Shape inspection method and system

Abstract

To meet the need to associate measurement data with each other in processing between measurement data from multiple directions, due to an influence applied by a measurement point having a large measurement error, when measurement data is compensated using a reference shape of an inspection object, and further, due to differences in coordinates according to postures of a measuring instrument at the measurement time, when compensation is conducted such that a reference shape overlaps measurement data.SOLUTION: There is provided, a method for conducting three-dimensional shape inspection of an inspection object, in which model data equal to or less than a measurement range is created according to a reference shape of an inspection object, a transformation matrix for transforming the model data to measurement data is calculated, and the measurement data is compensated according to an inverse matrix of the transformation matrix. In addition, a system for the method is provided.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a shape inspection method for inspecting the shape of an object to be inspected, and a system therefor. [Background technology]

[0002] Patent Document 1 describes a shape inspection method. It states that "the shape inspection method of the present invention is characterized by having an acquisition step of acquiring measurement point cloud data obtained by measuring an inspection object by a three-dimensional shape measuring device and three-dimensional CAD data in which a three-dimensional CAD part of the inspection object is arranged at a regular position, a generation step of generating moved CAD data in which the three-dimensional CAD part of the three-dimensional CAD data is arranged at the position of the part of the measurement point cloud data, a detection step of detecting a difference between the three-dimensional CAD part and the part of the inspection object using the data acquired by the acquisition step and the moved CAD data generated by the generation step, and a display processing step of displaying the difference detected by the detection step." In addition, in the embodiment "shape inspection system 10," it is described that "the shape inspection system 10 has a three-dimensional shape measuring device 20, a three-dimensional CAD system 30, and a shape inspection device 40." Regarding the matching unit 56 in the shape inspection device 40, it is described that "the matching unit 56 matches (image matches) the part of the divided measurement data selected by the selection unit 55 with the corresponding three-dimensional CAD part of the three-dimensional CAD data."

[0003] An example of a matching method is the ICP algorithm (see Non-Patent Document 1). The ICP algorithm searches for and associates each point of a point group A on one side with the nearest point on the other side, and adjusts the position and orientation of the coordinate system of the point group so as to minimize the difference between the associated points. If one side is CAD data, model data is created by generating points on the surface of the three-dimensional shape of the CAD data, and the model data and measurement data are matched. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2016-133347 A [Non-patent literature]

[0005] [Non-Patent Document 1] Paul J. Besl, and Neil D. McKay: “A Method for Registration of 3-D Shapes”, IEEETrans. on PAMI, vol. 14, no. 2, pp. 239-256, 1992. Summary of the Invention [Problem to be solved by the invention]

[0006] The present invention aims to solve the following two problems. First, when using the ICP algorithm to overlay part data and reference shapes such as CAD data, the search for corresponding points in the model data generated from the CAD data from each point in the measurement data is influenced by measurement points with large measurement errors in the measurement data. The matching accuracy when matching parts in divided measurement data with 3D CAD data is influenced by measurement points with large measurement errors in the measurement data when searching for corresponding points in the model data from each point in the measurement data. On the other hand, when searching for corresponding points in the measurement data from each point in the model data, data from parts of the model data that have not been measured are also matched to the measurement data, resulting in overlay errors.

[0007] The second point is that when performing coordinate conversion to overlay CAD data on measurement data during matching, the coordinates differ depending on the orientation of the measuring instrument during measurement, so when processing measurement data from multiple directions, it is necessary to match the measurement data with each other.

[0008] Therefore, an object of the present invention is to provide a shape inspection method and system that excludes measurement points with large measurement errors when calculating the overlay of data during measurement of the object to be inspected, and automatically performs shape inspection from measurement results from multiple directions. [Means for solving the problem]

[0009] In order to achieve the above object, the present invention provides a shape inspection method for performing arithmetic processing on a three-dimensional shape of an object to be inspected, the method comprising the steps of: generating model data from a reference shape of the object to be inspected and measurement area information; and calculating a transformation matrix for converting the model data into measurement data. and calculating the measurement area information from the reference shape of the object to be inspected and position and orientation information at the time of measurement. The present invention provides a method for inspecting the shape of a structure. In addition, the present invention provides a shape inspection method for inspecting a three-dimensional shape of a target object measured by a measuring instrument by arithmetic processing, the shape inspection method comprising the steps of generating model data from a reference shape and measurement area information of the target object, and calculating a transformation matrix for converting the model data into measurement data, and wherein the measurement area information is input by an operator.

