System and method for equipment inspection

Abstract

Method and system for device inspection. According to one example, the method includes receiving image data representing the appearance of a device component prior to an inspection process, and evaluating the image data using a machine learning model that defines baseline image data to determine whether the image data indicates that the device component is ready for inspection. When the evaluation indicates that the device component is not ready, the inspection process can be blocked. According to one example, the system includes a controller that receives image data representing the appearance of a device component obtained prior to an inspection process. The controller evaluates the image data relative to the baseline image data using a machine learning model to determine whether the device component is ready for the inspection process. Based on the evaluation indicating that the device component is not ready for the inspection process, the controller blocks the inspection process from proceeding.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims priority to U.S. Provisional Application No. 63 / 130,085, filed on December 23, 2020, and U.S. Non-Provisional Application No. 17 / 523,343, filed on November 10, 2021, the entire disclosures of which are incorporated herein by reference. Technical Field

[0003] The disclosed subject matter described herein relates to systems and methods for inspecting equipment. More specifically, the disclosed subject matter described herein relates to systems and methods for evaluating visual inspection criteria, providing feedback and / or evaluation of visual inspection criteria, and providing tracking and notification of visual inspection criteria. Background Art

[0004] Equipment, such as vehicle components, is inspected to detect components that may be damaged or defective, or are close to being damaged or defective. Inspections may be performed according to established procedures for each component. However, adherence to procedures may vary from inspector to inspector, which may result in inaccurate results. Inspectors may not have access to previous inspection results, which makes it difficult to determine whether previous inspections were performed according to established procedures. If a component is inspected and incorrectly identified as not damaged or defective, failure of the component may cause equipment (e.g., a locomotive) to malfunction. Conversely, if a component is inspected and incorrectly identified as damaged or defective, unnecessary component replacement may cause the equipment to be taken out of service and result in additional repair costs. Summary of the Invention

[0005] According to one embodiment, a method for inspecting a device may include receiving image data representing an appearance of a device component. The image data may be obtained before an inspection process is performed on the device component. The method may also include evaluating the image data using a machine learning model that defines baseline image data to determine whether the image data indicates that the device component is ready for the inspection process. The method may also include preventing the device component from undergoing the inspection process if the evaluation of the image data relative to the baseline image data indicates that the device component is not ready for the inspection process.

[0006] According to one embodiment, a system for inspecting equipment may include a controller that receives image data representing the appearance of an equipment component. The image data may be obtained before an inspection process is performed on the equipment component. The controller may use a machine learning model to evaluate the image data relative to baseline image data to determine whether the equipment component is ready for the inspection process. If the evaluation of the image data relative to the baseline image data indicates that the equipment component is not ready for the inspection process, the controller may prevent the inspection process from being performed on the equipment component.

[0007] According to one embodiment, a method for inspecting a device component may include receiving a first image of the device component prior to inspecting the device component. The method may include evaluating the first image relative to one or more of a second image or baseline image data using a machine learning model to determine whether the device component is ready for an inspection procedure. The method may also include preventing the device component from undergoing an inspection procedure based on the evaluation of the first image relative to one or more of the second image or baseline image data indicating that the device component is not ready for the inspection procedure. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] The subject matter of the present invention may be understood by reading the following description of non-limiting embodiments with reference to the accompanying drawings, in which:

[0009] Figure 1 is an image of the equipment component in a first state according to preparation for inspection;

[0010] Figure 2 is an image of the equipment component in a second state according to preparation for inspection;

[0011] Figure 3 is an image of the equipment component in a third state according to preparation for inspection;

[0012] Figure 4 Schematically illustrates a machine learning model according to one embodiment;

[0013] Figure 5 Schematically illustrates a machine learning model according to one embodiment;

[0014] Figure 6 A system for evaluating a device component to be inspected according to one embodiment is schematically shown;

[0015] Figure 7 Schematically illustrates an application for evaluating image data of a plant component according to one embodiment;

[0016] Figure 8 Schematically illustrates an application for evaluating image data of a plant component according to one embodiment;

[0017] Figure 9 Schematically illustrates an application for evaluating image data of a plant component according to one embodiment;

