Superfinishing spring defect detection method and system

Abstract

The invention relates to the technical field of defect detection, and provides a super-precision spring defect detection method and system, and the method comprises the steps: obtaining the surface deformation information of a to-be-detected super-precision diaphragm spring; extracting current defect feature data from the surface deformation information; storing the current defect feature data to a defect feature cache region; comparing the current defect feature data with historical defect feature data stored in a defect feature cache region to judge whether the current defect feature data has space-time constancy or not, and obtaining a space-time constancy judgment result; according to the time-space constancy judgment result, defect type marks corresponding to the current defects are judged as system artifacts or real product defects; and generating a detection report, and displaying the defect type mark based on a preset display rule for distinguishing and displaying the system artifacts and the real product defects. The method has the advantages that system artifacts and real product defects are distinguished, the detection accuracy and efficiency are improved, and unnecessary production interruption and verification cost are reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of defect detection, in particular to a super-precision spring defect detection method and system. BACKGROUND

[0002] In the field of precision manufacturing, residual stress detection of super-precision diaphragm springs is a key link to ensure product quality. Laser speckle interferometry is often used to evaluate the stress distribution within the spring and identify potential structural defects due to its non-contact and high sensitivity characteristics.

[0003] However, in practical applications, the optical elements of the measurement system, particularly the lens surface of the image acquisition camera, may be affected by trace contamination. This contamination may originate from non-standard cleaning agents used during routine maintenance, which leave a layer of transparent film on the lens surface that is difficult to detect with the naked eye and is unevenly distributed. This film introduces a fixed, position-dependent distortion to the phase information of the measurement laser beam, while the self-calibration mechanism of existing systems typically only compensates for the attenuation of the light signal intensity and cannot identify or correct this phase distortion. Therefore, the system may misjudge the fixed phase distortion introduced by optical element contamination as real deformation of the spring body, resulting in false defects in the detection report that do not match the actual situation. This false defect is extremely easy to misjudge as a systematic problem in the production process by experienced technicians due to its repeatability in form and location, thereby causing unnecessary production downtime and expensive verification costs.

[0004] The prior art needs to be improved in view of the above problems. SUMMARY

[0005] To solve the problems of the prior art, the present application provides a super-precision spring defect detection method and system, which has the advantages of improving the accuracy and efficiency of detection, reducing unnecessary production interruptions and verification costs.

[0006] The present application provides a super-precision spring defect detection method, and the technical points are as follows: A super-precision spring defect detection method, comprising the following steps: Obtaining surface deformation information of a super-precision diaphragm spring to be measured; Extracting current defect feature data from the surface deformation information; Storing the current defect feature data in a defect feature cache area; Comparing the current defect feature data with historical defect feature data stored in the defect feature cache area to determine whether the current defect feature data has spatial and temporal constancy, and obtaining a spatial and temporal constancy judgment result; According to the spatial and temporal constancy judgment result, the defect type corresponding to the current defect is marked as a system artifact or a real product defect; Generate a detection report, and based on the preset display rules for distinguishing between system artifacts and real product defects, display the defect type label.

[0007] Through the above scheme, system artifacts and real product defects can be effectively distinguished, false judgments can be avoided, and detection accuracy can be improved.

[0008] To further solve the problem, the application also proposes that the current defect feature data is compared with the historical defect feature data stored in the defect feature cache area to determine whether the current defect feature data has spatial and temporal constancy, and a spatial and temporal constancy judgment result is obtained. The step comprises: Obtain the laser power of the current detection; According to the laser power, the current defect feature data or the historical defect feature data stored in the defect feature cache area is adjusted; Compare the adjusted current defect feature data with the adjusted historical defect feature data to determine whether the current defect feature data has spatial and temporal constancy, and obtain a spatial and temporal constancy judgment result.

[0009] Through the above scheme, the influence of laser power on defect feature data is considered, and the accuracy of spatial and temporal constancy judgment is further improved.

[0010] To improve the solution, the application also proposes that the current defect feature data or the historical defect feature data stored in the defect feature cache area is adjusted according to the laser power. The step comprises: Based on the artifacts in the current defect feature data or the historical defect feature data stored in the defect feature cache area, the spatial distribution area of the artifacts in the image is identified; Determine the local laser power sensitivity of each position in the spatial distribution area of the artifacts; According to the laser power and the local laser power sensitivity, the current defect feature data or the historical defect feature data stored in the defect feature cache area is adjusted in a spatial position related manner.

[0011] Through the above scheme, the local laser power sensitivity is introduced to realize more fine spatial position related adjustment and improve the accuracy of artifact correction.

[0012] To further solve the problem, the application also proposes that the local laser power sensitivity of each position in the spatial distribution area of the artifacts is determined. The step comprises: Obtain the artifact feature data and the corresponding laser power data of each position in the spatial distribution area of the artifacts; According to the artifact feature data and the laser power data, the ratio of the artifact feature change amount to the laser power change amount of each position in the spatial distribution area of the artifacts is calculated, and the local laser power sensitivity of each position in the spatial distribution area of the artifacts is obtained; Based on the calculated results, the local laser power sensitivity of each position in the spatial distribution area of the artifact is updated and confirmed.

[0013] Through the above scheme, by calculating the ratio of the artifact feature change amount and the laser power change amount, the local laser power sensitivity is accurately quantified, which provides a basis for accurate adjustment.

[0014] To perfect the solution, the present application also proposes that the step of adjusting the current defect feature data or the historical defect feature data stored in the defect feature cache area in the spatial position correlation includes: The spatial distribution area of the artifact is spatially divided to form a plurality of micro space units; According to the local laser power sensitivity corresponding to the micro space unit and the corresponding laser power, the artifact feature correction amount of the micro space unit is calculated; The artifact feature correction amount is applied to the feature value of the corresponding micro space unit in the current defect feature data or the historical defect feature data stored in the defect feature cache area, so as to realize the spatial position correlation adjustment of the current defect feature data or the historical defect feature data stored in the defect feature cache area.

[0015] Through the above scheme, the artifact feature correction amount is applied to the micro space unit, realizing fine and localized adjustment, and further improving the correction effect.

[0016] To further solve the problem, the present application also proposes that the step of calculating the artifact feature correction amount of the micro space unit includes: Obtain the local laser power sensitivity corresponding to the micro space unit; Obtain the current laser power; According to the local laser power sensitivity corresponding to the micro space unit and the current laser power, the pre-stored correction amount lookup table or the pre-calibrated correction curve is consulted to determine the artifact feature correction amount of the micro space unit.

[0017] Through the above scheme, by consulting the pre-stored correction amount lookup table or applying the pre-calibrated correction curve, the determination process of the artifact feature correction amount is simplified, and the efficiency is improved.

[0018] To perfect the solution, the present application also proposes that the step of judging whether the current defect feature data has spatiotemporal constancy to obtain the spatiotemporal constancy judgment result includes: Quantify the similarity between the current defect feature data and the historical defect feature data stored in the defect feature cache area; According to the relationship between the similarity and the preset similarity threshold, it is judged whether the current defect feature data has spatiotemporal constancy, and the spatiotemporal constancy judgment result is obtained.

[0019] By the above scheme, by quantifying the similarity and comparing with the threshold value, a method for objectively judging the space-time constancy is provided, and the reliability of the judgment is enhanced.

[0020] To further solve the problem, the application further provides that the step of judging whether the current defect feature data has space-time constancy according to the relationship between the similarity and the preset similarity threshold value to obtain a space-time constancy judgment result includes: Obtaining defect feature data of a reference part; Comparing the defect feature data of the reference part with the pre-stored reference part standard data to obtain a reference part feature deviation; According to the reference part feature deviation, the initial threshold value is corrected to obtain a corrected preset similarity threshold value; According to the relationship between the similarity and the preset similarity threshold value, whether the current defect feature data has space-time constancy is judged, and a space-time constancy judgment result is obtained.

