Method for providing identification information of a component

Abstract

The invention relates to a method (100) for providing identification information of a component (20), comprising the steps of detecting, by means of a reading device (11), a mark region (21) of the component (20), which mark region comprises a dot-peened mark (22) bearing identification information, and providing the identification information contained in the dot-peened mark (22) for processing.

Description

[0001] Method for providing identification information of a component

[0002] The invention relates to a method for providing identification information for a component, comprising the steps of detecting a marking area having a pin mark with identification information using a reading device, evaluating image data of the marking area, and providing the identification information of the pin mark for further processing. Furthermore, the invention relates to a system for providing identification information for a component.

[0003] Components or parts of machines, such as an engine, must often be clearly identified for the purposes of traceability within a production process and / or during maintenance, and this clear identification must be documented. For this purpose, components often have a unique serial number that is applied to the corresponding component in the form of a needle mark. The manufacturing process of components, which particularly involves several different process steps such as heat treatments, abrasive processes, platinum plating, annealing, coating, ceramic coating and / or adverse environmental influences during operation such as erosion, temperature fluctuations or contamination, influences and changes the readability of a needle mark. In addition, the appearance of the needle mark varies depending on the viewing direction, lighting and ambient situation and / or reflections on metallic surfaces.

[0004] Since this can result in poor legibility, for example, for serial numbers in the form of a pin mark, identification information in the form of a pin mark is typically read by personnel or with the naked eye. The serial number read in this way must then be manually transferred into digital documentation and / or processing systems. This process is often error-prone, time-consuming, and ergonomically unfriendly.

[0005] Based on this, it is an object of the present invention to improve a method for providing identification information of a component in the form of a pin mark. In particular, the reading and / or identification of such identification information is to be improved in order to avoid incorrect provision of the information. This is achieved according to the invention by the teaching of the independent claims. Advantageous embodiments of the invention are the subject of the dependent claims.

[0006] To achieve this objective, a method for providing identification information of a component is proposed, comprising the steps of detecting a marking region having a needle mark with identification information by means of a reading device, evaluating image data of the marking region detected by the reading device, determining the identification information of the needle mark, the evaluation being carried out by an evaluation device trained with image data of needle marks, and providing the identification information of the needle mark for further processing. Identification information comprises, in particular, details that enable identification of an object or component, in particular by means of unique features and / or characteristics. A combination of alphanumeric characters, for example, can be used as identification information.

[0007] The evaluation device can be a computer, in particular a personal or industrial computer, or a mobile device, such as a smartphone, with an application installed thereon, which includes a model trained with the training data.

[0008] The trained or still untrained model can include or be a deep neural network (dNN), for example with the well-known YOLO architecture, for object detection.

[0009] The model is adapted to the trained model by training it with training data for the task in question, namely the recognition of pin marks on components. This means that, for example, for the task of recognizing an alphanumeric character string, the information is recognized exclusively as an alphanumeric character string, and no other characters or symbols can be recognized as included. In other words, the training allows the structure of the identification information, e.g., alphanumeric structure, to be set as a recognition condition for determination. In particular, the well-known "Y0L0-V9" architecture can be used for the above tasks.

[0010] Images of different needle markings, in particular in alphanumeric form, on components can be used as training data, for example in a number of at least 1000, 3000 or 5000, for example between 1000 and 10000, in particular between 6000 and 9000. In one embodiment, 8000 images are used.

[0011] The training images may exhibit different variations, i.e., different real-world appearances. The different appearances of the pin markers in the component images may occur at different production stages, particularly due to specific manufacturing steps such as heat treatment, coating, machining, etc.

[0012] Similar variations, e.g., from the same production stage, can be assigned to a common group. In one version, 8,000 images are used.

[0013] In one implementation, at least 800 images per group are used for training and / or there are between 10 and 20 different groups.

[0014] Machine learning can be performed on and / or with the help of a computer or mobile device. This can be done on-premises and / or in the cloud.

