Watch key fault diagnosis system based on intelligent detection

Abstract

The application discloses a watch key fault diagnosis system based on intelligent detection, which comprises a clamping positioning module, a driving execution module, a synchronous triggering module, an image acquisition module and a diagnosis processing module. The synchronous triggering module is used to control the image acquisition module to acquire the key area image including the key body, the connection boundary of the key and the watch case and the static reference area of the watch case, with the driving start time and the action stage switching time as the unified time reference. The watch key fault diagnosis system based on intelligent detection can detect the action change of the to-be-tested key in the pressing, holding, releasing and restoring process under the unified driving condition and the unified image acquisition condition, and realize the diagnosis of the mechanical state of the watch key by combining the reference registration, the action interval identification, the action feature extraction and the fault determination, compared with the detection mode which only pays attention to whether the key is triggered or whether the appearance is qualified.

Description

Technical Field

[0001] This invention relates to the field of visual inspection technology, specifically to a smart watch button fault diagnosis system based on intelligent detection. Background Technology

[0002] Existing testing methods for watch buttons or similar mechanical pressing components mostly focus on pressing life tests, pressing force tests, conduction trigger tests, or overall functional response tests. For example, repeatedly pressing the button through a drive mechanism to count the number of presses and determine whether a preset function can be triggered, or observing whether the screen interface changes as expected through a camera device. These methods can reflect whether the button can complete basic triggering actions to some extent, but they focus more on lifespan, conduction, or functional results, and are difficult to directly reflect the actual mechanical state of the button itself during the pressing, holding, releasing, and recovery processes.

[0003] In addition, there are technical solutions that use image processing to detect the appearance, contour, or local defects of buttons. These solutions are usually more suitable for static appearance inspection, such as judging surface defects, contour abnormalities, or whether the assembly appearance is qualified. However, they often lack targeted detection methods for changes in displacement, yaw, posture, and return state of buttons during dynamic stress processes. This is especially true for structures like watch buttons, which are small in size, have narrow boundaries, are highly reflective, and are tightly connected to the watch case. Relying solely on static images or a single result makes it difficult to accurately distinguish between different types of faults such as insufficient travel, abnormal yaw, and delayed return.

[0004] Even if the same button under test performs basically normally in a single test, it may still exhibit fluctuating issues such as intermittent sticking, early fatigue, or loose assembly. For these types of faults, relying solely on a single test result can easily lead to missed detections, thus affecting the accuracy of the test and the effectiveness of subsequent quality traceability. Summary of the Invention

[0005] The purpose of this invention is to provide a smart watch button fault diagnosis system to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a smart watch button fault diagnosis system based on intelligent detection, comprising a clamping and positioning module, a drive execution module, a synchronization triggering module, an image acquisition module, and a diagnostic processing module; The drive execution module is used to drive the button under test to complete the pressing, holding, releasing and restoring actions according to a preset action cycle; The synchronous triggering module is used to control the image acquisition module to acquire images of the button area, including the button body, the connection boundary between the button and the watch case, and the static reference area of ​​the watch case, with the drive start time and the switching time of each action stage as a unified time reference. The diagnostic processing module includes a reference registration unit, an action phase identification unit, an action feature extraction unit, and a fault determination unit connected in sequence. The reference registration unit is used to perform relative registration directly based on the natural contour features of the static reference area of ​​the watch case and the connection boundary between the button and the watch case, without additional markings. The action phase identification unit is used to identify the pressing, holding, releasing, and recovery phases. The action feature extraction unit is used to extract at least one action feature from the displacement, yaw, contour posture change, and return residual of the button under test during the action cycle. The fault determination unit is used to compare the extracted action features with standard reference data to output the button fault category and degree of abnormality.

[0007] As a further step, the clamping and positioning module includes a watch body clamp and an angle adjustment component. The angle adjustment component is used to adjust the spatial posture of the watch under test so that the motion axis of the button under test is aligned with the force direction of the drive execution module, and the button under test is within the field of view and depth of field of the image acquisition module.

[0008] As a further step, the drive execution module includes a servo drive and a pressure head positioned opposite to the button to be tested. The servo drive is used to set the pressing displacement, pressing speed, holding time, and release rhythm to form a repeatable action cycle.

[0009] As a further step, the synchronization trigger module is electrically connected to the drive execution module and the image acquisition module respectively. The synchronization trigger module is used to output a pre-trigger signal before the drive starts and to record the corresponding time information at the switching time of the pressing, holding, releasing and restoring phases.

[0010] As a further step, an illumination module is also included, which provides directional illumination to the key area under test. The directional illumination includes at least one of side backlighting and oblique strip light to enhance the imaging contrast of the key outline edge and the gap between the key and the case.

[0011] As a further step, the reference registration unit is used to extract the static reference area of ​​the watch case and the target area of ​​the buttons from the acquired image. First, a global reference coordinate system is established based on the static reference area of ​​the watch case, and then a local measurement coordinate system is established based on the target area of ​​the buttons to separate the micro-displacement of the entire watch from the movement of the buttons themselves. The reference registration unit and the motion feature extraction unit do not rely on adding displacement marks, identification pieces or reflective targets to the surface of the buttons or watch case to be measured. Instead, they directly use the natural contour features of the edge of the watch case, the boundary of the fixed decorative parts, the outer contour of the buttons, and the connection boundary between the buttons and the watch case for registration and measurement.

[0012] As a further step, the action phase recognition unit is used to perform phase recognition of the action process of the key under test based on the time information output by the synchronous trigger module and the displacement change in the key area image. The action phase recognition unit is used to first determine the initial interval of pressing, holding, releasing and restoring based on the time information output by the synchronous trigger module, and then correct the initial interval based on the actual starting position, stable position, retreating position and stabilizing position of the key under test in the key area image.

[0013] As a further step, the motion feature extraction unit employs at least one of subpixel edge fitting, feature point tracking, or optical flow analysis to extract at least two of the following parameters of the button under test during the motion cycle: axial displacement, lateral displacement, tilt angle, gap width, contour deformation, and return residual.

[0014] As a further step, the standard reference data is the motion feature data collected and extracted from the corresponding model of good quality buttons under multiple action cycles.

[0015] As a further step, the fault determination unit is also used to perform consistency analysis on the action characteristics corresponding to each action after performing multiple action cycles on the same key under test, so as to identify intermittent jamming, early fatigue or assembly loosening faults.

[0016] Compared with existing technologies, the beneficial effects of this invention are as follows: This intelligent detection-based watch button fault diagnosis system can detect the changes in the action of the button under test during the pressing, holding, releasing, and recovery processes under unified driving conditions and unified image acquisition conditions. By combining reference registration, action range recognition, action feature extraction, and fault determination, it can diagnose the mechanical state of the watch button. Compared with detection methods that only focus on whether the button is triggered or whether its appearance is qualified, this invention can more directly reflect the action state and fault characteristics of the button itself, and has better detection targeting and engineering application value.

