Fixture markings in the testing system

By automatically identifying fixtures in the testing system using image recognition and machine learning technologies, the problem of difficulty in identifying fixture types in existing technologies is solved, thereby improving the reliability and efficiency of the testing system.

CN114930147BActive Publication Date: 2026-11-17TA INSTRUMENTS WATERS LLC
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202180010244.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2020-01-22
Filing Date
2021-01-15
Publication Date
2026-11-17
Estimated Expiration
2041-01-15

AI Technical Summary

Technical Problem

In testing systems, existing technologies struggle to automatically identify and confirm the type of fixture used, leading to testing errors and equipment damage. Furthermore, existing marking technologies increase fixture complexity and cost.

Method used

Image recognition technology is used to capture fixture image data through imaging equipment, and machine learning models and neural networks are used to identify the fixture type on the processor, thus avoiding manual inspection.

Benefits of technology

It enables automatic identification of fixture types, reduces test errors and equipment damage, simplifies fixture management, and reduces hardware complexity and cost.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114930147B_ABST
    Figure CN114930147B_ABST
Patent Text Reader

Abstract

Systems and methods for fixture identification in a test system are described. One method for fixture identification in a test system may include capturing image data representing a first fixture of the test system using an imaging device. The method may further include transferring the image data representing the first fixture from the imaging device to a processor running an image recognition application. The method may also include identifying the first fixture based on the image data using the processor running the image recognition application.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] Related applications This application claims the benefit of the earlier filing date of U.S. Provisional Patent Application Serial No. 62 / 964,277, entitled "Fixture Identification in Test Systems," filed January 22, 2020, the entire contents of which are incorporated herein by reference. Technical Field

[0002] This invention generally relates to fixture marking. More specifically, this invention relates to systems and methods for marking fixtures in a testing system. Background Technology

[0003] Test systems, including mechanical testing equipment, can use actuators to apply loads or forces to samples or materials, or otherwise agitate them. For example, some testing applications involve using actuators to push or pull samples and measuring the resulting deflection and load. To use the measurement results to determine various properties of the sample (e.g., stress and strain), it may be necessary to know what type of tool or fixture was used to agitate the sample. Summary of the Invention

[0004] In this disclosure, a method for identifying a fixture in a test system is described. The method may include capturing image data representing a first fixture of the test system using an imaging device. The method may further include transferring the image data representing the first fixture from the imaging device to a processor running an image recognition application. The method may also include identifying the first fixture based on the image data using the processor running the image recognition application.

[0005] The method may include one or more of the following features. The method may include determining whether the first fixture corresponds to a second fixture of the test system, the second fixture being a anticipated fixture associated with test information of the test system. The method may also include performing a next operation of the test system in response to determining that the first fixture corresponds to the second fixture of the test system. The method may further include providing an indication that the first fixture does not correspond to the second fixture of the test system in response to determining that the first fixture does not correspond to the second fixture of the test system. The imaging device may be a camera.

[0006] In the implementation, the next operation may include at least one of the following: loading a test file associated with the test system and setting test parameters associated with the test system. The instruction may be included in a notification on the graphical user interface of the test system. The image recognition application may include at least one of a machine learning model and a neural network. The test system may be selected from the group consisting of: dynamic mechanical analyzers; fatigue testing systems, tensile testing systems, test systems incorporating load frames, thermal analyzers; rheometers; tissue engineering systems, and heart valve replacement testing systems. At least one of the first and second clamps may be a clamp type selected from the group consisting of: shear clamps, bending clamps, compression clamps, and tensile clamps. The test information may indicate the clamp type corresponding to the test selected for the test system.

[0007] In this disclosure, a fixture identification system for identifying fixtures in a test system is described. The fixture identification system may include an imaging device that can be focused on a test area of ​​the test system. The fixture identification system may also include a processor that communicates with the imaging device. The processor may be configured to run an image recognition application. The processor may also be configured to receive image data representing a first fixture from the imaging device. The processor may also be configured to identify the first fixture based on the image data.

[0008] It may include one or more of the following features. The processor may be configured to determine whether the first fixture corresponds to a second fixture of the test system, the second fixture being a anticipated fixture associated with test information of the test system. The processor may also be configured to determine, upon determining that the first fixture corresponds to the second fixture of the test system, to perform a next operation of the test system. The processor may also be configured to provide an indication that the first fixture does not correspond to the second fixture of the test system in response to determining that the first fixture does not correspond to the second fixture of the test system. The imaging device may be a camera.

[0009] In the implementation, the next operation may include at least one of the following: loading a test file associated with the test system and setting test parameters associated with the test system. The instruction may be included in a notification on the graphical user interface of the test system. The image recognition application may include at least one of a machine learning model and a neural network. The test system may be selected from the group consisting of: dynamic mechanical analyzers; fatigue testing systems, tensile testing systems, test systems incorporating load frames, thermal analyzers; rheometers; tissue engineering systems, and heart valve replacement testing systems. At least one of the first and second clamps may be a clamp type selected from the group consisting of: shear clamps, bending clamps, compression clamps, and tensile clamps. The test information may indicate the clamp type corresponding to the test selected for the test system.

[0010] In this disclosure, a fixture identification system for identifying mounted fixtures in a test system incorporated into a load frame is described. The fixture identification system may include a camera focused on a test area of ​​the test system incorporated into the load frame. The fixture identification system may also include a processor communicating with the camera. The processor may be configured to run an image recognition application including an artificial intelligence component. The artificial intelligence component may include at least one of a machine learning model and a neural network. The artificial intelligence component may be trained to identify the mounted fixture. The processor may also be configured to receive image data representing the mounted fixture from the camera. The processor may also be configured to identify the mounted fixture based on the image data and the artificial intelligence component trained to identify the mounted fixture. The processor may further be configured to determine whether the mounted fixture corresponds to a selected fixture of the test system associated with test information of the test system. Furthermore, the processor may also be configured to provide a notification on the graphical user interface of the test system indicating that the mounted fixture does not correspond to the selected fixture of the test system in response to determining that the mounted fixture does not correspond to the selected fixture of the test system.

[0011] In this disclosure, a method for fixture identification using image recognition is described. The method may include receiving image data representing a first fixture of a test system. The method may also include identifying the first fixture based on the image data using a processor running an image recognition application.

[0012] The method may include one or more of the following features. The method may include determining whether the first fixture corresponds to a second fixture of the test system, the second fixture being a anticipated fixture associated with test information of the test system. The method may also include performing a next operation of the test system in response to determining that the first fixture corresponds to the second fixture of the test system. The method may further include providing an indication that the first fixture does not correspond to the second fixture of the test system in response to determining that the first fixture does not correspond to the second fixture of the test system. The imaging device may be a camera.

[0013] In this implementation, the next operation may include at least one of the following: loading a test file associated with the test system and setting test parameters associated with the test system. The instruction may be included in a notification on the graphical user interface of the test system. The image recognition application may include at least one of a machine learning model and a neural network.

[0014] In this disclosure, a method for a security operation testing system is described. The method may include capturing image data representing a test area of ​​the testing system using an imaging device. The method may further include transmitting the image data representing the test area from the imaging device to a processor running an image recognition application. The method may also include using the processor running the image recognition application to detect the presence of a user in the test area based on the image data.