[0015] The proposed method enables digital or machine-assisted reading and provision of unique identification information, which can, for example, improve the documentation process for manufacturing processes and / or maintenance work.

[0016] According to a further aspect, a system for providing identification information of a component is proposed, comprising a reading device configured to detect a marking region of the component that has a needle mark with identification information, and a processing device configured to evaluate the image data of the marking region detected by the reading device and to determine the identification information of the needle mark in order to provide this identification information of the needle mark for further processing. The reading device and the processing device are, in particular, adapted to carry out or can carry out the steps of the method described herein.

[0017] A needle marking is specifically designed for durability or persistence and is, in particular, a marking permanently applied to the component by means of an oscillating hard metal needle. The needle marking comprises, in particular, alphanumeric characters for identifying a component, which are represented by individual dots adjacent to one another and / or continuously nested within one another, depending on the dot density. The method described herein is specifically designed to make needle markings on a component readable without technical post-processing or altering the surface to improve readability. For this purpose, the provision of suitable lighting and / or a suitable evaluation algorithm can be used, in particular.

[0018] The reading device is particularly configured, for example by means of optoelectronic means, to detect the marking area and to create a representation of the marking area in the form of computer-readable data or image data, for example as a raster or pixel graphic. In some embodiments, the reading device can have a detection device, for example in the form of a camera, to detect the marking area. The marking area is in particular a surface area of ​​the component comprising the needle marking to be detected, which is detected or recorded by the reading device. In this case, a recording can be recorded or recorded, for example, by means of the reading device having a camera device or a digital camera and stored in the form of data, for example as a pixel grid, in particular in a memory device of the reading device, or can be directly processed or evaluated.For this purpose, for example, the system and / or the reading device has the processing device which is configured to evaluate the image data of the marking area; in other embodiments, the evaluation step can be performed by a computing unit provided outside the reading device. When evaluating the captured image data of the marking area, for example, an observation or comparison of pixel properties of the raster or pixel graphic, such as a respective color, shading and / or contrast, can be used to recognize and / or analyze the needle marking. In this process, the identification information contained in the needle marking is determined and can be converted into a data set or identified as such. This identification information of the needle marking obtained in this way can then be filed and / or retained for further processing, storage and / or use.In some embodiments, the identification information can also be transmitted to another data processing device.

[0019] The invention is based, among other things, on the idea of ​​enabling machine-based reading of an individual identification based on a pin marking, particularly in an alphanumeric format, of a component, while also enabling digital provision of the read data. To this end, it is proposed to record an image using a reading device, for example, a camera, in order to improve the detection rate of the pin marking provided on the component. The reading device itself and / or a further data processing device can be configured to extract the identification information of the pin marking based on the image data thus acquired and to make this information available in digital form.

[0020] In one embodiment, the evaluation is carried out on the basis of known data, images, and / or image data of needle markings. Images of comparable needle markings, in particular those needle markings based on the same manufacturing principle as the needle marking to be detected or detected, can be stored and / or provided to serve for comparison, in particular for or during character recognition. This allows, for example, already known lighting conditions and / or appearances of the needle markings, which may be present after certain manufacturing processes, to be used to provide an improved comparison basis for the identification information to be evaluated.

[0021] In one embodiment, the method comprises the further step of performing machine learning for evaluating the image data based on the captured image data, in particular containing a mapping of an alphanumeric format. In this case, models, in particular in the form of so-called neural networks, can be created from the recorded or captured image data, for example, which can be used to recognize a characteristic pattern, in particular in alphanumeric form, in character recognition. The patterns and / or models held for comparison can thus be updated and / or supplemented by adding additional image data in order to improve the evaluation and thus the provision of the identification information.

[0022] Training or machine learning can be achieved using a deep neural network (dNN), for example, with the well-known YOLO architecture, for object detection, which can then form the trained model. In particular, the well-known "Y0L0-V9" algorithm can be used.

