Embodiments of the present invention relate to a method and a device for determining an optimized parameter set.
Complex measuring devices such as optical coherence tomography (OCT) measuring devices have numerous (setting) parameters which allow the user to adapt the measuring device to the respective measuring situation or processing situation of a measuring element. The multiplicity of parameters and their interactions result in a highly complex parameter space. The adjustment of the parameters of a parameter set therefore currently requires expert knowledge and is time-consuming.
Embodiments of the present invention provide a method for determining an optimized parameter set having a plurality of measurement parameters to carry out a measurement. The method includes: C) carrying out and storing n measurements of a measuring element, n being an integer greater than one. Each measurement has one parameter set. Each measurement has a multiplicity of measuring points. The method further includes: D) evaluating the n measurements with an evaluation function and storing the evaluation, E) generating new parameter sets from the parameter sets used in step C), F) carrying out steps C) to E) multiple times, and J) outputting at least one parameter set that is evaluated as good.
Subject matter of the present disclosure will be described in even greater detail below based on the exemplary figures. All features described and/or illustrated herein can be used alone or combined in different combinations. The features and advantages of various embodiments will become apparent by reading the following detailed description with reference to the attached drawings, which illustrate the following:
Embodiments of the present invention provide a method and a device for automated determination of an optimized parameter set.
According to embodiments of the present invention, a method with the following, in particular automatically carried out, method steps:
The measurement parameters are necessary in order to be able to adapt the measurement to the multiplicity of processing situations. The complex and asymmetric parameter space with the multiplicity of parameters and interactions can be completely hidden from the user. No specific knowledge is therefore required for the operation. However, the functional scope and robustness with regard to the change in measuring situations, in component characteristics, are completely retained. Consequently, the method according to embodiments of the invention allows even inexperienced users to carry out a measurement with very good measurement parameters.
A parameter set corresponds to a number of measurement parameters with which a measurement is possible. Two parameter sets differ if at least one parameter of a parameter set is different from the same parameter in the other parameter set.
In method step E), the new parameter sets are generated, in particular, by applying evolutionary operators, preferably in the form of crossover operators and/or mutation operators, to the parameter sets used in method step C). If a crossover operator is applied, two parent parameter sets are combined to form one next-generation parameter set. If a mutation operator is applied, a single part/parts of a parent parameter set is/are randomly changed.
The measuring element can be present in the form of a workpiece.
n and/or m can be greater than 1, in particular greater than 2, preferably greater than 5, preferably greater than 10.
The measurement in method step C) is preferably carried out in the form of a contactless scan. The scan can be carried out one-dimensionally (line scan) or multi-dimensionally.
The measurement in method step C) is preferably carried out in the form of an optical coherence tomography (OCT) measurement or in the form of a pyrometry measurement. The parameter set for carrying out OCT measurements and pyrometry measurements is effectively optimizable with the method according to embodiments of the invention.
The evaluation function can comprise an algorithm, in particular in the form of an image processing algorithm, for evaluating the recording quality of the measurement, and/or a deep convolutional neural network. The algorithm can evaluate a raw sensor signal, for example a Fast Fourier Transform (FFT) signal. The image processing algorithm can be designed to evaluate the image quality of the measurement. The image processing algorithm can evaluate, for example, edge sharpness and/or image noise.
The n new parameter sets can be generated in method step E) randomly (E1) or (E2) using artificial intelligence (AI) which adapts its target function by means of an online learning method and the evaluations of the evaluation function. The AI achieves a significantly faster optimization of the parameter set through the continuous (online) performance of the learning method, taking into consideration the evaluations.
In a further preferred embodiment, the following method steps are carried out after method step F) and before method step J):
The merging in method step G) can be carried out by averaging, by determining a median, and/or by determining other statistical values.
The ROI corresponds to the measurement area in which a measurement signal is received from the sampling element. The areas of the measurement in which no signal is received from the sampling element are thereby excluded from the optimization. The optimization is significantly improved as a result.
The ROI is preferably continuous.
o can be greater than 1, in particular greater than 2, preferably greater than 5, preferably greater than 10.
Following method step J), the parameter set output in method step J) can be stored in a method step K). In addition, one or more further parameter sets evaluated as good can be stored.
The following method step can be carried out before method step C):
In addition, the following method step can be carried out before method step B):
Embodiments of the invention further provide a device for determining an optimized parameter set with a method described here, wherein the device has a measuring device to carry out the measurements in method step C) and a computer to carry out the further method steps. The computer can be part of the measuring device.
The computer can have software with an algorithm to control the measuring device.
The measuring device is preferably designed in the form of an OCT measuring device or a pyrometry measuring device.