[0015] The method may include one or more of the following features. The method may include switching the test system to a safe mode in response to detecting the presence of the user in the test area of ​​the test system. The method may also include providing an indication that the user's presence has been detected in the test area of ​​the test system in response to detecting the user's presence in the test area of ​​the test system. This indication may include a notification on the graphical user interface of the test system. The method may further include using the processor running the image recognition application to detect that the user's presence is no longer in the test area based on the image data. Additionally, the method may include exiting the safe mode of the test system in response to detecting that the user's presence is no longer in the test area of ​​the test system. The imaging device may be a camera. Attached Figure Description

[0016] The above and other advantages of this disclosure can be better understood by referring to the following description in conjunction with the accompanying drawings, in which the same reference numerals indicate the same elements and features in each drawing. For clarity, not every element is labeled in every drawing. The drawings are not necessarily drawn to scale, but are intended to illustrate the principles of the invention.

[0017] Figure 1 An example system that can implement aspects of this disclosure is described.

[0018] Figure 2 An example process for facilitating image recognition is shown according to aspects of this disclosure.

[0019] Figure 3 An example test system according to aspects of this disclosure is described.

[0020] Figure 4 An example fixture according to aspects of this disclosure is depicted.

[0021] Figure 5 An example test system according to aspects of this disclosure is also described.

[0022] Figure 6 An example of the classification according to aspects of this disclosure is shown.

[0023] Figure 7 A diagram depicting images of fixtures at various resolutions according to aspects of this disclosure.

[0024] Figure 8 An example workflow according to aspects of this disclosure is shown.

[0025] Figure 9 An example test area of ​​a test system according to aspects of this disclosure is depicted.

[0026] Figure 10 An example graphical user interface according to aspects of this disclosure is depicted.

[0027] Figure 11 Examples of feature detection according to aspects of this disclosure are described.

[0028] Figure 12 An example implementation of an aspect of this disclosure is described. Detailed Implementation

[0029] In this specification, references to "an embodiment" or "an embodiment" mean that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of this teaching. References to a particular embodiment within this specification do not necessarily refer to the same embodiment.

[0030] As mentioned above, some testing applications of a test system may require knowledge of the type of tool or fixture used to agitate the specimen. Therefore, understanding the fixture geometry of the type of fixture used in the test system, or the type of fixture itself, can be beneficial. Using the techniques and features described in this disclosure, image recognition and artificial intelligence (including machine learning or neural networks) can be used to identify fixtures loaded, installed, or otherwise used in the test system.

[0031] Applications of testing systems, including mechanical testing equipment, can involve using actuators to push and / or pull samples or specimens and measuring the resulting deflection. Applications of the testing system can also include measuring the amount of load applied to the sample or specimen. Understanding the type of fixture or tool used to push or pull the sample or specimen can be an essential element in translating measurements of deflection, displacement, and / or load into useful information such as stress and strain in the sample or specimen. Understanding the geometry of the fixture can aid in unit conversions to calculate useful information.

[0032] Now for reference Figure 3 The diagram illustrates a test system 300, which may be a dynamic mechanical analyzer or other test system incorporated into a load frame. The test system 300 may include a frame base 302, an actuator assembly 304, an actuator output shaft 306, a force sensor assembly 308, a load response assembly 310, a lower fixture 312, an upper fixture 314, a sample 316, an environmental chamber 318, and a camera position 320. Reference is also made to... Figure 1 The test system 300 can communicate with client systems, server systems, and / or networks, and can transmit test data and imaging data to client systems (e.g., client electronic devices 28), server systems (e.g., server computers 20), or networks (e.g., network 22) for further analysis by processors or other users. Test data and imaging data can be transmitted together from the test system 300 to client systems, server systems, or networks, or transmitted separately using one or more communication protocols, including but not limited to USB, Ethernet, Wi-Fi (e.g., IEEE 802.11x), Bluetooth, TCP / IP, cellular (e.g., GSM, CDMA, LTE, 5G…), UDP, GigE, IEEE-1394, and MIPI / CSI. In various implementations, MIPI / CSI can be used for direct communication with imaging devices or cameras and processors, and the processor can send image data via Ethernet over TCP / IP.

[0033] It should be noted that the test system 300 is discussed herein for illustrative purposes only, and the techniques and features described in this disclosure can be used with or applied to a variety of test systems in various industries, including but not limited to: dynamic mechanical analyzers, fatigue testing systems, tensile testing systems, test systems incorporated into load frames, thermal analyzers, rheometers, tissue engineering systems, and heart valve replacement testing systems.

[0034] Still referencing Figure 4The test system 300 can use a variety of fixtures that can be swapped in and out for different testing applications. Some fixtures used with the test system 300 may include a shearing fixture 402, a three-point bending fixture 404, a compression fixture 406, a clamping bending fixture 408, and a tensile fixture 410. Depending on the test being performed and the sample being tested, these fixtures, etc., can be swapped in and out of the test system 300. The fixtures may have the same or similar connection geometry (e.g., fixing screws and / or sockets), and changing fixtures may take several minutes.

[0035] For example, test system 300 can use a bending fixture, in which a strip of sample material can be placed on the U-shaped portion of the bending fixture, and the top of the sample material can be pushed in opposite directions by a point. This configuration can be referred to as three-point bending, such as in three-point bending fixture 404. Other tests can use compression fixture 406, in which, for example, a rubber cylinder or a disc-shaped rubber sample can be placed on a disc, and another disc can be applied to the top of the sample in the opposite direction. Some tests can use tension fixture 410, in which, for example, a rod or strip of sample material can be clamped and pulled apart by grips on opposite sides.

[0036] It should be noted that, as described above, the techniques and features described in this disclosure can be implemented using one or more other testing systems, such as tissue engineering systems, heart valve replacement testing systems, and / or heart valve testing instruments, and it should be understood that other types of clamps can be used with those systems. For example, a tissue engineering system may use a clamp or well as a plate on or in which a sample can be placed without applying compression. Similarly, a heart valve testing instrument may use a clamp for mounting rings of a given size and type of heart valve (e.g., a 33 mm mitral valve or a 25 mm aortic valve).

[0037] For example, suppose researchers or other users have produced a batch of polymers or plastics. The researchers may be interested in confirming that the polymers or plastics possess the properties they expect. The researchers may have a set of criteria represented, for example, by displacement based on the reaction force applied to the polymer or plastic (e.g., from a fixture mounted in test system 300). This criterion may also include stress, strain, or modulus.

[0038] In many cases, researchers may begin testing a polymer or plastic in a testing system (e.g., testing system 300), and fixtures may already be installed in the system. Researchers may not necessarily know which fixture is installed. To run tests to determine if the polymer or plastic meets the researcher's criteria, the researcher may need to transform or convert the displacement or load generated on the polymer or plastic into a modulus value (e.g., the relationship between stress and strain). The researcher may need to know the type of fixture used in the testing system to perform the correct transformation and determine the modulus value. Furthermore, using unintended fixtures on samples such as polymers or plastics can lead to adverse results or damage to the testing system, the sample, or both.

[0039] If researchers know that the correct fixtures are already installed in the testing system, they can proceed to set up appropriate tests for the samples of interest. Researchers can open the software associated with the testing system. The software may have multiple settings or parameters that are entered by the user, or these settings may be pre-filled if the same tests are run regularly. Researchers can change some test parameters, such as the size of the polymer or plastic, and run the tests as usual using most of the pre-filled settings.