[0010] In order to achieve the above object, the present invention provides a shape inspection system, comprising a measuring device for measuring a three-dimensional shape of an object to be inspected, and a calculation processing device, the calculation processing device generating model data from a reference shape of the object to be inspected and measurement area information, and calculating a transformation matrix for converting the model data into measurement data. The arithmetic processing device calculates measurement area information from the reference shape of the object to be inspected and the position and orientation information at the time of measurement. The present invention provides a shape inspection system configured to: In addition, the present invention provides a shape inspection system comprising a measuring device that measures the three-dimensional shape of a target object, and an arithmetic processing device, wherein the arithmetic processing device generates model data from the reference shape and measurement area information of the target object, calculates a transformation matrix for converting the model data into measurement data, and further comprises an input unit that inputs the measurement area information by operation of an operator. Effect of the Invention

[0011] According to the present invention, even if the measurement data includes measurement points with large measurement errors, it is possible to calculate a matrix for adjusting the measurement data by excluding the measurement points with large errors. In addition, since the measurement data is adjusted to coordinates based on the model data, it is possible to evaluate inspection items that span multiple measurement data.

[0012] Problems, configurations, and effects other than those described above will become apparent from the following description of the embodiments. [Brief description of the drawings]

[0013] [Figure 1]5 is a flowchart showing a processing procedure when applied to a shape inspection of a part according to the first embodiment. [Diagram 2] FIG. 1 is a block diagram showing a configuration example of a measurement device and an inspection system according to a first embodiment. [Diagram 3] FIG. 11 is a diagram showing a data group stored in an information storage unit 235 according to the first embodiment. [Figure 4] 2 is a detailed flowchart of S110 in FIG. 1 according to the first embodiment. [Diagram 5] 2 is a detailed flowchart of S115 in FIG. 1 according to the first embodiment. [Figure 6] FIG. 2 is a block diagram showing the effect of the measurement device according to the first embodiment. [Figure 7] FIG. 11 is a diagram showing an input GUI when applied to part inspection in the second embodiment. [Figure 8] FIG. 13 is a block diagram showing an example of the configuration when a measuring instrument equipped with a position / attitude changing mechanism according to a third embodiment is applied to inspecting the shape of a part. [Figure 9] 11 is a detailed flowchart of S110 according to the third embodiment. [Figure 10] FIG. 11 is a diagram showing an input GUI according to the third embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. EXAMPLES

[0015] A case where the present invention is applied to inspecting the shape of a part is shown in Example 1. One of the preferred shape inspection methods and systems of the present invention creates model data that is equal to or smaller than the measurement range from the reference shape of the inspection target, calculates a transformation matrix for converting the model data into measurement data, and adjusts the measurement data using the inverse matrix of the aforementioned matrix.

[0016] FIG. 1 is a flow chart showing steps S100 to S125 of an embodiment of a processing procedure when the present invention is applied to inspecting the shape of a part.

[0017] 2 shows an inspection system 200 for carrying out the flowchart shown in FIG. 1, and includes a measuring instrument 215 and a processor 230 for inspecting an inspection object 205. The processor 230 of the inspection system 200 includes an input unit 210, a processing unit 225, and an information storage unit 235.

[0018] The processor 230 may be any device capable of calculating the measurement data obtained by the measuring instrument 215 for the inspection object 205, and may be configured with an electronic computer and electronic circuits. The measuring instrument 215 is a device that obtains shape data of the object surface used in this inspection.