[0023] Images of different needle markings, in particular in alphanumeric form, on components can be used as training data, for example in a number of at least 1000, 3000 or 5000, for example between 1000 and 10000, in particular between 6000 and 9000. In one embodiment, 8000 images are used.

[0024] The training images may exhibit different variations, i.e., different real-world appearances. The different appearances of the pin markers in the component images may occur at different production stages, particularly due to specific manufacturing steps such as heat treatment, coating, machining, etc.

[0025] Similar variations, e.g., from the same production stage, can be assigned to a common group. In one implementation, 8,000 images are used. In one implementation, at least 800 images per group are used for training and / or there are between 10 and 20 different groups.

[0026] Machine learning can be performed on and / or with the help of a computer or mobile device. This can be done on-premises and / or in the cloud.

[0027] In one embodiment, the evaluation is carried out using optical character recognition. For this purpose, the image data can be divided, for example, into relevant areas and irrelevant areas. Furthermore, corrections can be made based on algorithms, particularly by considering the proximity relationships between adjacent pixels. The relevant pixel patterns can be compared with known image data, models, and / or patterns in particular to recognize characters and / or generate a data set containing the identification information contained in the pin mark. The data set can, for example, be a text file or have another, particularly specific, output format.

[0028] In one embodiment, the method comprises the further step of illuminating the marking area during the detection of the marking area. For this purpose, in one embodiment, the reading device can comprise an illumination device configured to illuminate the marking area of ​​the component. The illumination device can be configured to enable homogeneous and / or contrast-enhancing illumination of the marking area. This allows comparable illumination situations to be created for recording the image data, enabling improved recognizability of the needle marking for evaluation or analysis.

[0029] In one embodiment, a plausibility check of the identification information, in particular a sequence of alphanumeric characters, is carried out in a further step. In particular, based on previously stored identification information, a format and / or a sequence of characters can be checked for the probability of their correct identification in order to indicate manual verification if necessary. This can reduce the probability of false documentation. In one embodiment, the marking area is illuminated in a bright-field arrangement for capturing the identification information and / or for creating training images. This is particularly advantageous for needle markings, which change the surface texture and influence light reflection.The automated and - in the case of a plausibility check by a human - also the manual (human-visual) recognition rate can be advantageously improved.

[0030] In one embodiment, to check the plausibility of the identification information, an image of the detected marking area is graphically displayed, in particular with digital magnification, alongside a visual representation of the specific identification information, in particular on a display device, monitor, or display. This is particularly advantageous for small markings that are difficult to detect with the naked eye and / or for fast, simple, and reliable human verification of correctness.

[0031] In one embodiment, the reading device is a mobile phone or device with a camera and / or a camera, wherein one or the camera is used to capture the marking area. The term mobile phone is intended to encompass all suitable mobile radio-capable devices, such as smartphones or tablets with an image capture device. In this case, the data processing can take place directly within the mobile phone or device, or can be transmitted at least partially to at least one remote location or external computing unit, such as a fog server, a cloud-based processing device, and / or another computing instance. In such an embodiment, the lighting device can be, for example, a lighting device that can be charged, in particular via a USB port, and which can be arranged or attached to the mobile phone or a mobile phone case by means of a holder.

[0032] In one embodiment of the system for providing identification information of a component, the reading device comprises an illumination device configured to illuminate the marking area of ​​the component. This improves the detection of the marking area by the reading device, in particular the quality of the captured image data. The illumination device can comprise, for example, bright-field illumination and / or a ring light. Such illumination can improve the readability of the needle marking, and interference effects based, for example, on a detection direction, a room lighting situation, and / or the ambient situation can thus be reduced.

[0033] In one embodiment, the system according to the invention is used to provide identification information of a component during the manufacturing, maintenance and / or repair process, for example during manufacture of the component or when the component is used as a spare or replacement part in the context of a maintenance or repair process.