[0018] Figure 10 schematically illustrates a method according to one embodiment; and

[0019] Figure 11 A method according to one embodiment is schematically illustrated. DETAILED DESCRIPTION

[0020] Embodiments of the subject matter described herein relate to systems and methods that improve and verify inspection quality and adherence to established inspection processes and procedures by providing inspection feedback concurrently during the inspection process. The systems and methods provide a way to track the inspection process by, for example, image, date, part, location, and inspector. The systems and methods can automatically track and identify noncompliant or at-risk inspections, allowing follow-up actions to mitigate the risk of noncompliant inspections.

[0021] Proper inspection of equipment components requires that inspection criteria be met before inspection. For example, components can be cleaned in the area to be inspected, clean areas can be marked or demarcated to indicate clean areas from other areas, and components can be numbered to ensure that inspection results are consistent with the inspected components. Failure to meet inspection criteria can lead to poor inspection performance and inaccurate results. Ensuring that inspection criteria are correctly met can reduce the occurrence of poor inspections. Providing inspectors with alerts that equipment components have not been properly prepared for inspection can prevent poor inspections.

[0022] Although one or more embodiments are described in conjunction with rail vehicle systems, not all embodiments are limited to rail vehicle systems. Unless expressly stated or otherwise indicated, the subject matter of the invention described herein extends to other types of vehicle systems, such as automobiles, trucks (with or without trailers), buses, ships, aircraft, mining vehicles, agricultural vehicles or other off-highway vehicles. The vehicle systems described herein (rail vehicle systems or other vehicle systems that do not travel on tracks or rails) can be formed by a single vehicle or multiple vehicles. For multi-vehicle systems, the vehicles can be coupled to each other mechanically (e.g., by couplers) or logically rather than mechanically. For example, when separate vehicles communicate with each other to coordinate each other's movement so that the vehicles (e.g., as a convoy) travel together, the vehicles can be logically rather than mechanically coupled.

[0023] refer to Figure 1 The equipment component 10 to be inspected includes an inspection area 12. The equipment component may be, for example, a fan blade for a locomotive radiator. The inspection area may be the middle portion of the fan blade, i.e., the portion between the ends of the fan blade. The inspection area may be any portion of the equipment component, such as an area where defects such as cracks may develop during use. However, any area of ​​the equipment component may be prepared for inspection and subsequently inspected.

[0024] According to other embodiments, equipment components that can be inspected include heat exchanger tubes, radiator fan blade welds, gearbox assemblies, compressor frames, cables, and wheels. Equipment components can also come from equipment other than locomotives, such as aircraft and aircraft engines, construction equipment, power generation equipment, or other vehicles (e.g., cars, buses, trucks, mining vehicles, ships, agricultural vehicles, etc.). The systems and methods disclosed herein are applicable to any equipment that requires periodic inspection.

[0025] Reference again Figure 1 The equipment component to be inspected can be marked with a unique component identifier 14 of the equipment component. The inspection area of ​​the equipment component can be marked with an inspection area mark 16. Figure 1 As shown, the test area is marked by two test area markers, but it should be understood that any number of test area markers, such as one marker or multiple markers, can be used to mark the test area.

[0026] Figure 1 The equipment component in the image is an example of inadequate inspection preparation. The equipment component shows no signs of cleaning. Defects such as cracks are difficult to detect with high accuracy using current technology. Lack of cleaning in the equipment component inspection area makes defect detection more difficult and can result in failure to detect defects that could lead to equipment component failure. Figure 1 Inspection of equipment components may be considered non-compliant inspection.

[0027] refer to Figure 2 , the equipment component to be inspected includes a component identifier for identifying the component and an inspection area mark for identifying the inspection area. Figure 2 The equipment part in Figure 1 is an example of an incomplete inspection preparation. The equipment part shows some signs of cleaning, but the inspection area is not sufficiently clean, preventing the inspection process from detecting defects such as cracks. Although the equipment part includes a part identifier and inspection area markings to identify the inspection area, incomplete cleaning of the inspection area can prevent the detection of defects. Figure 2 Inspection of equipment components may be considered risky or poor inspection.