[0017] 1. Directly acquire and analyze images of the key area under test, extract the axial displacement, lateral displacement, tilt angle and return residual of the key during the complete action process, which can more directly reflect whether the key under test has mechanical problems such as insufficient travel, abnormal wobble, delayed return or jamming, thereby improving the pertinence of fault identification.

[0018] 2. By extracting the static reference area of ​​the watch case and the target area of ​​the buttons, and establishing a global reference coordinate system and a local measurement coordinate system, the slight shaking of the entire watch is separated from the movement of the buttons under test. This reduces the influence of clamping error, platform micro-vibration and imaging drift on the detection results, and makes the extracted motion features closer to the real response state of the buttons under test, thereby improving the stability and consistency of the detection results.

[0019] 3. It can not only identify obvious faults such as insufficient travel, abnormal sway, and delayed return based on a single test result, but also identify fluctuating faults such as intermittent jamming, early fatigue, and loose assembly by repeatedly performing multiple tests on the same button under test and performing consistency analysis. It is not only suitable for factory screening, but also convenient for rework analysis and quality traceability, and has strong practical application value. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the overall processing flow of the system of the present invention; Figure 2 This is a schematic diagram of the reference registration and action range identification process of the present invention; Figure 3 This is a schematic diagram of the action feature extraction process of the present invention; Figure 4 This is a schematic diagram of the fault determination and consistency analysis process of the present invention; Figure 5 This is a schematic diagram of the detection link of the present invention; Figure 6 This is a schematic diagram of the timing of a single detection in this invention; Figure 7 This is a schematic diagram comparing the failure modes of the present invention; Figure 8 This is a schematic diagram illustrating the consistency comparison of multiple tests according to the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please see Figures 1-8 The present invention provides the following technical solution: This embodiment provides a smart watch button fault diagnosis system, including a clamping and positioning module, a drive execution module, a synchronization triggering module, an image acquisition module, and a diagnostic processing module. The drive execution module drives the button under test to complete the pressing, holding, releasing, and restoring actions according to a preset action cycle. The synchronization triggering module controls the image acquisition module to acquire images of the button area, including the button body, the connection boundary between the button and the watch case, and the static reference area of ​​the watch case, using the start time of the drive and the switching time of each action stage as a unified time reference. The diagnostic processing module performs reference registration, action stage recognition, action feature extraction, and fault determination on the acquired button area images to output the button fault category and degree of abnormality.

[0023] The clamping and positioning module includes a watch body clamp and an angle adjustment component. The angle adjustment component is used to adjust the spatial posture of the watch under test so that the motion axis of the button under test is aligned with the force direction of the drive execution module, and the button under test is within the field of view and depth of field of the image acquisition module. Specifically, the watch body clamp is used to stably clamp the watch under test to avoid significant shaking of the entire watch during the test. The angle adjustment component is used to fine-tune the tilt angle, orientation angle and exposed position of the side button of the watch to ensure that the button body, the connection area between the button and the watch case and the local static area of ​​the watch case can be stably acquired by the image acquisition module.

[0024] The drive execution module includes a servo drive and a pressure head positioned opposite the button under test. The servo drive is used to set the pressing displacement, pressing speed, holding time, and release rhythm to form a repeatable action cycle. The pressure head is adapted to the contact end face of the button under test to reduce the influence of eccentric force on the test results. By uniformly setting the pressing displacement, pressing speed, holding time, and release rhythm, different watches under test can complete consistent force actions under the same test conditions, which facilitates subsequent comparison of differences in action characteristics.

[0025] The synchronous trigger module is electrically connected to the drive execution module and the image acquisition module respectively. The synchronous trigger module is used to output a pre-trigger signal before the drive starts, and to record the corresponding time information at the switching time of the pressing, holding, releasing and restoring phases. The pre-trigger signal is used to enable the image acquisition module to enter the predetermined acquisition state before the key under test is subjected to force. The corresponding time information is used to characterize the switching node of the drive execution module under different action phases, so that the acquired key area image can be matched with the corresponding action phase in the future.

[0026] The image acquisition module is used to acquire images of the key area. Preferably, the image acquisition module adopts an industrial camera and imaging lens. The acquired key area images at least cover the key body under test, the connection boundary between the key and the watch case, and the static reference area of ​​the watch case. The image acquisition module acquires multiple frames of key area images within one action cycle to reflect the image changes of the key under test during the pressing, holding, releasing and recovery processes, providing an image basis for subsequent action stage recognition and action feature extraction.

[0027] The system also includes an illumination module, which provides directional illumination to the area of ​​the key under test. The directional illumination includes at least one of side backlighting and oblique strip light to enhance the imaging contrast of the key outline edge and the gap between the key and the case. The side backlighting is used to highlight the outer contour boundary of the key under test, and the oblique strip light is used to enhance the grayscale difference in the transition area of ​​the key surface edge and the gap area between the key and the case, thereby improving the stability of subsequent displacement recognition, yaw recognition and gap change recognition.

[0028] The diagnostic processing module includes a reference registration unit, an action phase identification unit, an action feature extraction unit, and a fault determination unit connected in sequence. The reference registration unit is used to perform relative registration directly based on the natural contour features of the static reference area of ​​the watch case and the connection boundary between the button and the watch case, without additional markings. The action phase identification unit is used to identify the pressing, holding, releasing, and recovery phases. The action feature extraction unit is used to extract at least one action feature from the displacement, yaw, contour posture change, and return residual of the button under test during the action cycle. The fault determination unit is used to compare the extracted action features with standard reference data to output the button fault category and degree of abnormality.

[0029] The reference registration unit is used to extract the static reference area of ​​the watch case and the target area of ​​the buttons from the acquired images. First, a global reference coordinate system is established based on the static reference area of ​​the watch case, and then a local measurement coordinate system is established based on the target area of ​​the buttons to separate the micro-displacement of the entire watch from the movement of the buttons themselves. The reference registration unit and the motion feature extraction unit do not rely on adding displacement marks, identification pieces or reflective targets to the surface of the buttons or watch case to be tested. Instead, they directly use the natural contour features of the edge of the watch case, the boundary of the fixed decorative parts, the outer contour of the buttons, and the boundary between the buttons and the watch case for registration and measurement. The static reference area of ​​the watch case can be selected from the edge of the watch case, the root of the lugs, the boundary of the fixed decorative parts, or a fixed structural area that does not change displacement with the button movement during the detection process. The target area of ​​the buttons is the button body to be tested and its adjacent boundary area. By establishing a global reference coordinate system and a local measurement coordinate system, it is possible to avoid misjudging slight shaking of the entire watch as abnormal displacement or sway of the button itself.