[0021] Through the above scheme, the reference part feature deviation is introduced to correct the preset threshold value, so that the judgment threshold value is more adaptive, and the accuracy of the judgment is improved.

[0022] To perfect the solution, the application further provides that the step of correcting the initial threshold value according to the reference part feature deviation to obtain a corrected preset similarity threshold value includes: Obtaining the position deviation, the shape deviation and the stress deviation in the reference part feature deviation; According to the combination of the position deviation, the shape deviation and the stress deviation, the pre-stored correction mapping table is consulted to determine the correction amount of the corresponding position, shape and stress part in the initial threshold value; Apply each determined correction amount to the corresponding position, shape and stress part in the initial threshold value to obtain a corrected preset similarity threshold value.

[0023] Through the above scheme, the influence of position, shape and stress deviation on the threshold value is considered, multi-dimensional and fine threshold correction is realized, and the judgment accuracy is further improved.

[0024] To further solve the problem, the application further provides a super-precision spring defect detection system for performing super-precision spring defect detection, comprising: A surface deformation acquisition module is configured to acquire surface deformation information of a super-precision diaphragm spring to be tested. A defect feature extraction module is configured to extract current defect feature data from the surface deformation information. A feature data storage module is configured to store the current defect feature data in a defect feature cache area. A spatiotemporal constancy judgment module is used to compare the current defect feature data with the historical defect feature data stored in the defect feature buffer area to determine whether the current defect feature data has spatiotemporal constancy and obtain a spatiotemporal constancy judgment result; The defect type determination module is used to determine the defect type label corresponding to the current defect as a system artifact or a real product defect based on the spatiotemporal constancy judgment result; The test result display module is used to generate a test report and display defect type marks based on preset display rules for distinguishing system artifacts from real product defects.

[0025] Through the above solution, a system for implementing the above defect detection method is provided, which is convenient for actual deployment and application and improves the automation and reliability of detection.

[0026] In summary, the present application provides a method and system for detecting defects in ultra-precision springs. By comparing the spatiotemporal constancy of current defect feature data with historical defect feature data, the defect type is marked as a system artifact or a real product defect, and displayed separately, thereby effectively distinguishing between system artifacts and real product defects, avoiding misjudgment, improving detection accuracy and efficiency, and reducing unnecessary production interruptions and verification costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a flow chart of a method for detecting defects in an ultra-precision spring in one embodiment of the present invention; Figure 2 This is one of the flow charts of a method for detecting defects in an ultra-precision spring according to another embodiment of the present invention; Figure 3 This is a second flow chart of a method for detecting defects in an ultra-precision spring according to another embodiment of the present invention; Figure 4 This is a third flow chart of a method for detecting defects in an ultra-precision spring according to another embodiment of the present invention; Figure 5 This is a fourth flow chart of a method for detecting defects in an ultra-precision spring according to another embodiment of the present invention; Figure 6 This is a fifth flow chart of a method for detecting defects in an ultra-precision spring according to another embodiment of the present invention; Figure 7 This is a sixth flow chart of a method for detecting defects in an ultra-precision spring according to another embodiment of the present invention; Figure 8 This is a system block diagram of a super-precision spring defect detection system in another embodiment of the present invention; Description of reference numerals: 1. The super-precision spring defect detection system; 11, surface deformation acquisition module; 12, defect feature extraction module; 13, feature data storage module; 14, space-time constant judgment module; 15, defect type judgment module; 16, detection result display module. DETAILED DESCRIPTION

[0028] The technical solutions in the present application will be described in detail below with reference to the drawings in the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all embodiments. The components of the present application described and shown in the drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0029] It should be noted that: similar numbers and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. Meanwhile, in the description of the present application, the terms "first", "second" and the like are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0030] The conventional existing super-precision spring defect detection method has the problem of misjudging the phase distortion introduced by the transparent film formed and distributed unevenly on the surface of the image acquisition optical element as the spring body deformation when evaluating the spring stress distribution by using laser speckle interferometry. The self-calibration mechanism of the existing system usually only compensates for the intensity attenuation of the light signal, and cannot identify or correct such phase distortion. This misjudgment leads to the appearance of non-real defects in the detection report that do not match the actual situation, and these non-real defects are easily misjudged as system problems in the production process due to their repeatability in form and position. When the problem root information is blocked by external factors, the troubleshooting path fails, making problem diagnosis stuck.

[0031] To this end, the present application proposes a super-precision spring defect detection method, which combines Figure 1 as shown, comprising the following steps: S1, obtaining the surface deformation information of the super-precision diaphragm spring to be tested; S2, extracting current defect feature data from the surface deformation information; S3, storing the current defect feature data to the defect feature cache area; S4, comparing the current defect feature data with the historical defect feature data stored in the defect feature cache area to determine whether the current defect feature data has space-time constancy, and obtaining a space-time constancy judgment result; S5, according to the spatiotemporal constancy judgment result, marking the defect type corresponding to the current defect as a system artifact or a real product defect; S6, generating a detection report, and displaying the defect type mark based on a preset display rule for distinguishing between system artifacts and real product defects.

[0032] Wherein, the surface deformation information refers to the change data of the surface geometry or microtopography of the to-be-tested ultra-precision diaphragm spring under force or no force state, which can be obtained by using non-contact optical detection technologies such as laser speckle interferometry, optical profilometry or white light interferometry, and its purpose is to provide data basis for subsequent defect recognition and analysis. Wherein, the current defect feature data refers to the quantitative data extracted from the surface deformation information, which can represent the defect attributes, including the position coordinates, geometric size, shape characteristics, intensity distribution or stress concentration degree of the defect, etc. Its main purpose is to convert the original deformation data into structured information, so as to facilitate subsequent defect recognition and classification. Wherein, the spatiotemporal constancy refers to the property that the spatial position and morphological characteristics of the defect feature data remain consistent or similar at different time points or in different detection batches, which can be judged according to the similarity between the current detection result and the historical detection result. Its purpose is to distinguish the artifacts caused by system itself errors (such as optical pollution) from the real product defects of the product itself. Wherein, the system artifact refers to the pseudo-defect mark presented in the detection report caused by the defects of the detection system itself or external environmental interference (such as optical element pollution), which usually has spatiotemporal constancy; the real product defect refers to the structure or performance abnormality of the ultra-precision diaphragm spring caused by material, process or use, etc., which usually does not have spatiotemporal constancy or has randomness, and its purpose is to accurately classify the detected abnormalities and avoid misjudgment.

[0033] In some embodiments, to further illustrate the working principle described above, the present application is implemented as follows. For example, when acquiring the surface deformation information of the ultra-precision diaphragm spring to be tested, a laser measurement system based on digital holography or speckle interference principle can be used. A laser beam is emitted to irradiate the surface of the spring, and a CCD camera with high resolution is used to capture the reflected or transmitted speckle image. Then, the three-dimensional deformation field data of the spring surface is calculated through a phase unwrapping algorithm. When extracting the current defect feature data from these surface deformation information, image processing software can be used. For example, by setting threshold, edge detection, connected component analysis and other algorithms, the abnormal areas in the deformation field are identified, and the geometric features (such as area, perimeter, shape factor) of these areas, position information (such as center coordinates) and deformation intensity are extracted as defect feature data. When storing the current defect feature data to the defect feature cache area, a structured data storage (such as SQL database or NoSQL database) can be used for management, where each defect record contains its feature data, detection timestamp and corresponding product batch information. When comparing the current defect feature data with the historical defect feature data stored in the defect feature cache area to determine the spatiotemporal constancy, a feature matching based algorithm can be used. For example, the similarity of spatial position and morphological features between the current defect feature data and the defect data of the past N times detection in the cache area is calculated. If the number of similar defects appearing in similar positions exceeds the frequency set in advance, it is determined to have spatiotemporal constancy. According to the spatiotemporal constancy judgment result, when marking the defect type corresponding to the current defect as system artifact or real product defect, a rule engine can be set. For example, if it is determined to have spatiotemporal constancy, it is marked as system artifact; otherwise, it is marked as real product defect. Finally, when generating the detection report and displaying the defect type marking, the detected defects can be visualized on the three-dimensional model of the spring through the graphical user interface (GUI) software, and different colors or icons can be used to distinguish system artifacts (for example, marked with blue) and real product defects (for example, marked with red). At the same time, the defect list and classification statistics are provided in the report.