[0028] refer to Figure 3 The equipment component to be inspected includes a component identifier for identifying the component and an inspection area marking for identifying the inspection area. The inspection area has been cleaned of materials or components that may make inspection difficult and defects difficult to detect. For example, the equipment component may be a fan blade for a locomotive radiator. Prior to preparation (cleaning) for inspection, the inspection area may have carbon or soot, grease, and / or rust on it. Figure 3 The inspection area of ​​the equipment component has been cleaned to remove any such materials or components. Figure 3 Inspection of equipment components in the equipment can be regarded as compliance inspection.

[0029] refer to Figure 4 According to one embodiment, the machine learning model 18 can be provided in the form of a neural network. The machine learning model includes an input layer 20, a hidden layer 22, and an output layer 24. The input layer receives image data representing the appearance of the device component to be inspected. The image data is obtained before the device component is subjected to the inspection process. The device component can be prepared for inspection, for example, Figures 1 to 3 Any individual piece of equipment to be inspected may be prepared as non-compliant, risky, or compliant.

[0030] In one embodiment, the machine learning model is a supervised machine learning model. Labeled training data is provided to the machine learning model. Image data of a device component that has been prepared for inspection, that is, image data of a device component with a component identifier, an inspection area mark, and an inspection area that has been at least partially cleaned, is provided to the machine learning model. Each image data of the training data is labeled as non-compliant, risky, or compliant based on the condition of the device component after preparation for inspection. The machine learning model uses the training data to establish a benchmark for the surface profile of the device component, which can be used to determine whether the input image data corresponds to a compliant surface profile. If the collected image data of the device component to be inspected is determined to correspond to a benchmark surface profile of a compliant surface profile, a message can be provided to the inspector that the inspection can be carried out. If the collected image data of the device component to be inspected does not correspond to the benchmark surface profile, for example, if the machine learning model determines that the collected image data indicates that the inspection preparation is non-compliant or risky, a message can be provided to the inspector that the device component has not yet been prepared for the inspection process, and the inspector can prevent the inspection.

[0031] Hidden layers are located between the input and output layers of a machine learning model's algorithm. The algorithm applies weights to the input (for example, pixels of image data) and outputs it through an activation function. Hidden layers perform nonlinear transformations on the inputs entering the network.

[0032] refer to Figure 5According to one embodiment, a machine learning model 26 includes an input layer 28, multiple hidden layers 30, 32, 34, 36, and an output layer 38. Due to the multiple hidden layers, the machine learning model can be referred to as a deep learning machine learning model. The hidden layers may vary depending on the functionality of the machine learning model, and the hidden layers may vary in their associated weights. The hidden layers allow the functionality of the machine learning model to be decomposed into specific transformations of the input data. Each hidden layer can be provided with functionality to produce a defined output. For example, one hidden layer can be used to identify the type of device component to be inspected. One hidden layer can identify the device component as a radiator fan blade, a gearbox assembly, or some other device component to be inspected. Another hidden layer can be provided to identify, for example, a component identifier, and another hidden layer can be provided to identify inspection region markings. Additional hidden layers can be provided to detect, for example, whether the captured image data represents the same device component, a blurred image, or the lighting of the device component. While the functionality of each hidden layer is insufficient to independently determine whether the captured image data represents a device component that is sufficiently prepared for inspection to provide compliance inspection, the multiple hidden layers work together in the machine learning model to determine the probability that the captured image data represents a correctly prepared device component.

[0033] refer to Figure 6 A system 40 for evaluating an equipment component to be inspected includes a mobile handheld device 41. The mobile handheld device may be a smartphone, a tablet, or a personal digital assistant (PDA). The mobile handheld device includes an image capture device, such as a camera, configured to capture images of the equipment component to be inspected. The mobile handheld device also includes a controller 42 that executes instructions stored in a memory 44 to use a machine learning model to determine whether the equipment component to be inspected represents a compliant inspection setup, a risky inspection setup, or a non-compliant inspection setup.