[0030] The action phase recognition unit is used to identify the phases of the action process of the key under test based on the time information output by the synchronous trigger module and the displacement changes in the key area image. The action phase recognition unit divides the acquired image into pressing, holding, releasing and recovery phases according to the corresponding time information before and after the driver starts, combined with the starting point, stable segment and return segment of the position change of the key under test in the key area image, thereby providing a basis for the extraction of action features in different phases.

[0031] The motion feature extraction unit employs at least one of subpixel edge fitting, feature point tracking, or optical flow analysis to extract at least two of the following parameters for the key under test during its motion cycle: axial displacement, lateral displacement, tilt angle, gap width, profile deformation, and return residual. Axial displacement characterizes the amount of movement of the key under test along the force direction; lateral displacement characterizes the amount of offset of the key under test in the non-force direction; tilt angle characterizes the degree of attitude deflection of the key under test during the motion; gap width characterizes the change in the boundary gap between the key under test and the case; profile deformation characterizes the degree of change in the profile shape of the key under test relative to its initial state; and return residual characterizes the remaining deviation of the key under test relative to its initial position after recovery. By extracting at least two of these motion features, the ability to distinguish faults such as insufficient travel, abnormal sway, delayed return, and loose assembly can be improved.

[0032] The standard reference data consists of motion feature data collected and extracted from the buttons of the corresponding model under multiple action cycles. The standard reference data is preferably obtained under the same clamping conditions, lighting conditions, driving parameters and image acquisition conditions as the sample to be tested, so as to ensure the consistency of the comparison basis. For the same model of watch, multiple good samples can be selected in advance to establish corresponding standard reference data.

[0033] The fault determination unit is also used to perform consistency analysis on the action characteristics corresponding to each action after performing multiple action cycles on the same key under test, in order to identify intermittent jamming, early fatigue or assembly loosening faults. When the axial displacement, lateral displacement, tilt angle, gap width, profile deformation or return residual of the same key under test under multiple action cycles show obvious discrete fluctuations, the fault determination unit can determine it as a consistency abnormality, and further combine it with standard reference data to determine whether it belongs to intermittent jamming, early fatigue or assembly loosening faults.

[0034] In summary, this embodiment, through the cooperation of the clamping and positioning module, drive execution module, synchronization triggering module, image acquisition module, illumination module, and diagnostic processing module, enables the button under test to complete image acquisition, motion recognition, motion feature extraction, and fault determination under unified action and acquisition conditions, thereby achieving intelligent detection of watch button faults. In this embodiment, the reference registration unit and motion feature extraction unit do not rely on adding displacement marks, recognition pieces, or reflective targets to the surface of the button under test or the watch case, nor do they rely on adding displacement sensors, rotation sensors, or TOF sensors inside the watch under test. Instead, they directly utilize the natural edge features of the static reference area of ​​the watch case and the natural contour features of the outer contour of the button under test and the connection boundary between the button and the watch case for relative registration and motion feature extraction. Through the above method, external non-destructive testing can be achieved without changing the structure of the watch under test, and the impact of additional mark installation errors on the test results can be reduced.

[0035] Based on the above, this embodiment further provides a method for diagnosing watch button malfunctions, including the following: S1. Fixing and preparation of the object to be tested The watch under test is installed in the clamping and positioning module, with the button under test facing the side where the drive execution module and image acquisition module are located. The exposed area of ​​the button under test is fully included in the imaging range of the image acquisition module. During the clamping process, the watch's posture is adjusted by the angle adjustment component to make the axial movement direction of the button under test as consistent as possible with the force direction of the drive execution module. At the same time, the watch case edge, fixed decorative part boundary, or other structural areas that remain stationary during the test are also included in the image acquisition range as a reference area for subsequent registration.

[0036] After clamping is completed, the control illumination module provides preset illumination conditions to the area of ​​the key under test, and adjusts the focal length, aperture, exposure parameters and acquisition magnification of the image acquisition module to ensure that the outer contour edge of the key under test, the transition boundary between the key and the watch case, and the static reference area of ​​the watch case have a clear and stable grayscale contrast relationship in the image. Subsequently, before the drive execution module applies any action to the key under test, the image acquisition module first acquires the initial image of the key under test in its natural static state, which is used to establish the initial position, initial posture of the key under test, and the reference imaging information of the static reference area of ​​the watch case.

[0037] In this embodiment, the purpose of step S1 is to unify the posture, field of view, and reference image of the watch under test, providing a consistent detection starting point for subsequent action recognition and fault determination. By simultaneously introducing a stationary reference area for the watch case and the target area of ​​the button under test at this stage, it is possible to distinguish between minor shaking of the entire watch and the movement of the button itself during subsequent processing, thereby avoiding misjudging clamping errors as button malfunctions. By establishing an initial image of the button under test in an unforced state, a unified reference can also be provided for subsequent judgment of button displacement, sway, return status, and contour changes.

[0038] Preferably, in S1, the area of ​​the button to be tested includes at least the outer contour of the button body, the connection position between the button and the watch case, and the adjacent watch case areas on both sides of the button; the watch case stationary reference area is preferably a fixed structure that does not deform or shift with the button action during the detection process. For crown-type buttons, in addition to acquiring the image of its axial force surface, the outer periphery texture boundary or outer edge contour of the crown can also be included simultaneously, so as to identify the sway and axial movement state of the crown during the pressing process.

[0039] S2, Driver Synchronization Trigger and Button Area Image Acquisition After completing the clamping and positioning, illumination adjustment and initial image acquisition of S1, the control drive execution module applies detection actions to the key under test according to the preset action cycle, so that the key under test sequentially completes pressing, holding, releasing and restoring. The servo drive in the drive execution module is used to set the pressing displacement, pressing speed, holding time and release rhythm to form a repeatable action cycle. The pressure head applies force along the motion axis of the key under test and is set relative to the outer surface of the key under test to ensure the consistency of the key under test's action during the force application process.

[0040] Before the drive execution module starts, the synchronous trigger module outputs a pre-trigger signal to enable the image acquisition module to enter the predetermined acquisition state. When the drive execution module starts to operate, the synchronous trigger module uses the start time of the drive as a unified time reference to control the image acquisition module to acquire images of the key area. When the key under test changes from being pressed to being held, from being held to being released, and from being released to being restored, the synchronous trigger module records the corresponding time information so that the acquired image can correspond to the action range of the key under test.

[0041] In this embodiment, the image acquisition module acquires multiple images of the button area within the same action cycle. Each button area image includes at least the button body under test, the connection boundary between the button and the watch case, and the static reference area of ​​the watch case. The button body under test is used to reflect the position and posture changes of the button during the force process; the connection boundary between the button and the watch case is used to reflect the gap changes and contour transition relationship changes; and the static reference area of ​​the watch case is used to provide a fixed reference for subsequent reference registration.

