Method and computer program product for filtering a measurement dataset

Abstract

The invention relates to a method and a computer program product for filtering a measurement dataset, the dataset being usable for specifying and / or verifying an internal feature (10) of a workpiece (100), the method comprising: providing a measurement dataset comprising a plurality of measurement points (P) of the internal feature (10); providing a helper feature (30) representing an ideal estimate of the internal feature (10) of the workpiece (100); mirroring each measurement point (P) of the measurement dataset with respect to a bounding element of the helper feature, thereby generating a first modified dataset comprising a plurality of first modified measurement points (P'); determining a convex hull (H) of the first modified measurement points (P') and projecting the first modified measurement points (P') to the determined convex hull (H), thereby generating a second modified dataset comprising a plurality of second modified measurement points (P"); and mirroring each second modified measurement point (P") with respect to a bounding element of the helper feature (30), thereby generating a filtered measurement dataset comprising a plurality of filtered measurement points (P'").

Description

Technical Field

[0001] This invention relates to a method for filtering measurement datasets that can be used to specify and / or verify internal features of a workpiece. Furthermore, this invention relates to a corresponding computer program product. Background Technology

[0002] In metrology, a crucial task is verifying that workpieces are manufactured according to the designer's specifications. Specifications are typically defined by CAD (Computer-Aided Design) models or drawings and are intended to describe functional requirements. Many specifications, such as directional tolerances or positional tolerances, are defined relative to reference features on the workpiece. These reference features are called "datums."

[0003] As an example: the functional requirement of a workpiece might be to assemble it with a corresponding component. The workpiece may include a flat surface with a hole, and the corresponding component may include a flat surface with a corresponding bolt. The flat surface with bolts of the corresponding component must fit the corresponding hole and flat surface of the workpiece. Therefore, the functional requirement of the specification might be to characterize the flat surface of the workpiece as perpendicular to the hole. In this case, the hole (geometrically a cylinder) serves as a reference.

[0004] To verify this specification, the reference must be measured using measuring equipment, and the obtained measurement points must be further processed. According to the current ISO 5459 standard (i.e., ISO 5459:2011), the obtained measurement points should be used to determine the optimal geometry of the reference (e.g., the optimal cylinder). However, since any measurement method has the possibility that some measurement points are outliers and do not reflect the true surface, Annex A of ISO 5459:2011 requires filtering the obtained measurement points before determining or calculating the associated cylinder based on the measurements. However, currently, ISO 5459:2011 does not explicitly specify which filter to use.

[0005] For external features (such as bolts on a workpiece), it is known that the obtained measurement points are filtered by determining the convex hull of the measurement points, and the filtered measurement points are obtained by projecting the measurement points onto the determined convex hull. This process satisfies the following conditions required for filtering the measurement data of external features:

[0006] - Outliers within features should be eliminated, as they should have minimal impact on baseline correlation processing;

[0007] - The outermost points should be retained because these are the relevant points used to calculate the baseline correlation; and

[0008] - The filter should behave similarly to a morphological filter, that is, it should smooth out deviations.

[0009] Therefore, this known convex hull-based filtering method is qualified as a candidate for the default filtering process for outer features, and has been proposed as the default method for preprocessing benchmark calculation (association) for the new GPS ISO 5459 standard.

[0010] However, the requirements and / or conditions for filtering measurement points on inner or internal features (such as holes in a workpiece) differ from those described above for outer features. Specifically, for methods of filtering measurement data on inner features, the following conditions must be met to be considered as a candidate for default filtering:

[0011] - Outliers outside the features should be eliminated, as they should have minimal impact on the baseline correlation process;

[0012] - The innermost points should be retained because these are the relevant points used to calculate the baseline correlation; and

[0013] - The filter should behave similarly to a morphological filter, that is, it should smooth out deviations.

[0014] Considering different conditions and / or requirements, the convex hull-based filtering method proposed for outer features cannot be directly applied to inner features. Therefore, considering inner features, the use of other completely different types of filters is currently being discussed, particularly morphological filters (e.g., closed spheres with finite radii). Summary of the Invention

[0015] Therefore, the problem of the present invention is to unify the filtering methods used for both inner and outer features. In particular, the problem of the present invention is to provide a convex hull-based filtering method for inner features of a workpiece. This problem is solved by the subject matter of the independent claims. Preferred embodiments are defined in the dependent claims.

[0016] According to one aspect of the present invention, a method for filtering a measurement dataset capable of specifying and / or verifying internal features of a workpiece is provided, the method comprising the steps of: providing a measurement dataset including a plurality of measurement points of the internal features; providing an auxiliary feature representing an ideal estimate of the internal features of the workpiece; mirroring each measurement point of the measurement dataset relative to boundary elements of the auxiliary feature, thereby generating a first modified dataset including a plurality of first modified measurement points; determining the convex hull of the first modified measurement points and projecting the first modified measurement points onto the determined convex hull, thereby generating a second modified dataset including a plurality of second modified measurement points; and mirroring each second modified measurement point relative to boundary elements of the auxiliary feature, thereby generating a filtered measurement dataset including a plurality of filtered measurement points.

