Mim electric clip blade flatness detection method and detection equipment

Abstract

The application relates to a MIM electric clip blade flatness detection method and a detection device, and relates to the field of electric clip blade processing detection.The application comprises the following steps: a workpiece conveying step, which is configured with a detection process identification strategy, is used for analyzing the detection process of a workpiece, and transmits an instruction to alternately convey the detected workpiece and the workpiece to be detected; a workpiece identification step, which is configured with a key area division strategy, is used for identifying the image features of the workpiece to determine the blade structure features of the workpiece, and dividing the workpiece detection areas with different detection priorities; a workpiece detection step, which is configured with a scanning parameter adaptation strategy, is used for adapting the sampling scanning parameters according to the detection priorities of the workpiece detection areas, and instructing a laser scanning unit to perform flatness detection to obtain flatness detection data; and a detection feedback step, which is configured with an identification feedback strategy, is used for analyzing and feedback adjusting the collection parameters of an image collection unit.The application has the effects of improving the flatness detection efficiency and reliability of the MIM electric clip blade.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric push blade machining detection, in particular to a MIM electric push shear blade flatness detection method and detection equipment. BACKGROUND

[0002] MIM is a metal injection molding process, which is widely used in the manufacturing of metal blades. The flatness of MIM electric push shear blades is directly related to their cutting sharpness, assembly compatibility and long-term stability. Flatness detection of MIM electric push shear blades is a key step to determine whether they meet the quality standards of finished products.

[0003] In related technologies, when detecting the flatness of MIM electric push shear blades, a feeler gauge is generally used in combination with manual operation: a detection personnel holds the feeler gauge, selects several points on the surface of the blade, and determines whether the flatness of the point is qualified by inserting the feeler gauge into the gap between the blade and the reference plane. The flatness of the whole blade is determined by comprehensively analyzing the detection results of each point.

[0004] For the above related technologies, the manual feeler gauge detection method has obvious limitations: on the one hand, the detection process completely depends on the experience of the operator, and the operation method and gap perception difference of different personnel will cause the detection accuracy to fluctuate, which is easy to cause the problems of misjudgment of qualified blades and missed detection of unqualified blades; on the other hand, the manual point-by-point detection is low in efficiency, and it takes a long time to detect a single blade, which is difficult to match the flow line detection demand of batch production of MIM electric push shear blades, which not only affects the overall production progress, but also cannot guarantee the consistency and reliability of the detection results. SUMMARY

[0005] In order to improve the flatness detection efficiency and reliability of MIM electric push shear blades, the present application provides a MIM electric push shear blade flatness detection method and detection equipment.

[0006] In the first aspect, the present application provides a MIM electric push shear blade flatness detection method:

[0007] A MIM electric push shear blade flatness detection method is configured with a flatness detection equipment, and the flatness detection equipment is provided with a laser scanning unit, an image acquisition unit and a workpiece conveying unit, characterized in that it comprises:

[0008] The workpiece conveying step is configured with a detection process identification strategy, which is used to analyze the detection process of the workpiece and issue a transmission instruction to alternately convey the detected workpiece and the workpiece to be detected;

[0009] The workpiece identification step is configured with a key area division strategy, which identifies the image features of the workpiece to determine the workpiece blade structure features, and divides the workpiece detection area with different detection priorities;

[0010] The workpiece detection step is configured with a scanning parameter adaptation strategy, and the sampling scanning parameter is adapted according to the detection priority of the workpiece detection area, and the planeness detection data is obtained by instructing the laser scanning unit to perform planeness detection according to the sampling scanning parameter;

[0011] The detection feedback step is configured with an identification feedback strategy, which is used to analyze whether the feature quantity of the workpiece image meets the preset reference feature quantity, and to feedback adjust the acquisition parameter of the image acquisition unit, so that the feature quantity of the workpiece image is kept within the preset reference quantity range.

[0012] Through the above technical solutions, the workpiece conveying, workpiece identification, workpiece detection and detection feedback steps cooperate with each other, so that the workpiece detection areas with different detection priorities are divided by image feature identification, so that the important structure area on the electric clipper blade can be scanned in detail when performing laser planeness scanning detection, and the number of workpiece image features is kept within the reference range through the identification feedback strategy, which helps to improve the accuracy of subsequent key area division, does not need to rely on the detection experience of personnel, and can increase the consistency of the detection result and improve the detection work efficiency.

[0013] Optionally, the scanning parameter adaptation strategy comprises:

[0014] According to the detection priority, the corresponding reference scanning parameter in the preset scanning parameter database is matched, and the reference scanning parameter reflects the accuracy of scanning the workpiece;

[0015] According to the image feature in the workpiece image feature identification process, the feature definition is analyzed, when the feature definition is lower than the preset reliable definition, the detection priority is calculated and updated by using the preset priority analysis model.

[0016] Through the above technical solutions, the scanning parameter adaptation strategy adapts the reference scanning parameters with different accuracies according to the detection priority of the workpiece detection area, and adjusts the detection priority of the detection area according to the feature definition of the image feature identified by the image, which can reduce the influence of inaccurate workpiece detection area priority division in the image identification process, thereby improving the reliability of the planeness detection result, and making the workpiece in different states can obtain a reasonable scanning detection scheme.

[0017] Optionally, the priority analysis model is calculated by using the following formula:

[0018] ;

[0019] wherein, is the updated detection priority, A priority of the initial matching reference is detected according to a preset scanning parameter database, A priority adjustment coefficient is dynamically adapted according to a hardness of the workpiece material, A preset reliable definition threshold, An actual recognized feature definition is determined through pixel analysis of the image.