[0040] In some testing systems, the software running alongside the system may instruct which fixture should be installed for a specific selected test, or the user may have to input the fixture to be installed. For example, the testing system may need to know whether the installed fixture can support tensile, compressive, or bidirectional loads. Attempting to apply a tensile load to a compressive sample may cause the fixture to lose contact with the sample, potentially leading to data errors, sample damage, or damage to the testing system. Furthermore, requiring the user to input the type of fixture to be installed (e.g., from a drop-down menu) is a manual step that may allow for errors or unintended changes.

[0041] Continuing the example above, researchers might not know what type of fixture is installed in the testing system, might assume the required fixture is installed even when it's not, or there might be other errors, such as installing the wrong fixture for the required test and sample. In these cases, researchers might enter the required fixture as installed in the software and run the test with the wrong fixture. For example, a researcher might open a test file designed for a bending fixture but install a compression fixture and continue running the test. This could lead to researchers wasting time using the wrong fixture, obtaining incorrect test results, contaminating or damaging the sample, or damaging the testing system.

[0042] Therefore, it may be desirable for researchers or other users of the testing system to know what types of fixtures are installed in the testing system, without having to manually inspect them. If users can more easily identify the types of fixtures installed in the testing system, various testing errors and damage to test equipment and samples can be prevented, and the burden of manually checking what fixtures are installed in the testing system can be avoided.

[0043] Some testing systems can use marking technologies, such as RFID tags or QR codes, to provide a unique identifier for each fixture, which the system can then read to identify the fixture. These systems introduce new hardware requirements for the testing system and the fixtures themselves (e.g., RFID tags and readers), while increasing the complexity and cost of manufacturing each fixture. They may also make the use of custom fixtures more difficult. Additionally, fixtures can be used over a wide temperature range (e.g., -150°C–600°C or higher), which can further complicate the use of marking / tag technologies.

[0044] Image identification or recognition techniques, such as machine learning, deep learning, and neural networks, can be used to facilitate fixture identification. Images of existing fixtures can be used to identify the type of fixture mounted on a test system via algorithms or algorithmic frameworks for classifying the type of fixture shown in the image. Once image identification or recognition has been performed, more precise details about the fixture (e.g., three-point bending span or stretching sample length) can be extracted using direct pixel measurements. Additionally, image identification or recognition techniques can measure depth information, which can be used to improve the fixture identification process (e.g., by isolating fixtures in the foreground of an image from elements that may be located in the background) or to improve the estimation / measurement of sample geometry from the image.

[0045] Using the techniques and features described in this disclosure, image identification or recognition systems can be used to implement fixture identification in test systems, which can provide several advantages or benefits over existing systems. For example, while the test system may require one or more imaging devices such as a camera, no additional hardware or changes may be needed on the fixtures. Furthermore, test systems with fixture identification capabilities (via image recognition) as described herein can be trained by the user to identify new or different fixtures. Therefore, the techniques and features described herein provide a technical solution to the technical problem (i.e., manual fixture identification or neglect of fixture identification) (i.e., automated fixture identification in test systems) and prevent test errors and damage to test equipment. Additionally, the imaging devices or cameras used for these techniques can also be used to provide other benefits to the user, such as motion capture during testing. In other words, the hardware used to implement fixture identification (via image recognition) in a test system may have other uses within the test system.

[0046] Still referencing Figure 1 and Figure 2 The diagram illustrates server application 10 and client applications 12, 14, 16, and 18. Server application 10 and / or one or more client applications 12, 14, 16, and / or 18 can execute one or more processes configured to perform one or more features described herein. Server application 10 may be referred to as a process configured to perform one or more features described herein, such as image recognition process 10. Furthermore, one or more of client applications 12, 14, 16, and 18 may be referred to as a process configured to perform one or more features described herein, such as image recognition processes 12, 14, 16, and / or 18.

[0047] Now for reference Figure 2 An example image recognition process 12 is shown. Image recognition process 12 may include capturing (202) image data representing a first fixture of the test system using an imaging device. Image recognition process 12 may also include transferring (204) the image data representing the first fixture from the imaging device to a processor running an image recognition application. Image recognition process 12 may also include identifying (206) the first fixture based on the image data using the processor running the image recognition application.

[0048] The image recognition process can be a server-side process (e.g., server-side image recognition process 10), a client-side process (e.g., client-side image recognition process 12, client-side image recognition process 14, client-side image recognition process 16, or client-side image recognition process 18), or a hybrid server-side / client-side process (e.g., a combination of server-side image recognition process 10 and one or more client-side image recognition processes 12, 14, 16, 18).

[0049] System Overview refer to Figure 1The server-side image recognition process 10 may reside on and be executed by a server computer 20, which may communicate with a network 22 (e.g., the Internet or a local area network). Examples of server computer 20 may include, but are not limited to, personal computers, server computers, a series of server computers, microcomputers, and / or mainframe computers. Server computer 20 may be a distributed system, and the operation of server computer 20 may be executed simultaneously and / or continuously on one or more processors. Server computer 20 may be a symbolic representation of a cloud computing site, cloud environment, or cloud platform running multiple servers, computers, or virtual machines. Server computer 20 may execute one or more operating systems, examples of which may include, but are not limited to, Microsoft Windows™; Red Hat Linux™, Unix, or custom operating systems. For example, server computer 20 may be a symbolic representation of a cloud-based image recognition system or server suitable for performing the operations described in this disclosure. In an implementation, server computer 20 may be a cloud-based image recognition system or server that may include, store, run, and / or execute image recognition process 10.

[0050] The instruction set and subroutines of the server-side image recognition process 10 can be stored on a storage device 24 coupled to the server computer 20 and can be executed by one or more processors (not shown) and one or more memory architectures (not shown) incorporated in the server computer 20. The storage device 24 may include, but is not limited to: hard disk drives; tape drives; optical disk drives; solid-state storage devices; RAID arrays; random access memory (RAM); and read-only memory (ROM).

[0051] Server computer 20 can execute web server applications that allow access to server computer 20 (via network 22) using one or more protocols, examples of which may include, but are not limited to, HTTP (i.e., Hypertext Transfer Protocol). Network 22 can communicate with one or more auxiliary networks (e.g., network 26), examples of which may include, but are not limited to, local area networks; wide area networks; or, for example, intranets.

[0052] For example, client-side image recognition processes 12, 14, 16, and 18 may reside on client electronic devices 28, 30, 32, and / or 34, and may be executed by these client electronic devices (respectively). Examples of these client electronic devices may include, but are not limited to, a personal computer 28, a television (not shown) embedded with or coupled to one or more processors, a laptop computer 30, a data-enabled mobile phone or smartphone 32, a laptop computer 34, a tablet computer (not shown), and a personal digital assistant (not shown). Client electronic devices 28, 30, 32, and / or 34 may each communicate with network 22 and / or network 26, and may each execute an operating system, examples of which may include, but are not limited to, Apple iOS™, Microsoft Windows™, Android™, Red Hat Linux™, or a custom operating system.