[0017] Measurement datasets can be obtained through experiments using measuring equipment. In particular, measurement datasets are obtained to specify and / or verify the internal characteristics of an object or workpiece.

[0018] In this invention, the term "feature" specifically refers to or means a geometric shape. An "internal feature" of a workpiece (also referred to herein as an "inner feature") includes voids, openings, and / or hollow spaces within the workpiece. In particular, an internal feature is the geometry of a cavity (such as a hole or aperture) within the workpiece. An internal feature or cavity of the workpiece can be a reference feature of the workpiece, i.e., a reference feature of the workpiece.

[0019] "Providing a measurement dataset" can include acquiring, determining, and / or capturing measurement data. The measurement dataset for an internal feature includes multiple measurement points of the internal feature. Specifically, the measurement points include, or are, the coordinates of measurement points of the internal feature. Measurement points can be surface or boundary points, i.e., points on the surface or boundary of the internal feature of the workpiece. Depending on the internal feature, the coordinates can include 2D coordinates and / or 3D coordinates; that is, the coordinates can involve two-dimensional and / or three-dimensional coordinate systems. The coordinate system can be predefined by the measuring device used to acquire the measurement data.

[0020] The measuring device used to obtain measurement data can be a coordinate measuring device. Coordinate measuring devices can be based on, for example, tactile, optical, and / or computed tomography measurements.

[0021] "Providing auxiliary features" can include determining and / or calculating auxiliary features. An "auxiliary feature" can be a specified (i.e., predetermined or determinable) ideal feature or geometry. Specifically, an "auxiliary feature" can be an ideal feature or geometric element (that can be mathematically described and / or defined). An auxiliary feature can be, for example, a circle, a sphere, a cylinder, or a cone. These features have a common radial parameter. In particular, an auxiliary feature is a least-squares geometric element (e.g., a least-squares circle, least-squares sphere, least-squares cylinder, or least-squares cone), and may also be referred to as a "Gaussian element" or "Gaussian feature." In other words, "providing auxiliary features" can include providing auxiliary (ideal) geometric data for the internal features of a workpiece. The auxiliary feature and / or auxiliary geometric data represent an ideal estimate of the internal features of the workpiece. An "ideal estimate" specifically refers to an ideal model with a perfect form.

[0022] The measurement points of the measurement dataset are mirrored on the boundary elements of the auxiliary feature. The boundary elements of the N-dimensional auxiliary feature are typically hyperplanes in n-dimensional Euclidean space, where the hyperplane forms the boundary of the auxiliary feature. Specifically, the boundary elements of the auxiliary feature can be the boundary surface (also called the reflection surface), boundary line (also called the reflection line), or boundary point (also called the reflection point), although the boundary point is not a hyperplane, it can be derived from the hyperplane. For example, when the auxiliary feature is a circle, the boundary element can be the circle's circular line; when the auxiliary feature is a sphere, the boundary element can be the surface (or curved region) of the sphere; when the auxiliary feature is a cylinder, the boundary element can be a curved region of the cylinder; and when the auxiliary feature is a cone, the boundary element can be a curved region of the cone.

[0023] The first modified dataset (i.e., the set of modified measurement points {P'}) is generated by mirroring each measurement point of the measurement dataset on the boundary elements of the auxiliary features. Since the first modified dataset includes the mirrored measurement data, it can also be called the mirrored measurement dataset.

[0024] The convex hull is determined or computed based on the first modified dataset. In other words, the convex hull is determined based on mirrored measurement points. Determining the convex hull means constructing an unambiguous and efficient representation of the desired convex shape for a finite set of points (here referring to the first modified dataset). This is preferably accomplished computationally, i.e., by means of a microprocessor and / or computer. In particular, any suitable standard convex hull algorithm can be used within the scope of this invention. For example, in the two-dimensional and three-dimensional cases, any of the following well-known convex hull algorithms can be used: the "Incremental brute force algorithm" (also known as "Gift wrapping," or "Jarvis march" in the two-dimensional case), "Quickhull," "Divide and Conquer," and "Chan's algorithm." Furthermore, at least in the two-dimensional case, any of the following well-known convex hull algorithms can be used: the "Monotone chain" (also known as "Andrew's algorithm") and the "ultimate planar convex hull algorithm." Since these and other algorithms for determining or computing the convex hull are well-known in the art, they will not be described further here. Note that the above list is not exhaustive, and therefore any other suitable convex hull algorithm may also be used within the scope of this invention.

[0025] Then, the points of the first modified dataset are projected onto the determined convex hull. Specifically, each point of the first modified dataset is projected onto the determined convex hull. By projecting the points of the first modified dataset onto the determined convex hull, a second modified dataset (i.e., the set {P”} of second modified measurement points that have been projected onto the convex hull) is generated. Points {P”} can be referred to as projected points, and the second modified dataset can be referred to as the projected dataset.

[0026] Then, each point in the second modified dataset (i.e., each second modified measurement point) is mirrored on the boundary elements of the auxiliary features. This generates the filtered measurement dataset (also known as the third modified dataset). The filtered measurement dataset is the set {P”'} of filtered measurement points (also known as third modified measurement points).