[0020] By adopting the above technical solution, the priority analysis model calculates the updated detection priority by a quantitative formula, rather than relying on manual experience adjustment, can be targeted to respond to the detection needs of workpieces of different materials, makes the priority update range and the defect degree of definition directly related, reasonably improves the detection priority of the fuzzy area, and further ensures the detection accuracy of such area, reduces the risk of inaccurate detection caused by improper priority setting.

[0021] Optionally, the workpiece detection step is further configured with a flatness error correction strategy, comprising:

[0022] Surface roughness features are recognized according to the workpiece image to determine the area roughness corresponding to the workpiece rough area, and the workpiece rough area is analyzed to determine the abnormal rough area;

[0023] The area roughness is recognized based on the abnormal rough area, and when the area roughness exceeds a preset allowable roughness, the actual flatness is determined by a preset flatness correction model combined with the flatness detection data.

[0024] By adopting the above technical solution, the flatness error correction strategy first recognizes the surface roughness features of the workpiece and determines the abnormal rough area, and then corrects the flatness based on the roughness of the abnormal area, solving the problem of misjudgment of surface roughness as flatness error in traditional manual detection and pure laser scanning. When the area roughness exceeds the allowable threshold, the actual flatness is calculated by the correction model to ensure that the final detection result truly reflects the true flatness of the blade, rather than a false error affected by the surface appearance, further improving the accuracy of flatness detection.

[0025] Optionally, the flatness correction model is calculated by the following formula:

[0026] ;

[0027] Wherein, The corrected actual flatness, The original flatness detection data is directly obtained by the laser scanning unit, and k is the roughness correction coefficient, Determined according to the hardness of the blade material, The actual area roughness of the abnormal rough area, The preset allowable roughness threshold, and the area of the abnormal rough area, The total area of the workpiece detection area.

[0028] By adopting the above technical scheme, the flatness correction model calculates the actual flatness by a quantitative formula instead of qualitative correction, can adapt to the influence difference of roughness of different materials on flatness, reduces the problem that the correction amplitude is not reliable enough to be set by experience, makes the flatness correction more accurate and targeted, ensures that workpieces of different materials and different roughness levels can be reasonably corrected, and finally the output actual flatness data is more consistent with the actual working condition, and the detection deviation caused by improper correction is reduced.

[0029] Optionally, a freely movable adjacent laser scanning unit is further provided to form two scanning units for simultaneously detecting the flatness of the workpiece, and is configured with a batch detection and viewing angle optimization strategy:

[0030] According to the feature analysis of the workpiece image to determine the current number of detected workpieces, and adjusting through a preset viewing angle sharing strategy to determine the minimum detection angle of the adjacent laser scanning unit;

[0031] Based on the minimum detection angle, the laser scanning of the adjacent laser scanning unit is adjusted, and a preset adjustment gain model is analyzed to obtain a multi-workpiece synchronous scanning gain;

[0032] When the workpiece surface flatness detection time corresponding to the multi-workpiece synchronous scanning gain is greater than the preset optimal time gain, a minimum detection angle adjustment instruction is triggered to reduce the total number of synchronous detection workpieces and the minimum detection angle.

[0033] By adopting the above technical scheme, the freely movable adjacent laser scanning unit forms double-unit synchronous detection, which can directly improve the number of detected workpieces per unit time compared with single-unit detection, and can adapt to match batch production. The viewing angle sharing strategy determines the minimum detection angle by analyzing the workpiece image, reduces the computational load of subsequent image feature analysis, improves the detection speed, and the gain model analyzes the multi-workpiece synchronous scanning gain. When the time consumption exceeds the optimal threshold, the angle and the number of workpieces are adjusted, which helps to balance the batch detection efficiency and detection accuracy, keeps the synchronous detection in an efficient and accurate state at all times, and further adapts to the detection needs of large-scale production of MIM electric clippers.

[0034] Optionally, the preset adjustment gain model includes:

[0035] The scanning gain time and the number of multi-workpiece synchronous scanning are calculated to determine the multi-workpiece synchronous scanning gain;

[0036] The adjustment gain model is calculated by the following formula:

[0037] ;

[0038] Wherein, G is the time-consuming gain of multi-workpiece synchronous scanning, N is the total number of synchronous detection workpieces, T is the reference time-consuming of a single laser scanning unit detecting 1 workpiece, T is the actual total time-consuming of two laser scanning units synchronously detecting N workpieces, is the preset angle influence coefficient, is the additional time-consuming generated by the minimum detection angle adjustment, The detection priority is the current minimum detection angle, is the set optimal reference detection angle.

[0039] By adopting the above technical scheme, the actual gain of synchronous detection is accurately quantified through the calculation model, which provides a clear basis for whether to adjust the angle and the number of workpieces, avoids the problem of whether the time-consuming is reasonable by subjective judgment, ensures that the double-unit synchronous detection always maintains in the optimal efficiency interval, maximizes the efficiency advantage of batch detection, and at the same time guarantees that the detection accuracy is not excessively affected by the angle deviation.

[0040] Optionally, the identification feedback strategy comprises:

[0041] The reference feature quantity is determined by matching the acquisition parameters of the image acquisition unit with the corresponding reference feature quantity in the preset standard feature quantity library.

[0042] When the feature quantity identified by the workpiece image feature identification is less than the reference feature quantity, the environmental feature analysis is performed to determine the surface light reflection area of the workpiece;

[0043] Based on the surface light reflection area and the preset light compensation sub-strategy, the analysis is performed to determine the auxiliary light source adjustment parameters of the surface light reflection area, including the light source irradiation angle and the light source brightness;

[0044] Based on the auxiliary light source adjustment parameters, the preset compensation light unit is adjusted.