[0034] Optionally, in combination with Figure 2 As shown, the step of S4 comparing the current defect feature data with the historical defect feature data stored in the defect feature cache area to determine whether the current defect feature data has spatiotemporal constancy, obtaining a spatiotemporal constancy judgment result, includes: S41, acquiring the laser power of the current detection; S42, adjusting the current defect feature data or the historical defect feature data stored in the defect feature cache area according to the laser power; S43, comparing the adjusted current defect feature data with the adjusted historical defect feature data to determine whether the current defect feature data has spatiotemporal constancy, obtaining a spatiotemporal constancy judgment result.

[0035] Wherein, the adjustment to the current defect feature data or the historical defect feature data stored in the defect feature cache area refers to correcting the defect feature data according to the change of the laser power, so as to eliminate or weaken the influence of the laser power fluctuation on the data feature, so that it has comparability under different power conditions. Specifically, it can be through establishing a mathematical model or a functional relationship between the laser power and the defect feature data, such as a linear regression model, a polynomial model or a nonlinear mapping relationship, and then normalizing or compensating the data according to the model. It can also be through pre-calibrating standard defect feature data under different laser powers to construct a lookup table, and correcting based on the lookup table according to the current laser power in actual detection, or using machine learning method to train a model to learn the nonlinear relationship between the laser power and the defect feature data, and using the model for predictive adjustment. The purpose is to ensure that the comparison basis of the defect feature data remains consistent in the case of laser power change, thereby improving the accuracy of the spatiotemporal constancy judgment.

[0036] In some preferred embodiments, the application is implemented as follows. In acquiring the current detected laser power, an optical power sensor integrated in the laser emitter or the optical path can be employed to monitor and output the power value of the laser beam in real time. The sensor can be a photodiode whose output voltage is proportional to the received optical power, and an analog-to-digital converter is used to convert the analog signal into a digital power value. In adjusting the current defect feature data or the historical defect feature data stored in the defect feature cache according to the laser power, system calibration can be performed in advance. For example, in a controlled environment, a known and stable artifact is used as the test object, and multiple detections are performed at multiple different laser power levels (e.g., from the lowest working power to the highest working power, with a fixed step increment), and the artifact feature data detected each time is recorded. Based on these calibration data, a correction model between laser power and artifact feature data can be established, for example, a polynomial function is fitted by least squares method, or a multi-dimensional lookup table is constructed. When the system acquires the current defect feature data in actual detection, the current laser power is also acquired. The system can use the pre-established correction model or lookup table to adjust the current defect feature data or the historical defect feature data stored in the defect feature cache to the equivalent feature value under a preset standard laser power according to the laser power corresponding to it. For example, if the standard power is set to P_std, the current power is P_curr, and the historical power is P_hist, the current data F_curr can be adjusted to F_curr_adj=F_curr*(P_std / P_curr_factor), and the historical data F_hist can be adjusted to F_hist*(P_std / P_hist_factor), where factor is the power-related correction factor calculated according to the correction model. In comparing the adjusted current defect feature data with the adjusted historical defect feature data, various similarity or difference measurement methods can be used. For example, the Euclidean distance, cosine similarity or Pearson correlation coefficient between the adjusted current defect feature data and the adjusted historical defect feature data can be calculated. If the calculated similarity is higher than a preset similarity threshold, or the difference is lower than a preset difference threshold, it is determined that the current defect feature data has spatiotemporal constancy, otherwise it does not.

[0037] Optionally, in combination with Figure 3 As shown in S42, the step of adjusting the current defect feature data or the historical defect feature data stored in the defect feature cache according to the laser power comprises: S421, based on the artifact in the current defect feature data or the historical defect feature data stored in the defect feature cache, identifying the spatial distribution region of the artifact in the image; S422, determine the local laser power sensitivity of each position in the spatial distribution region of the artifact; S423, according to the laser power and the local laser power sensitivity, make spatial position related adjustment on the current defect feature data or the historical defect feature data stored in the defect feature cache area.

[0038] Wherein, the spatial distribution region of the artifact refers to the abnormal region caused by the system artifact in the image data, which is located and delimited by analyzing the defect feature data, and the purpose is to clarify the specific range that needs to be corrected, avoid unnecessary processing of non-artifact regions, and ensure the accuracy of the adjustment. Specifically, it can be identified by image processing algorithms such as edge detection, region growing, threshold segmentation, etc. to identify the artifact region with specific morphological or intensity characteristics; it can also be identified based on a pre-established artifact pattern library through pattern matching or machine learning methods. The local laser power sensitivity refers to the degree of change of the feature value of the artifact at different spatial positions in the image with the change of the laser power, and the purpose is to reflect the response difference of the artifact in different regions to the change of the laser power, and to provide a basis for subsequent fine correction. Specifically, it can be calibrated by experiment, and the artifact data is collected at different laser powers, and then the ratio of the change amount of the artifact feature to the change amount of the laser power is calculated; it can also be a mathematical model, which predicts the sensitivity of different positions by fitting historical data. The spatial position related adjustment refers to the adjustment of the defect feature data or the historical defect feature data, and the adjustment amount or adjustment method is not globally uniform, but varies according to the specific position of the artifact in the image and the corresponding local laser power sensitivity, and the purpose is to adjust the characteristics of the artifact at different spatial positions, avoid the problem of overcorrection or undercorrection caused by global adjustment. Specifically, it can be divided into multiple sub-regions, and different adjustment parameters are applied to each sub-region; it can also be a continuous adjustment function, which takes spatial coordinates as input and outputs the corresponding adjustment amount.

[0039] In some preferred embodiments, the specific procedure of adjusting the current defect feature data or the historical defect feature data stored in the defect feature cache according to the laser power can be implemented as follows. First, identify the spatial distribution region of the artifact in the image. This can be done based on the analysis of the defect feature data. For example, the system can pre-store a series of feature templates of known artifacts, and when a pattern highly matching these templates is detected in the current defect feature data, it is identified as an artifact. Alternatively, image segmentation algorithms such as threshold-based segmentation or region growing can be used to label the regions in the image with abnormal brightness, texture or shape features as artifact regions. In addition, a deep learning model can also be trained to learn and identify specific spatial distribution patterns of artifacts from a large amount of historical data. Further, determine the local laser power sensitivity of each position within the spatial distribution region of the artifact. This can be done in steps. For example, the system can obtain the artifact feature data and the corresponding laser power data of each position within the spatial distribution region of the artifact. These data can come from historical detection records or be obtained through special calibration experiments. Then, the system can calculate the ratio of the change in artifact feature value to the change in laser power at each position within the spatial distribution region of the artifact, thereby obtaining the local laser power sensitivity of each position. For example, a regression analysis can be performed on the artifact feature value and the laser power of each pixel or each small region to obtain a sensitivity coefficient. Based on the calculated results, the system can update and confirm these local laser power sensitivities, for example, by smoothing or statistical averaging to improve their stability. Finally, adjust the current defect feature data or the historical defect feature data stored in the defect feature cache according to the laser power and the local laser power sensitivity in a spatial position-dependent manner. Specifically, the spatial distribution region of the artifact can be spatially divided into a number of small spatial units, for example, divided into a grid-like pixel block or small region. Then, for each small spatial unit, the system can calculate the artifact feature correction value of the small spatial unit according to its corresponding local laser power sensitivity and corresponding laser power. For example, a pre-stored correction value lookup table can be consulted, which records the correction values under different sensitivities and different laser powers; or a pre-calibrated correction curve can be applied, which describes the functional relationship between the correction value and the sensitivity and the laser power. Finally, the calculated artifact feature correction value is applied to the feature value of the corresponding small spatial unit in the current defect feature data or the historical defect feature data stored in the defect feature cache to achieve spatial position-dependent adjustment of the data. For example, if the artifact feature value of a small spatial unit is too high due to laser power fluctuations, a negative correction value is calculated according to its local sensitivity and current laser power, and it is subtracted from the feature value to correct the deviation.