[0019] The measuring instrument 215 may be any device capable of measuring multiple points on the surface shape of the inspection object 205, and examples of such devices include a scanning type light wave distance meter, a range finder, a laser scanner, and a stereo camera. The measuring instrument 215 may be equipped with a position / angle adjustment mechanism so that the inspection object 205 can be imaged from multiple angles. In addition, a mechanism capable of changing the position and orientation of the measurement object 205 using a stage, a jig, or the like may be attached to the measurement object. The position and inclination during measurement are represented by a coordinate system 220 with the center of gravity of the inspection object 205 as the origin.

[0020] 3 shows a data group 300 input to the processor 230 of the inspection system 200 when the flowchart S100 is performed, and is stored in the information storage unit 235 in the processor 230. The data group 300 is composed of a measurement data group 305 measured by the measuring instrument 215, reference shape data 330, and inspection information 335.

[0021] 3 is composed of a combination of a measurement number 310, measurement data 315 of the object's surface shape, position and tilt data 320 of the measuring instrument 215 at the time of measurement, and measurement position and tilt error data 325. The measurement data 315 is information including information on a plurality of points on the object surface, and each point includes three-dimensional coordinate data or a part thereof.

[0022] The reference shape data 330 in the middle of the figure is information regarding the expected surface shape when the inspection object 205 is a good product, and may be three-dimensional measurement data of the good product, or point cloud data in which points are evenly assigned to CAD data or mesh data of the measurement object.

[0023] Inspection information 335 in the lower part of the figure is a summary of the items to be inspected, and for each item, it is composed of an inspection number 340, an inspection area 345, and an inspection procedure 350. The inspection area 345 is information that specifies, by means of rectangular information, an area to be used for inspection among the reference shape data 330 and the measurement data 315. The inspection procedure 350 is a summary of the procedure for calculating the inspection result of the corresponding inspection item using the data of the inspection area 345 among the measurement data.

[0024] The operation based on the flowchart in FIG. 1 is as follows. Step (hereinafter abbreviated as S) 105: A measurement data set 305 regarding the surface shape of the inspection object 205 is acquired from a storage unit. S110: For each of the measurement data 315 in the measurement data group, model data 355 is created from the reference shape data 330, the measurement position / tilt 320, and the measurement position / tilt error 325. S115: Calculate a transformation matrix A for superimposing the model data 355 on the measurement data 315. The transformation matrix A transforms the coordinates of each point in the model data 355 using the following formula (1).

[0025]

number

[0026] S120: Coordinate transformation is performed using the inverse matrix of transformation matrix A so as to superimpose the measurement data on model data 355. The three-dimensional coordinates after transformation are expressed by the following equation (2).

[0027]

number

[0028] S125: Each item in the examination information 335 is calculated for the converted measurement data. FIG. 4 shows a processing procedure for creating the model data 355. The operation based on the flowchart in FIG. 4 is as follows. S405: If the measurement position and tilt, and the measurement position and tilt error are expressed by the following equation (3),

[0029]

number

[0030] The maximum X coordinate of the position is Xp+Xpdiv, and the minimum is Xp-Xpdiv. The set of all combinations where the X, Y, and Z coordinates and the inclinations around the X, Y, and Z axes θX, θY, and θZ are minimum and maximum, respectively, are calculated as set C. The positions and inclinations in set C are (Pi,Di) (i=1 to L), and the number of combinations is L. S410: For the set C of positions and inclinations, calculate the measurement area Ai of the object to be measured at each position and inclination (Pi, Di). The measurement area Ai is information that represents the area to be measured in the reference shape data 330 when a measurement is performed from a specified position and inclination, and may be data formats assumed for the reference shape data 330, such as mesh data and point cloud data. As a calculation method, when the shape data 330 is mesh data, a method in which a mesh group existing in the measurement range from the measuring device is set as the measurement area Ai, and when it is point cloud data, a method in which point cloud data within the measurement range is set as the measurement area Ai can be considered. In the case of CAD data, it can be calculated by the above method by converting it into mesh data or point cloud data. The set of measurement areas Ai is set as set D. S415: A common area of ​​the measurement areas Ai is calculated and set as the measurement area. The common area is an area included in all the measurement areas Ai in the set D. S420: Point cloud data within the measurement region is extracted from the shape data 330 and used as model data. If the shape data 330 is not point cloud data, it is converted to point cloud data. The area of ​​the model data generated by this process is included in, and is equal to, or a part of, the area of ​​the measurement data 315. This reduces errors when calculating the overlay matrix A in S115.