[0053] The instruction sets and subroutines of the client-side image recognition processes 12, 14, 16, and 18 can be stored (respectively) on storage devices 36, 38, 40, and 42 (respectively) coupled to client electronic devices 28, 30, 32, and 34, and can be executed by one or more processors (not shown) and one or more memory architectures (not shown) incorporated into client electronic devices 28, 30, 32, and 34. Storage devices 36, 38, 40, and 42 can include, but are not limited to: hard disk drives; magnetic tape drives; optical disk drives; solid-state storage devices; RAID arrays; random access memory (RAM); read-only memory (ROM); compact flash (CF) storage devices; secure digital storage (SD) devices; and memory stick storage devices.

[0054] Client-side image recognition processes 12, 14, 16, 18 and / or server-side image recognition process 10 may be processes that run within (i.e., are part of) a cloud computing site, cloud computing application, cloud platform, or cloud environment. Alternatively, client-side image recognition processes 12, 14, 16, 18 and / or server-side image recognition process 10 may be standalone applications that work in conjunction with a cloud computing site, cloud computing application, cloud platform, or cloud environment. One or more of client-side image recognition processes 12, 14, 16, 18 and server-side image recognition process 10 may be connected to each other (via network 22 and / or network 26).

[0055] For example, users 44, 46, 48, and 50 can directly access the server-side image recognition process 10 through devices executing client-side image recognition processes (e.g., client-side image recognition processes 12, 14, 16, and 18), namely client electronic devices 28, 30, 32, and 34. Users 44, 46, 48, and 50 can also directly access the server-side image recognition process 10 through network 22 and / or through auxiliary network 26. Furthermore, server computer 20 (i.e., the computer executing server-side image recognition process 10) can communicate with network 22 through auxiliary network 26, as shown by virtual link 52.

[0056] Various client electronic devices can be directly or indirectly coupled to network 22 (or network 26). For example, personal computer 28 is shown as directly coupled to network 22 via a hardwired network connection. Additionally, laptop computer 34 is shown as directly coupled to network 26 via a hardwired network connection. Laptop computer 30 is shown as wirelessly coupled to network 22 via a wireless communication channel 54 established between laptop computer 30 and a wireless access point (i.e., WAP) 56, which is shown as directly coupled to network 22. For example, WAP 56 can be an IEEE 802.11a, 802.11b, 802.11g, 802.11n, 802.11x Wi-Fi and / or Bluetooth device capable of establishing a wireless communication channel 54 between laptop computer 30 and WAP 56. Data-enabled mobile phone 32 is shown as wirelessly coupled to network 22 via a wireless communication channel 58 established between data-enabled mobile phone 32 and a cellular network / bridge 60, which is shown as directly coupled to network 22.

[0057] All IEEE 802.11x specifications can use Ethernet protocols and collision-avoiding Carrier Sense Multiple Access (CSMA / CA) for path sharing. For example, various 802.11x specifications can use Phase Shift Keying (PSK) modulation or Complementary Code Keying (CCK) modulation. Bluetooth is a telecommunications industry standard that allows devices such as mobile phones, computers, and personal digital assistants to interconnect using short-range wireless connections.

[0058] Image recognition process For illustrative purposes, the client-side image recognition process 12 will be described in the following discussion. It should be noted that the server-side image recognition process 10 may interact with the client-side image recognition process 12 and may be executed within one or more applications that allow communication with the client-side image recognition process 12. However, this is not intended to limit this disclosure, as other configurations are possible (e.g., standalone client-side image recognition processes and / or standalone server-side image recognition processes). For example, some implementations may include a server-side image recognition process 10 in place of the client-side image recognition process 12 or in addition to that client-side image recognition process. The image recognition process described herein can be run by an application or software for controlling a test system, such as test system 300 shown communicating with server computer 20, network 22, and client electronic device 28. Furthermore, while one or more operations of the image recognition process described herein may be run or executed by a processor of, for example, a client system (e.g., client electronic device 28) or a server system (e.g., server computer 20), other operations of the image recognition process may be executed by other processors or components, including those incorporating imaging devices (such as cameras or other components of the image recognition system).

[0059] refer to Figure 2 In an embodiment, a method for facilitating image recognition (i.e., for fixture identification in a test system) may be provided, and this can be implemented via an image recognition process 12. The image recognition process 12 may be configured to run continuously or may be activated in response to a user using the test system 300 or an associated graphical user interface. In an embodiment, the image recognition process 12 may use an imaging device to capture (202) image data representing a first fixture (e.g., lower fixture 312 and / or upper fixture 314) of the test system (e.g., test system 300). The imaging device may be a camera or other imaging device and may be located at camera position 320.

[0060] For example, it is still referenced now. Figure 5Test system 500, which may be similar to test system 300 and configured to include any of the technologies or features described in this disclosure, may include a lower fixture 502 and an upper fixture 504, and may be equipped with a camera 506. The imaging device or camera (e.g., camera 506) may be a webcam, a smartphone camera module, or other cameras (or systems of multiple cameras) that can provide depth-of-field (i.e., depth mapping) data and captured images to isolate the fixture from the background. In various implementations, the imaging device may use a range of light beyond the visible spectrum (e.g., infrared, ultraviolet, X-rays, etc.). As shown, camera 506 may transmit captured image or image data to a processor or software for processing to identify the fixture. The camera may be mounted such that one or more mounted fixtures are within the camera's field of view. The camera may also be mounted such that one or more mounted fixtures remain within the camera's available focus range and / or depth of field during any adjustment of the test system. For example, a load-response component (e.g., 310) may move up and down relative to the actuator during test setup (but may not typically move during testing).

[0061] In various implementations, one or more of the following configurations can be used. A single camera can be mounted to allow the upper and lower grippers to remain within the camera's field of view (and / or within the camera's focal range and / or depth of field) through the normal range of motion of the actuators and load response assembly. An array of two or more cameras can cover the entire focal range, where, for example, several cameras can be fixed to the test system frame and aimed at different locations. A fixed camera (e.g., 506) can aim at the upper gripper, and another camera can be mounted to the load response assembly such that the camera moves with the assembly and aims at the lower gripper. The camera can be mounted on a mechanism that allows the camera's field of view to move from one point to another (e.g., the camera can be mounted on one or more gimbals). This mechanism can be actuated (via an electric motor) and controlled by a computer. One or more cameras fixed to the test system can have lenses attached to flexible optical fibers. The flexible optical fibers can then be moved / positioned to provide the desired view and can be automated via miniature electromechanical actuators and computer control.

[0062] Some testing systems may have an oven, environmental chamber, or other enclosure surrounding the fixture during testing. In these systems, the fixture is accessible and visible during the test setup, and the camera can be positioned to make the setup area visible or to allow the camera to view through the enclosure. The enclosure may be a harsh environment (e.g., high temperature or high corrosiveness); however, in cases where the camera is remotely mounted, the aforementioned flexible fiber optic feature allows the target end of the flexible fiber to penetrate the wall of the enclosure, thus enabling visibility of the fixture.

[0063] While a testing system can typically have two fixtures for testing, including a lower fixture and an upper fixture, the lower and upper fixtures may or may not be the same for a given test. In some cases, the type of lower fixture can indicate the type of fixture required as the upper fixture for the test. For example, a three-point bending fixture mounted as the lower fixture may have a point contact fixture for the upper fixture. Tensile and compression related tests may require tensile fixtures and compression fixtures for both the lower and upper fixtures, respectively. In most cases, the lower and upper fixtures can be a matching kit, and the lower fixture can have more different elements. Therefore, if the lower fixture is identified, the appropriate corresponding upper fixture for a given test can be easily determined. Imaging equipment or a camera can then be arranged to focus on the lower fixture.