[0034] The system for evaluating an equipment component may also include a computer 46 including a controller 48 and a memory 50. The computer may be connected to the mobile handheld device, for example, wirelessly or via a hard connection (e.g., a cable). The memory of the computer may include training data for the machine learning model. Image data captured by the mobile handheld device may also be transferred from the mobile handheld device to the memory of the computer to be added to the training data of the machine learning model. The machine learning model may modify the reference surface profile image data based on the captured image data of the equipment component provided to the memory of the mobile handheld device and / or the memory of the computer.

[0035] The processor of the computer can also execute instructions in the memory of the computer to use a machine learning model to determine whether the equipment part to be inspected represents a compliant inspection preparation, a risky inspection preparation, or a non-compliant inspection preparation. In one embodiment, the mobile handheld device can be a digital camera that captures images of the equipment part to be inspected. The image data can be transferred to the memory of the computer, and the controller of the computer can use the machine learning model to determine whether the equipment part corresponds to the reference surface profile image data representing a compliant inspection preparation. Other image acquisition devices can be used to collect image data of the equipment part to be inspected. For example, an inspector preparing the equipment part for inspection can wear a camera. As another example, the camera can be set on a sight to be inserted into the interior of the equipment part or set on an unmanned aerial vehicle (e.g., a drone) to collect image data on a remote equipment part.

[0036] The system may also include a cloud computing network 52. The cloud computing network may store image acquisition data, including, for example, training data for a machine learning model and image acquisition data obtained during the inspection of a device component. The cloud computing network may include one or more cloud computing nodes with which a mobile handheld device and / or a computer may communicate. These nodes may communicate with each other in one or more networks and may be grouped physically or virtually. The cloud computing network may communicate with any type of computerized device, including, for example, a mobile handheld device and a computer, via any type of network and / or network addressable connection (e.g., using a web browser).

[0037] The cloud computing network may also use machine learning models to evaluate captured image data provided by a mobile handheld device and / or computer to determine whether the equipment part to be inspected represents a compliant inspection preparation, a risky inspection preparation, or a non-compliant inspection preparation.

[0038] The image data collected of the device component to be inspected may include one or more images or video frames of visible light reflected from the device component. The image data collected of the device component to be inspected may additionally or alternatively include one or more images or video frames of light outside the visible spectrum reflected from the device component. The machine learning model can identify blurred images based on the reflected light. The machine learning model can also determine whether multiple collected image data are reflected from the same or different device components.

[0039] refer to Figure 7, the mobile handheld device can execute a program or application for evaluating the equipment component to be inspected by operating an application installed on the mobile handheld device. According to other embodiments, the program or application can be performed on a computer. The program or application accepts input through, for example, a touch screen 54 and allows the inspector to indicate a specific equipment component to be inspected, such as a radiator fan, and provides a display 56 that the equipment component has been added to the inspection history of the equipment component. The program or application allows the inspector to enter identification information 58, which may include, for example, the name and / or location of the inspection agency, the serial number or tracking number of the equipment component to be inspected, and the inspection date. After confirming the identification information, the program or application provides input 60 to start the inspection.

[0040] refer to Figure 8 , an image 62 of the device part to be inspected is evaluated by a machine learning model. The machine learning model may conclude that the image data does not correspond to the machine learning model's baseline image data and that the preparation of the device part represents risky or non-compliant preparation. For example, the machine learning model may alert the inspector that an inspection area marking (e.g., a line) that defines the inspection area is not visible. The machine learning model may also determine that other features of the image data make the evaluation of the image data of the device part unreliable. For example, the machine learning model may determine that the image data is blurry, or that the image data is not of the same device part as previously acquired image data. The machine learning model may provide the inspector with evaluation information 64 of the image data to identify preparation or image data issues that may cause the inspection to be risky or non-compliant. The application also includes an input 66 to allow the inspector to exclude the image data from the final evaluation of the machine learning model or include the image data in the final evaluation of the machine learning model.