[0031] FIG. 5 shows a process procedure 500 of S110 for calculating a transformation matrix A for superimposing the model data 355 on the measurement data 315.

[0032] The operation based on the flowchart in FIG. 5 is as follows. S505: A transformation matrix A for superimposing the model data 355 on the measurement data 315 is generated as a four-dimensional unit matrix. S515: For each point Pmi in the model data 355, a point Pni having the smallest Euclidean distance among the points in the measurement data 315 is calculated and set as the corresponding point of Pmi. S520: The sum of squares of the distances between each point Pmi of the model data and the corresponding point Pni is calculated as the registration error E from the following formula (4), where N is the number of points of the model data.

[0033]

number

[0034] S525: If the registration error E is equal to or smaller than the threshold, this step ends. If it is greater than the threshold, the process proceeds to S530. S530: A transformation matrix M that minimizes the registration error E is calculated as shown in the following formula (5). M is composed of a rotation matrix R and a translation matrix T.

[0035]

number

[0036] S535: The transformation matrix A is multiplied with M as shown in the following formula, and the value is updated. Also, each point Pmi in the model data is transformed and updated by the transformation matrix M, and the process proceeds to S515.

[0037] This step calculates a transformation matrix A that performs coordinate transformation on the model data so that the overlay error E with the measurement data is equal to or less than a threshold value. In calculating the transformation matrix M in S530, in processing procedure 500, all points in the model data are used to calculate the correspondence with the point in the measurement data that has the smallest Euclidean distance, and the transformation matrix is ​​calculated, but the calculation may also be performed using some points in the model data.

[0038] In calculating the sum-of-squares registration error E of the distance between each point Pmi of the model data and the corresponding point Pni, the error E may be calculated by weighting the error E based on information about the measurement point, such as the brightness value, measurement distance, and angle of the corresponding measurement point. The distance between corresponding points may be calculated using a robust function, mutual information, etc., other than the Euclidean distance. The transformation matrix M may be calculated using singular value decomposition, gradient method, etc.

[0039] 6 shows an inspection for evaluating the parallelism between planes 610 and 615 using measurement data of inspection object 205. At this time, measurement data 630 by measurement position 620 and measurement data 635 at measurement position 630 are used. Also, for measurement position 620, if there is an error of distance V in the position where the measurement was actually performed with respect to position 621 stored in the measurement position / tilt information, then the coordinate values ​​of each point in measurement data 630 will have an error of distance V with respect to the actual measurement point.

[0040] At this time, the effect of the configuration of this embodiment will be explained by comparing the case where the position of the measurement data is adjusted by the conventional method and the proposed method.

[0041] In the ICP algorithm, which is a conventional method, for each point Pni in the measurement data 630, the correspondence 640 with the nearest point Psi in the three-dimensional shape 330 is calculated. In the ICP algorithm, the sum of squares E' of the Euclidean distances between corresponding points is taken as the overlay error as shown in the following formula (6), and a transformation matrix A' of the measurement data that minimizes the error is calculated. At this time, the number of points in the measurement data is taken as N'.

[0042]

number

[0043] At this time, if a point 645 with a large measurement error is present in the measurement data 630, the distance between the point 645 with the large error and the corresponding point on the three-dimensional shape 330 will be greater than the distance V. Therefore, the overlay error E' before conversion is N'V 2 Become bigger.

[0044] In the configuration of this embodiment, first, model data 650 is generated in S110. The area of ​​the model data 650 is generated so as to be equal to or a part of the area of ​​the measurement data 630. For each point Pmi in the model data 355, a point Pni having the smallest Euclidean distance among the points in the measurement data 315 is calculated, and the sum of squares of the distance between Pmi and the corresponding point Pni is calculated as the registration error E using the following formula (7). N is the number of points in the model data.