[0027] Specifically, through the first mirroring step, the measurement data of the workpiece's internal features are transformed to represent the corresponding virtual external features of the workpiece. Specifically, the first modified dataset represents the corresponding virtual external features of the workpiece. Through the second mirroring step, the transformation (or the first mirroring) is reversed. In other words, through the second mirroring step, the data representing the virtual external features is transformed into or mirrored back to data representing the actual internal features of the workpiece.

[0028] Specifically, the mirroring process (i.e., both the first and second mirroring steps) includes the transformation of measurement points with the properties of the outermost point becoming the innermost point and the innermost point becoming the outermost point. For this purpose, an auxiliary feature is used as a "mirror." For example, in the case where the auxiliary feature is a cylinder (e.g., a Gaussian cylinder), points inside / outside the cylinder are moved along a direction orthogonal to the cylinder's surface and / or axis to the outer / inner side of the cylinder. Specifically, these points are moved such that the distance from the moved points to the cylinder's surface is the same as before the movement.

[0029] By means of the method according to the invention, measurement data of external features of a workpiece can be filtered not only by determining the convex hull, but also by filtering measurement data of internal features of the workpiece. Therefore, the invention provides a convex hull-based filtering method applicable to one or more internal features of a workpiece, and thus advantageously unifies filtering methods used for both internal and external features of the workpiece.

[0030] In a preferred embodiment, providing the measurement dataset includes capturing measurement data (and / or measurement points included in the measurement data) using a coordinate measuring device or coordinate measuring machine. The coordinate measuring device may be based on tactile measurement, such as touch-triggered measurement. Alternatively, the coordinate measuring device may be based on optical measurement. Alternatively, the coordinate measuring device may be based on computed tomography (CT) measurement. Specifically, the coordinate measuring device may be a tactile coordinate measuring machine, an optical measuring device, a computed tomography scanner, or a combination thereof.

[0031] In another preferred embodiment, the auxiliary feature is a predetermined ideal feature defined based on the design data of the internal features and / or the design data of the workpiece. In other words, the auxiliary feature can represent the ideal internal features of the workpiece. For example, the design data may include, or may be, the CAD data of the internal features and / or the workpiece.

[0032] In another preferred embodiment, the auxiliary features are Gaussian features determined based on the measurement dataset. "Gaussian features" (also referred to as geometric compensation elements according to Gauβ) are geometric elements determined or calculated based on the measurement data using the well-known least squares method. According to this method, the sum of squared distances of all measurement points relative to an ideal geometry (e.g., a circle, sphere, cylinder, or cone) is minimized. The ideal geometry can be a specified ideal geometry, i.e., it can be predetermined or determinizable.

[0033] In another preferred embodiment, the auxiliary feature is a circle, sphere, cylinder, or cone. However, it should be understood that the auxiliary feature can generally be any geometric element (especially 2D or 3D) that can be mathematically described or defined.

[0034] In another preferred embodiment, the step of mirroring each measurement point of the measurement dataset on the boundary element of the auxiliary feature (i.e., the first mirroring step) includes: defining a corresponding first reflection point on the boundary element of the auxiliary feature for each measurement point; determining a corresponding first distance between the measurement point and the corresponding first reflection point for each measurement point; and generating the first modified dataset by determining mirrored measurement points for each measurement point, wherein the mirrored measurement points are obtained by moving the measurement point through the corresponding first reflection point by twice the determined corresponding first distance.

[0035] The corresponding first reflection point is specifically defined for the first mirroring step. The mirroring measurement point corresponds to the first modified measurement point.

[0036] For each measurement point, a corresponding first reflection point is defined such that: the corresponding first reflection point is located on the boundary element of the auxiliary feature, and the virtual line between the measurement point and the corresponding first reflection point is perpendicular to the tangent or tangent plane of the auxiliary feature at the corresponding first reflection point.

[0037] Whether the virtual line between the measurement point and the corresponding first reflection point is perpendicular to the tangent or the tangent plane depends on the dimension of the auxiliary feature. When the auxiliary feature is two-dimensional, the virtual line between the measurement point and the corresponding first reflection point is perpendicular to the tangent. And when the auxiliary feature is three-dimensional, the virtual line between the measurement point and the corresponding first reflection point is perpendicular to the tangent plane.

[0038] In another preferred embodiment, the step of projecting the points of the first modified dataset onto the determined convex hull is performed orthogonally to the convex hull. In other words, the projection is performed such that each point of the first modified dataset (i.e., each first modified measurement point) moves along a line perpendicular to the convex hull until it reaches the convex hull. Optionally or additionally, the projection of the points of the first modified dataset onto the determined convex hull is performed radially from the center or axis of the auxiliary feature. In other words, the projection is performed such that each point of the first modified dataset (i.e., each first modified measurement point) moves radially from the center of the auxiliary feature (particularly when the auxiliary feature is a circle or sphere) or the axis (particularly when the auxiliary feature is a cylinder or cone) until the moved point reaches the convex hull.