[0064] Furthermore, the image recognition process 12 can transmit (204) image data representing the first clamp (e.g., lower clamp 312 and / or upper clamp 314) from an imaging device (such as camera 506) to a processor running the image recognition application. The processor can be integrated with a client system (e.g., client electronics 28), a server system (e.g., server computer 20), or otherwise accessed via a network (e.g., network 22). As discussed above, image data can be transmitted to the client system or server computer via an interface through one or more communication protocols.

[0065] Various machine learning models can be configured to analyze images of typical size, typical pixel count, and / or typical format. Before transmitting the image to the processor or processing unit, the image recognition process 12 can crop, resize, or reformat the image for combinability purposes. The processor or processing unit can be located with the test system itself, or the image can be sent to a client system or server computer for processing. Image cropping, rescaling, or reformatization can be performed on the imaging device or camera itself or at the processor or processing unit.

[0066] Image recognition process 12 can use a processor running an image recognition application to identify (206) the first fixture (e.g., lower fixture 312 and / or upper fixture 314) based on image data. The image recognition application can perform one or more operations of the image recognition process described herein, including image recognition process 12. The image recognition application may include artificial intelligence components, including machine learning, machine learning models, neural networks, or some combination thereof. The machine learning model may reside in memory associated with the processor or processing hardware.

[0067] Machine learning models can be trained during or before installing the test system 300 and / or image recognition applications or processes. The type of fixture used in the test system (e.g., such as...) Figure 4A set of training data (as shown) can be used to train a machine learning model. The training data can be image data representing each gripper of the test system. In some implementations, the training data may include 300 to 1000 images per gripper or approximately 10 to 30 seconds of video per gripper at 30fps. In some cases, 50,000 images or more (approximately 10,000 images per gripper) may be used. It should be noted that the image set of the grippers may include most or all of the expected variations in gripper position and orientation. Therefore, more complex systems (e.g., systems with more types of grippers, or grippers that can rotate and can be viewed from different angles) may require more training data and images.

[0068] Various models, tools, and languages ​​can be used to generate and / or train machine learning models. Once trained, machine learning models can often be converted to different formats for use in different applications. For example, a model can be trained using TensorFlow in Python, but a TensorFlow model can be converted to ONNX format for native use in Windows code.

[0069] The time required to train a machine learning model can vary depending on the chosen model and training code. Training can be significantly accelerated using a graphics processing unit (GPU). For example, training with a 4-core 3.5 GHz Xeon and an unsupported GPU might take approximately 300 minutes per epoch. Training with a 4-core 2.9 GHz i7 laptop using the CPU might take approximately 360 minutes per epoch. Training with a 4-core 2.9 GHz i7 laptop using an Nvidia M1200 CPU might take approximately 30 minutes per epoch. Training with a 4-core 3.5 GHz Xeon and an Nvidia RTX 2080 Ti GPU might take approximately 3 minutes per epoch. Cloud resources and / or computing resources from third parties can also be used to train machine learning models.

[0070] Machine learning can combine various statistical methods. Groups or clusters of data (e.g., image data) can be identified based on features of fixtures found in the data. New points can be identified by determining the groups or clusters that are geometrically closest. Each group can be defined by a multidimensional statistical distribution (e.g., mean and standard deviation). For each group, there may be new points that are part of that group. These points can be identified as belonging to the group or cluster with the highest probability.

[0071] In this sense, a point can be an array of pixel values ​​representing an image of a fixture. A point can be defined by two or three (or more) values. For example, a 2D point can be a pair of (x, y), and a 3D point can be a triplet of (x, y, z) (with 1×2 and 1×3 arrays respectively). Other sequences can be used (e.g., instead of (x, y, z), (x1, x2, x3) can be used). Reference Figure 6 A simple example of (X,Y) coordinate pairs can be used for two-class classification: Class A (smaller points) and Class B (larger points). For example, Class A could correspond to the bending fixture class, and Class B could correspond to the compression fixture class. As shown in this example, X values ​​range from 0 to 1, and Y values ​​range from 0 to 10. If only the X value is known for each point, the categories can be difficult to distinguish. However, when the Y value is added, it becomes easy to see that points with Y > 5 are most likely in Class A, and points with Y < 5 are most likely in Class B. This distinction may not apply to all points, as some Class A points may be below the Y=5 line, while a Class B point may be above it; however, using this evaluation allows for the classification of new, unknown points with high confidence.

[0072] Digital images consist of a large number of pixels, each defined by a digital value (or the value of a color image). A typical image on a computer monitor, telephone, or television may consist of millions of dots. Smaller images can be 4×4 pixel black and white images. A 4×4 pixel black and white image (e.g., with pixel values ​​from 0 to 9) can be represented by a 4×4 matrix: This 4×4 matrix can be converted into a 1×16 array: image 1 Array = [1 0 0 1 2 3 3 2 0 6 6 0 0 4 4 0] Where, as with the previous (x, y, z) or (x1, x2, x3) coordinates, the point can now be: (x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, x11, x12, x13, x14, x15, x16). Another image can be represented by another 4×4 matrix: This 4×4 matrix can also be converted into a 1×16 array: image 2 Array = [0 0 0 0 0 2 2 0 0 6 6 0 0 4 4 0].

[0073] Still referencing Figure 7Image 1 can be a highly pixelated image of a bending fixture, and image 2 can be a highly pixelated image of a compression fixture. Figure 7 The table at the bottom shows the coordinates x1 to x16 corresponding to two images with significant differences. These differences are what machine learning models can use to detect discrepancies between one clamp and another. Based on this data, the machine learning model can learn that images or points with x1, x4, x5, and x8 values ​​> 0 may correspond to a bending clamp rather than a compression clamp.

[0074] Image recognition applications may alternatively or additionally include one or more neural networks. A neural network may include multiple layers or nodes, including an input layer, hidden layers, and an output layer. In a general sense, a neural network can be a complex mathematical function that takes input and returns an output. A neural network can be constructed such that the model contains many adjustable parameters that can be tuned (or “trained”) so that the desired output corresponds to a given input. For example, to determine whether a clamp is a bending clamp or a compression clamp, an image (or image data) of the compression clamp can be passed as input to the neural network, and the trained model can return as output an indication that the image (or image data) corresponds to the compression clamp (e.g., a numerical value corresponding to the compression clamp). The mathematical function can have a general format and can be constructed hierarchically, with each layer consisting of one or more nodes.

[0075] For example, the input layer can take data in a given data format. One or more intermediate layers can pass data from the input layer and perform various computations (e.g., applying different types of convolutional filters). Each layer takes input from the layer above and passes its output to the next layer that connects its output. Each layer can have multiple parameters that can be adjusted during training to match the actual output with the desired output. The number of trainable parameters in intermediate layers can be large. For a typical model, there may be 1 million to 5 million trainable parameters, but some models may have even more. Furthermore, the output layer can receive the output after the data has been passed through the intermediate layers. The output layer can return the results in a defined format. For example, this might lead to classifying an image as a curved clamp or a compressed clamp.