[0041] refer to Figure 9 , the program or application presents a plurality of image data 68 of the device part to be inspected on a display screen. The machine learning model determines whether each image is from the same device part. The machine learning model also determines whether the amount of image data indicates that the device part to be inspected is prepared for compliance, and determines whether the amount of compliant image data accounts for a specified percentage of all image data of the device part. If the amount of image data indicating that a compliant inspection can be performed is insufficient and / or the amount of image data indicating that a compliant inspection can be performed is insufficient as a percentage of all image data, the program or application can automatically send an alert to the account or inspection agency, for example, via email, that the device part should not be inspected because the device part has not been properly prepared and the device part indicates that the inspection is risky or non-compliant. The alert can prevent the inspection of the device part. If the machine learning model determines that the image data corresponds to baseline image data that will provide a compliant inspection, the program or application includes an input 70 that allows the inspector to continue or pause the inspection.

[0042] refer to Figure 10 Method 100 includes, at step 110, receiving image data representing an appearance of a device component, the image data being obtained before the device component is subjected to an inspection process; and, at step 120, evaluating the image data using a machine learning model that defines baseline image data to determine whether the image data indicates that the device component is ready for the inspection process. The method also includes, at step 130, preventing the device component from being subjected to the inspection process when the evaluation of the image data relative to the baseline image data indicates that the device component is not ready for the inspection process.

[0043] refer to Figure 11 Method 200 includes, at step 210, receiving a first image representing a device component. The first image may be obtained before the device component is subjected to an inspection process. The method further includes, at step 220, evaluating the first image relative to one or more of a second image or baseline image data using a machine learning model to determine whether the image data indicates that the device component is ready for the inspection process. The method further includes, at step 230, preventing the device component from being subjected to the inspection process if the evaluation of the first image relative to one or more of the second image or baseline image data indicates that the device component is not ready for the inspection process.

[0044] A method may include receiving image data representing an appearance of a device component, the image data obtained before subjecting the device component to an inspection procedure, and evaluating the image data using a machine learning model that defines baseline image data to determine whether the image data indicates that the device component is ready for the inspection procedure. The method may also include preventing subjecting the device component to the inspection procedure when the evaluation of the image data relative to the baseline image data indicates that the device component is not ready for the inspection procedure.

[0045] The image data may be evaluated relative to baseline image data to determine one or more of whether the equipment component has been cleaned, marked with an inspection mark, or marked with an identification mark. The baseline image data may include or be based on one or more historical images of the equipment component.

[0046] The method may further include modifying the baseline image data based on usage of the equipment component. The method may further include evaluating different portions of the image data relative to each other to determine whether the different portions of the image data are images of the same area of ​​the equipment component, and, in response to the different portions of the image data being images of the same area of ​​the equipment component, preventing the equipment component from undergoing an inspection process.

[0047] The method may also include receiving an inspection result of an inspection procedure performed on the equipment component and modifying the inspection result based on the received image data. In response to the image data indicating that the equipment component is not ready for the inspection procedure, the inspection result may be modified from a first conclusion of no damage or an acceptable amount of damage to a different second conclusion of damage or an unacceptable amount of damage.

[0048] The image data may include one or more images or video frames of visible light reflected from a device component.The image data may include one or more images or video frames of light outside the visible spectrum.

[0049] A system may include a controller that receives image data representing an appearance of a device component. The image data may be obtained before an inspection process is performed on the device component. The controller may utilize a machine learning model to evaluate the image data relative to baseline image data to determine whether the device component is ready for the inspection process. The controller may prevent the device component from undergoing the inspection process based on the evaluation of the image data relative to the baseline image data indicating that the device component is not ready for the inspection process.

[0050] The controller may evaluate the image data relative to baseline image data to determine whether the equipment component has been cleaned, marked with an inspection mark, or marked with an identification mark. The baseline image data may include or be based on one or more historical images of the equipment component.

[0051] The controller may further modify the baseline image data based on usage of the equipment component. The controller may also evaluate different portions of the image data relative to each other to determine whether the different portions of the image data are images of the same area of ​​the equipment component, and, in response to the different portions of the image data being images of the same area of ​​the equipment component, prevent the equipment component from undergoing an inspection process.