[0045]

number

[0046] At this time, since the point 645 with a large measurement error cannot be associated with each point of the model data, the overlay error E is calculated by the conventional method, and is calculated by N'V, which is the sum of squares of the distance error V when measuring each point, which is the true value. 2 In both the present invention and the conventional method, a transformation matrix that minimizes the overlay error is calculated, so that the coordinates of the measurement data corrected by the transformation matrix are closer to the actual measurement point in the present invention.

[0047] In this embodiment, if the area of ​​the model data is equal to or larger than the measurement data, points in the model data outside the area of ​​the measurement data are also associated with points in the measurement data, resulting in large errors during adjustment. For this reason, the model data is generated in accordance with the procedure in step S110 so that the area of ​​the model data is equal to or smaller than the measurement data.

[0048] The effect of S120 of this embodiment will be described. When evaluating the parallelism between the plane 610 and the plane 615, multiple measurement data from measurement positions different from the measurement positions 620 and 630 are used in combination because the planes span multiple planes. In this case, if the measurement data 630 and 635 at the measurement positions 620 and 630 are transformed so that the respective model data 650 and 660 are superimposed on the measurement data, the coordinate relationship between the measurement data cannot be determined, making it difficult to automatically evaluate the inspection item by specifying the area. In the present invention, the measurement data is superimposed on the model data by the inverse transformation matrix in S115, and the measurement data is adjusted based on the coordinate system 220, and the adjusted measurement data is extracted by specifying the area in S125 to evaluate the inspection item. EXAMPLES

[0049] Example 2 shows a case where the present invention is applied to inspecting the shape of a part. The difference between this example and Example 1 is the model data creation method in S110. In this example, the measurement area of ​​the reference shape data is manually input, and the reference shape data within the measurement area is calculated as model data. This makes it possible to correct the measurement data and perform inspection even if measurement position / tilt data and measurement position / tilt error data do not exist.

[0050] 7 shows a GUI 700 for inputting a measurement region. The measurement region is input through the input unit 210. The GUI 700 at the time of input comprises a display section 705 of the mesh of the reference shape 330 to be measured, a measurement area selection section 710, a viewpoint position / tilt input section 715 of the display section 705 and selection section 710, a measurement number input section 720 and a measurement data display section 725 of the corresponding number, and a viewpoint position / tilt input section 730 of the display section 725. In the GUI 700, the measurement area is input by selecting mesh data created from the reference shape data as the measurement area.

[0051] In this embodiment, in S110, it is sufficient to calculate the model data from the measurement area manually input, and the measurement area may be input by area designation, or the measurement area may be input by directly instructing whether each point in the reference data is within the measurement area or not. EXAMPLES

[0052] In the third embodiment, the present invention is applied to inspection of the shape of a part using a measuring instrument equipped with a position / attitude changing mechanism.

[0053] 8 shows an inspection system 800 in this embodiment when applied to inspecting the shape of a part using a measuring machine equipped with a position / attitude changing mechanism. It is composed of a measuring machine 810 equipped with a position / tilt changing mechanism and a processor 230. The processor 230 of this system has an input unit 210, a processing unit 225, and an information storage unit 235.

[0054] The measuring instrument 810 may be any device capable of measuring the surface shape of the measurement object 815 from multiple measurement positions, and may be a measuring instrument 215 attached to a robot hand or a measuring instrument 215 attached to a rotating stage in addition to the measuring instrument 215 attached to the mobile robot shown in FIG. 8.

[0055] The difference between this embodiment and the first embodiment is the method of acquiring the surface shape data in S105. Fig. 9 is a flowchart showing the processing procedure of S105. S910: A union region 920 of the examination regions of all examination items in the examination information is calculated. S915: A combination of the measurement position and tilt at which the union region 920 is measured is manually input. S920: The measurement target 840 is measured at the input measurement position and inclination.

[0056] FIG. 10 shows a GUI 1000 for inputting a combination of measurement positions and orientations in S915. The GUI 1000 is composed of an inspection area display section 1010 that displays the union 920 of the inspection areas on the reference shape, a measurement number input section 1015, a measurement position / tilt input section 1020 for the corresponding input number, a measurement position / tilt error input section 1025, a model data display section 1030, a display viewpoint input section 1035 of the input section 1020, and a display viewpoint input section 1040 of the display section 1030.