[0039] In another preferred embodiment, the step of mirroring each second modified measurement point of the second dataset on the boundary element of the auxiliary feature includes: defining a corresponding second reflection point on the boundary element of the auxiliary feature for each second modified measurement point; determining a corresponding second distance between the second modified measurement point and the corresponding second reflection point for each second modified measurement point; and generating the filtered measurement dataset by determining a mirror point for each second modified measurement point, wherein the mirror point is obtained by moving the second modified measurement point through the corresponding second reflection point by twice the determined corresponding second distance.

[0040] The corresponding second reflection point is specifically defined for the second mirroring step. The corresponding second reflection point can correspond to the corresponding first reflection point. The mirrored point corresponds to the filtered measurement point.

[0041] In another preferred embodiment, for each second modified measurement point, a corresponding second reflection point is defined such that the corresponding second reflection point is located on the boundary element of the auxiliary feature, and the virtual line between the second modified measurement point and the corresponding second reflection point is perpendicular to the tangent or cross-section of the auxiliary feature at the corresponding second reflection point.

[0042] Whether the virtual line between the measurement point and the corresponding second reflection point is perpendicular to the tangent or the tangent plane depends on the dimension of the auxiliary feature. When the auxiliary feature is two-dimensional, the virtual line between the second modified measurement point and the corresponding second reflection point is perpendicular to the tangent. And when the auxiliary feature is three-dimensional, the virtual line between the second modified measurement point and the corresponding second reflection point is perpendicular to the tangent plane.

[0043] In another preferred embodiment, the method further includes specifying and / or verifying internal features of the workpiece based on the filtered measurement dataset and / or filtered measurement points.

[0044] In another preferred embodiment, specifying and / or verifying the internal features of the workpiece includes determining least-squares geometric elements (Gaussian elements) based on a filtered measurement dataset. The determined least-squares geometric elements may be specifications of the measured internal features. Specifically, the determined least-squares geometric elements may be compared with predefined ideal internal features and / or design data for the internal features and / or design data for the workpiece. For example, the design data may include, or may be, CAD data for the internal features and / or the workpiece. Through the above comparison step, the internal features can be verified. Specifically, through the above comparison step, it can be verified whether the investigated or measured internal features of the workpiece are within predetermined tolerances.

[0045] The method according to the invention particularly relates to a computer-implemented method. Accordingly, a computer having a processor, memory, and a display can be provided to perform the method according to the invention. More specifically, a computer or computer system including a processor capable of implementing or executing machine-readable instructions that perform some or all of the methods, functions, and other processes described herein can be provided. Commands and data from the processor are communicated via a communication bus. The computer or computer system may also include main memory such as random access memory (RAM) and auxiliary data memory, wherein the processor's machine-readable instructions and data may reside in the main memory during operation, and the auxiliary data memory may be non-volatile and store machine-readable instructions and data. Memory and data memory are examples of computer-readable media.

[0046] According to another aspect of the invention, a computer program product is provided, comprising computer-readable instructions that, when loaded into a computer's memory and executed by the computer, cause the computer to perform the method described according to the invention. In particular, the computer program product may relate to a program stored on a computer-readable medium. Optionally, the computer program product may relate to a computer-readable medium storing a program or corresponding program code including computer-readable instructions.

[0047] For the other independent aspects described above, and particularly for the preferred embodiments thereof, the explanations given above or below relating to the embodiments of the first aspect also apply. In particular, for one independent aspect of the invention and its preferred embodiments, the explanations given above and below relating to the embodiments of the various other aspects also apply.

[0048] Individual embodiments for solving the problem are described below by way of example with reference to the accompanying drawings. In this context, the described individual embodiments have some features that are not absolutely necessary for achieving the claimed subject matter, but provide the required properties in a particular application. In this respect, embodiments that do not have all the features of the embodiments described below are also intended to be disclosed in a manner consistent with the described technical teachings. Furthermore, to avoid unnecessary repetition, only specific features of the individual embodiments described below are mentioned. Therefore, it should be noted that these individual embodiments are not only intended to be considered individually, but also collectively. Based on this common consideration, those skilled in the art will recognize that individual embodiments can also be modified by including one or more features of other embodiments. It should be noted that systematic combinations of individual embodiments with the one or more features described with respect to other embodiments may be desirable and advantageous, and are therefore intended to be considered and are also considered to be included in the specification. Attached Figure Description

[0049] The above and other objects, features, and advantages of the present invention will become more apparent after reading the following description of preferred embodiments and the accompanying drawings. Other features and advantages of the subject matter described herein will become apparent from the description, the accompanying drawings, and the claims. It should be understood that even though embodiments are described separately, individual features and functions of the embodiments can be combined without prejudice to other embodiments. This disclosure is illustrated by way of example and is not limited to the accompanying drawings.

[0050] Preferred embodiments of the invention are described by way of example with reference to the following figures:

[0051] Figure 1a shows a schematic representation of a workpiece with internal features and a corresponding corresponding component of the workpiece, wherein the corresponding corresponding component has external features that correspond to the internal features of the workpiece.