[0034] During the manufacture of a component, it may be necessary to provide identification information for the component during one or more manufacturing steps. During these steps, the component may be subjected to processes that alter the appearance of its external surface as well as the identification marking, particularly needle marking. These processes may include, for example, heat treatments, abrasive processes, platinizing, annealing, coating, and ceramic coatings. Particularly for components such as turbine parts, there may be a need to identify the component not only during the production process but also throughout the entire product life cycle, including maintenance and repair.

[0035] During the lifetime of a component, its surface may undergo changes due to environmental influences such as erosion, temperature fluctuations, or contamination. These changes can also lead to changes in the appearance of the component identification information, thus complicating its correct evaluation.

[0036] The system according to the invention enables efficient provision of component identification information and can reduce erroneous technical documentation. Implementing this system improves the accuracy and reliability of identification throughout the entire life cycle of the component, leading to improved traceability and quality assurance. Further features, advantages, and possible applications of the invention will become apparent from the following description in conjunction with the figures. In general, features of the various exemplary aspects and / or embodiments described herein, in particular of the device and the method, can be combined with one another, unless this is clearly excluded in the context of the disclosure.

[0037] In the following part of the description, reference is made to the figures shown to illustrate specific aspects and embodiments of the present invention. It is understood that other aspects may be utilized and structural or logical changes to the illustrated embodiments are possible without departing from the scope of the present invention. The following description of the figures is therefore not to be understood as limiting. It shows

[0038] Fig. 1 is a schematic representation of an exemplary system according to the invention for providing identification information of a component;

[0039] Fig. 2 is a schematic representation of another exemplary system according to the invention for providing identification information of a component; and

[0040] Fig. 3 is a flowchart of an exemplary method according to the invention for providing identification information of a component.

[0041] Fig. 1 shows an exemplary representation of an embodiment of a system 10 according to the invention for providing identification information of a component 20 in a schematic representation.

[0042] The system 10 comprises a reading device 11, in particular a portable one, which is configured to detect a marking area 21 of the component 20, which has a needle mark 22 with identification information. The reading device 11 itself can be configured to detect the marking area 21 and / or can have a detection device 111 or camera configured for this purpose to record the marking area 21 and create image data. This enables flexible, digital detection of the needle mark 22. Furthermore, needle marks 22 at different component positions and / or different components 20 can be detected spatially independent, in particular without the component 20 having to be moved. In the illustrated embodiment, the reading device 11 is designed as a mobile phone and uses its camera 111 to detect the marking area 21.

[0043] The system 10 further comprises a processing device 12 configured to evaluate the image data of the marking area 21 acquired by the reading device 11, to determine the identification information of the needle marking 22, and to provide this identification information of the needle marking 22 for further processing. In the illustrated embodiment, the processing device 12 is integrated into the reading device 11, but in other embodiments, it may also be an externally provided processing device 12, for example, a cloud-based processing device 12a indicated here.

[0044] An illumination device 13 is arranged on the reading device 11 by means of a holder 113, which is configured to illuminate the marking area 21 of the component 20. The illumination device 13 can, for example, comprise bright-field illumination and / or a ring light. By illuminating or exposing the marking area 21, the legibility of the needle marking 22 is improved and interference is reduced.

[0045] Fig. 2 shows a perspective view of a further embodiment of a system 10 according to the invention for providing identification information of a component 20.

[0046] The system 10 has a reading device 11, which in the illustrated embodiment is embodied as a smartphone. An integrated camera 111 of the smartphone or mobile phone 11 (not visible here due to the perspective view) can be used to capture a marking area 21 of the component 20, which has a pin mark 22 with identification information. An illumination device 13 is arranged on the smartphone 11 by means of a holder 113, which is configured to illuminate the marking area 21 of the component 20. Furthermore, the smartphone can be used as a processing device 12 or connected externally to such a device in order to evaluate the image data of the marking area 21 captured by the reading device 11.The identification information of the needle marking 22 is determined, the identification information is provided for further processing and / or displayed on a display device 14, here a display of the smartphone 11.