The measurement with the measuring device 12 is carried out with the setting of a plurality of parameters. The measurement parameters are predefined by the computer 14. Before and/or during the measurement, the measuring device 12 communicates the measurement result (“the measurement”) with a multiplicity of measuring points to the computer 14.
The measuring device 12 is designed in the form of an optical coherence tomography (OCT) measuring device. The measuring device 12 has an OCT scanner 24 for the measurement of a measuring element 22. In addition, a laser processing optical element 28 can be provided for the processing of the measuring element 22. An OCT measuring beam 30 is injected into the measuring device 12. A processing laser beam 32 can be injected in addition to this. The measuring device 12 can have deflection mirrors and/or beam splitters as an alternative or in addition to the devices shown.
The measurement parameters (“parameter set”) used in the measurement are optimized with the method according to embodiments of the invention. This is explained in
in method step A), value ranges are defined for measurement parameters of the parameter sets, wherein the parameters of the parameter sets are generated in method steps B) and E) within these value ranges.
In method step B), n parameter sets are generated within the previously defined value ranges. This is performed B1) randomly; or B2) by means of a default initial parameterization; or B3) through measurement of the measuring element 22 with a default initial parameterization and determination of one or more nearest neighbours of the default initial parameterization.
In method step C), n measurements of the measuring element 22 are carried out and stored in each case with one parameter set, wherein each measurement has a multiplicity of measuring points.
In method step D, the n measurements from method step C) are evaluated with an evaluation function. The evaluations are stored.
In method step E), a loop is executed. In method step E), n new parameter sets are generated. The generation is carried out by applying crossover operators and/or mutation operators to the parameter sets used in method step C). The generation is carried out E1) randomly; or E2) using artificial intelligence which adapts its target function by means of online learning methods and the previously produced evaluations.
The loop of method steps C), D) and E) is repeated m times according to method step F).
After m repetitions, all measurements carried out in method step C) are averaged in method step G).
The smallest possible region of interest (ROI) in the averaged measurement is then determined in method step H). In this ROI, the evaluation function exceeds a defined threshold value.
In method step I), and further loop is executed: method steps C), D), E) and F) are carried out o times, wherein method step D) is modified in such a way that the evaluation function is applied within the ROI only.
In method step J), after the o repetitions, the optimization of the parameter sets is ended and a parameter set evaluated as good is output.
In method step K), at least this parameter set evaluated as good is stored. This parameter set can be used in a subsequent method in method step B).
To provide a synopsis of both figures of the drawing, embodiments of the invention relate to a method and a device 10 for determining an optimized parameter set for a measurement with a measuring device 12. For this purpose, n m measurements in particular are carried out and evaluated, are preferably averaged, and a region of interest (ROI) is determined in the averaged measurement. Subsequently, n*m measurements can be carried out o times, wherein the evaluation is performed in the ROI only. After o repetitions at the latest, the optimization can be ended and a evaluated parameter set evaluated as good can be output and used for a measurement.
While subject matter of the present disclosure has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive. Any statement made herein characterizing the invention is also to be considered illustrative or exemplary and not restrictive as the invention is defined by the claims. It will be understood that changes and modifications may be made, by those of ordinary skill in the art, within the scope of the following claims, which may include any combination of features from different embodiments described above.
The terms used in the claims should be construed to have the broadest reasonable interpretation consistent with the foregoing description. For example, the use of the article “a” or “the” in introducing an element should not be interpreted as being exclusive of a plurality of elements. Likewise, the recitation of “or” should be interpreted as being inclusive, such that the recitation of “A or B” is not exclusive of “A and B,” unless it is clear from the context or the foregoing description that only one of A and B is intended. Further, the recitation of “at least one of A, B and C” should be interpreted as one or more of a group of elements consisting of A, B and C, and should not be interpreted as requiring at least one of each of the listed elements A, B and C, regardless of whether A, B and C are related as categories or otherwise. Moreover, the recitation of “A, B and/or C” or “at least one of A, B or C” should be interpreted as including any singular entity from the listed elements, e.g., A, any subset from the listed elements, e.g., A and B, or the entire list of elements A, B and C.
Number | Date | Country | Kind |
---|---|---|---|
10 2021 201 806.8 | Feb 2021 | DE | national |
This application is a continuation of International Application No. PCT/EP2022/051444 (WO 2022/179776 A1), filed on Jan. 24, 2022, and claims benefit to German Patent Application No. DE 10 2021 201 806.8, filed on Feb. 25, 2021. The aforementioned applications are hereby incorporated by reference herein.
Number | Date | Country | |
---|---|---|---|
Parent | PCT/EP2022/051444 | Jan 2022 | US |
Child | 18448203 | US |