[0052] Figure 1b shows a schematic representation of a filter that should be applied to a nominally flat surface with an outer feature according to ISO 5459:2011.

[0053] Figure 1c shows a schematic representation of a filter that should be applied to the nominal cylindrical surface of the inner features according to ISO 5459:2011 standard;

[0054] Figure 2 A flowchart illustrating a method according to a preferred embodiment of the present invention is shown;

[0055] Figure 3a A schematic diagram illustrating the method according to the invention is shown by way of a first example;

[0056] Figure 3bA schematic diagram illustrating the method according to the invention is shown by way of a second example;

[0057] Figure 4 Show Figure 3a An enlarged view of the schematic diagram is provided to illustrate further details of the method according to the preferred embodiment;

[0058] Figure 5 The measurement points and filtered points are shown based on the internal characteristics of the workpiece in the first example;

[0059] Figure 6 The measurement points and filtered points are shown based on the internal characteristics of the workpiece according to the second example.

[0060] Explanation of reference numerals in the attached figures

[0061] 1 Surface

[0062] 2 Surface

[0063] 3. True and complete surface

[0064] 4. Filtered features

[0065] 10. Internal Characteristics

[0066] 20. External features (external characteristics)

[0067] 30 auxiliary features

[0068] 100 workpieces

[0069] Corresponding components of workpiece 200

[0070] d1 First distance

[0071] d2 Second distance

[0072] H convex hull

[0073] P represents the points (measurement points) in the measurement dataset.

[0074] P' is the point in the first modified dataset (the first modified measurement point).

[0075] P” is the point in the second modified dataset (the second modified measurement point).

[0076] P”' Points in the filtered dataset (filtered measurement points)

[0077] R1 First Reflection Point

[0078] R2 Second Reflection Point

[0079] T-tangent / section Detailed Implementation

[0080] The following detailed description relates to exemplary embodiments of the invention. Other embodiments of the invention may be within the scope of the invention as defined in the appended claims. Throughout the drawings, the same reference numerals are used for the same or similar elements.

[0081] Figure 1a shows a schematic representation of an exemplary workpiece 100 having an internal or inner feature 10 (a hole in the example shown) and a corresponding corresponding component 200 of the workpiece 100. The corresponding component 200 has an external or outer feature 20 (a bolt in the example shown) corresponding to the internal feature 10 of the workpiece 100. The functional requirement of the workpiece 100 is to assemble it with the corresponding component 200. The flat surface 2 of the corresponding component 200 with the bolt 20 must be fitted into the corresponding hole 10 and flat surface 1 of the workpiece 100.

[0082] The functional requirement regarding the specifications could be that the flat surface 1 of the workpiece 100 is characterized as perpendicular to the hole 20 of the corresponding component 200. In this case, the hole 10, which is geometrically cylindrical, is used as a reference.

[0083] To verify this specification, the first step could be to establish a reference. Therefore, workpiece 100 can be placed on a coordinate measuring machine (not shown in Figure 1) to capture measurement points from the inside of hole 10 (e.g., by contacting it from the inside with a probe). These measurement points should geometrically represent hole 10 as a cylinder. According to the current ISO 5459:2011 standard, the measurement points are used to determine the optimal cylinder with the largest diameter capable of fitting into hole 10 and possessing perfect form. This means that all measurement points must be located on the outside of the cylinder. The calculated cylinder (the associated cylinder) serves as the reference.

[0084] To capture measurement points, various methods using a coordinate measuring machine (CMM) can be used, such as scanning or touch triggering. Alternatively, optical measuring machines and / or computed tomography (X-ray) based machines can be used. With any method, there is a possibility that some measurement points are outliers and do not reflect the true surface. Therefore, Annex A of ISO 5459:2011 specifies a filtering step before calculating the associated cylinder. However, it does not explicitly specify which filter to use.

[0085] Figure 1b provides a schematic representation of what filtering relative to the outer feature 20 should typically look like according to the ISO 5459:2011 standard. In the example shown, corresponding to Figure A.3 in Annex A of the ISO 5459:2011 standard, the outer feature has a nominally flat surface. The continuous black line 3 in Figure 1b represents the true, intact feature, while the dashed line 4 represents the filtered feature.

[0086] Figure 1c shows a schematic representation illustrating how filtering relative to the inner feature 10 should typically look according to the ISO 5459:2011 standard. In the example shown, corresponding to Figure A.3 in Annex A of the ISO 5459:2011 standard, the inner feature 10 has a nominal cylindrical surface. As in Figure 1b, the continuous black line 3 in Figure 1c represents the true, complete feature, while the dashed line 4 represents the filtered feature.

[0087] As described above, for external features like bolt 20 in Figure 1a, it is known to filter measurement points by determining the convex hull of the measurement point. More specifically, the general process for filtering external features can be described as follows: for a measurement point {P}, the convex hull is found and the point is projected onto the convex hull to generate a point {P*}, where point {P*} represents the filtered point. However, considering internal features such as hole 10 of workpiece 100 as shown in Figure 1a, a convex hull-based filter with the same conditions has not been known until now. The present invention fills this gap and provides a filtering method applicable to internal features of workpieces, which is also a convex hull-based filtering method. Therefore, the present invention enables a unified filtering method for internal and external features.