[0042] To ensure the comparability of subsequent test results, in S2, the focal length, exposure parameters, acquisition magnification, and acquisition frequency of the image acquisition module are kept consistent with the imaging conditions determined in S1. Throughout the entire process, the directional illumination output by the illumination module is kept stable. The side backlight is used to enhance the clarity of the outline edge of the button under test, and the oblique strip light is used to enhance the grayscale difference between the button and the case connection area, thereby improving the stability of subsequent displacement recognition, sway recognition, and return status recognition.

[0043] In this embodiment, the function of S2 is to establish the correspondence between driving actions and image acquisition, so that the image changes of the button under test during the pressing, holding, releasing and recovery processes can accurately correspond to the corresponding action range. By outputting a pre-trigger signal before the drive starts, the image acquisition module can be prevented from entering the working state only after the drive starts, thus avoiding the loss of initial action information. By recording the time information of the switching time of each action range, a basis can be provided for subsequent identification of action ranges, extraction of action features and fault determination based on time information and image changes.

[0044] Preferably, for watch buttons of different models or structures, the pressing displacement, pressing speed, holding time and release rhythm can be adaptively set according to the nominal travel of the button, the exposed size of the button and the differences in the local structure of the watch case; however, for the same model of watch, the above parameters should be kept consistent when establishing standard reference data and testing the sample to be tested, so as to ensure that the basis for subsequent comparison is consistent.

[0045] S3, Reference Registration and Action Range Recognition After completing the drive synchronization triggering and key area image acquisition of S2, the reference registration unit in the diagnostic processing module processes the acquired key area images. First, the case stationary reference area and the key target area are extracted from each key area image. The case stationary reference area is preferably a fixed structure area that does not change displacement with the action of the key under test during the detection process. The key target area is preferably the key under test body and the adjacent connection area between the key and the case. By introducing the above two areas at the same time, the slight shaking of the whole watch can be distinguished from the actual movement of the key under test in subsequent processing, avoiding misjudgment caused by clamping error, platform micro-vibration or slight optical imaging drift.

[0046] After extracting the static reference area of ​​the watch case, the reference registration unit first establishes a global reference coordinate system based on the static reference area of ​​the watch case. This global reference coordinate system is used to unify the positional relationship of the entire watch in each image, so that the fixed structure of the watch case at different acquisition times can be aligned to the same reference base. Subsequently, a local measurement coordinate system is established based on the target area of ​​the button. The local measurement coordinate system takes the area where the button to be tested is located as the measurement object. Its axis is set to correspond to the movement direction of the button to be tested, and the other direction intersects with the movement direction, preferably in a perpendicular relationship. By establishing the global reference coordinate system first and then establishing the local measurement coordinate system, on the one hand, the influence of the slight changes in the overall position of the watch under test on the detection results can be eliminated, and on the other hand, the true displacement, yaw and attitude changes of the button under test in the local area can be highlighted.

[0047] In this embodiment, the reference registration unit can use one or more of the following methods to locate and align the static reference area of ​​the watch case: contour matching, edge alignment, corner tracking, or feature region matching. For watches with clear edges and fixed structures, edge contour matching can be used first. For watches with fixed decorative parts boundaries, lug boundaries, or regular case features, corner or feature region matching can also be used. Through the above processing, the static reference area of ​​the watch case in each image can be kept as consistent as possible, thereby providing a unified spatial reference for the extraction of subsequent action features of the button under test.

[0048] After the reference registration is completed, the action phase recognition unit distinguishes the action process of the key under test. On the one hand, the action phase recognition unit reads the time information recorded by the synchronous trigger module before the driver starts and at each switching node of pressing, holding, releasing and restoring. On the other hand, it combines the changes of the key under test relative to the initial position in the key area image to identify the action range of the key under test.

[0049] Specifically, the interval before the drive starts and before the button under test has undergone significant positional change can be considered the initial stable interval; when the button under test begins to continuously deviate from its initial position under the action of the drive execution module, it can be identified as the pressing interval; when the button under test reaches the position after being subjected to force and remains relatively stable within a short period of time, it can be identified as the holding interval; when the drive execution module releases the force and the button under test begins to retract towards its initial position, it can be identified as the release interval; when the button under test gradually approaches its initial position and the magnitude of change tends to decrease, it can be identified as the recovery interval.

[0050] To improve the accuracy of action interval recognition, in this embodiment, the action phase recognition unit does not divide the interval solely based on the time information output by the synchronous trigger module. Instead, it combines the time information with the displacement changes in the key area image. In other words, the time information recorded by the synchronous trigger module is used to provide the approximate start and end range of each action interval, while the actual position changes in the key area image are used to correct the corresponding positions when the key under test actually starts moving, actually enters the holding state, actually starts to retreat, and finally tends to stabilize. Through the above dual judgment method, the error caused by relying solely on the drive control timing can be reduced, making the subsequently extracted action features closer to the actual mechanical response process of the key under test.

[0051] In this embodiment, after completing S3, the diagnostic processing module can obtain the registration image under a unified reference basis, the clear target area of ​​the key, and the action range information corresponding to pressing, holding, releasing, and returning. The above results provide a reliable image basis for subsequent action feature extraction and also provide a range basis for subsequent fault judgment, so that different fault types can be more accurately distinguished within the corresponding action range. For example, insufficient travel is usually reflected in the pressing range, abnormal swaying is usually reflected in the pressing and holding ranges, and delayed return or jamming is usually reflected in the release and returning ranges.

[0052] Preferably, for buttons with small size, narrow boundaries, or strong surface reflectivity, the contour enhancement effect formed by the illumination module can be combined in S3 to prioritize the extraction of the boundary area between the button and the watch case, as well as the corner area of ​​the button's outer contour, as a local measurement basis. This improves the stability of subsequent motion feature extraction. For crown-type buttons, the area containing the outer edge contour of the crown or the regular texture of the outer periphery can also be included in the button target area to simultaneously identify its axial movement state and posture changes. In this embodiment, the motion stage recognition unit does not divide the pressing, holding, releasing, and recovery intervals solely based on the preset motion rhythm of the drive execution module. Instead, it first provides the initial boundaries of each motion interval based on the time information recorded by the synchronous trigger module, and then corrects the initial boundaries based on the actual starting position, stable position, retreating position, and stabilizing position of the button in the image. Through the above method, the stage error caused by relying solely on the drive control rhythm can be avoided, making the subsequent motion feature extraction closer to the real mechanical response process of the button under test.

[0053] S4. Action Feature Extraction After completing the reference registration and action range recognition in S3, the action feature extraction unit extracts the action features of the button under test in the local measurement coordinate system. The direction corresponds to the movement axis of the button under test. The axis of motion is intersected by the axis of motion, preferably perpendicular to it. Based on this coordinate relationship, the actual motion of the button under test during the force process can be broken down into axial changes along the axis of motion and lateral changes deviating from the axis of motion, thereby reflecting the travel state and yaw state of the button respectively.

[0054] In this embodiment, the motion feature extraction unit employs at least one of subpixel edge fitting, feature point tracking, or optical flow analysis to quantify the image changes of the button under test during the pressing, holding, releasing, and recovery processes. To balance creative support and feasibility, this embodiment prioritizes extracting four types of motion features: axial displacement, lateral displacement, tilt angle, and return residual. These four types of motion features directly correspond to fault manifestations such as insufficient travel, abnormal swaying, delayed return, intermittent jamming, and loose assembly, and facilitate subsequent comparison with standard reference data. Specifically, during extraction, the feature center of the button under test is first determined within the target area of ​​the button, and the corresponding feature center position in the reference image is used as the initial reference position. Let the currently acquired image number be... The reference image number is The axial displacement of the button under test in the currently acquired image. It can be represented as , Indicates the first The center of the feature of the key to be tested in the acquired image is located at Coordinates in direction This indicates that the center of the feature of the key to be tested in the reference image is located at... By calculating the coordinates along the direction, we can obtain the displacement change of the button under test relative to the reference state along the motion axis. This amount can directly reflect the actual travel change of the button under test during the pressing process, which is the basis for subsequent judgment on whether the travel is insufficient.

[0055] While extracting the axial displacement, the motion feature extraction unit also extracts the lateral displacement of the key under test in the direction perpendicular to the motion axis, to reflect the degree of offset of the key under test during the force process. The calculation formula is as follows: In the formula Indicates the first The lateral displacement corresponding to each acquired image. Indicates the first The center of the feature of the key to be tested in the acquired image is located at Upward coordinates This indicates that the center of the feature of the key to be tested in the reference image is located at... The coordinates in the direction, calculated as described above, can reflect whether the button under test has a significant offset during operation. When the button under test has assembly eccentricity, local interference, or uneven force, the lateral displacement will usually increase significantly. Therefore, this amount can serve as an important basis for identifying abnormal sway.

[0056] To further characterize the posture deflection of the button under test during the action, two relatively stable contour feature points were selected within the target area of ​​the button, and denoted as feature points. and feature points Based on the change in the direction of the line connecting two feature points in the current acquired image and the reference image, the tilt angle of the button under test can be obtained, and the calculation formula is:

[0057] In the formula, Indicates the first The tilt angle corresponding to each acquired image and They represent the first Feature points in a sampled image The horizontal and vertical coordinates, and They represent the first Feature points in a sampled image The horizontal and vertical coordinates, , , as well as These represent the horizontal and vertical coordinates of the corresponding feature points in the reference image, respectively.

[0058] The tilt angle is used to characterize the degree of posture deflection of the button under test during the action. Compared with relying solely on lateral displacement, the tilt angle can better reflect whether the button under test has problems such as local skewness, unilateral obstruction, or eccentric pressing. Therefore, it is more effective in identifying abnormal swaying and assembly problems. After extracting the action features of the entire process of pressing, holding, releasing, and returning, it is also necessary to extract the remaining deviation of the button under test relative to the reference state after the return is completed to determine whether the button under test has fully returned to its original position. Let the image number corresponding to the end of the return and the point of stabilization be denoted as . , then the return residual It can be represented as: , Indicates the first The axial displacement return corresponding to the acquired image is used to characterize whether the button under test still has a residual deviation after the external force is removed and it has not fully returned to the reference position. When the button under test has insufficient rebound, slight jamming, abnormal internal return components or friction resistance, this value is usually too large. Therefore, it can be used as an important basis for judging the return delay and jamming faults.

[0059] In this embodiment, the motion feature extraction unit prioritizes extracting the above four types of motion features because: axial displacement directly reflects the effective travel of the button, lateral displacement and tilt angle together reflect button wobble and abnormal posture, and return residual directly reflects the sufficiency of return. By combining these four types of motion features, a relatively complete characterization basis for common mechanical failures of watch buttons can be formed without introducing too much complex calculation. This balances the stability of the specification and the support requirements for creativity and practicality during subsequent review. Preferably, in some watch buttons with more complex structures or narrower boundaries, the gap change or contour deformation between the button and the watch case can be further extracted as supplementary motion features to participate in subsequent judgment.

[0060] S5. Fault Determination and Consistency Analysis After the motion feature extraction in S4 is completed, the fault determination unit determines the motion features of the button under test. In this embodiment, the fault determination is not based solely on a single image or a single moment, but rather on a comprehensive analysis of the motion features extracted throughout the entire process of pressing, holding, releasing, and recovering. To facilitate comparison with standard reference data, representative features that represent the state of a single detection are first extracted from the single detection process, including maximum axial displacement, maximum lateral displacement, and maximum tilt angle. The maximum axial displacement of the button under test during a single detection process is... It can be represented as: , Indicates the currently acquired image number. Indicates the reference image number. This indicates the image number acquired when the recovery process has ended and the image has stabilized. Indicates the first The axial displacement corresponding to the acquired image; The maximum lateral displacement of the button under test during a single test. It can be represented as: , Indicates the first Lateral displacement corresponding to each acquired image; The maximum tilt angle of the button under test during a single test. It can be represented as: , Indicates the first The tilt angle corresponding to each acquired image.

[0061] After obtaining the aforementioned representative features, they are compared with the corresponding reference values ​​in the standard reference data. The standard reference data is preferably motion feature data extracted from a compliant button of the corresponding model under the same clamping conditions, driving parameters, lighting conditions, and image acquisition conditions. Specifically, the standard reference data includes at least the maximum axial displacement reference value, the maximum lateral displacement reference value, the maximum tilt angle reference value, and the return residual reference value, denoted as follows: , , and Meanwhile, to accommodate normal dispersion during actual processing, assembly, and testing, allowable deviations for axial displacement, lateral displacement, tilt angle, and return residual are set, denoted as follows: , , and When the maximum axial displacement of the button under test is lower than the reference value of the maximum axial displacement minus the allowable deviation of the axial displacement, it can be determined that there is a risk of insufficient travel. The determination relationship is as follows: When the maximum lateral displacement of the button under test exceeds the maximum lateral displacement reference value plus the allowable lateral displacement deviation, or the maximum tilt angle exceeds the maximum tilt angle reference value plus the allowable tilt angle deviation, it can be determined that there is a risk of abnormal sway. The determination relationship is as follows: When the return residual of the button under test exceeds the reference value of the return residual plus the allowable deviation of the return residual, it can be determined that there is a risk of return lag or jamming. The determination relationship is as follows: In this embodiment, the above-mentioned determination method has good engineering feasibility, wherein the maximum axial displacement This directly reflects the effective travel of the button under test during the force application process. A significantly smaller travel usually indicates insufficient button pressing, limited internal travel, or obstruction in the local structure; maximum lateral displacement. With maximum tilt angle Together, they reflect the degree of yaw and attitude deflection of the button under test during the force application process. When these values ​​are significantly larger, it usually indicates that the button under test has problems such as assembly misalignment, unilateral obstruction, incorrect relative position of the pressure head, or tilting of the button body; return residual It directly reflects the adequacy of the button's recovery after the external force is removed. When it is significantly larger, it usually indicates that the button has problems such as slight sticking, insufficient rebound, or abnormal operation of the internal return component.

[0062] In addition to single-test judgment, to further identify faults such as intermittent jamming, early fatigue, or assembly loosening that may not be consistently detected in a single test, this embodiment performs multiple tests on the same button under test and analyzes the consistency of the multiple test results. Let the current repeated test number be . The total number of repeated tests is Consistency indices can then be constructed based on the maximum axial displacement, maximum lateral displacement, maximum tilt angle, and return residual obtained from multiple tests. Its expression is:

[0063] in Indicates the first The maximum axial displacement obtained from the second test express The average value of the maximum axial displacement obtained from each test; Indicates the first The maximum lateral displacement obtained from the second test; express The average value of the maximum lateral displacement obtained from the detection; Indicates the first The maximum tilt angle obtained from the second test express The average maximum tilt angle obtained from the second test Indicates the first The return residual obtained from the second detection; express The average value of the return residuals obtained from the first test. , , and These represent the allowable deviations for axial displacement, lateral displacement, tilt angle, and return residual, respectively, based on the above consistency indicators. It can quantify the dispersion between multiple detection results. A smaller value indicates that the same button under test performs relatively stably under multiple tests; when... A significant increase indicates that the behavior of the same button under test fluctuates greatly across different tests. Even if a single test result does not exceed the range of a single judgment, the button under test can still be considered to have a risk of intermittent jamming, early fatigue, or loose assembly. Compared to judging solely based on a single test result, the above consistency analysis method can more effectively identify hidden faults and early abnormal states, thereby improving the accuracy and stability of fault diagnosis.

[0064] In this embodiment, the fault determination unit can first rely on the results of a single detection. , , and Make a preliminary judgment on obvious faults, and then combine them with consistency indicators. To supplement the judgment of hidden faults, by combining single feature comparison with multiple consistency analysis, the system can identify not only obviously out-of-tolerance button faults, but also potential faults that are still in the early stages but have already shown abnormal fluctuations. In this embodiment, the fault judgment unit does not only output qualified or unqualified results, but distinguishes at least one fault category among insufficient output stroke, abnormal fluctuation, delayed return, and jamming based on the correspondence between the maximum axial displacement and effective stroke, the correspondence between the maximum lateral displacement and maximum tilt angle and the yaw state, and the correspondence between the return residual and the return sufficiency. For multiple test results of the same button under test, normalized consistency analysis is further combined to identify fluctuating faults such as intermittent jamming, early fatigue, or loose assembly.

[0065] S6. Diagnostic Result Output After completing the fault determination and consistency analysis in S5, the fault determination unit outputs the final diagnostic result of the button under test. The diagnostic result includes at least the fault category and the degree of abnormality. Preferably, it also includes key images, key motion features and test conclusion descriptions corresponding to the determination result, so that the tester can intuitively understand the action performance and abnormality source of the button under test in this test. In this embodiment, the fault category may include one or more of the following: normal, insufficient travel, abnormal sway, delayed return, intermittent jamming, early fatigue and loose assembly.

[0066] Among them, when the maximum axial displacement of the button under test When the value is significantly lower than the corresponding standard reference data range, insufficient stroke can be output; when the maximum lateral displacement... or maximum tilt angle When the value is significantly larger than expected, an abnormal yaw can be output; when the return residual is significantly larger than expected, an abnormal yaw can be output. When the value is significantly larger than expected, an abnormal return delay or jamming can be output; when the results of a single test may not all be significantly out of tolerance, but the consistency index is high... When the value increases significantly, the system can output conclusions of fluctuating faults such as intermittent jamming, early fatigue, or loose assembly. In this way, overt and covert faults can be distinguished, thereby improving the completeness of the diagnostic results.

[0067] To make the output results more readable and traceable, in this embodiment, in addition to outputting the fault category, the fault determination unit can also simultaneously output the key action features corresponding to the fault category.

[0068] For example, for faults related to insufficient travel, the maximum axial displacement can be output as the key indicator. And its deviation from the standard reference data; for yaw anomaly faults, the maximum lateral displacement can be output as the key point. and maximum tilt angle For faults such as delayed or stuck return, the return residual can be output as the main focus. For intermittent jamming, early fatigue, or assembly loosening faults, consistency indicators can be the focus of output. By simultaneously outputting the fault category and key action characteristics, as well as the fluctuations between multiple test results, the diagnostic results can not only remain at the conclusion level, but also reflect the main test basis that led to the conclusion.

[0069] Furthermore, to facilitate on-site verification, rework analysis, and subsequent quality traceability, key images corresponding to the fault determination can be output simultaneously when outputting diagnostic results.

[0070] The key images preferably include a reference image, an image corresponding to maximum displacement, an image corresponding to maximum yaw, and an image corresponding to the end of recovery. The reference image reflects the initial position and initial posture of the button under test in an unforced state; the image corresponding to maximum displacement reflects the image performance of the button under test when it reaches its maximum pressed state during the pressing process; the image corresponding to maximum yaw reflects the image performance of the button under test when its offset or tilt is most obvious during the action; and the image corresponding to the end of recovery reflects the image performance of the button under test in its final return state after the external force is removed. By outputting the above key images, the testing personnel can directly observe the image changes of the button under test at different key moments, thereby improving the interpretability of the results.

[0071] In this embodiment, the degree of anomaly can be determined based on the deviation of the action characteristics of the button under test from the standard reference data and the consistency index. The size is graded and output accordingly.

[0072] For example, when the single action characteristic of the button under test only slightly exceeds the corresponding allowable range, and the consistency index... When the level is low, the abnormality can be marked as mild; when multiple action characteristics of the button under test exceed the tolerance simultaneously, or the consistency index... When the abnormality is significant and accompanied by substantial fluctuations in multiple test results, the degree of abnormality can be marked as moderate or severe. By grading the degree of abnormality, the diagnostic results can be used not only to determine whether a fault exists, but also to measure the degree of fault development and the priority of handling.

[0073] Preferably, when outputting the final result, the fault determination unit can also generate a detection conclusion description based on the correspondence between the fault category and the key action features.

[0074] For example, when Smaller and Under normal conditions, it can indicate that the button under test may have a limited effective travel distance; when and When the value is too high, it may indicate that the button under test may have misalignment during assembly or improper force application; when Large and If the value is too high, it may indicate that the button under test may be intermittently stuck or that the internal return mechanism is malfunctioning. By outputting the above conclusions, a more direct correspondence can be established between the test results and the actual mechanical condition, facilitating subsequent manual verification and repair.

[0075] Through S6, this embodiment can uniformly output the image acquisition results, action feature extraction results, fault determination results, and consistency analysis results of the button under test under controlled action, thereby forming a complete detection closed loop. Compared with detection methods that only output pass or fail conclusions, this embodiment can not only provide the fault category, but also the degree of abnormality, key action features, and key images. Therefore, it is more suitable for application scenarios such as watch button production inspection, repair screening, and quality traceability.

[0076] The following description, using a common scenario in actual production—the testing of side function buttons on a smartwatch—further illustrates the working process of this invention: The test object is a rectangular-cased smartwatch with a single function button on its right side. This button is used for short presses to wake up and return to the menu. It is a typical axially compressed metal button, characterized by the following features: a small exposed area, a narrow gap between the button and the watch case, and a reflective surface. Furthermore, during assembly, it is prone to problems such as insufficient travel, excessive wobbling, incomplete return, or intermittent jamming due to misalignment during pressing, misalignment of internal elastic components, or increased local friction. Therefore, using the system of this invention to test this type of watch button effectively demonstrates the application value of this solution.

[0077] In this example, the clamping and positioning module uses an adjustable watch body clamp to fix the watch under test. The clamp has elastic clamping parts on both sides to avoid hard compression of the watch case. The spatial posture of the watch under test is adjusted by the angle adjustment component so that the movement axis of the button under test is aligned with the force direction of the drive execution module. At the same time, the button body, the connection boundary between the button and the watch case, and the fixed edge area on the right side of the watch case are completely within the field of view and depth of field of the image acquisition module. The fixed edge area on the right side of the watch case does not change with the movement of the button under test throughout the detection process, so it is selected as the stationary reference area of ​​the watch case.

[0078] The drive execution module adopts a combination structure of servo drive and arc-shaped pressure head. The contact end face of the arc-shaped pressure head is adapted to the outer surface of the button under test to reduce local eccentric force. For the structural dimensions and nominal stroke of the button of this model of watch, the pressing displacement is set to 1.10mm, the pressing speed is set to 16mm / s, the holding time is set to 0.12s, the release time is set to 0.06s, and the recovery observation time is set to 0.18s. The above parameters are kept consistent in the testing of the same model to ensure the comparability between the standard reference data and the test data.

[0079] The image acquisition module uses an industrial camera with a long working distance lens, while the illumination module employs a combination of side backlighting and oblique strip lighting. The side backlighting enhances the outer contour edge of the button under test, while the oblique strip lighting enhances the grayscale contrast in the transition area between the button and the case. To ensure imaging stability during the detection process, the exposure time, acquisition magnification, and focal length of the image acquisition module remain constant throughout the entire detection process in this example.

[0080] Before testing began, 10 identical good-quality watches, whose operation had been manually verified, were used as a sample to establish standard reference data. The test button on each good-quality watch was tested five times, and the maximum axial displacement, maximum lateral displacement, maximum tilt angle, and return residual were extracted. Statistical reference values ​​were then calculated. The final standard reference data for the good-quality buttons of this model are as follows: Maximum axial displacement reference value. The maximum lateral displacement reference value is 0.96 mm. The maximum tilt angle reference value is 0.018mm. The return residual is 0.72°, which is the reference value. The allowable deviation of axial displacement is 0.012mm, taking into account the fluctuation range of good products. The allowable deviation for lateral displacement is set at 0.06 mm. Set to 0.012mm, allowable deviation of tilt angle. Set to 0.38°, allowable deviation of return residual. Set to 0.015mm.

[0081] The following is a complete test description using a single sample as an example.

[0082] In step S1, the watch under test is installed in the clamping and positioning module, and its spatial orientation is adjusted so that the button under test is located in the imaging center area of ​​the image acquisition module. Subsequently, the illumination module and image acquisition module are adjusted to ensure that the outer contour of the button under test, the boundary between the button and the watch case, and the stationary reference area of ​​the watch case all have clear outlines in the image. Afterwards, a reference image is acquired without the drive execution module activating, and the corresponding number of this reference image is recorded as follows. .

[0083] In S2, the synchronous trigger module outputs a pre-trigger signal before the drive execution module starts, causing the image acquisition module to enter the predetermined acquisition state. Subsequently, the drive execution module performs pressing, holding, releasing, and restoring actions on the key under test according to the preset action cycle. The image acquisition module synchronously acquires the key area image during the entire action process. The synchronous trigger module simultaneously records the time information corresponding to the start of pressing, holding, releasing, and restoring, which is used for subsequent action interval recognition.

[0084] In S3, the reference registration unit first extracts the static reference area of ​​the watch case from each image and establishes a global reference coordinate system; then, based on the key body under test and its adjacent boundary area, it establishes a local measurement coordinate system. In this local measurement coordinate system, The direction is consistent with the movement axis of the button under test. direction and With the direction perpendicular, after registration, the action phase recognition unit combines time information and position changes in the image to distinguish the action process of the button under test. After recognition, the button under test showed a relatively clear pressing range and holding range in this detection, but there was a slight trailing phenomenon in the process of returning to the initial position after release, and the position change within the recovery range tended to stabilize slowly.

[0085] In S4, the motion feature extraction unit uses a combination of subpixel edge fitting and feature point tracking to extract the axial displacement, lateral displacement, tilt angle, and return residual of the button under test.

[0086] For example, in the current detection, the reference image number is =0, the acquired image number when the recovery ends and the situation stabilizes is . =126, at the moment of maximum indentation, the corresponding image acquisition number is =48, at this time the center of the feature of the button under test is at... The coordinate change in the direction resulted in a calculated maximum axial displacement of 0.84 mm; at the end of the hold period, the corresponding acquired image number was... =71, at which point the lateral offset is most pronounced, corresponding to a maximum lateral displacement of 0.041 mm; during the transition process before and after release, the corresponding acquired image number is =77, at which point the direction of the line connecting the two contour feature points changes the most, corresponding to a maximum tilt angle of 1.46°; at the end of the recovery, the corresponding acquired image number is =126, and the calculated return residual is 0.043mm.

[0087] Therefore, the key motion characteristic obtained in this test is: maximum axial displacement. =0.84mm, maximum lateral displacement =0.041mm, maximum tilt angle =1.46°, return residual error =0.043mm.

[0088] Compare the above results with the standard reference data.

[0089] Regarding the travel status, because =0.90mm, while in this test =0.84, which is less than 0.90mm, therefore it can be determined that there is a risk of insufficient stroke; For the yaw state, because =0.030mm, while in this test... =0.041mm, greater than 0.030mm; at the same time =1.10°, while in this test =1.46°, which is greater than 1.10°, therefore it can be determined that there is a risk of abnormal sway. For the return state, because =0.027mm, while in this test... =0.043mm, which is greater than 0.027mm, therefore it can be determined that there is a risk of delayed return or slight jamming.

[0090] Considering that some button malfunctions are intermittent, a single test is insufficient to fully reflect their stability. Therefore, in S5, the test is repeated five times consecutively for the button under test. The results of the five tests are as follows: First test: =0.84mm, =0.041mm, =1.46°, =0.043mm; Second test: =0.87mm, =0.036mm, =1.21°, =0.031mm; Third test: =0.82mm, =0.045mm, =1.58°, =0.051mm; Fourth test: =0.88mm, =0.033mm, =1.18°, =0.028mm; Fifth test: =0.83mm, =0.047mm, =1.63°, =0.056mm.

[0091] The data above shows that the maximum axial displacement of the button under test was consistently low across the five tests, while the maximum lateral displacement and maximum tilt angle were generally high, and the return residual fluctuated significantly between different tests. Substituting the results of the five tests into the aforementioned consistency index... After calculating the formula, the consistency index of the button under test is obtained. =5.37. Based on the statistical results of the good sample of this model, this embodiment uses the consistency index of 2.00 as the upper limit of the reference for abnormal fluctuations, because the button under test... It is significantly higher than 2.00, so it can be further concluded that there is a significant anomaly in consistency.

[0092] In S6, the fault determination unit outputs the final diagnostic result: the button under test has three types of obvious faults: insufficient travel, abnormal sway, and delayed return, and is accompanied by the risk of intermittent jamming.

[0093] Regarding the degree of abnormality, considering that the maximum axial displacement, maximum lateral displacement, maximum tilt angle, and return residual of the button under test have all significantly deviated from the standard reference data, and the consistency index... The value is significantly higher than normal, therefore its overall abnormality level can be marked as moderate to severe abnormality.

[0094] Simultaneously, the system outputs key images corresponding to the diagnostic result, including the baseline image, the maximum axial pressing image, the maximum sway image, and the recovery end image, and outputs the detection conclusion: the button under test has insufficient effective stroke during controlled pressing, obvious sway during pressing and holding, and insufficient return and large fluctuations after release. It is speculated that this is related to button assembly eccentricity, local friction resistance, or abnormal condition of internal return components. It is recommended to proceed to the rework and inspection station for disassembly and inspection confirmation.

[0095] As can be seen from the above complete example, the present invention does not only determine whether the button under test can be pressed or rebound, but also comprehensively analyzes the motion state and stability of the button under test throughout the entire action process through unified driving, unified acquisition, unified registration and unified judgment. This enables it to more accurately distinguish different fault types such as insufficient travel, abnormal sway, delayed return and intermittent jamming, and provides a clear basis for subsequent rework, quality traceability and process optimization.

[0096] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A smartwatch button fault diagnosis system based on intelligent detection, characterized in that: It includes a clamping and positioning module, a drive execution module, a synchronization triggering module, an image acquisition module, and a diagnostic processing module; The drive execution module is used to drive the button under test to complete the pressing, holding, releasing and restoring actions according to a preset action cycle; The synchronous triggering module is used to control the image acquisition module to acquire images of the button area, including the button body, the connection boundary between the button and the watch case, and the static reference area of ​​the watch case, with the drive start time and the switching time of each action stage as a unified time reference. The diagnostic processing module includes a reference registration unit, an action phase identification unit, an action feature extraction unit, and a fault determination unit connected in sequence. The reference registration unit is used to perform relative registration directly based on the natural contour features of the static reference area of ​​the watch case and the connection boundary between the button and the watch case, without additional markings. The action phase identification unit is used to identify the pressing, holding, releasing, and recovery phases. The action feature extraction unit is used to extract at least one action feature from the displacement, yaw, contour posture change, and return residual of the button under test during the action cycle. The fault determination unit is used to compare the extracted action features with standard reference data to output the button fault category and degree of abnormality.

2. The smart watch button fault diagnosis system based on intelligent detection according to claim 1, characterized in that: The clamping and positioning module includes a watch body clamp and an angle adjustment component. The angle adjustment component is used to adjust the spatial posture of the watch under test so that the motion axis of the button under test is aligned with the force direction of the drive execution module, and the button under test is within the field of view and depth of field of the image acquisition module.

3. The smart watch button fault diagnosis system based on intelligent detection according to claim 1, characterized in that: The drive execution module includes a servo drive and a pressure head positioned opposite to the button to be tested. The servo drive is used to set the pressing displacement, pressing speed, holding time, and release rhythm to form a repeatable action cycle.

4. The smart watch button fault diagnosis system based on intelligent detection according to claim 1, characterized in that: The synchronization trigger module is electrically connected to the drive execution module and the image acquisition module respectively. The synchronization trigger module is used to output a pre-trigger signal before the drive starts and to record the corresponding time information at the switching time of the pressing, holding, releasing and restoring phases.

5. The smart watch button fault diagnosis system based on intelligent detection according to claim 1, characterized in that: It also includes an illumination module for providing directional illumination to the key area to be tested. The directional illumination includes at least one of side backlighting and oblique strip light to enhance the imaging contrast of the key outline edge and the gap between the key and the case.

6. The smart watch button fault diagnosis system based on intelligent detection according to claim 1, characterized in that: The reference registration unit is used to extract the static reference area of ​​the watch case and the target area of ​​the buttons from the acquired image. First, a global reference coordinate system is established based on the static reference area of ​​the watch case, and then a local measurement coordinate system is established based on the target area of ​​the buttons to separate the micro-displacement of the entire watch from the movement of the buttons themselves. The reference registration unit and the motion feature extraction unit do not rely on adding displacement marks, identification pieces or reflective targets to the surface of the buttons or watch case to be measured. Instead, they directly use the natural contour features of the edge of the watch case, the boundary of the fixed decorative parts, the outer contour of the buttons, and the connection boundary between the buttons and the watch case for registration and measurement.

7. The smart watch button fault diagnosis system based on intelligent detection according to claim 1, characterized in that: The action phase recognition unit is used to identify the phases of the action process of the key under test based on the time information output by the synchronous trigger module and the displacement changes in the key area image. The action phase recognition unit is used to first determine the initial interval of pressing, holding, releasing and restoring based on the time information output by the synchronous trigger module, and then correct the initial interval based on the actual starting position, stable position, retreating position and stabilizing position of the key under test in the key area image.

8. The smart watch button fault diagnosis system based on intelligent detection according to claim 1, characterized in that: The motion feature extraction unit uses at least one of subpixel edge fitting, feature point tracking, or optical flow analysis to extract at least two of the following parameters of the button under test during the motion cycle: axial displacement, lateral displacement, tilt angle, gap width, contour deformation, and return residual.

9. The smart watch button fault diagnosis system based on intelligent detection according to claim 1, characterized in that: The standard reference data is the motion feature data collected and extracted from the corresponding model of good quality buttons under multiple action cycles.

10. The smart watch button fault diagnosis system based on intelligent detection according to claim 1, characterized in that: The fault determination unit is also used to perform consistency analysis on the action characteristics corresponding to each action after performing multiple action cycles on the same key under test, so as to identify intermittent jamming, early fatigue or assembly loosening faults.

Images