[0045] By adopting the above technical scheme, the identification feedback strategy determines the reference feature quantity by matching the preset standard feature quantity library, and analyzes the surface light reflection area of the workpiece when the feature quantity is insufficient, accurately adjusts the irradiation angle and brightness of the auxiliary light source in combination with the light compensation sub-strategy, effectively reduces the problem of insufficient image feature identification caused by the surface light reflection of the workpiece, ensures that the feature quantity of the workpiece image is stable within the reference feature quantity range, provides a reliable image basis for accurate division of the key area in the subsequent workpiece identification step, and improves the accuracy and stability of the overall flatness detection.

[0046] Optionally, the light compensation sub-strategy comprises:

[0047] According to the image light reflection feature analysis, the exposure parameters and the shadow parameters of the light reflection feature are determined, and the image light reflection feature is divided to obtain a light reflection analysis image;

[0048] Based on the preset image noise reduction parameters, the exposure parameters and the shadow parameters of the light reflection analysis image are adjusted to obtain an optimized light reflection image, and the effective image noise reduction parameters are determined when the image feature quantity of the workpiece is within the preset reference quantity range;

[0049] Based on the effective image noise reduction parameters, the corresponding auxiliary light source adjustment parameters in the preset light source database are matched.

[0050] By adopting the above technical solutions, the light compensation sub-strategy performs fine analysis on the light reflection feature, optimizes the light reflection image by dividing the light reflection analysis image and combining the image noise reduction parameters, determines the effective image noise reduction parameters to match the auxiliary light source adjustment parameters, this process can solve the interference of different light reflection features on image quality, improve the accuracy and effectiveness of light compensation, ensure that the image feature quantity of the workpiece is stable within the reference quantity range in the complex light reflection environment, provide support for efficient implementation of the identification feedback strategy, and further improve the adaptability and reliability of the MIM electric clip blade flatness detection method.

[0051] In a second aspect, the present application provides a MIM electric clip blade flatness detection device, which adopts the following technical solutions:

[0052] A MIM electric clip blade flatness detection device, comprising:

[0053] A workpiece conveying module configured with a detection process identification strategy, for analyzing the detection process of the workpiece, and issuing a transmission instruction to alternately convey the detected workpiece and the workpiece to be detected;

[0054] A workpiece identification module configured with a key area division strategy, for image feature identification of the workpiece to determine the workpiece blade structure features, and dividing the workpiece detection area into different detection priority levels;

[0055] A workpiece detection module configured with a scanning parameter adaptation strategy, for adapting the sampling scanning parameters according to the detection priority of the workpiece detection area, and instructing the laser scanning unit to perform flatness detection according to the sampling scanning parameters to obtain flatness detection data;

[0056] A detection feedback module configured with an identification feedback strategy, for analyzing whether the feature quantity of the workpiece image acquisition meets the preset reference feature quantity, and feedback adjusting the acquisition parameters of the image acquisition unit.

[0057] By adopting the technical scheme, the workpiece conveying module realizes efficient alternating flow of the workpiece, solves the problem of low efficiency of manual handling; the workpiece identification module accurately divides the key area and detects the priority, avoids the indiscriminate operation of manual detection; the workpiece detection module adapts the scanning parameter and drives the laser scanning unit, replaces the experience judgment of manual gauge, and improves the detection accuracy; the detection feedback module adjusts the acquisition parameter, ensures that the image feature meets the reference requirement, and solves the problem of poor consistency of manual detection results.

[0058] To sum up, the present application includes at least one of the following beneficial technical effects:

[0059] 1. The workpiece conveying, workpiece identification, workpiece detection and detection feedback steps cooperate with each other, so as to divide the workpiece detection area of different detection priorities through image feature recognition, so that the important structure area on the electric clipper blade can be scanned in detail when laser flatness scanning detection is performed, and the number of workpiece image features is further improved within the reference range through identification feedback strategy, which helps the accuracy when the key area is divided subsequently, does not need to rely on the detection experience of personnel, can increase the consistency of detection results and improve the detection work efficiency;

[0060] 2. The scanning parameter adaptation strategy adapts the reference scanning parameter of different precision according to the detection priority of the workpiece detection area, and adjusts the detection priority of the detection area according to the feature clarity of the co-built image feature recognition, which can reduce the influence of inaccurate workpiece detection area priority division when the image recognition process is not clear enough, thereby improving the reliability of the flatness detection result, so that the workpiece in different states can obtain a reasonable scanning detection scheme;

[0061] 3. The adjacent laser scanning units moving freely form double-unit synchronous detection, compared with single-unit detection, which can directly improve the number of detection workpieces per unit time, can adapt to match batch production, the view angle sharing strategy determines the minimum detection angle by analyzing the workpiece image, reduces the operation amount of subsequent image feature analysis, improves the detection speed, the gain model analyzes the gain of multi-workpiece synchronous scanning, and when the time consumption exceeds the best threshold, the angle and the number of workpieces are adjusted, which helps to balance the batch detection efficiency and detection accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0062] Figure 1 is a method flowchart of steps S100 to S400 in the present application.

[0063] Figure 2 is a method flowchart of steps S301 to S302 in the present application.

[0064] Figure 3 is a method flowchart of steps S303 to S304 in the present application.

[0065] Figure 4 is a method flowchart of steps S500 to S502 in the present application.

[0066] Figure 5 is a method flowchart of steps S401 to S404 in the present application.

[0067] Figure 6 is a method flowchart of steps S4031 to S4033 in the present application.

[0068] Figure 7 is a structural schematic diagram of a MIM electric clip blade flatness detection device in the present application. DETAILED DESCRIPTION

[0069] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application. Figures 1-7 In order to make the purpose, technical scheme and advantages of the present application more clear, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0070] The embodiments of the present application will be further described in detail below with reference to the accompanying drawings.

[0071] The embodiments of the present application disclose a MIM electric clip blade flatness detection method, the workpiece conveying step realizes the alternate transmission of the detected workpiece and the to-be-detected workpiece by means of a detection process identification strategy, avoids manual handling and time-consuming waiting, and significantly improves the workpiece circulation efficiency; the workpiece identification step accurately locates the blade structure features and distinguishes the detection priority by means of a key area division strategy, can focus on the high importance area compared with the undifferentiated point-by-point detection in the manual plug gauge detection, reduces the invalid detection action; the workpiece detection step matches the scanning parameters corresponding to the accuracy for different priority areas by means of a scanning parameter adaptation strategy, replaces the manual experience judgment, and fundamentally avoids the accuracy fluctuation caused by the operation difference of different personnel; the detection feedback step ensures that each collected image meets the reference feature requirement by analyzing the image feature quantity and adjusting the acquisition parameters, and solves the problems of missed detection, misjudgment and poor result consistency in manual detection.

[0072] Referring to Figure 1 , a method flow of a MIM electric clip blade flatness detection method includes the following steps:

[0073] Step S100: workpiece conveying step, configured with a detection process identification strategy, for analyzing the detection process of the workpiece, and issuing a transmission instruction to alternately convey the detected workpiece and the to-be-detected workpiece;

[0074] Workpiece detection process refers to the stage of the workpiece in the flatness detection process, which can reflect whether the workpiece needs to enter or leave the detection station. For example, the workpiece is in the undetected state or the detected state.

[0075] The transmission instruction is an instruction signal for controlling the action of the workpiece conveying unit, which can trigger the start, stop or adjustment of the moving direction of the conveying unit to ensure the orderly movement of the workpiece along the preset path.

[0076] The detected workpiece refers to the MIM electric push shear blade that has completed flatness detection and obtained detection data. The undetected workpiece refers to the MIM electric push shear blade that has not yet undergone flatness detection and needs to enter the detection station.

[0077] The implementation process of this step is as follows: after starting the workpiece conveying unit, the detection process recognition strategy is used to analyze the detection process of each workpiece in real time. For example, the detection station sensor is used to determine whether there is a detected workpiece in the detection station, and whether there is an undetected workpiece in the detection area. When it is detected that the workpiece in the detection station has completed the detection and becomes a detected workpiece, a transmission instruction is immediately sent to control the conveying unit to move the detected workpiece out of the detection station and transport it to the subsequent sorting area. At the same time, the conveying unit transports the undetected workpiece in the detection area to the vacant detection station, realizing the alternating transmission of the detected workpiece and the undetected workpiece, avoiding the situation of vacancy or congestion of the workpiece in the detection station, and ensuring the continuous and efficient operation of the entire detection process.

[0078] Step S200: Workpiece recognition step, configure key area division strategy, and determine the workpiece blade structure features according to the image feature recognition of the workpiece, and divide the workpiece detection area with different detection priorities;

[0079] Image feature recognition refers to the process of feature extraction and analysis of the MIM electric push shear blade image obtained by the image acquisition unit, which can identify the contour, line, texture and other features in the image, and provide a basis for accurately determining the blade structure. The workpiece blade structure feature refers to the key structure of the MIM electric push shear blade for realizing the shearing function, including the main blade edge, the auxiliary blade edge, the corner connecting the blade edge and the blade body, etc. The flatness of these structures directly determines the shearing sharpness and service life of the blade, and is the core object of detection.

[0080] The detection priority is a detection level divided according to the influence degree of each area of the workpiece on the product quality. The higher the influence degree, the higher the detection priority. For example, the blade area has a significant influence on the shearing performance, so its detection priority is higher than that of the blade body flat area.

[0081] The workpiece detection area refers to the different detection parts of the MIM electric push shear blade divided according to the structure features, and each area corresponds to a specific structure feature and detection requirement.

[0082] The implementation process is: first, the complete image of the MIM electric clipper blade is acquired by the image acquisition unit; then the image is subjected to image feature recognition by means of a key area division strategy, and the contour details and structural lines in the image are extracted, so as to accurately determine the main blade edge, secondary blade edge, corner and other blade structure characteristics of the workpiece; according to the influence degree of each blade structure characteristic on the shearing performance of the blade, different detection priorities are divided, for example, the main blade edge and the corner are divided into a high detection priority area, and the blade body plane is divided into a conventional detection priority area; finally, the detection order and accuracy requirements of each workpiece detection area are determined, which lays a foundation for subsequent adaptation of corresponding sampling scanning parameters.

[0083] Step S300: workpiece detection step, configured with a scanning parameter adaptation strategy, adapting sampling scanning parameters according to the detection priority of the workpiece detection area, and instructing the laser scanning unit to perform flatness detection according to the sampling scanning parameters to obtain flatness detection data;

[0084] The sampling scanning parameters are parameters directly related to the laser scanning detection accuracy, including scanning resolution, scanning speed, scanning range, etc. Different detection priority areas adapt different sampling scanning parameters. High priority areas usually adapt higher scanning resolution and slower scanning speed to ensure detection accuracy. The scanning resolution of the conventional priority area can be appropriately reduced and the scanning speed can be increased to balance the detection efficiency.

[0085] The laser scanning unit is the core component for realizing the flatness detection of the MIM electric clipper blade, which can emit a laser beam and receive the laser signal reflected from the workpiece surface. Through analysis and calculation of the laser signal, the three-dimensional position information of the workpiece surface is obtained.

[0086] The flatness detection data refers to the data reflecting the flatness condition obtained after the laser scanning unit scans the workpiece detection area, including the height deviation of each sampling point relative to the preset reference plane, etc., which is the core basis for judging whether the workpiece flatness is qualified.

[0087] In specific implementation, after the detection priorities of the detection regions of the workpiece are determined, the scanning parameter adaptation strategy is used to match the sampling scanning parameters corresponding to the detection regions with different detection priorities, for example, the main blade edge region with high priority is adapted to a scanning resolution of 0.001 mm and a scanning speed of 50 mm / s, and the blade body plane region with normal priority is adapted to a scanning resolution of 0.002 mm and a scanning speed of 100 mm / s. The matched sampling scanning parameters are sent to the laser scanning unit, the laser scanning unit starts scanning work according to the parameters, the emitted laser beam uniformly acts on the surfaces of the detection regions of the workpiece, and the reflected laser signals are received by the laser scanning unit and converted into electrical signals. After the electrical signals are processed by the data processing module, the height deviation of each sampling point relative to the reference plane and other information are obtained, and finally complete flatness detection data are formed to provide data support for subsequent flatness determination.

[0088] Step S400: detection feedback step, configured with an identification feedback strategy, used to analyze whether the feature quantity of the collected workpiece image meets the preset reference feature quantity, and to feedback adjust the collection parameters of the image collection unit.

[0089] The feature quantity of the collected workpiece image refers to the number of features reflecting the structure of the workpiece that can be effectively identified from the workpiece image obtained by the image collection unit, including the number of blade edge contour features and the number of corner features, which directly affects the accuracy of subsequent workpiece identification and detection analysis.

[0090] The preset reference feature quantity is a standard set in advance according to the structural complexity and detection accuracy requirement of the MIM electric push shear blade, and only when the feature quantity of the collected image meets or exceeds the standard, the reliability of subsequent workpiece identification, region division and detection analysis can be ensured.

[0091] The collection parameters of the image collection unit refer to the parameters affecting the image collection quality, including exposure time, focal length, aperture size, etc., the adjustment of these parameters can change the sharpness and contrast of the image, and then affect the identification effect of the image features.

[0092] The feedback adjustment refers to the process of adjusting the collection parameters of the image collection unit according to the comparison result of the feature quantity of the collected image and the reference feature quantity, the purpose is to make the subsequently collected image meet the feature quantity requirement and ensure the stability of the detection process.

[0093] The implementation process of this step is as follows: after the image collection unit completes single workpiece image collection, the collected image is analyzed by the identification feedback strategy, and the number of features that can be effectively identified is counted, for example, the number of main blade edge contours and the number of corners that can be clearly identified are counted.

[0094] The number of features obtained by statistics is compared with the preset reference feature number. If the number of features of the collected image reaches or exceeds the reference feature number, it indicates that the current collection parameter adapts to the image collection requirement, and the image collection unit continues to collect subsequent images according to the current parameter.

[0095] If the number of features of the collected image is less than the reference feature number, it indicates that the current collection parameter has deviation, for example, the exposure time is too short to cause image blur and the features are difficult to identify, or the focal length is not appropriate to cause part of the features not to be clearly imaged. At this time, an adjustment signal is immediately sent out to feedback adjust the collection parameter of the image collection unit, such as appropriately increasing the exposure time, adjusting the focal length to the appropriate position, or increasing the aperture to improve the image brightness. After the adjustment is completed, the image collection unit re-collects the workpiece image, and the number of features is analyzed again until the number of features of the collected image meets the preset reference feature number requirement, ensuring the accuracy of subsequent workpiece recognition and detection analysis.

[0096] Referring to Figure 2 , the scanning parameter adaptation strategy includes:

[0097] Step S301: According to the detection priority, the corresponding reference scanning parameter in the preset scanning parameter database is matched. The reference scanning parameter reflects the accuracy of scanning the workpiece;

[0098] The preset scanning parameter database is a pre-constructed structured data set, which internally stores the association relationship between different detection priorities and corresponding scanning parameters. These parameters are obtained based on the detection accuracy requirements of different regions of the MIM electric push scissors, the performance of the laser scanning unit, and the historical detection data calibration, to ensure the rationality and adaptability of the parameters. The reference scanning parameter is the standard scanning parameter bound with a specific detection priority in the preset scanning parameter database, including scanning resolution, scanning speed, laser power, etc. The parameter value is directly related to the scanning accuracy. The higher the detection priority, the higher the scanning resolution and the slower the scanning speed corresponding to the reference scanning parameter, to ensure the detection accuracy of the high-priority area; the lower the detection priority, the reference scanning parameter can appropriately reduce the resolution and increase the speed to balance the overall detection efficiency.

[0099] The implementation process of this step is: after completing the detection priority division of the workpiece detection area in step S200, the system extracts the priority information of each detection area, such as the high priority of the main blade edge area and the normal priority of the blade body plane area; then the system calls the preset scanning parameter database and matches according to the extracted priority information; the high priority area matches the high resolution reference scanning parameter, such as scanning resolution 0.001 mm / point, scanning speed 50 mm / s; the normal priority area matches the adaptive resolution reference scanning parameter, such as scanning resolution 0.002 mm / point, scanning speed 100 mm / s; after matching, the system temporarily stores the corresponding reference scanning parameters of each area to provide parameter basis for the subsequent scanning operation of the laser scanning unit, and ensures that the scanning accuracy of different priority areas meets the preset requirements.

[0100] Step S302: According to the image features in the workpiece image feature recognition process, the feature clarity is analyzed, and when the feature clarity is lower than the preset reliable clarity, the preset priority analysis model is calculated and the detection priority is updated.

[0101] The feature clarity refers to the clarity of each structural feature in the workpiece image, such as the blade edge contour and the corner line, which can be judged by the sharpness of the feature edge and the gray scale contrast difference. The sharper the edge, the more obvious the gray scale contrast, the higher the feature clarity, and the more conducive to subsequent detection analysis; otherwise, the feature clarity is lower, which may cause detection data deviation.

[0102] The preset reliable clarity is a critical value for judging whether the feature clarity meets the detection requirements, which is determined based on the recognition accuracy requirement of the key features of the MIM electric push cutter. Only when the feature clarity reaches or exceeds this value, can the reliability of the subsequent flatness detection data be ensured.

[0103] The implementation process of this step is: after the image acquisition unit acquires the workpiece image and completes the preliminary feature recognition, the system analyzes the feature clarity of the recognized image features, judges the clarity level by calculating the gray scale gradient value of the feature edge, for example, the higher the gray scale gradient value of the blade edge contour edge, the clearer the blade feature. Compare the analyzed feature clarity with the preset reliable clarity, if the feature clarity reaches or exceeds the reliable clarity, it means that the current image feature meets the detection requirements, and the original detection priority divided in step S200 remains unchanged; if the feature clarity is lower than the reliable clarity, it means that the current image feature may be blurred, such as the blade edge contour edge blur may affect the subsequent detection accuracy, at this time the system calls the preset priority analysis model, and updates the original detection priority through model calculation; the updated detection priority will be fed back to step S301 to re-match the corresponding reference scanning parameter, so as to improve the feature clarity of subsequent image acquisition by adjusting the scanning parameter, and ensure the detection accuracy.

[0104] The priority analysis model uses the following formula for calculation:

[0105] ;

[0106] in, The updated detection priority, The initial matching baseline detection priority is determined based on a preset scanning parameter database. The priority adjustment coefficient is dynamically adapted based on the workpiece material hardness. The preset reliable resolution threshold, The sharpness of the features to be actually identified is determined through pixel analysis of the image.

[0107] Reference Figure 3 The workpiece inspection process also includes a flatness error correction strategy, including:

[0108] Step S303: Perform surface roughness feature recognition based on the workpiece image to determine the regional roughness corresponding to the rough area of ​​the workpiece, and analyze the rough area of ​​the workpiece to determine the abnormal rough area.

[0109] Surface roughness feature recognition refers to the process of extracting and judging the features that reflect the surface roughness in a workpiece image. By analyzing information such as texture density and gray value fluctuation in local areas of the image, smooth and rough areas of the workpiece surface can be distinguished. For example, areas with dense texture and large gray value fluctuation usually correspond to rough areas; areas with sparse texture and stable gray value usually correspond to smooth areas.

[0110] Regional roughness refers to the roughness of a specific area on the surface of a workpiece. It is an indicator reflecting the microscopic unevenness of that area. The regional roughness of different areas may vary due to differences in MIM molding processes and subsequent machining precision. A rough area on the workpiece surface refers to an area with microscopic unevenness and a roughness higher than the normal level. An abnormally rough area refers to an area selected from the rough areas of the workpiece whose roughness far exceeds the allowable range of normal processes. The roughness of such areas may interfere with the accuracy of flatness test data and needs to be marked and analyzed separately.

[0111] The implementation process of this step is: first, call the high-definition image of the workpiece obtained in step S200, focus on each detection area in the image, such as the blade edge area and the blade body plane area; through the surface roughness feature recognition method, analyze the texture features and gray scale fluctuations of the image region by region, for example, in the blade edge area, if the texture lines of a certain section of the blade edge in the image are dense and the gray scale value frequently changes in a short distance, it can be preliminarily determined that this section is a rough area; for the rough area preliminarily determined, further calculate the area roughness thereof, and establish the correlation between each rough area and the corresponding roughness; then compare the area roughness of each rough area with the conventional roughness range allowed by the MIM electric clip blade forming process, and screen out the areas whose area roughness exceeds the conventional range, mark these areas as abnormal rough areas, and record their specific positions on the workpiece, such as the 10-15mm section on the left side of the blade edge and the 5-8mmx3-5mm area in the middle of the blade body, to provide the basis for the target area for subsequent flatness error correction.

[0112] Step S304: Based on the abnormal rough area, the area roughness is identified, and when the area roughness exceeds the preset allowable roughness, the actual flatness is determined by calculating through the preset flatness correction model combined with the flatness detection data.

[0113] The preset allowable roughness is a roughness upper limit value set in advance according to the use requirements of the MIM electric clip blade, the subsequent assembly precision and the industry quality standard. Only when the area roughness is lower than or equal to this value, can the surface state of the workpiece be ensured not to significantly interfere with the flatness detection result, and the shearing performance and service life requirements of the blade can also be met. The actual flatness refers to the real flatness state of the workpiece excluding the interference of the abnormal rough area of the workpiece surface. Since the original flatness detection data may misjudge the micro convex or concave of the abnormal rough area as flatness deviation, it is necessary to obtain the actual flatness which is more consistent with the actual flatness of the workpiece through correction. This data is the core basis for finally determining whether the flatness of the workpiece is qualified.

[0114] The implementation process of this step is: first, for the abnormally rough area marked in step S303, the area roughness is identified again, and more accurate area roughness data is obtained through more detailed image analysis; the obtained abnormal area roughness data is compared with the preset allowed roughness, if the roughness of the abnormal area is lower than or equal to the preset allowed roughness, it means that the roughness of the area is within the acceptable range of interference to the flatness detection data, at this time, the flatness detection data obtained in step S300 can be directly used as the actual flatness of the workpiece; if the roughness of the abnormal area exceeds the preset allowed roughness, it means that the roughness of the area has obviously interfered with the original flatness detection data, and the correction process needs to be started, that is, the preset flatness correction model is called, the roughness data of the abnormal area, the position information of the abnormal area and the flatness detection data obtained in step S300 are input into the model, the influence of roughness interference on flatness detection result is stripped through model operation, and finally the actual flatness reflecting the real flatness of the workpiece is calculated, ensuring the accuracy of the flatness determination result.

[0115] The flatness correction model calculates the actual flatness as follows:

[0116] ;

[0117] Wherein, is the corrected actual flatness, is the original flatness detection data, which is directly obtained by the laser scanning unit, k is the roughness correction coefficient, determined according to the hardness of the blade material, is the actual area roughness of the abnormal rough area, is the preset allowed roughness threshold, and is the area of the abnormal rough area, is the total area of the workpiece detection area.

[0118] Referring to Figure 4 , there is also a freely movable adjacent laser scanning unit for forming two scanning units for simultaneously detecting the flatness of the workpiece, and a batch detection and viewing angle optimization strategy is configured:

[0119] Step S500: according to the feature analysis of the workpiece image to determine the number of current detection workpieces, and adjusting through the preset viewing angle sharing strategy to determine the minimum detection angle of the adjacent laser scanning unit;

[0120] The freely movable adjacent laser scanning units refer to two laser scanning components with position and angle adjustment functions, which can move within a preset range according to detection requirements to realize simultaneous scanning of multiple workpieces, which can significantly improve batch detection efficiency compared to a single scanning unit. The workpiece image feature analysis refers to feature extraction of the image containing multiple workpieces obtained by the image acquisition unit, and the number of workpieces that can be included in the current detection range is counted by identifying the contour, size and other features of the workpiece in the image, such as identifying the number of blade contours of MIM electric hair scissors in the image to determine the total number of workpieces to be detected.

[0121] The current detection workpiece quantity refers to the number of workpieces that can be jointly covered and detected by the two laser scanning units in a single simultaneous detection, which needs to be determined in combination with the scanning range of the scanning unit, the workpiece size and the work position space to avoid excessive number leading to incomplete scanning coverage.

[0122] The view angle sharing strategy is a method for coordinating the view angles of two adjacent laser scanning units, which integrates the scanning fields of view of the two units to ensure that the scanning view angle of a single unit is minimized under the premise of covering all workpieces to be detected, thereby reducing redundant view fields.

[0123] The minimum detection angle refers to the minimum angle required by a single laser scanning unit to completely cover the assigned detection workpiece. An excessively large angle is easy to introduce background areas outside the workpiece, increasing the computational load of image feature analysis; an excessively small angle may result in incomplete scanning of some areas of the workpiece, affecting the detection integrity, so the angle needs to be accurately determined through the view angle sharing strategy.

[0124] The implementation process of this step is as follows: first, the overall image containing multiple workpieces to be detected is obtained by the image acquisition unit, the image is analyzed for features, the complete contour of each workpiece is identified, and the number of workpieces that can enter the simultaneous detection range is counted, for example, 3 complete MIM electric hair scissors contours are identified, and it is determined that the current detection workpiece quantity is 3; then the preset view angle sharing strategy is started, the best coverage range of the two laser scanning units is calculated according to the determined workpiece quantity and the position distribution of each workpiece, for example, 3 workpieces are distributed as left, middle and right, the left scanning unit is responsible for scanning the left and middle workpieces, and the right scanning unit is responsible for scanning the middle and right workpieces, and the view angle is shared by overlapping the middle workpiece.

[0125] The minimum detection angle of each scanning unit is further calculated based on the coverage range to ensure that the angle is just enough to completely cover the assigned workpiece and does not include excessive background areas, for example, for a blade with a length of 100 mm and a width of 20 mm, the minimum detection angle of the left scanning unit is calculated to be 45°, and the right scanning unit is 40°, completing the determination of the minimum detection angle.

[0126] Step S501: Adjust the laser scanning of the adjacent laser scanning units based on the minimum detection angle, and analyze through a preset adjustment gain model to obtain the multi-workpiece synchronous scanning gain.

[0127] The laser scanning adjustment refers to adjusting the physical positions and laser emission angles of the two adjacent laser scanning units according to the determined minimum detection angle, including moving the scanning units horizontally to align with the assigned workpieces, rotating the scanning units to match the minimum detection angle, ensuring that the laser beams can accurately cover all detection areas of the workpieces, and avoiding scanning blind areas.

[0128] The multi-workpiece synchronous scanning gain is an index for measuring the synchronous detection efficiency of the two scanning units, reflecting the efficiency improvement degree of synchronous detection compared with single scanning unit detecting the same number of workpieces. The higher the gain, the more obvious the efficiency advantage of synchronous detection.

[0129] The adjustment gain model is an algorithm model for calculating the gain, which analyzes the multi-workpiece synchronous scanning gain by inputting parameters such as synchronous detection time consumption, single unit detection time consumption, and detection workpiece quantity. This step only needs to complete the gain calculation through the model, without disclosing the model details.

