turbine blade of aero-engine | |
| Spitz et al. | Accessibility analysis using computer graphics hardware | |
| Zhang et al. | Optimisation of camera positions for optical coordinate measurement based on visible point analysis | |
| CN100462672C (en) | Image measuring system and method | |
| US20020065637A1 (en) | Method and apparatus for simulating the measurement of a part without using a physical measurement system | |
| Krotova et al. | Development of a trajectory planning algorithm for moving measuring instrument for binding a basic coordinate system based on a machine vision system | |
| Sun et al. | Precision work-piece detection and measurement combining top-down and bottom-up saliency | |
| Wang et al. | A binocular vision method for precise hole recognition in satellite assembly systems | |
| CN116368349A (en) | Semantic segmentation of inspection targets | |
| Irgenfried et al. | Image formation simulation for computer-aided inspection planning of machine vision systems | |
| CN113393450B (en) | Data quality inspection method and system for digital line drawing and readable storage medium | |
| CN109470138A (en) | The On-line Measuring Method of part | |
| CN110411372A (en) | It is a kind of that room collecting method is tested based on total station | |
| CN114485468B (en) | Multi-axis linkage composite measurement system and micro-part full-contour automatic measurement method | |
| Peng et al. | An Improved Monocular-Vision-Based Method for the Pose Measurement of the Disc Cutter Holder of Shield Machine | |
| CN117629117A (en) | Blade air film hole measuring and evaluating method | |
| CN207319046U (en) | A kind of laser is in machine centering and inter process measuring device | |
| Garcia et al. | Machine Learning Based Position Prediction of a Target Tracked by Multi-Aperture Positioning System |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |