Self-adaptive detection method, device and equipment of detection tool and storage medium

Abstract

The invention provides a self-adaptive detection method and device of a detection tool, equipment and a storage medium. The method comprises the steps that when contour data of a to-be-detected part cannot be matched in a pre-stored part contour library, a system marks the type of the to-be-detected part; a number threshold value is set, parameters of unknown parts of the same type are compared through a clustering algorithm, a qualified product set is automatically screened out, and therefore standard parameters of the new parts are objectively determined. And finally, associating the newly identified contour data with the autonomously generated standard parameters, and importing the data into a database to complete automatic updating of the knowledge base. It can be understood that the whole process enables the gauge to be converted from a passive detection tool into an intelligent system capable of actively accumulating knowledge and continuously expanding the detection range of the gauge, the problem that due to part updating, manual intervention needs to be frequently involved in updating of a gauge database is fundamentally solved, and on the premise that production continuity is not interrupted, the production efficiency is greatly improved. And dynamic and accurate matching between the detection capability and the production demand is realized.

Description

Technical Field

[0001] This application relates to the field of detection technology, and specifically to an adaptive detection method, device, equipment and storage medium for a checking fixture. Background Art

[0002] In the manufacturing process of complex products such as vehicles, checking fixtures are key detection tools widely used in the production site for quickly controlling product dimensions, such as hole diameter, profile tolerance, position tolerance, etc. It usually serves as a supplement or alternative to professional measuring equipment and can quickly judge the dimensions and shapes of parts before assembly, thus effectively ensuring the production rhythm and final product quality.

[0003] In the related art, an operator places the vehicle part to be detected on the corresponding special checking fixture, fixes it through structures such as positioning pins and clamping devices on the checking fixture, and then uses measuring tools such as dial indicators and plug gauges to detect each of the multiple preset measuring points on the part one by one, and compares the readings with the standard tolerance range to determine whether the part is qualified. The whole process depends on the checking fixture specially designed and manufactured for this part, and its detection items and standards are fixed after the checking fixture is manufactured.

[0004] However, with the accelerating speed of product update and iteration and the increasing demand for personalized customization, the types of parts faced on the production line are becoming increasingly diverse. When a new part that is not in the pre - stored part profile library appears, the checking fixture cannot automatically identify its profile, let alone make an autonomous decision on how to detect it. It must rely on manual intervention to measure, analyze and manually update the part profile library and detection program, which is time - consuming and laborious, and seriously restricts the rapid response ability and intelligent level of the production line.

[0005] It should be noted that the information disclosed in the background art part of this application is only intended to deepen the understanding of the general background art of this application, and should not be regarded as an admission or any form of暗示 that this information constitutes the prior art known to those skilled in the art. Summary of the Invention

[0006] In view of this, this application provides an adaptive detection method, device, equipment and storage medium for a checking fixture, which helps to solve the problem in the prior art that when a part that is not in the pre - stored part profile library appears, it must rely on manual intervention to measure, analyze and manually update the part profile library and detection program, which is time - consuming and laborious and seriously restricts the rapid response ability and intelligent level of the production line.

[0007] In a first aspect, this application provides an adaptive detection method for a checking fixture. The checking fixture includes a positioning part, and the positioning part is used to adapt to and characterize the profile shapes of multiple parts to be detected. The method includes: When the part to be inspected is placed on the positioning part of the fixture, the contour data of the part to be inspected is identified; Determine whether the contour data matches the pre-stored component contour library; When the contour data does not match the pre-stored component contour library, the type of the unknown component corresponding to the contour data is marked, and the detection parameters of the unknown component are collected. When the number of unknown parts of the same type exceeds a preset threshold, the detection parameters of each unknown part are compared horizontally, and a clustering algorithm is used to determine the set of qualified parts among the unknown parts. Based on the detection parameters of the qualified parts set among the unknown parts, the standard parameters of this type of part are determined, and the contour data corresponding to this type of part is associated with the standard parameters and then imported into the pre-stored part contour library.

[0008] In this embodiment, when the contour data of the part to be inspected cannot be matched in the pre-stored part contour library, the system does not simply issue an alarm or reject the part, but actively marks it as an unknown type and initiates a learning process. By setting a quantity threshold and using a clustering algorithm to compare the parameters of unknown parts of the same type, the system can automatically filter out a set of qualified parts from a batch of new parts, thereby objectively establishing the standard parameters of the new part. Finally, the newly identified contour data is associated with the self-generated standard parameters and imported into the database, completing the automatic update of the knowledge base. It can be understood that the above-mentioned entire process transforms the inspection tool from a passive inspection tool into an intelligent system that can actively accumulate knowledge and continuously expand its inspection range, fundamentally solving the pain point of needing frequent manual intervention to update the inspection tool database due to part updates, and realizing a dynamic and accurate match between inspection capabilities and production needs under the premise of uninterrupted production continuity.

[0009] In one possible implementation, the positioning part includes a plate and a plurality of movable elements, wherein the movable elements movably protrude from the upper surface of the plate. When the component to be inspected is placed on the positioning part of the fixture, the process of identifying the contour data of the component to be inspected includes: When the component to be inspected is placed on the upper surface of the plate, the contour data of the component to be inspected is identified based on the position of the movable element.

[0010] In this embodiment, when a component is placed on the plate, its contact surface presses down on the corresponding movable element, and the positions of each element naturally form a spatial mapping of the bottom contour of the component. This combination of mechanical structure and detection principle allows components of different shapes to automatically generate corresponding contour data through the same set of devices, which not only eliminates the need for specially manufactured matching molds in traditional inspection tools, but also provides accurate and reliable contour information input for subsequent intelligent recognition and learning processes.

[0011] In one possible implementation, the positioning unit further includes a visual sensor or a pressure sensor array; When the component to be inspected is placed on the upper surface of the plate, the contour data of the component to be inspected is identified based on the position of the movable element, including: When the component to be inspected is placed on the upper surface of the plate, the contour data is obtained based on the position data of the movable element collected by the vision sensor and / or the pressure sensor array.

[0012] In this embodiment, the vision sensor can quickly capture the overall image of the part's contour, avoiding wear and errors caused by physical contact; while the pressure sensor array accurately reconstructs the contact surface shape of the part by sensing the force state of each movable element. These two approaches not only enhance adaptability to complex contours but also provide an accurate and reliable digital foundation for subsequent intelligent recognition and data analysis by directly converting physical shapes into electronic signals, effectively improving the automation level and data accuracy of the entire detection system.

[0013] In one possible implementation, obtaining the contour data based on the position data of the movable element collected by the vision sensor and the pressure sensor array includes: The contour data of the movable element collected by the vision sensor is compared with the contour data of the movable element collected by the pressure sensor array to obtain contour difference data. Determine whether the contour difference data is less than or equal to the preset difference threshold; When the contour difference data is less than or equal to the preset difference threshold, the contour data of the movable element collected by the sensor and the contour data of the movable element collected by the pressure sensor array are fused together to obtain the contour data of the component to be inspected.