[0076] The image can first be preprocessed to match the input format of a given neural network. Some neural networks used for image recognition can use an input size definition of 224×224×3, which represents a 224×224 pixel image with three color values ​​(red, green, and blue). Therefore, if the original image is 400×600×3, it can be cropped and rescaled to a 224×224×3 digital format before being passed to the input layer. The data can then be processed from one layer to the next, with each node applying different computations, such as convolution. Each node can pass its output to nodes in subsequent layers. The outputs of layers can have abstract dimensions. The final layer can be designed to transform the output into a more useful format. This could be, for example, a 1×2 vector where: (1,0) = curved; and (0,1) = compressed.

[0077] Existing neural networks can be incorporated into models pre-trained on very large datasets, and these neural networks can be reused for fixture identification using transfer learning. For example, the output layer of a neural network can be replaced and retrained to operate using fixture image data. The output layer can be replaced and retrained to work with, for example, fixtures from a test system 300. Figure 4 Use them together with the associated fixtures.

[0078] In this implementation, logistic regression can be used on multiple images / image data associated with one or more grippers. A neural network with a transfer learning model (e.g., VGG-16 based) can be trained using multiple images / image data associated with one or more grippers in multiple configurations with various backgrounds, and this neural network can be implemented. Typical validation and training accuracy levels can reach over 99%, which translates to approximately 3 errors per year if the grippers are changed approximately three times per day. Training accuracy can be improved as more images / image data are added to the database associated with the learning model.

[0079] In this manner, the image recognition process 12 can use a trained machine learning model or neural network to identify (206) the lower fixture 312 or the upper fixture 314 based on image data. In an embodiment, the image recognition process 12 can also determine (208) whether the first fixture (e.g., the lower fixture 312) corresponds to a second fixture of the test system (e.g., test system 300). The second fixture may be an expected fixture associated with test information of the test system. The test information may indicate the type of fixture to be used with a specific test or material selected by the user. The first fixture and the second fixture may be fixtures of the same type or different types. For example, the first fixture may be installed in the test system (e.g., from a previous test), and the second fixture may be selected by the user at a graphical user interface (e.g., for a new test). In other words, the first fixture can be an actual, physical fixture of the test system (e.g., an installed fixture or otherwise a fixture in the focus of the imaging device), and the second fixture may not be a physical fixture, but rather a selected, planned, anticipated, or configured fixture, as the second fixture is associated with the test system software or loaded test files.

[0080] Now for reference Figure 8 This illustrates an example workflow associated with the image recognition process described herein. The workflow may include one or more actions performed by a user or the image recognition process to initiate testing on a test system. A user (e.g., one or more users among users 44, 46, 48, and 50) may select (802) a fixture to be used with the test system. The fixture may be... Figure 4 One of the fixtures 402 to 410 corresponds to fixtures A to E, respectively. The user can select, for example, fixture D and install (804) fixture in the test system (e.g., test system 300). For example, the image recognition process 12 can use the techniques and features described herein to identify or recognize the fixture.

[0081] For example, a user can interact with the test system 300 and / or associated image recognition applications via a graphical user interface accessible from the client electronic device 28. The user can set (806) the tests to be run via the graphical user interface. As previously described, in this implementation, the image recognition process 12 may identify the first fixture in part based on image data representing the first fixture and may also preset or pre-fill various test settings or parameters.

[0082] Once the first fixture is identified and compared with the second fixture, further operations of the testing system can proceed. For example, in response to determining that the first fixture corresponds to the second fixture of the testing system, image recognition process 12 can proceed (210) or allow the testing system to perform the next operation of the testing system. This may include pre-filling various test settings or parameters of the testing system, loading the test file of the testing system, identifying / indicating that the force sensors of the testing system are ready to set tare weight, applying any preload to the sample, and / or verifying the parameters of a predefined test. The testing system may also determine the required force and / or displacement (e.g., the applied load) based on the fixture type identified for the first fixture and / or the geometry of the input test file or read from the test file. If the type of fixture installed in the testing system is incorrect for the loaded test file, the applied load may be incorrect.

[0083] Once the test system software is configured, the test can be run (808), or the user can select to run the test or experiment by applying appropriate forces and displacements through the fixture at the desired rate or frequency and temperature. As the test runs, the system can collect force and displacement data and convert the data into stress and strain based on the fixture type and geometry input during the test setup. The stress and strain values ​​can be reported to the user as test results.

[0084] In an implementation, in response to determining that a first fixture (e.g., upper fixture 314 or lower fixture 312) does not correspond to a second fixture of the test system (e.g., selected at a graphical user interface associated with the test system 300), the image recognition process 12 can provide an indication that the first fixture does not correspond to a second fixture of the test system. This indication may include a notification on the graphical user interface of the test system. Referring now to... Figure 9 An image of an example test area 900 of the test system is shown. The image recognition process 12 can capture (202) image data representing a first fixture (e.g., fixture 902) of the test system using an imaging device, transfer (204) the image data representing the first fixture (e.g., fixture 902) from the imaging device to a processor running an image recognition application, and use the processor running the image recognition application to identify (206) the first fixture (e.g., fixture 902) based on the image data. Figure 9 As shown, clamp 902 can be a clamping bending clamp.

[0085] Now for reference Figure 10An example graphical user interface 1000 of a test system according to this disclosure is shown. The graphical user interface 1000 can be accessed from a client electronic device or a server computer, for example, running an image recognition application (e.g., client electronic device 28). The graphical user interface 1000 may include multiple graphical controls, including checkboxes, radio buttons, text fields, drop-down menus, etc., which allow the user to input test information into the graphical user interface. The user can select a fixture (e.g., a second fixture) from, for example, a drop-down menu 1002, which can then become test information used by the image recognition process 12 and / or the test system (e.g., test system 300) to proceed to the next operation. For example, as... Figure 10 As shown, the user can select a three-point bending fixture (e.g., a second fixture) from the drop-down menu 1002, since the three-point bending fixture can be a fixture corresponding to the sample type being tested. The test information can indicate the fixture type (e.g., the second fixture) corresponding to the test selected for the test system.

[0086] Continuing the example above, image recognition process 12 can determine (208) whether the first fixture (e.g., fixture 902, clamped bending fixture) corresponds to the second fixture of the test system (e.g., the three-point bending fixture selected at drop-down menu 1002). The second fixture (e.g., the three-point bending fixture selected at drop-down menu 1002) can be associated with test information of the test system (e.g., input at graphical user interface 1000). Image recognition process 12 can determine that the first fixture (e.g., fixture 902, clamped bending fixture) does not correspond to the second fixture of the test system (e.g., the three-point bending fixture selected at drop-down menu 1002). In response to determining that the first fixture (e.g., fixture 902, clamped bending fixture) does not correspond to the second fixture of the test system (e.g., the three-point bending fixture selected at drop-down menu 1002), image recognition process 12 can provide (212) an indication that the first fixture does not correspond to the second fixture of the test system.

[0087] Instructions can be provided, for example, via the graphical user interface 1000 of the test system. The graphical user interface 1000 can show an indication (e.g., text 1004) that the first clamp (e.g., clamp 902, clamped bending clamp) does not correspond to the second clamp (e.g., the three-point bending clamp selected at drop-down menu 1002). Instructions can also be provided in the form of pop-up windows or other graphical indicators or via audio.

[0088] As discussed above, the techniques and feature identification fixtures described in this disclosure can be performed using a neural network, which may be part of image recognition process 12 or an image recognition application. Neural networks can be used for machine learning and can analyze images or image data into digital data. Depending on the configuration of the neural network, digital data can be filtered and processed in various ways.