[0088] Figure 2 A flowchart illustrating a method for filtering a measurement dataset that can be used to specify and / or verify internal features 10 of a workpiece 100 according to a preferred embodiment of the present invention is shown. In step S1, a measurement dataset including multiple measurement points of the internal features is provided. In step S2, an auxiliary feature representing an ideal estimate of the internal features of the workpiece is provided. In step S3, each measurement point of the measurement dataset is mirrored on the boundary elements of the auxiliary feature, thereby generating a first modified dataset including multiple first modified measurement points. In step S4, the convex hull of the first modified measurement points is determined, and the first modified measurement points are projected onto the determined convex hull, thereby generating a second modified dataset including multiple second modified measurement points. And in step S5, each second modified measurement point is mirrored on the boundary elements of the auxiliary feature, thereby generating a filtered measurement dataset including multiple filtered measurement points.

[0089] Specifically, a measurement dataset is provided and a filtered measurement dataset is generated as output. The filtered measurement dataset is the result of a smoothing process. For this smoothing process, auxiliary features (i.e., auxiliary geometric elements) such as least-squares cylinders are provided (specifically, computed). The distance from the measurement point to this auxiliary feature is the value to be smoothed. After smoothing these distance values, the original measurement point is moved from its original distance (relative to the auxiliary geometric element) to the smoothed distance value. This produces a filtered set of points. Typically, each point in the measurement data affects the computation of associated features (e.g., associated cylinders). In particular, for least-squares-based associations, measurement points with larger distances relative to the auxiliary features have a greater impact on the results. Filtering is performed to avoid irrelevant measurement points (so-called "outliers") from affecting the results. For example, this filtering has an effect similar to mechanically removing protrusions from bolts by adjusting gauges (holes) on bolts. Therefore, the filtering method provided by this invention can replace (expensive, cumbersome, and time-consuming) mechanical inspection with low-cost, simple, and fast measurements combined with algorithmic computation.

[0090] Specifically, the measurement points of the workpiece's inner features are simplified to those of the outer features by mirroring the measurement points at the reference feature, applying a convex hull filter to the outer features, and mirroring the result back. This preserves the properties of the convex hull (e.g., high points) for both the outer and inner features. It is well known that convex hulls cannot be directly applied to inner features. However, this invention makes it possible to apply convex hulls to inner features as well. Therefore, both inner and outer features can be filtered by the same filter type.

[0091] exist Figure 3a and Figure 3b The principles of the invention are illustrated in two different examples. More specifically, Figure 3a The diagram illustrates the inner feature filtering process used for a measurement point having two extreme peaks on the inner side of the inner feature 10 facing the workpiece to be measured. Figure 3b The diagram illustrates the inner feature filtering process used for measurement points having two extreme peaks on the inner side of the inner feature 10 facing the workpiece to be measured. In both examples, the inner feature 10 is circular, and therefore the auxiliary feature 30 is circular. Figure 3a and Figure 3b In the diagram, continuous circular lines represent a portion of auxiliary feature 30, specifically a portion of the boundary elements of auxiliary feature 30. Therefore, note that... Figure 3a and Figure 3bOnly some of the inner features 10 and auxiliary features 30 are shown or indicated. Points with reference marker P are measurement points obtained from the coordinate measuring device; points with reference marker P' are mirror measurement points; points with reference marker P” are points projected onto the convex hull H; and points with reference marker P”’ are mirror points of the convex hull H. Therefore, measurement point P forms a measurement dataset, mirror measurement point P' forms a first modified dataset, projected measurement point P” forms a second modified dataset, and point P”’ (also called the filtered measurement point) forms a third modified dataset (also called the filtered dataset). Figure 3a , Figure 3b and Figure 4 In the diagram, the measurement point P is represented by a solid circle (disk), and the corresponding continuous profile is represented by a short dashed line. The mirror measurement point P' is represented by a solid square (frame), and the corresponding continuous profile is represented by a dotted line. The projection point P” is represented by a hollow square, and the corresponding continuous profile is represented by a dashed dotted line. Point P”' is represented by a hollow circle, and the corresponding continuous profile is represented by a dotted line.

[0092] As per the above reference Figure 2 As already described, a measurement dataset including multiple measurement points P of the internal feature 10 is provided, specifically, measured using a coordinate measuring device. Specifically, the measurement points P are the coordinates of the internal feature 10. An auxiliary feature 30 is provided representing an ideal estimate of the internal feature 10 of the workpiece 100. Then, in Figure 3a and Figure 3b In the example shown, each measurement point P of the measurement dataset is mirrored on the boundary elements of the auxiliary feature 30 of the circle. This first mirroring step generates a first modified dataset including multiple first modified measurement points P'. Based on these first modified measurement points P', a convex hull H is determined or calculated using a convex hull algorithm, and the first modified measurement points P' are projected onto the determined convex hull H, thereby generating a second modified dataset including multiple second modified measurement points P''. Therefore, each second modified measurement point P'' lies on the determined convex hull. Subsequently, each second modified measurement point P'' is mirrored on the boundary elements of the auxiliary feature 30. This second mirroring step generates a filtered measurement dataset including multiple filtered measurement points P'''. As from... Figure 3a and Figure 3b As can be seen, the filtered measurement dataset with the filtered measurement point P”’ is a smoothed dataset based on the measurement dataset with the measurement point P.