[0014] In this embodiment, by comparing the differences between the contour data acquired by the visual sensor and the pressure sensor array, the system can automatically identify and eliminate false detections or noise interference that may exist in a single sensor. When the difference is within the allowable range, the two contour data are fused. In essence, this combines the advantages of the global accuracy of the visual sensor and the contact perception accuracy of the pressure sensor array, thereby generating a contour model that is more accurate and reliable than any single data source. This effectively avoids erroneous judgments caused by occasional sensor errors and provides a high-quality data foundation for subsequent recognition and decision-making.

[0015] In one possible implementation, the movable element is provided with an electromagnetic locking unit, which is used to lock the movable element to its current position after the movable element has been reset.

[0016] In this embodiment, when the component completes its reset, the electromagnetic locking unit immediately and firmly locks it in the reference position, forming a stable measurement reference plane. This eliminates the minute displacement of the component caused by external vibration or impact during component placement, providing a more reliable data foundation for the entire detection system.

[0017] In one possible implementation, the positioning part further includes a plurality of piezoelectric ceramic micro-actuators, which are correspondingly disposed below the movable element; before placing the part to be inspected into the positioning part of the fixture, the method further includes: A depth scan is performed on the movable elements in the positioning part to obtain the reset deviation value of each movable element in the current state; Based on the reset deviation value, the piezoelectric ceramic micro actuator is controlled to adjust the position of the corresponding movable element, and the position of the movable element is compensated.

[0018] In this embodiment, after obtaining the reset deviation of the movable element through depth scanning, the piezoelectric ceramic micro-actuator is directly driven to generate a high-precision (e.g., nanometer-level precision) displacement, which physically pushes the movable element that has not been fully reset to the standard zero position, thereby eliminating the reset error from the root and ensuring the absolute accuracy of the contour acquisition reference surface. This provides a more reliable physical basis for subsequent detection and significantly improves the long-term stability of the system and the absolute accuracy of the measurement.

[0019] In one possible implementation, determining the set of qualified parts among the unknown components using a clustering algorithm includes: The detection parameters are calculated using K-means or DBSCAN clustering algorithms, and the clusters in the parameter distribution set are identified as the set of qualified parts. In this embodiment, the K-means algorithm effectively identifies the densest clusters in the parameter space, while the DBSCAN algorithm adapts to more complex distribution patterns and automatically removes outliers. The application of these two algorithms enables the system to objectively determine the main clusters conforming to manufacturing patterns as the set of qualified parts based on the natural aggregation characteristics of part parameters, avoiding subjective bias caused by relying on fixed thresholds set by human experience. This data-driven approach not only ensures that the qualification standards originate from actual production but also adaptively follows subtle changes in the manufacturing process, thereby continuously maintaining the accuracy and reliability of the judgment standards.

[0020] In one possible implementation, after associating the contour data corresponding to the type of component with the standard parameters and importing it into the pre-stored component contour library, the method further includes: Using the updated pre-stored component contour library, the contour data of this type of unknown component is re-acquired; The contour data of the unknown component of this type is compared with the corresponding standard parameters in the pre-stored component contour library, and the qualification of the unknown component is determined based on the comparison result. In this embodiment of the application, by re-inspecting the previously marked unknown components, not only are the "historical legacy issues" generated during the learning process effectively resolved, avoiding the backlog and waste of these parts, but more importantly, this step substantially verifies the accuracy and practicality of the newly established standard parameters, ensuring that the system's learning results can be instantly transformed into reliable productivity, and significantly improving the continuity and intelligence level of the entire inspection process.

[0021] In one possible implementation, Before placing the component to be inspected into the positioning part of the fixture, the method further includes: performing a depth scan on the movable elements in the positioning part to obtain the reset deviation value of each movable element in the current state; Before determining whether the contour data matches the pre-stored component contour library, the method further includes: compensating and correcting the contour data based on the reset deviation value.

[0022] In this embodiment, a deep scan of the reset state of the movable component is performed before inspecting the part. This allows for the timely detection of reset deviations caused by spring fatigue or mechanical wear. Based on this deviation, the subsequently acquired contour data is compensated and corrected in real time, establishing a dynamic calibration benchmark for the inspection system. This design enables the fixture to self-diagnose and correct accuracy losses caused by mechanical wear, ensuring the accuracy of contour data from the source. This provides a reliable data foundation for subsequent part identification, cluster analysis, and standard establishment, significantly improving the stability of the entire system and the reliability of the inspection results during long-term use.

[0023] In one possible implementation, the detection parameters include surface profile, position, height, and perpendicularity.

[0024] In this embodiment, the surface profile reflects the overall shape accuracy of the part, the positional accuracy ensures the accuracy of its mounting reference, and the height and perpendicularity jointly control the spatial orientation of the part. This combination of key geometric parameters covers the most critical dimensional requirements of the part during assembly, providing rich and accurate data features for subsequent cluster analysis. This enables the system to establish a more complete and reliable standard parameter model, thereby ensuring the comprehensiveness and accuracy of the conformity assessment.

[0025] One possible implementation also includes: When the contour data matches the pre-stored component contour library, the detection parameters corresponding to the component to be inspected are obtained based on the standard parameter types in the pre-stored component contour library. Based on the detection parameters corresponding to the component to be inspected and the corresponding standard parameters, the component to be inspected is subjected to a conformity test.

[0026] In this embodiment, when the system identifies a part as a known type, it can immediately retrieve the corresponding standard parameters and inspection scheme from the pre-stored part contour library, automatically completing the entire process from parameter acquisition to conformity judgment. This knowledge-based rapid response mechanism enables the inspection of known parts to be executed accurately without any manual configuration, fully leveraging the guiding value of pre-stored standards while ensuring the standardization and consistency of the inspection process, significantly improving the automation level and processing efficiency of routine inspection operations.

[0027] Secondly, this application provides an adaptive inspection device for a gauge, the gauge including a positioning part for adapting to and characterizing the contour shape of various parts to be inspected, the device comprising: The contour data acquisition module is used to identify the contour data of the part to be inspected when it is placed on the positioning part of the fixture. The judgment module is used to determine whether the contour data matches the pre-stored component contour library; The type marking and parameter detection module is used to mark the type of unknown parts corresponding to the contour data and collect the detection parameters of the unknown parts when the contour data does not match the pre-stored part contour library; The qualified parts set determination module is used to determine the qualified parts set among the unknown parts by comparing the detection parameters of each unknown part horizontally when the number of unknown parts of the same type is greater than a preset threshold, and by using a clustering algorithm. The pre-stored component contour library import module is used to determine the standard parameters of a component type based on the detection parameters of the qualified parts set among the unknown components, and then import the contour data corresponding to the component type into the pre-stored component contour library after associating it with the standard parameters.