[0089] For example, a phase or layer of a neural network can apply a convolutional filter to digital data. Various image processing techniques can use convolutional filters, filter kernels, or other matrix-related mathematical applications to enhance or detect various features in an image. A convolutional filter can extract features or details of a vertically oriented fixture. Another layer of the neural network can detect features or details of a horizontally oriented fixture. The neural network can then be trained to learn or recognize relationships between these features and associate a combination of features with a given fixture type.

[0090] Now for reference Figure 11 Figure 1100 illustrates an example of feature detection according to aspects of this disclosure. Figure 1100 shows a shearing fixture associated with the actual image / numerical data described herein. Figure 1102 illustrates what the shearing fixture might appear when viewed after the actual image / digital data has been run through a convolution with a filter kernel capable of detecting horizontal edges. The actual image / digital data associated with Figure 1102 can be used to identify the shearing fixture from image data representing an unknown fixture captured by an imaging device focused on a test area of ​​a test system as described herein.

[0091] Similarly, image 1104 illustrates a three-point bending fixture associated with the actual image / numerical data described herein. Figure 1106 illustrates what a three-point bending fixture might appear when viewed after the actual image / digital data has been run through a filter kernel that detects vertical edges. The actual image / digital data associated with image 1106 can be used to identify a three-point bending fixture from image data representing an unknown fixture captured by an imaging device focused on a test area of ​​a test system as described herein.

[0092] Filter kernels can be used to detect other features, such as sharp corners and their position relative to other features. In this way, the machine learning model can ultimately learn the features that distinguish one fixture from another.

[0093] Figure 12 An example implementation of an aspect of this disclosure is described. Figure 12The code shown can be adapted or modified to be implemented as a program, function, or application via one or more programming languages ​​or software tools (e.g., Mathematica™, Matlab™, and Python™), which can be used or called by the image recognition process 12. Figure 12 As shown, functions can be defined to preprocess images and prepare them for sending to the classifier. For example, a bicubic resampling function can be used to resize the image to the desired size. Image data arrays can be manipulated into a format that matches the input to the machine learning model, such as 1×3×224×224 (1 image × 3 colors × 224×224 pixels). Functions can be called to load test files based on image / fixture classification. For example, bending fixture test files, compression fixture test files, stretching fixture test files, or shearing fixture test files can be loaded. If a fixture is not found or is not identified, a warning can be issued indicating that the fixture is not in the correct location or is not identified.

[0094] The desired pixel size can be defined (perhaps only one value is needed, as the image is expected to be square), and a trained fixture identification neural network can be loaded. An image of the first fixture (e.g., the installed fixture) can be loaded from, for example, a file system. The image data of the first fixture can also be passed via another process or returned by a function call. The image and the desired input size can be passed to a previously defined image processing function. The image data can be passed to a machine learning model to generate a prediction of what the fixture is like (i.e., the first fixture can be identified). If the first fixture matches the second fixture (e.g., the expected or selected fixture), the correct test file for testing the system can be loaded. If necessary, the results can be printed to the screen. If the fixtures do not match, a notification can be provided, or a function can be called to take action (e.g., stop the test).

[0095] In implementations, the image recognition process 12 and / or the image recognition application may include one or more operations or features for the safe operation test system. As previously discussed, the image recognition process 12 may use an imaging device to capture image data representing a test area of ​​the test system, and may also transfer the image data representing the test area from the imaging device to a processor running the image recognition application. The image recognition process 12 may also use the processor running the image recognition application to detect the presence of a user (e.g., a hand or finger) in the test area based on the image data.

[0096] Furthermore, the image recognition process 12 can adjust the test system to a safe mode (e.g., power-off or low-power mode) in response to detecting the presence of a user in the test area of ​​the test system. The image recognition process 12 can also provide an indication that a user has been detected in the test area of ​​the test system in response to detecting the presence of a user in the test area of ​​the test system. This indication may include a notification on the graphical user interface of the test system. The indication may also include a pop-up window or an audio alarm. Additionally, the image recognition process 12 can use a processor running an image recognition application to detect that the user presence (e.g., a hand or finger) is no longer in the test area based on image data. In response to detecting that the user is no longer in the test area of ​​the test system, the image recognition process 12 can put the test system out of safe mode.

[0097] Software tools / packages such as Mathematica™, Matlab™, and Python™ can be used to implement machine learning models. These software programs may be open-source or proprietary.

[0098] While the invention has been shown and described with reference to specific embodiments, those skilled in the art will understand that various changes in form and detail may be made without departing from the spirit and scope of the invention as set forth in the appended claims. For example, various forms of processes shown above may be used, in which steps are reordered, added, or deleted. Therefore, other implementations are within the scope of the following claims.

[0099] Implementations of this disclosure and all functional operations provided herein can be implemented in digital electronic circuit systems, or in computer software, firmware, or hardware including the structures disclosed herein and their structural equivalents, or in a combination of one or more of these. Implementations of this disclosure can be implemented as one or more computer program products, i.e., one or more modules of computer program instructions encoded on a computer-readable medium for execution by or control of the operation of a data processing apparatus. The computer-readable medium can be a machine-readable storage device, a machine-readable storage substrate, a memory device, or a combination of one or more of these. The term "data processing apparatus" encompasses all means, devices, and machines for processing data, including, for example, a programmable processor, a computer, or multiple processors or computers. In addition to hardware, the apparatus may also include code that creates an execution environment for the computer program in question, for example, code constituting processor firmware, a protocol stack, a database management system, an operating system, or a combination of one or more of these.

[0100] Computer programs (also known as programs, software, software applications, scripts, or code) can be written in any form of programming language, including compiled or interpreted languages, and can be deployed in any form, including as standalone programs or as modules, components, subroutines, or other units suitable for use in a computing environment. A computer program does not necessarily correspond to a file in a file system. A program can be stored in a file portion that holds other programs or data (e.g., one or more scripts stored in a markup language document), a single file dedicated to the program in question, or multiple coordinating files (e.g., files storing one or more modules, subroutines, or portions of code). A computer program can be deployed to execute on a single computer or on multiple computers located at a single site or distributed across multiple sites and interconnected via a communication network.

[0101] The processes and logic flows described in this disclosure can be executed by one or more programmable processors, which execute one or more computer programs to perform functions by manipulating input data and generating outputs. The processes and logic flows can also be executed by special-purpose logic circuit systems such as FPGAs (Field-Programmable Gate Arrays) or ASICs (Application-Specific Integrated Circuits), and the apparatus can also be implemented as a special-purpose logic circuit system.

[0102] Processors suitable for executing computer programs include, for example, both general-purpose microprocessors and special-purpose microprocessors, and any one or more processors in any type of digital computer. Typically, the processor receives instructions and data from read-only memory or random access memory, or both. The basic components of a computer are a processor for executing instructions and one or more memory devices for storing instructions and data. Typically, a computer will also include one or more mass storage devices for storing data, such as magnetic disks, magneto-optical disks, or optical disks, or the computer will be operatively coupled to such one or more mass storage devices to receive data from or transfer data to or from them, or both. However, a computer does not need to have such devices. Furthermore, a computer may be embedded in another device, such as a mobile phone, a personal digital assistant (PDA), a mobile audio player, a global positioning system (GPS) receiver, to name just a few. Computer-readable media suitable for storing computer program instructions or computer program products and data include all forms of non-volatile memory, media, and memory devices, including, for example, semiconductor memory devices, such as EPROM, EEPROM, and flash memory devices; magnetic disks, such as internal hard disks or removable disks; magneto-optical disks; and CD-ROM and DVD-ROM disks. These may also be referred to as computer-readable storage media. Processors and memory can be supplemented by or incorporated into dedicated logic circuitry systems.