[0040] Optionally, in combination with Figure 4As shown, the step of determining the local laser power sensitivity of each position in the spatial distribution region of the artifact includes: S4221, obtaining artifact feature data and corresponding laser power data of each position in the spatial distribution region of the artifact; S4222, calculating the ratio of the artifact feature change amount and the laser power change amount of each position in the spatial distribution region of the artifact according to the artifact feature data and the laser power data, to obtain the local laser power sensitivity of each position in the spatial distribution region of the artifact; S4223, updating and confirming the local laser power sensitivity of each position in the spatial distribution region of the artifact based on the calculated results.

[0041] Wherein, the artifact feature data of each position in the spatial distribution region of the artifact refers to the artifact-related numerical value exhibited by each specific pixel point or micro region in the identified artifact region in the image, such as its brightness, contrast, texture, phase distortion degree or deformation amplitude, etc., which can be realized by pixel-level analysis or regional statistics of the artifact region using image processing algorithms, and its purpose is to quantify the specific performance of the artifact at different spatial positions. The corresponding laser power data refers to the instantaneous output power value of the laser beam measured synchronously with each artifact position when the artifact feature data is obtained, which can be realized by real-time monitoring on the laser path using a high-precision optical power meter and synchronously recording with the image acquisition system, and its purpose is to establish the correlation between the artifact feature and the input of the laser energy causing the artifact. The artifact feature change amount refers to the numerical difference of the artifact feature data of the same spatial position under different laser power conditions, which can be realized by difference calculation or ratio calculation of the current artifact feature data and the historical or reference artifact feature data, and its purpose is to reflect the response degree of the artifact feature to the laser power fluctuation. The laser power change amount refers to the numerical difference of the laser power data under different measurement times or different test conditions, which can be realized by difference calculation or ratio calculation of the current laser power data and the historical or reference laser power data, and its purpose is to quantify the fluctuation of the input of the laser energy causing the artifact feature change. The local laser power sensitivity of each position in the spatial distribution region of the artifact refers to the ratio between the artifact feature change amount and the corresponding laser power change amount of each specific position in the spatial distribution region of the artifact, which represents the response strength of the artifact feature of that position to the laser power fluctuation, which can be realized by mathematical model fitting, regression analysis or direct ratio calculation, and its purpose is to accurately quantify the differential sensitivity of the artifact to the laser power at different spatial positions, providing a basis for subsequent accurate correction.

[0042] In some preferred embodiments, the local laser power sensitivity of each position within the spatial distribution region of artifacts can be determined as follows: first, in the system calibration or a specific test mode, the laser power can be actively changed in a certain range, for example from P_min to P_max, in steps, and the artifact images corresponding to each power point can be recorded. For each pixel point or predefined micro-grid unit within the spatial distribution region of artifacts, its artifact feature data, such as pixel intensity, phase distortion value or deformation amplitude, can be extracted from these images. At the same time, the current laser power data can be accurately obtained and recorded synchronously with each image acquisition. Then, for each position within the spatial distribution region of artifacts, at least two sets of artifact feature data and corresponding laser power data under different laser powers can be selected. For example, for a certain pixel point, when the laser power changes from P1 to P2, its artifact feature data changes from F1 to F2. At this time, the ratio of the artifact feature change (F2-F1) to the laser power change (P2-P1) of the pixel point can be calculated, thereby obtaining the local laser power sensitivity of the position. This process can be repeated multiple times to obtain more stable sensitivity values. Finally, based on the results obtained by calculation, the local laser power sensitivity can be updated and confirmed. For example, a time series averaging method can be used, i.e. the sensitivity values calculated at different time points for the same position are averaged to eliminate random noise; or a spatial smoothing algorithm, such as two-dimensional Gaussian filtering, can be applied to smooth the sensitivity values of adjacent positions, thereby obtaining a more stable and reliable local laser power sensitivity map of each position within the spatial distribution region of artifacts.

[0043] Optionally, in combination with Figure 5 As shown in FIG. 4, the step of adjusting the current defect feature data or the historical defect feature data stored in the defect feature cache area in step S423 includes the following steps: S4231, spatially dividing the spatial distribution region of artifacts to form a plurality of micro-space units; S4232, calculating the artifact feature correction amount of the micro-space unit according to the local laser power sensitivity of the micro-space unit and the corresponding laser power; S4233, applying the artifact feature correction amount to the feature values of the corresponding micro-space unit in the current defect feature data or the historical defect feature data stored in the defect feature cache area, to realize the spatial position-related adjustment of the current defect feature data or the historical defect feature data stored in the defect feature cache area.

[0044] The spatial division refers to dividing a continuous or discrete region into multiple smaller and independent sub-regions according to preset rules or algorithms, which can be grid division, feature-based region growing or clustering division, and the purpose is to perform more detailed and localized processing on the artifact-affected region. The micro space unit refers to the smallest processing unit formed by spatial division, which has a specific size and location, and can be a pixel, a pixel block, a predefined small region or an irregular local region, and the purpose is to realize fine correction of the artifact feature. The artifact feature correction amount refers to a numerical value or vector used to eliminate or reduce the influence of artifacts on defect feature data, which can be a scalar correction value, a proportion factor or a complex transformation matrix, and the purpose is to correct the defect feature data disturbed by artifacts. The feature value refers to the quantitative value corresponding to a specific spatial position or micro space unit in the defect feature data, which can be a pixel intensity value, a stress value, a deformation value or a statistical feature value, and the purpose is to represent the defect feature information of the position.

[0045] In some preferred embodiments, when the system identifies a circular artifact region caused by lens contamination in the image, in order to perform fine correction on the artifact, the system can first perform spatial division on the circular artifact region. Specifically, the circular region can be divided into several micro space units with a size of 10x10 pixels. For each such micro space unit, the system obtains its pre-calibrated local laser power sensitivity, for example, by measuring the unit multiple times under different laser powers and recording the ratio of the artifact feature change amount to the laser power change amount. At the same time, the system obtains the real-time laser power during the current defect detection. Subsequently, the system can calculate the artifact feature correction amount of the micro space unit according to the local laser power sensitivity of the micro space unit and the current laser power by consulting a pre-stored correction amount lookup table. For example, if the sensitivity of a certain unit is high and the current laser power deviates greatly from the standard value, the correction amount can be large. Finally, the system applies the calculated artifact feature correction amount to the feature values in the current defect feature data or historical defect feature data corresponding to the micro space unit. For example, if the feature value is pixel intensity, the correction amount can be an intensity value increment or a proportion factor, which is added to or multiplied by the original feature value to correct the defect feature data of the unit and make it closer to the feature in the true no-artifact state. By performing independent and local sensitivity-based correction on each micro space unit in the artifact region, the interference of artifacts on defect features can be effectively eliminated, and the accuracy of defect detection can be significantly improved.

[0046] Optionally, the step of calculating the artifact feature correction amount of the micro space unit comprises: obtaining a local laser power sensitivity corresponding to the micro space unit; obtaining a current laser power; determining the artifact feature correction amount of the micro space unit according to the local laser power sensitivity corresponding to the micro space unit and the current laser power, by consulting a pre-stored correction amount lookup table or applying a pre-calibrated correction curve.

[0047] The local laser power sensitivity corresponding to the micro space unit refers to the degree of change of the artifact feature value in a specific micro space region in a laser speckle interferometry system with the change of the laser power. The local laser power sensitivity can be represented by the ratio of the artifact feature change amount to the laser power change amount. The purpose is to quantify the response difference of different spatial positions to the laser power fluctuation. The current laser power refers to the instantaneous laser energy intensity actually output by the laser during the defect detection. The laser power can be measured in real time by a laser power meter or obtained from a system control unit. The purpose is to provide real-time input parameters for artifact feature correction calculation. The pre-stored correction amount lookup table refers to a data structure pre-established and stored in the system memory or database, which contains the corresponding artifact feature correction amount under different combinations of local laser power sensitivity and laser power. The lookup table can be stored in the form of a two-dimensional array or a hash table. The purpose is to achieve efficient and accurate correction amount acquisition by fast lookup. The pre-calibrated correction curve refers to a mathematical expression describing the functional relationship between the local laser power sensitivity, the laser power and the artifact feature correction amount, which can be in the form of a polynomial function, an exponential function or a piecewise linear function. The purpose is to accurately calculate the artifact feature correction amount by function calculation. The artifact feature correction amount refers to the value used to correct or compensate the defect feature data deviation introduced by the system artifact. The artifact feature correction amount can be in the form of deformation value, stress value or pixel intensity value. The purpose is to eliminate the interference of the artifact on the defect detection result and improve the detection accuracy.

[0048] In some preferred embodiments, the calculation of the artifact feature correction amount of the micro space unit can be specifically implemented as follows. First, during the system initialization or periodic calibration phase, a series of experimental calibration can be performed. For example, in a controlled environment, the laser power is changed, and the artifact feature data of different micro space units under different laser powers are measured. These data can be used to construct the correction amount lookup table or the correction curve.

[0049] Specifically, the artifact feature measurement can be performed in advance at a plurality of discrete laser power points (e.g., P1, P2, P3,...), for each micro space unit (e.g., unit A, unit B, unit C,...), and combined with its local laser power sensitivity, to calculate the corresponding artifact feature correction amount. These correction amounts can be stored in a two-dimensional lookup table, where the rows represent different micro space units or their local laser power sensitivity ranges, and the columns represent different laser power ranges. The values in the table are the corresponding artifact feature correction amounts. When actual detection is performed, the system obtains the local laser power sensitivity of the current micro space unit and the current laser power. Then, the system can quickly look up the corresponding artifact feature correction amount in the lookup table according to the obtained local laser power sensitivity and the current laser power, through interpolation or direct matching.

[0050] As another specific embodiment, the artifact feature correction amount can also be determined by a pre-calibrated correction curve. For example, through regression analysis or curve fitting, the experimental data is fitted into a mathematical function curve, which describes the relationship between the artifact feature correction amount and the local laser power sensitivity and the laser power. For example, a multivariate function F(S, P) = M can be established, where S represents the local laser power sensitivity, P represents the laser power, and M represents the artifact feature correction amount. In the actual detection process, the system obtains the current local laser power sensitivity S and the laser power P, and then substitutes these values into the pre-calibrated correction curve function F to calculate the accurate artifact feature correction amount M. Through these two ways, the system can ensure accurate correction of the artifact feature under different operating conditions, thereby improving the reliability of defect detection.

[0051] Optionally, in combination with Figure 6 As shown in FIG. 7, the step of judging whether the current defect feature data has spatio-temporal constancy to obtain a spatio-temporal constancy judgment result includes: A1, quantifying the similarity between the current defect feature data and the historical defect feature data stored in the defect feature cache area; A2, judging whether the current defect feature data has spatio-temporal constancy according to the relationship between the similarity and the preset similarity threshold, to obtain a spatio-temporal constancy judgment result.

[0052] The similarity between the current defect feature data and the historical defect feature data stored in the defect feature cache area refers to the degree of similarity or difference between two or more data objects, which is converted into a numerical value through mathematical methods. Specifically, various measurement methods can be used, such as calculating the Euclidean distance, Manhattan distance, cosine similarity, Pearson correlation coefficient, or structural similarity index (SSIM), etc. The purpose is to convert the complex defect feature data difference into a comparable numerical value to provide a basis for subsequent automated judgment.

[0053] The preset similarity threshold refers to a critical value for distinguishing "similar" and "not similar" when judging similarity. It can be obtained through historical data analysis, expert experience, or machine learning model training, such as setting a certain percentage value or a specific numerical range. The purpose is to provide an objective judgment standard to avoid errors caused by subjective judgment and allow the system to adjust the strictness of the judgment according to actual needs.

[0054] In some preferred embodiments, the application is implemented as follows. Assuming that the current defect feature data and the historical defect feature data stored in the defect feature cache area are represented as multi-dimensional vectors, for example, each vector contains the position coordinates of the defect, the shape parameters (such as area, perimeter, aspect ratio), and the stress distribution characteristics (such as maximum stress value, stress gradient). When judging whether the current defect feature data has spatial and temporal constancy, first, the Euclidean distance can be used to quantify the similarity between the current defect feature data and the historical defect feature data stored in the defect feature cache area. Specifically, the current defect feature data is represented as vector A, and the historical defect feature data is represented as vector B. The Euclidean distance between them can be calculated as the square root of the sum of the squares of the differences in each corresponding dimension. The smaller the Euclidean distance, the more similar the two defect feature data. Then, the system will judge whether the current defect feature data has spatial and temporal constancy according to the relationship between the calculated Euclidean distance and the preset similarity threshold. For example, a Euclidean distance threshold can be preset. If the calculated Euclidean distance is less than or equal to the threshold, it is considered that the current defect feature data has spatial and temporal constancy, i.e., the defect is a repeated systematic artifact; if the Euclidean distance is greater than the threshold, it is considered that the current defect feature data does not have spatial and temporal constancy, and it may be a real product defect or random noise. This preset similarity threshold can be determined by statistical analysis of the Euclidean distance distribution obtained by measuring the artifact feature data at the same position multiple times under the condition of no defect, for example, setting it as a certain percentile of the distribution to ensure that it can effectively distinguish between artifacts and real defects.

[0055] Optionally, in combination with Figure 7As shown, A2 determines whether the current defect feature data has spatial and temporal constancy according to the relationship between the similarity and the preset similarity threshold, and obtains a spatial and temporal constancy determination result. The step of obtaining the spatial and temporal constancy determination result includes: A21, obtaining defect feature data of a reference piece; A22, comparing the defect feature data of the reference piece with pre-stored reference piece standard data to obtain reference piece feature deviation; A23, correcting the preset initial threshold according to the reference piece feature deviation to obtain a corrected preset similarity threshold; A24, determining whether the current defect feature data has spatial and temporal constancy according to the relationship between the similarity and the preset similarity threshold, and obtaining a spatial and temporal constancy determination result.

[0056] The defect feature data of the reference piece refers to defect-related information obtained by measuring a reference piece in a known state, which can be surface deformation data, stress distribution data, or derived geometric feature data of the reference piece. The purpose is to provide a reference for subsequent threshold correction. The reference piece standard data refers to pre-stored feature information of the reference piece representing an ideal or known qualified state, which can be a theoretical design value, a gold sample data after precise calibration, or a statistical average data. The purpose is to serve as a reference for comparison to quantify the actual deviation of the reference piece. The reference piece feature deviation refers to the difference between the defect feature data of the reference piece and the reference piece standard data, which can be positional deviation, shape deviation, stress deviation, or a combination thereof. The purpose is to reflect the manufacturing tolerance of the reference piece itself, the measurement system error, or the influence of environmental factors on the measurement result. The preset initial threshold refers to a fixed similarity limit used to determine the spatial and temporal constancy of defects before any correction, which can be an empirical value, a statistical average value, or a fixed value set based on a theoretical model. The purpose is to provide a preliminary judgment standard. The corrected preset similarity threshold refers to a new similarity limit obtained by adjusting the preset initial threshold after considering the reference piece feature deviation, which can be a dynamic value obtained by linear adjustment, nonlinear mapping, or lookup table method. The purpose is to make the similarity judgment standard more suitable for the actual detection environment and the characteristics of the reference piece, and improve the accuracy of the judgment.

[0057] In some preferred embodiments, the application is implemented as follows: first, a precisely calibrated, known defect-free ultra-precision diaphragm spring can be selected as a reference piece. Through a laser speckle interferometric measurement system, the surface deformation information of the reference piece is obtained, and its defect feature data is extracted therefrom, for example, the stress distribution characteristics, geometric profile characteristics or texture characteristics of a specific area in the surface deformation map can be extracted. Then, the reference piece defect feature data obtained is compared with the reference piece standard data pre-stored in the system database. The reference piece standard data can be the theoretical design deformation data of the spring of this type, or the average deformation characteristics obtained by statistical analysis of a large number of qualified products. The comparison process can use various algorithms, for example, calculating the Euclidean distance, correlation coefficient or structural similarity index between the two sets of data, so as to quantitatively obtain the reference piece feature deviation. This deviation can specifically represent the difference in position, shape or stress value between the actual deformation of the reference piece and the standard deformation. Subsequently, according to the calculated reference piece feature deviation, the initial preset similarity threshold is corrected. For example, if the reference piece feature deviation is large in a certain direction (such as deformation amplitude), a preset correction function such as a linear function or a piecewise function can be used to appropriately magnify the initial threshold; on the contrary, if the deviation is small, it can be fine-tuned or kept unchanged. The corrected preset similarity threshold will be more in line with the actual performance of the current detection system and the inherent characteristics of the reference piece. Finally, when actually detecting the spring to be tested, the system calculates the similarity between the current defect feature data and the historical defect feature data stored in the defect feature cache. Then, this similarity is compared with the preset similarity threshold just corrected. If the similarity is higher than the corrected threshold, it is judged that the current defect has spatial and temporal constancy; otherwise, it is judged that it does not have spatial and temporal constancy, thereby obtaining the final spatial and temporal constancy judgment result.

[0058] Optionally, the step of correcting the preset initial threshold according to the reference piece feature deviation to obtain the corrected preset similarity threshold comprises: obtaining the position deviation, shape deviation and stress deviation in the reference piece feature deviation; according to the combination of the position deviation, shape deviation and stress deviation, consulting the pre-stored correction mapping table to determine the correction amount of the corresponding position, shape and stress part in the preset initial threshold; applying each determined correction amount to the corresponding position, shape and stress part in the initial threshold to obtain the corrected preset similarity threshold.

[0059] The correction mapping table refers to a lookup table pre-established for storing the corresponding relationship between different types of reference part characteristic deviation combinations and corresponding threshold correction values, which can be implemented in the form of a database, a lookup matrix or a multi-dimensional array, and the purpose is to quickly and accurately determine the required threshold adjustment parameter according to the actual deviation of the reference part. The correction value is calculated or looked up according to the specific characteristic deviation (such as position deviation, shape deviation, stress deviation) of the reference part, and is used to adjust the value of the corresponding part of the preset initial threshold, which can be a positive value, a negative value or zero, and the purpose is to finely and dynamically adjust the initial threshold to adapt to the actual state of the reference part, thereby improving the accuracy of the space-time constancy judgment.

[0060] In some preferred embodiments, in order to correct the preset initial similarity threshold, the surface deformation data of the reference part can be first obtained by a high-precision sensor array, and combined with three-dimensional scanning technology, the small displacement of the reference part in space position (i.e. position deviation) and the difference between its geometric shape and the standard model (i.e. shape deviation) can be extracted. At the same time, the laser speckle interferometry technology can be used to measure the residual stress distribution of the reference part and compare it with the standard stress distribution, so as to quantify the internal stress deviation. These deviation data can be digitized and stored. Subsequently, a processing unit can receive the combined data of the position deviation, the shape deviation and the stress deviation. The processing unit can have a pre-stored correction mapping table, which can be a multi-dimensional lookup table, and the index can be different ranges of deviation combinations, and the corresponding value can be the correction value of the corresponding part of the preset initial threshold. For example, if the position deviation is in a certain range, the shape deviation is in another range, and the stress deviation is in a third range, the processing unit can consult the table to determine the specific correction value for the position-sensitive part, the shape-sensitive part and the stress-sensitive part of the initial threshold. Finally, the processing unit applies the determined correction value to the corresponding position, shape and stress part of the preset initial threshold by weighted average or direct superposition, thereby generating a dynamically adjusted preset similarity threshold. For example, if the initial threshold is defined as multiple components (such as position threshold, shape threshold, stress threshold), each component can be added or subtracted by the corresponding correction value to obtain the final corrected threshold. This corrected threshold will be used for subsequent defect feature data similarity judgment to improve the accuracy of the judgment.

[0061] It should be noted that in the ultra-precision spring defect detection method, the dynamically adjusted threshold is to improve the accuracy and reliability of the defect judgment, especially when the reference part itself may have manufacturing deviations or measurement errors.

[0062] The specific embodiment of dynamically adjusting the threshold can be as follows: First, the defect feature data of a reference piece needs to be acquired. This reference piece can be a standard spring known as a qualified product, or a specific sample with known and stable deviation. The reference piece is detected by a laser speckle interferometric measurement system to obtain its surface deformation information, and the data representing its features are extracted therefrom. These feature data can include but are not limited to the deformation amplitude of the reference piece at a specific location, the stress concentration degree, the geometric profile parameters, etc.

[0063] Next, the acquired defect feature data of the reference piece is compared with the pre-stored reference piece standard data. The reference piece standard data is the feature value of the reference piece in an ideal state or after strict calibration, which can be established by theoretical calculation, CAD model data or statistical average of a large number of historical qualified products. By comparison, the difference between the actual features of the reference piece and the standard features can be quantified, thereby obtaining the reference piece feature deviation. This deviation can be decomposed into multiple dimensions, such as position deviation (offset between actual position of the reference piece and standard position), shape deviation (difference between actual shape of the reference piece and standard shape), and stress deviation (difference between actual stress distribution of the reference piece and standard stress distribution).

[0064] Then, the initial similarity threshold is modified according to the decomposed reference piece feature deviation. The initial similarity threshold is a fixed standard used by the system to judge the defect spatiotemporal constancy before any dynamic adjustment. In order to modify it, a modification mapping table can be established in advance. This modification mapping table is a multi-dimensional lookup table containing the threshold modification amount corresponding to different combinations of position deviation, shape deviation and stress deviation. For example, if the position deviation of the reference piece is large, the shape deviation is small, and the stress deviation is moderate, the system will look up the modification mapping table according to this specific deviation combination to find the adjustment values for the "position sensitive part", "shape sensitive part" and "stress sensitive part" of the initial threshold respectively.

[0065] Finally, these modification amounts determined from the modification mapping table are applied to the corresponding position, shape and stress parts of the preset initial threshold. For example, if the initial threshold is defined to include multiple components such as position threshold, shape threshold and stress threshold, each component will be adjusted according to its corresponding modification amount. In this way, a modified preset similarity threshold that is finely adjusted and more consistent with the current detection system state and the actual characteristics of the reference piece can be obtained. This modified threshold will be used for subsequent defect spatiotemporal constancy judgment, i.e. when comparing the similarity between the defect feature data of the current spring to be tested and the historical defect feature data, it will be compared with this dynamically adjusted threshold, thereby obtaining a more accurate spatiotemporal constancy judgment result.

[0066] In the defect detection method of ultra-precision springs, the preliminary setting of the threshold value is a key step to ensure the accuracy of detection. According to the understanding of those skilled in the art, the preliminary setting of the threshold value can be based on the following implementation: One implementation is based on statistical analysis of historical data. In actual production environment, a large amount of detection data of known qualified and known defective ultra-precision springs can be collected. For these data, the corresponding defect feature data can be extracted, and the similarity between the current defect feature data and the historical defect feature data can be calculated. By statistical analysis of these similarity data, such as calculating the mean, standard deviation or fitting the distribution, a similarity threshold value can be determined which can effectively distinguish between qualified and defective products. For example, a statistical confidence interval can be set, and the similarity values falling outside the interval can be used as the preliminary judgment basis for defects.

[0067] Another implementation is based on expert experience setting. In the absence of a large amount of historical data, experienced quality control engineers or technical experts can rely on their long-term observation and understanding of ultra-precision spring defects, combined with the tolerance of defects in actual production, to directly set an initial similarity threshold value. For example, experts can give a preliminary similarity range based on the visual features of the defects, the degree of influence on product performance, and past detection experience, as a basis for judging the constancy of space and time.

[0068] In addition, reference calibration method can also be used for preliminary setting. By selecting one or more reference pieces of known state (for example, a standard defect-free spring or a spring with typical artifacts), multiple detections are carried out under controlled conditions. Record the defect feature data of these reference pieces, and calculate the similarity between them and their standard data. According to these similarity values, an initial threshold value can be set, which can reflect the inherent fluctuation of the system under ideal or known deviation conditions. For example, if a standard defect-free reference piece has a similarity between its defect feature data and standard data that remains at a certain high level in multiple detections, a value near this level can be used as the initial similarity threshold value.

[0069] In actual operation, these preliminary set threshold values may need to be verified and adjusted in small batch trial production or actual operation to ensure their effectiveness and robustness in real production environment. For example, after the preliminary threshold value is set, a small amount of known defective and qualified products can be detected, and the threshold value can be fine-tuned according to the detection results to achieve the best detection performance.

[0070] An ultra-precision spring defect detection system, combining Figure 8 As shown in the figure, the structure of the ultra-precision spring defect detection system 1 includes: The surface deformation acquisition module 11 is used to acquire the surface deformation information of the ultra-precision diaphragm spring to be tested. a defect feature extraction module 12 configured to extract current defect feature data from the surface deformation information; a feature data storage module 13 configured to store the current defect feature data into a defect feature cache; a space-time constancy judgment module 14 configured to compare the current defect feature data with historical defect feature data stored in the defect feature cache to determine whether the current defect feature data has space-time constancy, and obtain a space-time constancy judgment result; a defect type judgment module 15 configured to mark a defect type corresponding to the current defect as a system artifact or a real product defect according to the space-time constancy judgment result; a detection result display module 16 configured to generate a detection report and display the defect type mark based on a preset display rule for distinguishing between system artifacts and real product defects.

[0071] The surface deformation acquisition module refers to a unit for capturing the microscopic deformation of the surface of the spring to be tested. It can be implemented by using optical measurement equipment such as a laser speckle interferometer, a white light interferometer, a confocal microscope, or a three-dimensional scanner. Its purpose is to obtain high-precision surface deformation data as the original input for defect analysis. The defect feature extraction module refers to a unit for identifying and quantifying defect-related features from the original deformation information. It can be implemented by using image processing algorithms, signal processing techniques, or machine learning models. Its purpose is to convert the original deformation data into structured feature data for subsequent judgment. The feature data storage module refers to a unit for persistently or temporarily storing defect feature data. It can be implemented by using a cache, a solid-state drive, a database system, or a distributed storage system. Its purpose is to establish a historical record of defect features to support time-series analysis and comparison. The space-time constancy judgment module refers to a unit for evaluating whether the defect features remain consistent at different time points and spatial locations. It can be implemented by using statistical methods, pattern recognition algorithms, or rule-based logical judgments. Its purpose is to distinguish between artifacts caused by system inherent characteristics and real product defects. The defect type judgment module refers to a unit for classifying and marking the identified defects according to the space-time constancy judgment result. It can be implemented by using classification algorithms such as decision trees, support vector machines, neural networks, or expert systems based on preset logical rules. Its purpose is to clearly distinguish between system artifacts and real product defects to provide accurate basis for subsequent processing. The detection result display module refers to a unit for presenting the defect detection results to users in a visualized manner. It can be implemented by using graphical user interface software, data visualization tools, or report generation engines. Its purpose is to visually display the detection results, facilitate user understanding and decision-making, and highlight the differences between artifacts and real defects.

[0072] The scheme of the present application realizes the automation and intelligence of ultra-precision spring defect detection through modular system design, especially in distinguishing between system artifacts and real product defects, it provides an effective technical means. Specifically, the surface deformation acquisition module is responsible for obtaining the surface deformation information of the ultra-precision diaphragm spring to be tested first, which provides the original data input for the entire detection process. Then, the defect feature extraction module processes these original deformation information, identifies and extracts the feature data of the current detected defect from it, and converts the unstructured deformation information into quantifiable feature representation. These extracted current defect feature data are received by the feature data storage module and stored in the defect feature cache area, which not only saves the current data, but also accumulates the defect feature data of historical detection, laying a foundation for subsequent time sequence analysis. The space-time constant judgment module is the core of the system, which compares the current defect feature data with the historical defect feature data stored in the defect feature cache area to judge whether the current defect feature data has constancy in time and space. This comparison mechanism can identify those "defects" that repeatedly appear in the same position and have similar morphology in multiple detections, which are usually artifacts caused by optical pollution or environmental interference of the system itself. In this way, the system can get a space-time constancy judgment result. Then, the defect type determination module classifies the current defect according to the space-time constancy judgment result. If the judgment result shows that the defect has space-time constancy, it is marked as a system artifact; otherwise, it is marked as a real product defect. This judgment mechanism based on space-time constancy enables the system to automatically and accurately distinguish between false signals introduced by the detection equipment itself and real defects of the product, avoiding misjudgment. Finally, the detection result display module generates a detailed detection report according to the output of the defect type determination module. Based on the preset display rules, the system artifact and the real product defect are displayed differently, for example, different colors, icons or marks can be used to intuitively present the two different types of defects. This enables users to immediately identify which are the product quality problems that need attention and which are the system itself interference, thereby avoiding production stoppage and resource waste caused by misjudging artifacts. Through the above system structure, the present application upgrades the original ultra-precision spring defect detection method from pure algorithm logic to a solution with physical support and automation capability. The system can efficiently perform each step in the method, especially through the cooperation of the space-time constant judgment module and the defect type determination module, solving the problem that the traditional method is difficult to automatically and accurately distinguish between system artifacts and real product defects in actual application. This systematic implementation makes the detection process more reliable, reduces the risk of manual intervention and misjudgment, and significantly improves the efficiency and accuracy of detection.