[0052] The controller may also receive inspection results of an inspection process performed on a device component and modify the inspection results based on the received image data, wherein, in response to the image data indicating that the device component is not yet ready for the inspection process, the inspection result may be modified from a first conclusion of no damage or an acceptable amount of damage to a different second conclusion of damage or an unacceptable amount of damage.

[0053] The image data may include one or more images or video frames of visible light reflected from a device component.The image data may include one or more images or video frames of light outside the visible spectrum.

[0054] A method may include receiving a first image of a device component before subjecting the device component to an inspection procedure, and evaluating the first image relative to one or more of a second image or baseline image data using a machine learning model to determine whether the device component is ready for the inspection procedure. The method may also include preventing subjecting the device component to the inspection procedure if the evaluation of the first image relative to one or more of the second image or baseline image data indicates that the device component is not ready for the inspection procedure.

[0055] The first image can be evaluated relative to the baseline image data to determine whether the equipment part has been cleaned, is marked with an inspection mark, or is marked with an identification mark. The first image can be evaluated relative to the second image to determine whether the first image and the second image show the same equipment part, and in response to determining that the first image and the second image show the same equipment part, the inspection process can be prevented from proceeding.

[0056] The method may also include receiving an inspection result of an inspection procedure performed on the equipment component and modifying the inspection result based on the received image data. In response to the image data indicating that the equipment component is not ready for the inspection procedure, the inspection result may be modified from a first conclusion of no damage or an acceptable amount of damage to a different second conclusion of damage or an unacceptable amount of damage.

[0057] As used herein, an element or step described in the singular and beginning with the word "a" or "an" does not exclude the presence of plural elements or operations unless such exclusion is explicitly stated. In addition, reference to "one embodiment" of the present invention does not exclude the existence of other embodiments that include the described features. In addition, unless explicitly stated to the contrary, references to "comprising," "including," or "having" an element or elements having a particular property in an embodiment may also include other elements that do not have such properties. In the appended claims, the terms "comprising" and "wherein" are used as the plain English equivalents of the corresponding terms "including" and "wherein." In addition, in the following claims, terms such as "first," "second," and "third" are used merely as labels and do not impose numerical requirements on their objects. In addition, the limitations of the following claims are not written in a means-plus-function format and are not intended to be interpreted under 35 U.S.C. §112(f) unless and until such claim limitation explicitly uses the phrase "means for," followed by a functional statement without further structure.

[0058] The above description is illustrative, not restrictive. For example, the above embodiments (and / or aspects thereof) can be used in combination with each other. In addition, without departing from the scope of the present invention, many modifications can be made to adapt specific situations or materials to the teachings of the present invention. Although the size and type of the material described herein limit the parameters of the present invention, it is an exemplary embodiment. After reading the above description, other embodiments will be apparent to those of ordinary skill in the art. Therefore, the scope of the present invention should be determined with reference to the complete scope of the equivalents authorized by the appended claims and such terms.

[0059] This written description uses examples to disclose several embodiments of the inventive subject matter, including the best mode, and to enable one of ordinary skill in the art to practice the embodiments of the inventive subject matter, including making and using any devices or systems and performing any incorporated methods. The patentable scope of the inventive subject matter is defined by the claims, and may include other examples that occur to one of ordinary skill in the art. Such other examples are intended to be within the scope of the claims if they have structural elements that do not differ from the literal language of the claims, or if they include equivalent structural elements with insubstantial differences from the literal language of the claims.

Claims

1. A method for testing a device, characterized in that The method comprises: receiving image data representing an appearance of a device component, the image data being obtained before an inspection process is performed on the device component; evaluating the image data using a multi-layer neural network that transforms the image data using a machine learning model that defines baseline image data to determine whether the image data indicates that the device part is ready for the inspection procedure; and In response to the neural network's evaluation of the image data relative to the baseline image data indicating that the equipment component is not yet ready for the inspection process, preventing the equipment component from being subjected to the inspection process; wherein the image data is evaluated relative to the baseline image data to determine one or more of whether the equipment component has been cleaned, is marked with an inspection mark, or is marked with an identification mark.