[0103] To provide interaction with the user, the implementation described herein can be carried out on a computer having a display device for displaying information to the user (such as a CRT (cathode ray tube) or LCD (liquid crystal display) monitor), and a keyboard and pointing device (such as a mouse or trackball) for the user to provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback, such as visual feedback, auditory feedback, or tactile feedback; and input from the user can be received in any form, including acoustic input, voice input, or tactile input.

[0104] Implementations of this disclosure can be implemented in a computing system including back-end components (e.g., as a data server), or in a computing system including middleware components (e.g., an application server), or in a computing system including front-end components (e.g., a client computer with a graphical user interface or a web browser through which a user can interact with an implementation of this disclosure), or in any combination of one or more such back-end components, middleware components, or front-end components. Components of the system can be interconnected via any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include local area networks (“LANs”) and wide area networks (“WANs”), such as the Internet.

[0105] Computing systems can include clients and servers. Clients and servers are typically geographically separated and usually interact through communication networks. The client-server relationship arises from the computer programs running on the respective computers and the client-server relationship between them.

[0106] While this disclosure contains numerous details, these details should not be construed as limiting the scope of this disclosure or the scope of possible claims, but rather as descriptions of features specific to particular implementations of this disclosure. Certain features described in the context of individual implementations may also be provided in combination within a single implementation. Conversely, various features described in the context of a single implementation may also be provided individually in multiple implementations or in any suitable sub-combination. Furthermore, although features may be described above as functioning in certain combinations and even initially claimed in this manner, one or more features from a claimed combination may, in some cases, be removed from that combination, and the claimed combination may be for sub-combinations or variations thereof.

[0107] Similarly, although operations are depicted in a specific order in the accompanying drawings, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order, or requiring all the operations shown to achieve the desired result. In some cases, multitasking and parallel processing may be advantageous. Furthermore, the separation of various system components in the above-described implementation should not be interpreted as requiring this separation in all implementations, and it should be understood that the program components and systems described can generally be integrated into a single software product or packaged into multiple software products.

Claims

1. A method for fixture identification in a testing system, the method comprising: Image data representing the first fixture of the test system is captured using an imaging device; The image data representing the first fixture is transmitted from the imaging device to a processor running an image recognition application; The processor running the image recognition application identifies the first fixture based on the image data; Determine whether the first fixture corresponds to a second fixture of the test system, wherein the second fixture is an expected fixture associated with test information of the test system; and In response to determining that the first fixture does not correspond to the second fixture of the test system, an indication is provided that the first fixture does not correspond to the second fixture of the test system.

2. The method according to claim 1, further comprising: In response to determining that the first fixture corresponds to the second fixture of the test system, the next operation of the test system is performed.

3. The method of claim 2, wherein the next step includes at least one of: loading a test file associated with the test system; and setting test parameters associated with the test system.

4. The method of claim 1, wherein the indication comprises a notification on the graphical user interface of the test system.

5. The method of claim 1, wherein the image recognition application comprises at least one of the following: a machine learning model; and a neural network.

6. The method of claim 1, wherein the testing system is selected from the group consisting of: dynamic mechanical analyzers; fatigue testing systems; tensile testing systems; testing systems incorporated into a load frame; thermal analyzers; rheometers; tissue engineering systems; and heart valve replacement testing systems.

7. The method of claim 1, wherein at least one of the first clamp and the second clamp is a clamp type selected from the group consisting of: shearing clamps; bending clamps; compression clamps; and tension clamps.

8. The method of claim 1, wherein the test information represents the fixture type corresponding to the test selected for the test system.

9. A fixture marking system for marking fixtures in a testing system, the fixture marking system comprising: An imaging device, which is focusable on the test area of ​​the test system; A processor, which communicates with the imaging device, is configured to: Run an image recognition application; Receive image data representing the first fixture from the imaging device; The first clamp is identified based on the image data; Determine whether the first fixture corresponds to a second fixture of the test system, wherein the second fixture is an expected fixture associated with test information of the test system; and In response to determining that the first fixture does not correspond to the second fixture of the test system, an indication is provided that the first fixture does not correspond to the second fixture of the test system.

10. The fixture marking system according to claim 9, wherein the processor is further configured to: In response to determining that the first fixture corresponds to the second fixture of the test system, the next operation of the test system is performed.

11. The fixture marking system of claim 10, wherein the next operation includes at least one of: loading a test file associated with the test system; and setting test parameters associated with the test system.

12. The fixture marking system of claim 9, wherein the indication includes a notification on the graphical user interface of the test system.

13. The fixture marking system of claim 9, wherein the image recognition application includes at least one of the following: a machine learning model; and a neural network.

14. The fixture marking system of claim 9, wherein the testing system is selected from the group consisting of: dynamic mechanical analyzers; fatigue testing systems; tensile testing systems; testing systems incorporated into a load frame; thermal analyzers; rheometers; tissue engineering systems; and heart valve replacement testing systems.

15. The clamp identification system of claim 9, wherein at least one of the first clamp and the second clamp is a clamp type selected from the group consisting of: shearing clamps; bending clamps; compression clamps; and tension clamps.

16. A fixture identification system for identifying mounted fixtures in a test system incorporated into a load frame, the fixture identification system comprising: A camera that is focused on the test area of ​​the test system incorporated into the load frame; A processor, which communicates with the camera, is configured to: An image recognition application that includes an artificial intelligence component, the artificial intelligence component comprising at least one of the following: a machine learning model; and a neural network, wherein the artificial intelligence component is trained to identify the mounted fixture; Receive image data representing the installed clamp from the camera; The installed fixture is identified based on the image data and the artificial intelligence component trained to identify the installed fixture; Determine whether the installed fixture corresponds to the selected fixture of the test system, the selected fixture being associated with the test information of the test system; as well as In response to determining that the installed fixture does not correspond to the selected fixture of the test system, a notification indicating that the installed fixture does not correspond to the selected fixture of the test system is provided on the graphical user interface of the test system.

17. The method of claim 1, wherein the image recognition application comprises a machine learning model trained with image data representing a plurality of fixture types of a test system, the plurality of fixture types including at least two of the following: shearing fixtures, bending fixtures, compression fixtures, and tension fixtures.

18. The method of claim 1, further comprising: The testing system is stopped in response to the determination that the first fixture does not correspond to the second fixture.

19. The method of claim 1, wherein the first fixture is mounted on the testing system and the second fixture is selected by a user at a graphical user interface.

20. The fixture identification system of claim 9, wherein the first fixture is mounted on the testing system and the second fixture is selected by a user at a graphical user interface.

21. The fixture identification system of claim 16, wherein the installed fixture is installed in the testing system, and the selected fixture is selected by a user at a graphical user interface.

Citation Information

Patent Citations

  • Material testing machine

    JP2002365188A