[0047] Fig. 3 shows an exemplary flow diagram of a method 100 according to the invention for providing identification information of a component 20 in a schematic representation.

[0048] In a first step a, a marking area 21, which has a needle mark 22 with identification information, is detected by a reading device 11. Image data can be created, for example, using a camera device 111 of the reading device 11. In an optional step a1, the marking area 22 can be illuminated by an illumination device 13 during the detection of the marking area 22.

[0049] In a further step b, image data acquired by the reading device 11 is evaluated, wherein, in particular on the basis of known image data of needle markings 22, the identification information of the needle marking 22 is determined. For this purpose, for example, a processing device 12 can be used, wherein the processing device 12 can, for example, use an artificial neural network for data processing and, in particular, can perform optical character recognition. In a further step b1, machine learning can be carried out on the basis of the acquired image data. In a further optional step b2, the plausibility of the identification information can be checked in order to increase the probability of correct evaluation and / or data provision.

[0050] In a further step c, the thus extracted identification information of the needle marking 22 is provided for further processing, particularly in a predetermined data format. This identification information can be used, for example, for documentation purposes during product development and / or maintenance work.

[0051] LIST OF REFERENCE SYMBOLS

[0052] 10 System for providing identification information

[0053] 11 Reading device 12 Processing device

[0054] 13 Lighting equipment

[0055] 14 Display device

[0056] 20 components

[0057] 21 Marking area 22 Marking

[0058] 100 procedures

[0059] 111 Detection device / camera

[0060] 113 bracket ac steps

Claims

CLAIMS 1. Method (100) for providing identification information of a component (20), comprising the steps of: a) detecting a marking area (21) of the component (20) which has a needle marking (22) with identification information, by means of a reading device (11), b) evaluating image data of the marking area (21) detected by the reading device (11), wherein the identification information of the needle marking is determined, wherein the evaluation is carried out with an evaluation device which is trained with image data of needle markings, and c) providing the identification information of the needle marking (22) for further processing.

2. The method (100) according to claim 1, wherein the evaluation is carried out on the basis of known image data of needle markings (22).

3. Method (100) according to at least one of the preceding claims, comprising the further step: bl) performing machine learning or training the evaluation of the image data on the basis of the acquired image data.

4. Method (100) according to at least one of the preceding claims, wherein the evaluation is carried out using optical character recognition.

5. Method (100) according to at least one of the preceding claims, comprising the further step: a1) illuminating the marking area (21) during the detection of the marking area (21).

6. Method (100) according to at least one of the preceding claims, comprising the further step: b2) checking a plausibility of the identification information.

7. The method (100) according to at least one of the preceding claims, wherein the marking area is illuminated in a bright field arrangement for detecting the identification information and / or for creating training images.

8. The method (100) according to at least one of the preceding claims, wherein, for checking a plausibility of the identification information, an image of the detected marking area is graphically displayed alongside a visual representation of the specific identification information, in particular with a digital magnification.

9. System (10) for providing identification information of a component (20), comprising a reading device (11) which is configured to detect a marking area (21) of the component (20) which has a needle marking (22) with identification information, and a processing device (12) which is configured to evaluate the image data of the marking area (21) detected by the reading device (11) and to determine the identification information of the needle marking (22) and to provide this identification information of the needle marking (22) for further processing.

10. System (10) according to claim 9, wherein the reading device (11) comprises an illumination device (13) which is configured to illuminate the marking area (21) of the component (20).

11. System (10) according to at least one of claims 9 or 10, wherein the reading device (11) and the processing device (12) are adapted to carry out the steps of the method according to at least one of claims 1 to 6.

12. System (10) according to at least one of claims 9 to 11, wherein the reading device (11) is a mobile phone or device with a camera and / or a camera (111) and / or one or the camera (111) is used to detect the marking area (22).

13. Uses of a system according to one of claims 9 to 12 for providing identification information of a component during a manufacturing, maintenance and / or repair process.

Images