[0057] A measurement area display unit 1030 creates and displays the model data of S110 from the measurement position / tilt input unit 1020 and the position / tilt error input unit 1025. The model data shows the surface of the target inspection object that is always measured at the input position and tilt, regardless of the position and tilt error, as long as it is within the range of the input position and tilt error. This allows the inspector, when specifying the measurement position and tilt, to determine whether the union area 920 of the inspection areas that needs to be measured for inspection can be sufficiently measured at the specified measurement position and tilt.

[0058] The present invention is not limited to the above-mentioned embodiment, and various modifications are included. For example, the above-mentioned embodiment has been described in detail to easily explain the present invention, and is not necessarily limited to those having all the configurations described. It is also possible to replace a part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. It is also possible to add, delete, or replace a part of the configuration of each embodiment with another configuration. In addition, each of the above-mentioned configurations, functions, processing units, processing means, etc. may be realized by hardware, for example, by designing them as integrated circuits. In addition, each of the above-mentioned configurations, functions, etc. may be realized by software by a processor interpreting and executing a program that realizes each function. Information such as a program, table, file, etc. that realizes each function can be placed in a memory, a recording device such as a hard disk or SSD (Solid State Drive), or a recording medium such as an IC card, SD card, or DVD. [Explanation of symbols]

[0059] 200, 800 Inspection System 205 Inspection subjects 210 Input section 220 Coordinate System 230 Processing Machine 305 Measurement Data Set 315 Measurement Data 330 Reference shape of the object to be inspected 335 Test Information 400 Flowchart showing the processing procedure of S105 of Flowchart 100 500 Flowchart showing the processing procedure of S110 of the flowchart 100 700 In step S110 of the flowchart 100, a GUI for manually selecting a measurement area 900 Flowchart showing the processing procedure of S105 of the flowchart 100 in the third embodiment 1000 GUI for inputting combinations of measurement positions and orientations

Claims

1. A shape inspection method for inspecting a three-dimensional shape of an object to be inspected by arithmetic processing of the three-dimensional shape of the object to be inspected measured by a measuring device, comprising: generating model data from the reference shape and measurement area information of the target object; A step of calculating a transformation matrix for the model data to the measurement data, Based on the reference shape of the object to be inspected and the position and orientation information at the time of measurement, Calculating the measurement area information, A shape inspection method comprising:

2. A shape inspection method for inspecting a three-dimensional shape of an object to be inspected by arithmetic processing of the three-dimensional shape measured by a measuring device, comprising: generating model data from the reference shape and measurement area information of the target object; A step of calculating a transformation matrix for the model data to the measurement data, The measurement area information is input by an operator's operation. A shape inspection method comprising:

3. 3. The shape inspection method according to claim 1, further comprising: a step of adjusting the measurement data by an inverse matrix of a transformation matrix of the model data to the measurement data; A shape inspection method comprising:

4. A shape inspection system, comprising: The present invention includes a measuring device for measuring a three-dimensional shape of an object to be inspected, and a calculation processing device, the arithmetic processing device generates model data from the reference shape and measurement area information of the target inspection object, and calculates a transformation matrix for the model data to the measurement data; the calculation processing device calculates the measurement area information from the reference shape of the target object and position / posture information at the time of measurement; A shape inspection system comprising:

5. 1. A shape inspection system, comprising: The present invention includes a measuring device for measuring a three-dimensional shape of an object to be inspected, and a calculation processing device, the arithmetic processing device generates model data from the reference shape and measurement area information of the target inspection object, and calculates a transformation matrix for the model data to the measurement data; An input unit that inputs the measurement area information by an operation of an operator is further provided. A shape inspection system comprising:

6. A shape inspection system according to claim 4 or claim 5, the calculation processing device adjusts the measurement data using an inverse matrix of a transformation matrix for transforming the model data into the measurement data; A shape inspection system comprising:

Images