2. The method according to claim 1, characterized in that The baseline image data includes or is based on one or more historical images of the equipment component.

3. The method according to claim 1, characterized in that The method further comprises: The baseline image data is modified based on usage of the equipment component.

4. The method according to claim 1, wherein The method further comprises: evaluating different portions of the image data relative to each other to determine whether the different portions of the image data are images of the same area of ​​the device component; and In response to the different portions of the image data being images of the same area of ​​the equipment component, the equipment component is prevented from undergoing the inspection procedure.

5. The method according to claim 1, wherein The method further comprises: receiving an inspection result of the inspection process performed on the equipment component; and The inspection result is modified based on the received image data, wherein, in response to the image data indicating that the device component is not yet ready for the inspection process, the inspection result is modified from a first conclusion of no damage or an acceptable amount of damage to a different second conclusion of damage or an unacceptable amount of damage.

6. The method according to claim 1, characterized in that The image data includes one or more images or video frames of visible light reflected from the device component.

7. The method according to claim 1, characterized in that The image data includes one or more images or video frames of light outside the visible spectrum.

8. A system for testing equipment, characterized in that The system comprises: A controller configured to receive image data representing an appearance of a device component, the image data being obtained before an inspection process is performed on the device component, the controller including a multi-layer neural network configured to: transforming the image data using a machine learning model to evaluate the image data relative to baseline image data to determine whether the device component is ready for the inspection procedure; and Based on the evaluation of the image data relative to the baseline image data, it is indicated that the equipment component is not yet ready for the inspection process, and the inspection process is prevented from being performed on the equipment component; wherein, the controller is configured to evaluate the image data relative to the baseline image data to determine one or more of whether the equipment component has been cleaned, whether it is marked with an inspection mark, or whether it is marked with an identification mark.

9. The system according to claim 8, characterized in that The baseline image data includes or is based on one or more historical images of the equipment component.

10. The system according to claim 8, wherein: The controller is further configured to: The baseline image data is modified based on usage of the equipment component.

11. The system according to claim 8, wherein: The controller is further configured to: evaluating different portions of the image data relative to each other to determine whether the different portions of the image data are images of the same area of ​​the device component; as well as In response to the different portions of the image data being images of the same area of ​​the equipment component, the equipment component is prevented from undergoing the inspection procedure.

12. The system according to claim 8, wherein: The controller is further configured to: receiving an inspection result of the inspection process performed on the equipment component; and The inspection result is modified based on the received image data, wherein, in response to the image data indicating that the device component is not yet ready for the inspection process, the inspection result is modified from a first conclusion of no damage or an acceptable amount of damage to a different second conclusion of damage or an unacceptable amount of damage.

13. The system according to claim 8, wherein: The image data includes one or more images or video frames of visible light reflected from the device component.

14. The system according to claim 8, wherein: The image data includes one or more images or video frames of light outside the visible spectrum.

15. A method for inspecting a device component, characterized in that: The method comprises: receiving a first image of the equipment component prior to performing an inspection process on the equipment component; evaluating the first image using a multi-layer neural network that utilizes a machine learning model to transform the first image relative to one or more of a second image or reference image data to determine whether the device part is ready for the inspection procedure; and When an evaluation based on the first image relative to one or more of the second image or the reference image data indicates that the equipment part is not yet ready for the inspection process, the inspection process is prevented from being performed on the equipment part; wherein the first image is evaluated relative to the reference image data to determine whether the equipment part has been cleaned, whether it is marked with an inspection mark or whether it is marked with an identification mark.

16. The method according to claim 15, characterized in that The first image is evaluated relative to the second image to determine whether the first image and the second image show the same equipment part, and in response to determining that the first image and the second image show the same equipment part, the inspection procedure is prevented from proceeding.

17. The method according to claim 15, characterized in that The method further comprises: receiving an inspection result of the inspection process performed on the equipment component; and The inspection result is modified based on the received image data, wherein, in response to the image data indicating that the device component is not yet ready for the inspection process, the inspection result is modified from a first conclusion of no damage or an acceptable amount of damage to a different second conclusion of damage or an unacceptable amount of damage.

Images