[0028] Thirdly, this application provides an electronic device, comprising: processor; Memory; And a computer program, wherein the computer program is stored in the memory, the computer program including instructions that, when executed by the processor, cause the electronic device to perform the method described in any one of the first aspects.

[0029] Fourthly, this application provides a computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device on which the computer-readable storage medium is located to perform the method described in any one of the first aspects.

[0030] Understandably, the adaptive detection device for the gauge provided in the second aspect, the electronic device provided in the third aspect, and the computer-readable storage medium provided in the fourth aspect are all used to perform some or all of the methods provided in this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here. Attached Figure Description

[0031] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 This is a schematic diagram of an application scenario provided by an embodiment of this application.

[0033] Figure 2 This is a schematic diagram of the structure of an adaptive detection method for a gauge provided in an embodiment of this application.

[0034] Figure 3 This is a partial structural diagram of an inspection tool provided in an embodiment of this application.

[0035] Figure 4 This is a partial structural schematic diagram of another inspection tool provided in an embodiment of this application.

[0036] Figure 5 This is a schematic diagram of the structure of an adaptive detection device for a gauge provided in an embodiment of this application.

[0037] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0038] To better understand the technical solution of this application, the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0039] It should be understood that the described embodiments are merely some, not all, of the embodiments in this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive effort are within the scope of protection of this application.

[0040] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. The singular forms “a,” “the,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0041] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0042] Inspection fixtures are indispensable specialized inspection tools in industrial production, especially in precision manufacturing fields such as automobiles and aerospace. Their core function is to replace general-purpose measuring tools, enabling rapid and efficient conformity assessments of specific dimensions, shapes, contours, and positional relationships of products. Unlike high-precision but complex equipment such as coordinate measuring machines (CMMs), inspection fixtures are designed to serve the production floor. They do not aim to output specific measurement values, but rather to quickly provide qualitative conclusions of "qualified" or "unqualified," thus providing immediate quality feedback for fast-paced assembly line operations. They are a crucial link in controlling product manufacturing quality and ensuring assembly interchangeability. To facilitate understanding, specific application scenarios will be illustrated below.

[0043] See Figure 1 This is a schematic diagram illustrating an application scenario provided by an embodiment of this application. For example... Figure 1 As shown, this application scenario includes: a door trim panel skeleton gauge 100 and a door trim panel skeleton 103 to be inspected. Specifically, the door trim panel skeleton gauge 100 includes: a positioning pin 101 and a clamping device 102.

[0044] It should be noted that, Figure 1 The type of fixture shown is merely illustrative and this application does not impose any specific limitations on it. Furthermore, fixture 100 may also include other components, which are not detailed here for the sake of brevity. Figure 1 As shown in the image.

[0045] In practical applications, the operator places the door trim panel rib 103 to be inspected on the door trim panel rib inspection fixture 100 and fixes it with the positioning pin 101, clamping device 102 and other structures on the fixture. Then, the operator uses measuring tools such as dial indicator and plug gauge to inspect multiple pre-set measuring points on the part one by one and compares the readings with the standard tolerance range to determine whether the part is qualified.

[0046] Understandably, the entire inspection process for the door trim panel frame relies on a door trim panel frame fixture 100 specifically designed and manufactured for this part, and its inspection items and standards are fixed after the fixture is manufactured.

[0047] However, with the accelerating pace of product updates and the growing demand for personalized customization, the variety of parts on production lines is increasing. Even with inspection fixtures capable of detecting multiple types of parts, manual assistance is still required. Specifically, when a new part not in the pre-stored part contour library appears, the fixture cannot automatically identify its contour, nor can it autonomously decide how to inspect it. Manual intervention is necessary to measure, analyze, and manually update the part contour library and inspection program. This process is time-consuming and labor-intensive, severely restricting the production line's rapid response capability and level of intelligence.

[0048] To address the aforementioned issues, in this embodiment, the inspection fixture is equipped with a positioning section, which adapts to and characterizes the contour shapes of various parts to be inspected. When the part to be inspected is placed on the positioning section of the fixture, its contour data is identified. When the contour data of the part to be inspected cannot be matched in the pre-stored part contour library, the system does not simply issue an alarm or reject the part, but actively marks it as an unknown type and initiates a learning process. By setting a quantity threshold and using a clustering algorithm to compare the parameters of unknown parts of the same type laterally, the system can automatically select a set of qualified products from a batch of new parts, thereby objectively establishing the standard parameters of the new part. Finally, the newly identified contour data is associated with the self-generated standard parameters and imported into the database, completing the automatic update of the knowledge base. It can be understood that the above entire process transforms the inspection fixture from a passive inspection tool into an intelligent system that can actively accumulate knowledge and continuously expand its inspection range, fundamentally solving the pain point of frequent manual intervention to update the inspection fixture database due to part updates, and achieving dynamic and accurate matching between inspection capabilities and production needs under the premise of uninterrupted production continuity. Specifically, detailed descriptions are provided below in conjunction with the accompanying drawings and specific embodiments.

[0049] See Figure 2 This is a schematic diagram of the structure of an adaptive detection method for a gauge provided in an embodiment of this application. This method can be applied to... Figure 1 In the application scenarios shown, such as Figure 2 As shown, it mainly includes the following steps.

[0050] Step S201: When the part to be inspected is placed on the positioning part of the fixture, the contour data of the part to be inspected is identified.

[0051] In this embodiment, the inspection fixture includes a positioning part, which is used to adapt to and characterize the contour shape of various parts to be inspected. When the part to be inspected is placed on the positioning part of the inspection fixture, the contour information of the part can be automatically captured by the physical structure of the inspection fixture.

[0052] Specifically, in one possible implementation, the positioning unit includes a plate and multiple movable elements, wherein the movable elements movably protrude from the upper surface of the plate. When the part to be inspected is placed on the upper surface of the plate, the contour data of the part to be inspected is identified based on the position of the movable elements.

[0053] It is understood that this positioning part is not a fixed surface in the traditional sense, but rather consists of a plate body covered with movable elements. These movable elements protrude uniformly from the plate surface in their natural state, forming a flexible contact interface with a basic supporting plane. For easier understanding, see [link to relevant documentation]. Figure 3 The figure shows a partial structural diagram of a gauge provided in an embodiment of this application. As shown in the figure, the plate 301 and multiple movable elements 302 are illustrated.

[0054] When an operator places a component to be inspected, such as an irregularly shaped vehicle door panel, onto the positioning part, the bottom contour of the door panel will come into contact with the densely arranged movable elements 302 below. Under the weight of the component itself, those movable elements 302 in contact with the solid part of the component will be pressed down, while the surrounding movable elements 302 that are not in contact will maintain their original protruding height. For easier understanding, see [link to documentation]. Figure 4 This is a partial structural schematic diagram of another inspection fixture provided in an embodiment of this application. As shown in the figure, the figure shows a plate 301, multiple movable elements 302, and a component 400 to be inspected.