[0073] In some preferred embodiments, the present application is implemented as follows: the surface deformation acquisition module can use a measurement device based on the principle of laser speckle interferometry. For example, a high-power laser can be configured as the light source, a beam splitter can divide the laser into a reference beam and an object beam, the object beam can be incident on the surface of the ultra-precision diaphragm spring to be measured, and after reflection, the reference beam and the object beam can form an interference speckle pattern on a CCD or CMOS image sensor. The image sensor can be a high-resolution industrial camera, for example, a camera with high frame rate and low noise characteristics can be selected to capture the small deformation information of the spring surface. The defect feature extraction module can be an embedded processor or an industrial computer running image processing software. The software can first preprocess the acquired speckle pattern, for example, noise filtering and image enhancement. Subsequently, a digital image correlation algorithm or a phase unwrapping algorithm can be applied to calculate the full-field deformation data of the spring surface from the speckle pattern. Then, feature extraction algorithms such as gradient-based, curvature-based, or local texture analysis methods can be used to identify abnormal areas from the deformation data and extract their geometric and deformation features. The feature data storage module can be a high-speed solid-state drive or a dedicated data server with a database management system. The current defect feature data extracted each time, including the defect location coordinates, geometric features, deformation features, and detection timestamp, can be stored in the database in a structured manner. The database can maintain a defect feature cache area to store historical defect feature data of the last period or the last batch of products for quick retrieval and comparison. The spatiotemporal constancy judgment module can be implemented by a high-performance central processing unit or a graphics processing unit, and a comparison algorithm is run on it. The algorithm can retrieve historical defect feature data similar to the current defect location from the defect feature cache area. During comparison, the similarity between the current defect feature data and the historical defect feature data can be calculated, for example, the Euclidean distance, cosine similarity, or structural similarity index can be used for quantification. If the similarity exceeds the preset threshold and the defects are highly coincident in spatial position, it can be judged to have spatiotemporal constancy. The defect type determination module can be a separate software module that receives the spatiotemporal constancy judgment result. This module can have a built-in decision logic, for example, if the spatiotemporal constancy judgment module outputs "constant", the current defect is marked as "system artifact"; if it outputs "non-constant", it is marked as "real product defect". This logic can be adjusted according to the actual application scenario, for example, confidence assessment can be introduced to improve the robustness of the judgment. The detection result display module can be a display connected to the industrial computer and running a user-friendly graphical interface program. The program can display the detection progress and results in real time. When generating the detection report, the program can mark the system artifacts with blue or dashed lines and the real product defects with red or solid lines according to the preset display rules, and provide detailed defect list and statistical information in the report.The user can view the images, feature data and determination results of the defects through the interface, and can export a report for archiving or further analysis.

[0074] The above merely provides an example of the present application, and is not intended to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for detecting defects in ultra-precision springs, characterized in that: The following steps are involved: Obtaining surface deformation information of the ultra-precision diaphragm spring to be tested; extracting current defect feature data from the surface deformation information; storing the current defect feature data in a defect feature buffer; Comparing the current defect feature data with historical defect feature data stored in a defect feature buffer area to determine whether the current defect feature data has spatiotemporal constancy, and obtaining a spatiotemporal constancy determination result; According to the spatiotemporal constancy judgment result, a defect type label corresponding to the current defect is determined as a system artifact or a real product defect; An inspection report is generated and the defect type mark is displayed based on preset display rules for distinguishing and displaying system artifacts and real product defects.

2. A method for detecting defects in super-precision springs according to claim 1, characterized in that: The step of comparing the current defect feature data with the historical defect feature data stored in the defect feature buffer to determine whether the current defect feature data has spatiotemporal constancy and obtaining a spatiotemporal constancy determination result includes: Get the currently detected laser power; adjusting the current defect feature data or the historical defect feature data stored in the defect feature buffer according to the laser power; The adjusted current defect feature data is compared with the adjusted historical defect feature data to determine whether the current defect feature data has spatiotemporal constancy, and a spatiotemporal constancy determination result is obtained.

3. A method for detecting defects in super-precision springs according to claim 2, characterized in that: The step of adjusting the current defect feature data or the historical defect feature data stored in the defect feature buffer according to the laser power includes: Based on the artifacts in the current defect feature data or the historical defect feature data stored in the defect feature buffer, identifying a spatial distribution area of ​​the artifacts in the image; Determine the local laser power sensitivity at each location within the spatial distribution area of ​​the artifact; According to the laser power and the local laser power sensitivity, spatial position-dependent adjustments are performed on the current defect feature data or the historical defect feature data stored in the defect feature buffer.

4. A method for detecting defects in super-precision springs according to claim 3, characterized in that: The step of determining the local laser power sensitivity of each position within the spatial distribution area of ​​the artifact comprises: Acquire artifact feature data and corresponding laser power data at each position within the spatial distribution area of ​​the artifact; Calculating, based on the artifact feature data and the laser power data, a ratio of an artifact feature change to a laser power change at each position within the artifact's spatial distribution area to obtain a local laser power sensitivity at each position within the artifact's spatial distribution area; Based on the calculated results, the local laser power sensitivity of each position within the spatial distribution area of ​​the artifact is updated and confirmed.

5. The method for detecting defects in super-precision springs according to claim 3, characterized in that: The step of adjusting the current defect feature data or the historical defect feature data stored in the defect feature buffer area in relation to the spatial position includes: Divide the spatial distribution area of ​​the artifact into several small spatial units; Calculating an artifact feature correction amount of the tiny space unit according to the local laser power sensitivity and the corresponding laser power corresponding to the tiny space unit; The artifact feature correction amount is applied to the feature value corresponding to the tiny spatial unit in the current defect feature data or the historical defect feature data stored in the defect feature cache, so as to achieve spatial position-related adjustment of the current defect feature data or the historical defect feature data stored in the defect feature cache.

6. A method for detecting defects in super-precision springs according to claim 5, characterized in that: The step of calculating the artifact feature correction amount of the small space unit includes: Obtaining the local laser power sensitivity corresponding to the micro-space unit; Get the current laser power; According to the local laser power sensitivity corresponding to the micro-space unit and the current laser power, the artifact feature correction amount of the micro-space unit is determined by consulting a pre-stored correction amount lookup table or applying a pre-calibrated correction curve.

7. The method for detecting defects in super-precision springs according to claim 2, characterized in that: The step of determining whether the current defect feature data has spatiotemporal constancy and obtaining a spatiotemporal constancy determination result comprises: quantifying the similarity between the current defect feature data and the historical defect feature data stored in the defect feature buffer; According to the relationship between the similarity and a preset similarity threshold, it is determined whether the current defect feature data has spatiotemporal constancy, and a spatiotemporal constancy determination result is obtained.

8. The method for detecting defects in super-precision springs according to claim 7, characterized in that: The step of determining whether the current defect feature data has spatiotemporal constancy based on the relationship between the similarity and a preset similarity threshold, and obtaining a spatiotemporal constancy determination result comprises: Obtain defect characteristic data of the reference part; Comparing the defect feature data of the reference part with the pre-stored reference part standard data to obtain the reference part feature deviation; According to the characteristic deviation of the reference part, the preset initial threshold is corrected to obtain a corrected preset similarity threshold; According to the relationship between the similarity and the preset similarity threshold, it is determined whether the current defect feature data has spatiotemporal constancy, and a spatiotemporal constancy determination result is obtained.

9. A method for detecting defects in super-precision springs according to claim 8, characterized in that: The step of correcting the preset initial threshold value according to the reference part feature deviation to obtain a corrected preset similarity threshold value includes: Obtaining position deviation, shape deviation and stress deviation in the characteristic deviation of the reference part; According to the combination of the position deviation, shape deviation and stress deviation, a pre-stored correction mapping table is consulted to determine the correction amount corresponding to the position, shape and stress part of the preset initial threshold; The determined correction amounts are applied to the corresponding position, shape and stress parts of the initial threshold to obtain the corrected preset similarity threshold.

10. A super-precision spring defect detection system for performing super-precision spring defect detection, characterized in that: include: A surface deformation acquisition module is used to obtain surface deformation information of the ultra-precision diaphragm spring to be tested; A defect feature extraction module, configured to extract current defect feature data from the surface deformation information; A feature data storage module, configured to store the current defect feature data in a defect feature buffer area; a spatiotemporal constancy judgment module, configured to compare the current defect feature data with the historical defect feature data stored in the defect feature buffer area to judge whether the current defect feature data has spatiotemporal constancy, and obtain a spatiotemporal constancy judgment result; a defect type determination module, configured to determine, based on the spatiotemporal constancy determination result, a defect type label corresponding to the current defect as a system artifact or a real product defect; The test result display module is used to generate a test report and display the defect type mark based on a preset display rule for distinguishing and displaying system artifacts and real product defects.

Images