[0093] Figure 3a and 3b Both examples shown illustrate the performance of inward feature filters, where the following properties are satisfied with the filter relative to the inward features:

[0094] - Eliminate outliers outside the features, as they should have a smaller impact on baseline correlation processing;

[0095] - Retain the innermost points, as these are the relevant points used to calculate the baseline correlation; and

[0096] - This filter behaves similarly to a morphological filter, i.e., it smooths out biases, which is required for candidates to be processed as the default filter.

[0097] Notice, Figure 3a and 3b The sketch is for illustrative purposes only and does not show the exact location of the points actually calculated.

[0098] Figure 4 Show Figure 3a An enlarged view of the schematic diagram is provided to illustrate further details of the method according to the preferred embodiment. Specifically, according to the preferred embodiment, the following steps define a convex hull-based method for inner features:

[0099] - Mirror each point P of the measurement dataset {P} on the auxiliary features to obtain the first modified dataset {P'};

[0100] - For the first modified dataset {P'}, find the convex hull and project the points P' of the first modified dataset {P'} onto the convex hull to generate the second modified dataset {P”}, where the projection can be orthogonal to the convex hull or in the radial direction from the center / axis of the auxiliary feature; and

[0101] - Mirror each point P” of the second modified dataset {P”} on the auxiliary features to obtain the third modified dataset {P”’}, where the points P”’ of the third modified dataset {P”’} are the filtered points.

[0102] The auxiliary feature 30 can be a predetermined ideal feature defined based on the design data of the internal feature 10 and / or the design data of the workpiece 100. Alternatively, the auxiliary feature 30 can be a Gaussian feature determined based on the measurement point P.

[0103] For each measurement point P, a corresponding first reflection point R1 is defined on the boundary element of the auxiliary feature 30. The corresponding first reflection point R1 is located on the boundary element of the auxiliary feature 30. Furthermore, the corresponding first reflection point R1 is defined such that the virtual line between the measurement point P and the corresponding first reflection point R1 is perpendicular to the tangent or tangent plane T of the auxiliary feature 30 at the corresponding first reflection point R1. Moreover, for each measurement point P, a corresponding first distance d1 is defined between the measurement point P and the corresponding first reflection point R1.

[0104] A first modified dataset is generated by determining mirror measurement points for each measurement point P, wherein the mirror measurement points are obtained by moving the measurement point P through the corresponding first reflection point R1 by twice the corresponding first distance d1.

[0105] For each second modified measurement point P", a corresponding second reflection point R2 is defined on the boundary element of the auxiliary feature 30. Similar to the first reflection point R1, the corresponding second reflection point R2 is also located on the boundary element of the auxiliary feature 30. Furthermore, the corresponding second reflection point R2 is defined such that the virtual line between the second modified measurement point P" and the corresponding second reflection point R2 is perpendicular to the tangent or tangent plane T of the auxiliary feature 30 at the corresponding second reflection point R2. Moreover, for each second modified measurement point P", a corresponding second distance d2 is determined between the second modified measurement point P" and the corresponding second reflection point R2.

[0106] Then, a filtered measurement dataset is generated by determining the mirror point for each second modified measurement point P”, wherein the mirror point is obtained by moving the second modified measurement point P” through the corresponding second reflection point R2 by twice the corresponding second distance d2.

[0107] exist Figure 4 In this context, P / P" means that the measured point P and the filtered measured point P" overlap. Correspondingly, P' / P" means that the first modified measured point P' and the second modified measured point P" overlap. Furthermore, R1 / R2 means that the first reflection point R1 and the second reflection point R2 overlap. In other words, R1 / R2 means that the second reflection point R2 corresponds to the first reflection point R1. d1 / d2 means that the first distance d1 and the second distance d2 are equal. In other words, d1 / d2 means that the second distance d2 corresponds to the first distance d1. Note that this configuration, i.e., the overlap of points P and P"' and the overlap of points P' and P"', only applies to points based on the innermost measured point.

[0108] Figure 5 and Figure 6 An illustrative result is shown of an implementation of the convex hull-based filtering method for inner features according to the present invention. More specifically, Figure 5 The diagram shows the measured point P (represented by a black dot) and the filtered point P”' (represented by a hollow circle) of the internal feature of a workpiece according to the first example, wherein the contour of the internal feature corresponds to the example disclosed in Figure A.3 of Appendix A of ISO 5459:2011 (see Figure 1c). Figure 6The diagram shows the measurement points P (represented as black dots) and the filtered P'' (represented as hollow circles) of the internal features of a workpiece according to the second example, where the contour of the internal features is a circle with random deviations. Specifically, based on the measurement points P of the circular features, the filter mirrors each measurement point P in the radial direction at a reference circle, applies a convex hull to these points, projects the points onto the convex hull, and finally mirrors them back. Note Figure A.3 in Appendix A of ISO 5459:2011 (see Figure 1c), which shows how the filtered distribution should look, obtained by using the method according to the invention.