[0055] It should be pointed out that, Figure 3 and Figure 4 The diagram shows a partial structural schematic of the inspection tool and Figure 4 The component 400 shown is merely a functional example and is not intended to impose any specific limitations on it.

[0056] The physical process described above causes the three-dimensional bottom contour of the part to be instantly and accurately "transferred" into a relative height distribution map of the group of movable elements 302 on the surface of the positioning part, that is, a contour pattern formed by a recessed area composed of pressed-down elements.

[0057] In this embodiment, when a component is placed on the plate 301, its contact surface presses down on the corresponding movable element 302, and the positions of each element naturally form a spatial mapping of the bottom contour of the component. This combination of mechanical structure and detection principle allows components of different shapes to automatically generate corresponding contour data through the same set of devices, which not only eliminates the need for specially manufactured matching molds in traditional inspection tools, but also provides accurate and reliable contour information input for subsequent intelligent recognition and learning processes.

[0058] Contour data identification can be accomplished through a sensing system. In one possible implementation, when the component 400 to be inspected is placed on the upper surface of the plate 301, contour data is acquired based on the position data of the movable element 302 collected by a vision sensor and / or a pressure sensor array.

[0059] Specifically, a vision sensor is mounted above the positioning section, acting like a high-speed camera to capture images of the entire surface of the positioning section from top to bottom. Through image processing technology, the system can clearly identify which areas of the components have been pressed down, thereby quickly extracting the part contour data outlined by the boundaries of these pressed components. Simultaneously, a miniature pressure sensor can be integrated into the bottom of each movable component 302, triggering a signal when the component is pressed down. These pressure sensors form a pressure sensor array. By reading the pressure status signals of all components, the system can reconstruct the precise shape of the part contour in digital space.

[0060] In this embodiment, the vision sensor can quickly capture the overall image of the part's contour, avoiding wear and errors caused by physical contact; while the pressure sensor array accurately reconstructs the contact surface morphology of the part by sensing the force state of each movable element 302. These two approaches not only enhance adaptability to complex contours but also provide an accurate and reliable digital foundation for subsequent intelligent recognition and data analysis by directly converting physical shapes into electronic signals, effectively improving the automation level and data accuracy of the entire detection system.

[0061] In practical applications, contour data can be acquired simultaneously using both a vision sensor and a pressure sensor array. In one possible implementation, the contour data of the movable element 302 acquired by the vision sensor is compared with the contour data of the movable element 302 acquired by the pressure sensor array to obtain contour difference data; it is then determined whether the contour difference data is less than or equal to a preset difference threshold; when the contour difference data is less than or equal to the preset difference threshold, the position data of the movable element 302 acquired by the vision sensor and the contour data of the movable element 302 acquired by the pressure sensor array are fused to obtain the contour data of the component 400 to be inspected.

[0062] Understandably, in this embodiment, the system does not simply select data from a single sensor. Instead, it first performs a detailed comparison of the contour data acquired by the vision sensor and the contour data obtained by the pressure sensor array. This step aims to identify inconsistencies that may arise when the two types of sensors perceive the same physical phenomenon. For example, when detecting a vehicle door panel with a complex curvature, the vision sensor may misjudge the height of a movable element 302 due to light reflection, while the pressure sensor may experience uneven force due to slight tilting of the part. This comparison process quantifies these perceptual differences.

[0063] Next, the system compares the calculated contour difference data with a preset difference threshold. This threshold is pre-calibrated based on sensor accuracy and system fault tolerance requirements. When the difference is within the threshold range, it indicates that both types of sensor data are relatively reliable, and the system initiates data fusion processing. Specifically, for example, when processing the contour of an engine block part, the system assigns different weights to visual data and pressure data. These weights can be set empirically or dynamically allocated using a weight prediction model trained on historical data. Specifically, using a machine learning model trained on historical detection data, such as a regression model, the system dynamically calculates and allocates the optimal weights based on the real-time characteristics of the current data, such as signal-to-noise ratio and confidence level.

[0064] After obtaining the weights, the system processes the original data using data fusion algorithms such as weighted averaging or Kalman filtering, and finally outputs an optimized contour dataset.

[0065] Ultimately, this contour dataset, generated through data fusion processing, was used to reconstruct the contour data of the 400 parts to be inspected. It can be understood that the above mechanism essentially constructs a mutually verifying and complementary perception system. It effectively suppresses the perception errors of a single sensor, thereby outputting a contour model with stronger anti-interference capabilities and richer details, providing a more solid and reliable data foundation for subsequent part recognition and autonomous learning.

[0066] Conversely, if the positional contour difference data is greater than or equal to the preset difference threshold, it indicates a significant discrepancy between the perception results of the visual sensor and the pressure sensor array, suggesting that at least one of them may have experienced transient interference or malfunction. In this case, the system will not perform data fusion and will automatically trigger the exception handling mechanism.

[0067] In its specific implementation, the anomaly handling mechanism includes: first, recording the current abnormal data and generating a system alarm to prompt the operator to intervene and check; second, the system can control the actuator to move the currently inspected part 400 out of the workstation, and can initiate one or more repeated inspections of the workstation according to a preset strategy to identify and eliminate instantaneous random interference; if the discrepancy still exists after repeated inspections, the system can determine that the sensor system has a persistent fault and request maintenance.

[0068] In practical applications, a spring is typically provided between the movable element 302 and the plate 301. After the part to be inspected 400 is removed, the spring force causes the movable element to return to its protruding position on the upper surface. Alternatively, a flexible latch is provided between the movable element 302 and the plate 301. When subjected to force, the flexible latch deforms, allowing the movable element 302 to move from top to bottom. A reset plate is provided on the lower surface of the positioning plate 301. After the part is removed, the reset plate is moved from bottom to top by manual or related automatic drive devices to fit the lower surface of the plate 301, causing the flexible latch to deform and the movable element protruding from the lower surface to return to its protruding position on the upper surface.

[0069] After prolonged continuous use, the movable element 302 in the positioning section may fail to fully reset to its initial zero position due to mechanical wear, spring fatigue, or impurity intrusion. If this minute reset deviation goes undetected and uncorrected, it will directly lead to systematic errors in subsequent contour recognition data. For example, if an element that should be fully depressed to the zero position only rebounds to a position a certain distance above the reference surface due to fatigue, the system will incorrectly determine that there is a physical support for the part at that position, resulting in a non-existent "protrusion" or boundary distortion in the reconstructed contour data. To ensure the accuracy and reliability of the detection system during long-term operation, this application embodiment further introduces a preprocessing procedure—namely, the calibration of the fixture's own condition.

[0070] Specifically, before placing the part to be inspected 400 into the positioning part of the fixture, the method further includes: performing a depth scan on the movable element 302 in the positioning part to obtain the reset deviation value of each movable element in the current state; and compensating and correcting the contour data based on the reset deviation value.

[0071] Specifically, a laser depth scanner mounted above the positioning section rapidly scans the top height of all movable components 302 on the positioning section plate 301, generating an actual height distribution map of each component in the current state. The system compares this distribution map with a pre-stored reference height map obtained when the fixture is intact, thereby accurately calculating the current reset deviation value of each movable component 302.

[0072] Based on this reset deviation value, the system can establish a dynamic data compensation mechanism. When performing part contour recognition later, the system will subtract the measured reset deviation value of each component position from the original acquired contour data.

[0073] For example, if the scan finds that the components in a certain area generally have a deviation that is not completely reset, then when recognizing the outline of the door panel later, the system will intelligently compensate and correct the height data collected in that area, thereby filtering out the error introduced by the state of the inspection tool itself and restoring the true and accurate outline shape of the door panel.

[0074] In this embodiment, a depth scan of the reset state of the movable element 302 is performed before inspecting the part, which can promptly detect reset deviations caused by spring fatigue or mechanical wear. Based on this deviation, the subsequently acquired contour data is compensated and corrected in real time, establishing a dynamic calibration benchmark for the inspection system. This design enables the fixture to self-diagnose and correct the accuracy loss caused by mechanical wear, ensuring the accuracy of the contour data from the source. This provides a reliable data foundation for subsequent part identification, cluster analysis, and standard establishment, significantly improving the stability of the entire system and the reliability of the inspection results during long-term use.

[0075] Of course, in one possible implementation, the positioning unit also includes multiple piezoelectric ceramic micro-actuators, which are correspondingly disposed below the movable element 302. A depth scan is performed on the movable element 302 in the positioning unit to obtain the reset deviation value of each movable element in the current state; based on the reset deviation value, the piezoelectric ceramic micro-actuators are controlled to perform position compensation on the movable element 302.

[0076] Specifically, based on the measured reset deviation value, the control unit sends a precise drive signal to the piezoelectric ceramic micro-actuator below the movable element 302 where the deviation exists. Piezoelectric ceramic materials undergo minute deformation under voltage; utilizing this characteristic, the actuator can push the movable element 302 above it with nanometer-level precision to make micro-displacements. For example, when a scan detects that a component in a certain area has experienced spring fatigue due to long-term load-bearing, resulting in an overall low reset height, the system will control the corresponding piezoelectric ceramic micro-actuator to extend by the appropriate amount, precisely lifting the component to the standard zero-position height.

[0077] Understandably, the above-described method and embodiments eliminate systematic errors accumulated due to factors such as mechanical wear and fatigue at the physical level, ensuring that the positioning unit can provide an absolutely accurate reference plane at the beginning of each inspection, providing the highest precision physical basis for the subsequent acquisition of contour data, and greatly improving the long-term measurement stability and lifespan of the gauge.

[0078] After detailing the reset function of the movable element 302, this application further discloses a precision locking mechanism for improving the operational stability of the positioning part. This mechanism aims to solve the problem of minute displacement that may occur in the movable element 302 after reset during subsequent testing.

[0079] The core of this precision locking mechanism lies in the electromagnetic locking unit configured for each movable element 302. When the system returns the movable element 302 to its initial protruding position via the reset mechanism, the electromagnetic locking unit is immediately energized and activated, generating a strong electromagnetic attraction or mechanical locking action on the movable element 302, locking the movable element 302 to its current position.

[0080] Among the two methods, electromagnetic attraction involves the unit generating a strong magnetic field upon power-up, which directly attracts and fixes the movable element 302 to its guide shaft or a specific attraction surface. Mechanical locking involves an electromagnetic force driving a miniature mechanical structure (such as a pin, chuck, or wedge), which creates a direct, rigid mechanical interference, physically restricting the degrees of freedom of the movable element 302. For example, an electromagnetically driven pin can precisely insert into its pinhole after the element is reset; or an electromagnetically driven wedge can extend and press against the locking surface of the element. This method typically maintains the locked state even after the system is powered off, offering stronger resistance to vibration interference.

[0081] For example, when inspecting a large vehicle side panel, the part may experience slight impacts or vibrations during placement. Without a locking function, the partially reset movable element 302 may experience slight, unintended sinking or wobbling, causing its protrusion height to no longer precisely equal to its initial zero position. This deviation is sufficient to introduce noise in subsequent contour recognition and even distort contour features. The electromagnetic locking unit eliminates this possibility at the physical level, acting as if it applies a "stay in place" command to each element, making it an absolutely stable reference point during data acquisition.

[0082] By introducing an electromagnetic locking unit, measurement errors caused by the instability of the components themselves are fundamentally eliminated, providing a highly reliable physical basis for visual sensors or pressure sensor arrays. This ensures that the acquired contour data has extremely high repeatability and accuracy, laying a solid technical foundation for subsequent intelligent recognition and learning.

[0083] Step S202: Determine whether the contour data matches the pre-stored component contour library.

[0084] In this embodiment, the system rapidly compares the collected contour data with digital contour templates of various standard parts pre-stored in a pre-stored parts contour library. It is understood that this comparison does not require pixel-level perfect consistency, but rather uses a specific contour comparison algorithm to calculate the geometric similarity between the current contour and each standard template. The system presets a similarity threshold; when the similarity between the current contour and a standard contour in the library exceeds this threshold, it is determined to be a "match"; conversely, if the similarity with all standard contours is below the threshold, it is determined to be a "mismatch".

[0085] For example, suppose the pre-stored parts outline library already contains standard outlines for "vehicle front door inner panel" and "vehicle hood". When an operator introduces a completely new part not recorded in the library, such as a newly designed "vehicle fender", and its outline data cannot achieve sufficient similarity with any template in the library, the system will determine it as a "mismatch". This judgment indicates that the system has identified an unknown type of part, thus automatically triggering the subsequent autonomous learning process, rather than simply reporting an error or stopping operation.

[0086] Conversely, if the part being inspected is a vehicle front door inner panel, the system identifies its contour data and compares it with templates in the library. Due to the high similarity in shape, the contour data will successfully match the "vehicle front door inner panel" template. At this point, the system confirms the known identity of the part and can immediately call upon the associated standard inspection parameters to perform subsequent rapid conformity checks.

[0087] Furthermore, in one possible implementation, when the contour data matches the pre-stored component contour library, the detection parameters corresponding to the component 400 to be inspected are obtained based on the standard parameter types in the pre-stored component contour library; based on the detection parameters corresponding to the component 400 to be inspected and the corresponding standard parameters, the component 400 to be inspected is subjected to a conformity test.

[0088] In this embodiment, when the system identifies a part as a known type, it can immediately retrieve the corresponding standard parameters and inspection scheme from the pre-stored part contour library, automatically completing the entire process from parameter acquisition to conformity judgment. This knowledge-based rapid response mechanism enables the inspection of known parts to be executed accurately without any manual configuration, fully leveraging the guiding value of pre-stored standards while ensuring the standardization and consistency of the inspection process, significantly improving the automation level and processing efficiency of routine inspection operations.

[0089] Step S203: When the contour data does not match the pre-stored component contour library, mark the type of the unknown component corresponding to the contour data and collect the detection parameters of the unknown component.

[0090] In this embodiment, after determining that the contour data of the component 400 to be inspected does not match the pre-stored component contour library, the system enters the cognitive learning stage for the new component. This step first requires establishing an identity for this new model and comprehensively collecting its key feature data.

[0091] The system creates a temporary type tag for this unrecognized component. This tag can be a descriptive name, such as "Unknown Type - Door Structural Component," and is associated with its unique profile data. For example, when a newly designed vehicle battery box tray first appears on the production line, its unique rectangular, ribbed profile cannot match any existing door or frame template in the library. The system will then tag it as "Unknown Type I - Chassis Structural Component" and begin recording its characteristics.

[0092] Regarding the definition of "type" in step S203, this application embodiment further explains that the "type" is automatically classified based on the shape features of the contour data (such as contour curvature, key point positions, geometric dimensions, etc.) using a similarity algorithm. Specifically, the system calculates the similarity between the newly acquired contour data of the unknown component and the contour data corresponding to all existing temporary type markers; if the similarity with a certain existing type exceeds a preset threshold, it is classified into that type; otherwise, the system creates a new temporary type marker and uses the contour data as the baseline feature of that type. After completing the temporary marker, the system will initiate the depth detection parameter acquisition process for the unknown component. At this time, the control unit will instruct the scanning mechanism of the detection department to perform a comprehensive automated measurement of the component 400 to be inspected.

[0093] In one possible implementation, the detection parameters typically include the profile of the critical surface, the position of each mounting hole group, the height of the positioning datum, and the verticality of the sidewalls.

[0094] It is understandable that the surface profile reflects the overall shape accuracy of the part, the positional accuracy ensures the accuracy of its mounting reference, and the height and perpendicularity together control the spatial orientation of the part.

[0095] In the specific example of the battery box tray, the scanning system will accurately measure a series of key geometric parameters, such as the flatness of its mounting surface, the relative positions of each bolt mounting hole, and the overall height of the entire tray.

[0096] In this embodiment, the combination of the above-mentioned detection parameters covers the most critical dimensional requirements of the parts during assembly, providing rich and accurate data features for subsequent cluster analysis. This enables the system to establish a more complete and reliable standard parameter model, thereby ensuring the comprehensiveness and accuracy of the conformity judgment.

[0097] Step S204: When the number of unknown parts of the same type is greater than the preset threshold, the detection parameters of each unknown part are compared horizontally, and the set of qualified parts among the unknown parts is determined by clustering algorithm.

[0098] In practical applications, the system continuously tracks the number of unknown parts marked with the same temporary type. When the number of unknown parts of the same type exceeds a preset threshold, for example, when 20 parts with the same "Unknown Type I - Chassis Structural Part" label appear on the production line, the system will automatically trigger the subsequent intelligent analysis process.

[0099] Understandably, this threshold mechanism ensures that the analysis is based on a sufficient sample of data, avoiding erroneous conclusions that are not representative due to accidental manufacturing deviations or measurement errors of a single part.

[0100] Then, the system will perform a comparative analysis of all detection parameters for this batch of unknown parts of the same type. The "comparative analysis" mentioned here refers to a group parameter comparison analysis of multiple unknown parts marked as the same temporary type within the same production batch context.

[0101] Specifically, this comparison is not a time-series tracking, but rather focuses on examining the distribution of all part parameters (such as contour, position, and height) within a set of unknown parts of the same temporary type, treating all the detection parameters of each part as a data direction in a multi-dimensional space. The aim is to identify, from a batch of samples, the group of parts with the highest clustering in the parameter space that represents the current stable manufacturing level.

[0102] After conducting a horizontal comparative analysis of all the test parameters of the same type of unknown parts in a batch, clustering algorithms will be used to further explore these multi-dimensional test parameters.

[0103] Understandably, clustering algorithms can automatically group parts with similar parameter characteristics into the same cluster. Taking the inner panel of a car door as an example, the algorithm will identify groups of parts with highly concentrated parameter values ​​in a data space consisting of multiple parameters such as contour, position, and height, and classify them into a main cluster.

[0104] The main clusters of highly concentrated parameters identified by the algorithm are then classified by the system as the "set of qualified parts." Under stable production conditions, the vast majority of parts conforming to design specifications should exhibit a natural clustered distribution of their inspection parameters. Individual parts with significantly deviated parameters due to manufacturing defects or random errors are automatically identified as outliers and excluded from the qualified set. This process is entirely data-driven, requiring no manual pre-setting of tolerance ranges, thus achieving an intelligent transformation of inspection standards from manually defined to adaptive calibration.

[0105] In determining the set of qualified parts through group analysis, the choice of clustering algorithm directly affects the accuracy and adaptability of the analysis results. Therefore, in one possible implementation, K-means or DBSCAN clustering algorithms are used to calculate the detection parameters, and clusters with concentrated parameter distributions are identified as the set of qualified parts.

[0106] It's important to note that clustering algorithms aim to reveal the inherent distribution structure of data, and their results may form one or more clusters. In this case, the system does not consider all clusters as acceptable; instead, it identifies the truly acceptable set of parts based on a predefined strategy. Specifically, when multiple clusters exist, the cluster with the largest number of data points and the densest distribution is identified as the acceptable set of parts. This is based on the principle that most parts produced under stable manufacturing conditions should converge within the dominant quality level range. Alternatively, the center or boundary of the cluster is compared with the known product tolerance range, and clusters that fall entirely within the acceptable tolerance zone are determined to be the acceptable set of parts.

[0107] Of course, a combined judgment principle can also be adopted, that is, combining the two principles mentioned above. For example, first select the clusters that fall within the tolerance range, and then select the cluster with the densest data points from them as the final qualified set.

[0108] Among them, the K-means algorithm is suitable for scenarios where the parameter distribution is relatively concentrated and the number of clusters is clear. This system projects all detection parameters of unknown parts of the same type into a multi-dimensional feature space and finds the central region with the densest data points through iterative calculation. For example, when analyzing the height and contour parameters of a batch of new car door inner panels, the algorithm automatically identifies the core cluster formed by the aggregation of most part parameters. This core cluster represents the most stable output quality state under the current manufacturing process level, and is thus confirmed as a set of qualified parts.

[0109] For cases with more complex parameter distributions or obvious outliers, the DBSCAN algorithm can be used. The advantage of this algorithm lies in its ability to automatically identify clusters of arbitrary shapes and its robustness to discrete outliers. When inspecting a batch of more complex engine mounts, their positional parameters may exhibit a non-spherical distribution pattern. The DBSCAN algorithm can accurately capture this complex distribution pattern and automatically mark defective parts that significantly deviate from the main group as noise points, ensuring the purity of the obtained set of qualified parts.

[0110] In this embodiment, the application of the two algorithms described above enables the system to objectively determine the main clusters that conform to manufacturing rules as the set of qualified parts based on the natural clustering characteristics of part parameters, avoiding subjective bias caused by relying on human experience to set fixed thresholds. This data-driven method not only ensures that the qualification standard originates from actual production, but also adaptively follows subtle changes in the manufacturing process, thereby continuously maintaining the accuracy and reliability of the judgment standard.

[0111] Step S205: Based on the detection parameters of the qualified parts set in the unknown parts, determine the standard parameters of this type of parts, and after associating the contour data of this type of parts with the standard parameters, import them into the pre-stored parts contour library.

[0112] In this embodiment, the system first performs statistical analysis on the detection parameters of all samples that have been classified into the qualified parts set. Specifically, the system calculates the statistical characteristics of these qualified samples in each detection parameter dimension, thereby generating a set of standard parameters representing the quality standard of this type of part.

[0113] For example, for the set of qualified parts previously marked as "Unknown Type I - Chassis Structural Components", the system will calculate the typical range of its mounting plane profile, the concentrated area of ​​each mounting hole position, and the common value range of its overall height. These statistical results together constitute the standard parameters of the battery box tray.

[0114] Then, the system binds the initially collected contour data, which uniquely identifies this type of part, to the newly established standard parameters. Through this association, the contour data is no longer just a graphic for identification, but becomes the sole key to retrieve the entire set of inspection standards from the pre-stored part contour library.

[0115] Finally, the system associates the contour data corresponding to this type of component with standard parameters and imports it into the pre-stored component contour library.

[0116] When the same type of part 400 appears on the production line next time, the system can immediately identify its identity through contour matching and automatically call the standard parameters established in this learning process to perform accurate and efficient conformity testing.

[0117] More importantly, the "perception-decision-execution" intelligent closed loop constructed by this system can reverse the analysis results of detection data and apply them to the production line to achieve continuous optimization of the manufacturing process. Specifically, during long-term monitoring and cluster analysis, when the system finds that the pass rate of a specific dimensional parameter (such as the position accuracy of mounting holes) of a known type of part is consistently low, and the parameter distribution of non-conforming products is significantly concentrated on one side of the tolerance zone, the system can automatically analyze potential systematic deviations in the equipment. Based on this, the system will generate specific equipment adjustment suggestions (e.g., "It is recommended to positively compensate the B-axis coordinate of the drilling machine by 0.05 mm"), and automatically send them to the manufacturing execution system through an interface, directly guiding operators to adjust the relevant processing equipment, thereby eliminating batch quality risks at the source.

[0118] Meanwhile, the system possesses early warning capabilities for production process drift. For known types of parts already entered into the pre-stored component contour library, the system continuously monitors the statistical changes in the center position and dispersion of all qualified clusters of its detection parameters while performing routine conformity checks. When trend analysis reveals that the center position of a certain parameter cluster is undergoing statistically significant unidirectional drift, even if all produced parts are still within the acceptable tolerance zone, the system will issue an early process capability warning, indicating the need for preventative maintenance. For example, if the system detects that the center of the qualified cluster for the hood height parameter is slowly and continuously approaching the upper tolerance limit, it will issue an early warning that the stamping die may be at risk of accelerated wear, thereby enabling predictive maintenance and preventing future batch deviations.

[0119] In one possible implementation, after importing the pre-stored component contour library, the method further includes: using the updated pre-stored component contour library to re-acquire the contour data of the unknown component of this type; comparing the contour data of the unknown component of this type with the corresponding standard parameters in the pre-stored component contour library, and determining whether the unknown component is qualified based on the comparison result.

[0120] Understandably, by re-inspecting previously marked unknown parts, not only are the "legacy issues" generated during the learning process effectively resolved, and the backlog and waste of these parts avoided, but more importantly, this step substantially verifies the accuracy and practicality of the newly established standard parameters, ensuring that the system's learning results can be transformed into reliable productivity in an instant, and significantly improving the continuity and intelligence level of the entire inspection process.

[0121] In this embodiment, when the contour data of the component 400 to be inspected cannot be matched in the pre-stored component contour library, the system does not simply issue an alarm or reject the component, but actively marks it as an unknown type and initiates a learning process. By setting a quantity threshold and using a clustering algorithm to compare the parameters of unknown parts of the same type horizontally, the system can automatically filter out a set of qualified products from a batch of new parts, thereby objectively establishing the standard parameters of the new part. Finally, the newly identified contour data is associated with the self-generated standard parameters and imported into the database, completing the automatic update of the knowledge base. It can be understood that the above-mentioned entire process transforms the inspection tool from a passive inspection tool into an intelligent system that can actively accumulate knowledge and continuously expand its inspection range, fundamentally solving the pain point of needing frequent manual intervention to update the inspection tool database due to component updates, and realizing a dynamic and accurate match between inspection capabilities and production needs under the premise of uninterrupted production continuity.

[0122] Corresponding to the above embodiments, this application also provides an adaptive detection device for a gauge. Specifically, see [link to relevant documentation]. Figure 5 This is a schematic diagram of the structure of an adaptive detection device for a gauge provided in an embodiment of this application. As shown in the figure, the adaptive detection device 500 for the gauge is illustrated. Specifically, the adaptive detection device 500 for the gauge includes: a contour data acquisition module 501, a judgment module 502, a type marking and parameter detection module 503, a qualified parts set determination module 504, and a pre-stored parts contour library import module 505. The module includes the following components: a contour data acquisition module 501, which identifies the contour data of the part to be inspected when it is placed on the positioning part of the fixture; a judgment module 502, which judges whether the contour data matches the pre-stored part contour library; a type marking and parameter detection module 503, which marks the type of the unknown part corresponding to the contour data and collects the detection parameters of the unknown part when the contour data does not match the pre-stored part contour library; a qualified parts set determination module 504, which determines the qualified parts set among the unknown parts by comparing the detection parameters of each unknown part horizontally and using a clustering algorithm when the number of unknown parts of the same type exceeds a preset threshold; and a pre-stored part contour library import module 505, which determines the standard parameters of the part of that type based on the detection parameters of the qualified parts set among the unknown parts, and imports the contour data of the part of that type into the pre-stored part contour library after associating it with the standard parameters.

[0123] For specific details, please refer to the embodiments described above. For the sake of brevity, this application will not elaborate further. Corresponding to the above embodiments, this application also provides a schematic diagram of the structure of an electronic device. See also Figure 6This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device 600 may include a processor 601, a memory 602, and a communication unit 603. These components communicate through one or more buses. Those skilled in the art will understand that the structure of the electronic device shown in the figure does not constitute a limitation on the embodiment of this application. It may be a bus topology or a star topology, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0124] The communication unit 603 is used to establish a communication channel, enabling the electronic device to communicate with other devices. It receives user data from other devices or sends user data to other devices.

[0125] The processor 601 serves as the control center of the electronic device, connecting various parts of the device via various interfaces and lines. It executes software programs, instructions, and / or modules stored in the memory 602, and calls data stored in the memory to perform various functions and / or process data. The processor may be composed of integrated circuits (ICs), such as a single packaged IC or multiple packaged ICs with the same or different functions connected together. For example, the processor 601 may consist only of a central processing unit (CPU). In this embodiment, the CPU may have a single processing core or include multiple processing cores.

[0126] The memory 602 is used to store the execution instructions of the processor 601. The memory 602 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0127] When the execution instructions in memory 602 are executed by processor 601, the electronic device 600 is able to perform operations. Figure 2 Some or all of the steps in the illustrated embodiments.

[0128] In a specific implementation, this application also provides a computer storage medium, wherein the computer storage medium may store a program, and when the program is executed, it may include some or all of the steps of the simulation scene generation method provided in this application. The storage medium may be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0129] In this application embodiment, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, the simultaneous existence of A and B, or the existence of B alone. A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects have an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, and c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0130] Those skilled in the art will recognize that the units and algorithm steps described in the embodiments disclosed herein can be implemented using electronic hardware, computer software, or a combination of electronic hardware and software. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0131] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0132] In the several embodiments provided in this application, any function, if implemented as a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0133] The same or similar parts between the various embodiments in this specification can be referred to mutually. In particular, the device embodiments and terminal embodiments are basically similar to the method embodiments, so the description is relatively simple, and the relevant parts can be referred to the description in the method embodiments.

Claims

1. An adaptive inspection method for a gauge, characterized in that, The inspection fixture includes a positioning part, which is used to adapt to and characterize the contour shape of various parts to be inspected. The method includes: When the part to be inspected is placed on the positioning part of the fixture, the contour data of the part to be inspected is identified; Determine whether the contour data matches the pre-stored component contour library; When the contour data does not match the pre-stored component contour library, the type of the unknown component corresponding to the contour data is marked, and the detection parameters of the unknown component are collected. When the number of unknown parts of the same type exceeds a preset threshold, the detection parameters of each unknown part are compared horizontally, and a clustering algorithm is used to determine the set of qualified parts among the unknown parts. Based on the detection parameters of the qualified parts set among the unknown parts, the standard parameters of this type of part are determined, and the contour data corresponding to this type of part is associated with the standard parameters and then imported into the pre-stored part contour library.

2. The method according to claim 1, characterized in that, The positioning part includes a plate and a plurality of movable elements, wherein the movable elements movably protrude from the upper surface of the plate; The step of identifying the contour data of the component to be inspected when it is placed on the positioning part of the fixture includes: identifying the contour data of the component to be inspected based on the position of the movable element when it is placed on the upper surface of the plate.

3. The method according to claim 2, characterized in that, The positioning unit also includes a visual sensor or a pressure sensor array; When the component to be inspected is placed on the upper surface of the plate, the contour data of the component to be inspected is identified based on the position of the movable element, including: when the component to be inspected is placed on the upper surface of the plate, the contour data is obtained based on the position data of the movable element collected by the vision sensor and / or the pressure sensor array.

4. The method according to claim 3, characterized in that, The process of obtaining the contour data based on the position data of the movable element collected by the vision sensor and the pressure sensor array includes: The contour data of the movable element collected by the vision sensor is compared with the contour data of the movable element collected by the pressure sensor array to obtain contour difference data. Determine whether the contour difference data is less than or equal to a preset difference threshold; When the contour difference data is less than or equal to the preset difference threshold, the contour data of the movable element collected by the vision sensor and the contour data of the movable element collected by the pressure sensor array are fused together to obtain the contour data of the component to be inspected.

5. The method according to claim 2, characterized in that, Before placing the component to be inspected into the positioning part of the fixture, the method further includes: performing a depth scan on the movable elements in the positioning part to obtain the reset deviation value of each movable element in the current state; Before determining whether the contour data matches the pre-stored component contour library, the method further includes: compensating and correcting the contour data based on the reset deviation value.

6. The method according to claim 2, characterized in that, The positioning section further includes multiple piezoelectric ceramic micro-actuators, which are correspondingly disposed below the movable element; before placing the part to be inspected into the positioning section of the fixture, it also includes: A depth scan is performed on the movable elements in the positioning part to obtain the reset deviation value of each movable element in the current state; Based on the reset deviation value, the piezoelectric ceramic micro actuator is controlled to adjust the position of the corresponding movable element, and the position of the movable element is compensated.

7. The method according to claim 1, characterized in that, The step of determining the set of qualified parts among the unknown components using a clustering algorithm includes: The detection parameters are calculated using K-means or DBSCAN clustering algorithms, and the clusters in the parameter distribution set are identified as the set of qualified parts.

8. The method according to claim 1, characterized in that, After associating the contour data corresponding to this type of component with the standard parameters and importing it into the pre-stored component contour library, the method further includes: Using the updated pre-stored component contour library, the contour data of this type of unknown component is re-acquired; The contour data of the unknown component of this type is compared with the corresponding standard parameters in the pre-stored component contour library, and the unknown component is judged to be qualified based on the comparison result.

9. The method according to claim 1, characterized in that, The detection parameters include surface profile, position, height, and perpendicularity.

10. An adaptive detection device for a gauge, characterized in that, The inspection fixture includes a positioning part for adapting to and characterizing the contour shape of various parts to be inspected. The device includes: The contour data acquisition module is used to identify the contour data of the part to be inspected when it is placed on the positioning part of the fixture. The judgment module is used to determine whether the contour data matches the pre-stored component contour library; The type marking and parameter detection module is used to mark the type of unknown parts corresponding to the contour data and collect the detection parameters of the unknown parts when the contour data does not match the pre-stored part contour library; The qualified parts set determination module is used to determine the qualified parts set among the unknown parts by comparing the detection parameters of each unknown part horizontally when the number of unknown parts of the same type is greater than a preset threshold, and by using a clustering algorithm. The pre-stored component contour library import module is used to determine the standard parameters of a component type based on the detection parameters of the qualified parts set among the unknown components, and then import the contour data corresponding to the component type into the pre-stored component contour library after associating it with the standard parameters.

11. An electronic device, characterized in that, include: processor; Memory; And a computer program, wherein the computer program is stored in the memory, the computer program including instructions that, when executed by the processor, cause the electronic device to perform the method of any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device on which the computer-readable storage medium is located to perform the method according to any one of claims 1 to 9.

Images