Claims

1. A method for filtering a measurement dataset, the measurement dataset being usable for specifying and / or verifying internal features (10) of a workpiece (100), the method comprising the steps of: Provide a measurement dataset including multiple measurement points (P) of the internal feature (10); Auxiliary features (30) are provided to represent an ideal estimate of the internal features (10) of the workpiece (100); Each measurement point (P) of the measurement dataset is mirrored on the boundary element of the auxiliary feature, thereby generating a first modified dataset including a plurality of first modified measurement points (P'), wherein, relative to the boundary element of the auxiliary feature, each measurement point (P) is on the opposite side and equidistant from the first modified measurement point (P') generated by mirroring the measurement point (P) on the boundary element. The convex hull (H) of the first modified measurement point (P') is determined, and the first modified measurement point (P') is projected onto the determined convex hull (H), thereby generating a second modified dataset including multiple second modified measurement points (P''); and Each second modified measurement point (P'') is mirrored on the boundary element of the auxiliary feature (30), thereby generating a filtered measurement dataset including multiple filtered measurement points (P'''), wherein, relative to the boundary element of the auxiliary feature, each second modified measurement point (P'') is on the opposite side and equidistant from the filtered measurement point (P''') generated by mirroring the second modified measurement point (P'') on the boundary element.

2. The method according to claim 1, wherein, Providing measurement datasets includes capturing measurement data using coordinate measuring devices.

3. The method according to claim 1 or 2, in, The auxiliary feature (30) is a predetermined ideal feature defined based on the design data of the internal feature (10) and / or the design data of the workpiece (100); or The auxiliary feature (30) is a Gaussian feature determined based on the measurement dataset.

4. The method according to claim 1 or 2, wherein, The auxiliary feature (30) is a circle, a sphere, a cylinder or a cone.

5. The method according to claim 1 or 2, wherein, Mirroring each measurement point (P) of the measurement dataset on the boundary elements of the auxiliary feature (30) includes: For each measurement point (P), the corresponding first reflection point (R1) is defined on the boundary element of the auxiliary feature (30); For each measurement point (P), determine the corresponding first distance (d1) between the measurement point (P) and the corresponding first reflection point (R1); and The first modified dataset is generated by determining mirror measurement points for each measurement point (P), wherein the mirror measurement points are obtained by moving the measurement point (P) twice the determined corresponding first distance (d1) through the corresponding first reflection point (R1).

6. The method according to claim 5, wherein, For each measurement point (P), the corresponding first reflection point is defined such that: The corresponding first reflection point (R1) is located on the boundary element of the auxiliary feature (30), and The virtual line between the measurement point (P) and the corresponding first reflection point (R1) is perpendicular to the tangent or plane (T) of the auxiliary feature (30) at the corresponding first reflection point (R1).

7. The method according to claim 1 or 2, wherein, The first modified measurement point (P') of the first modified dataset is projected onto the determined convex hull (H) in the following manner: Orthogonal to the convex hull (H), and / or In the radial direction relative to the center or axis of the auxiliary feature (30).

8. The method according to claim 1 or 2, wherein, Mirroring each of the second modified measurement points (P'') of the second modified dataset on the boundary elements of the auxiliary feature (30) includes: For each second modified measurement point (P''), a corresponding second reflection point (R2) is defined on the boundary element of the auxiliary feature (30); For each second modified measurement point (P''), determine the corresponding second distance (d2) between the second modified measurement point (P'') and the corresponding second reflection point (R2); and The filtered measurement dataset is generated by determining a mirror point for each second modified measurement point (P''), wherein the mirror point is obtained by moving the second modified measurement point (P'') by twice the determined corresponding second distance (d2) through the corresponding second reflection point (R2).

9. The method according to claim 8, wherein, The corresponding second reflection point (R2) corresponds to the corresponding first reflection point (R1) according to claim 6.

10. The method according to claim 8, wherein, For the second modified measurement point (P'') of the second modified dataset, define the corresponding second reflection point (R2) such that: The corresponding second reflection point (R2) is located on the boundary element of the auxiliary feature (30), and The virtual line between the second modified measurement point (P'') and the corresponding second reflection point (R2) is perpendicular to the tangent or plane (T) of the auxiliary feature (30) at the corresponding second reflection point (R2).

11. The method according to claim 1 or 2, further comprising the step of: The internal features (10) of the workpiece (100) are specified and / or verified based on the filtered measurement dataset.

12. The method according to claim 11, wherein, Specifying and / or verifying the internal features (10) of the workpiece (100) includes determining least-squares geometric elements based on the filtered measurement dataset.

13. A computer program product comprising computer-readable instructions, which, when loaded into the memory of a computer and executed by the computer, cause the computer to perform the method according to any one of claims 1 to 12.

Images