[0130] The implementation process of this step is: according to the minimum detection angle determined in step S500, adjust the two adjacent laser scanning units, for example, move the left scanning unit to the left side of the workpiece by 150 mm and rotate it to the minimum detection angle of 45°; move the right scanning unit to the right side of the workpiece by 150 mm and rotate it to the minimum detection angle of 40°. During the adjustment process, the scanning range is confirmed in real time through the image acquisition unit to ensure that the laser beams of the two units cover all workpieces to be detected without obvious overlapping blind areas.

[0131] After the adjustment is completed, start the two scanning units to perform flatness detection on the workpieces at the same time, and record the total time consumption of synchronous detection. Input parameters such as the total time consumption of synchronous detection, the reference time consumption of a single laser scanning unit detecting a single workpiece, and the current number of detection workpieces into the preset adjustment gain model, and calculate the multi-workpiece synchronous scanning gain through model analysis, which provides a basis for subsequent judgment of whether the detection efficiency meets the standard.

[0132] Step S502: When the workpiece surface flatness detection time consumption corresponding to the multi-workpiece synchronous scanning gain is greater than the preset optimal time consumption gain, trigger the minimum detection angle adjustment instruction to reduce the total number of synchronous detection workpieces and the minimum detection angle.

[0133] The detection time consumption corresponding to the multi-workpiece synchronous scanning gain refers to the actual detection time consumption obtained by inversely calculating the synchronous scanning gain. This time consumption directly reflects the efficiency level of synchronous detection. The shorter the corresponding time consumption, the higher the efficiency; otherwise, the longer the time consumption, the lower the efficiency.

[0134] The preset optimal time consumption gain is an efficiency threshold set in advance according to the batch production demand of the MIM electric clippers, which represents the minimum efficiency requirement meeting the production rhythm. If the actual detection time consumption exceeds the time consumption corresponding to the threshold, it indicates that the current synchronous detection efficiency is not up to standard, and parameter adjustment is required.

[0135] The minimum detection angle adjustment instruction is a control signal for triggering the adjustment of the scanning unit angle and the number of detected workpieces. After the instruction is issued, the system automatically reduces the total number of workpieces for synchronous detection and narrows the minimum detection angle of the two scanning units to reduce the working load of the scanning units and improve the detection efficiency.

[0136] In the specific implementation process, the detection time consumption corresponding to the multi-workpiece synchronous scanning gain obtained in step S501 is compared with the standard time consumption corresponding to the preset optimal time consumption gain. If the actual detection time consumption is less than or equal to the standard time consumption, it indicates that the current synchronous detection efficiency meets the production demand, and the current detection workpiece quantity and minimum detection angle are maintained unchanged, and the synchronous detection of the subsequent batch of workpieces is continued. If the actual detection time consumption is greater than the standard time consumption, for example, the standard time consumption is 10 seconds / 3 workpieces, and the actual time consumption is 15 seconds / 3 workpieces, it indicates that the current efficiency is not up to standard, and the system immediately triggers the minimum detection angle adjustment instruction. According to the instruction, the total number of workpieces for synchronous detection is first reduced, for example, from 3 to 2, to reduce the coverage demand of the scanning units; then the minimum detection angle of the two scanning units is recalculated and narrowed, for example, the left scanning unit angle is narrowed from 45° to 35°, and the right scanning unit angle is narrowed from 40° to 30°, to reduce the scanning range of the background area; after the adjustment is completed, step S501 is executed again to detect and judge whether the time consumption meets the standard, until the detection time consumption meets the preset optimal time consumption gain requirement, to ensure that the batch detection efficiency is stably adapted to the production rhythm.

[0137] The preset adjustment gain model includes:

[0138] The multi-workpiece synchronous scanning gain is calculated according to the scanning gain time consumption and the number of multi-workpiece synchronous scanning.

[0139] The adjustment gain model is calculated by the following formula:

[0140] ;

[0141] Wherein, G is the time consumption gain of multi-workpiece synchronous scanning, N is the total number of workpieces for synchronous detection, T is the reference time consumption of detecting 1 workpiece by a single laser scanning unit, T is the actual total time consumption of detecting N workpieces by two laser scanning units synchronously, is the preset angle influence coefficient, is the additional time consumption caused by the minimum detection angle adjustment, The detection priority is the current minimum detection angle, The set optimal reference detection angle.

[0142] Referring to Figure 5 The feedback strategy includes:

[0143] Step S401: According to the acquisition parameters of the image acquisition unit, the corresponding reference feature quantity in the preset standard feature quantity library is matched;

[0144] The acquisition parameters of the image acquisition unit refer to the core parameters that affect the image acquisition quality, including exposure time, lens focal length, aperture size, etc. These parameters directly determine the brightness, depth of field and detail rendering capability of the image. Under different parameter combinations, the number of features that can be clearly identified in the workpiece image differs. The preset standard feature quantity library is a pre-constructed structured data set that internally stores the association between different acquisition parameter combinations and corresponding reference feature quantities. These reference values are obtained based on the standard structure of MIM electric scissors and a large number of acquisition experiments, for example, the parameter combination of "exposure time 50μs, focal length 100mm, aperture F8" corresponds to a reference feature quantity of 8 (including blade edge contour, corner, mounting hole edge, etc.). The reference feature quantity refers to the minimum feature quantity threshold that can ensure the accuracy of subsequent workpiece recognition and key area division under a specific acquisition parameter. Only when the actual recognized feature quantity reaches this value, can a reliable basis be provided for subsequent detection.

[0145] The implementation process of this step is as follows: first, extract the real-time acquisition parameters of the current image acquisition unit, for example, obtain the parameter combination of "exposure time 60μs, focal length 95mm, aperture F10"; then the system calls the preset standard feature quantity library and retrieves the parameter entry closest to the current combination through the parameter matching algorithm. If there is an entry corresponding to "exposure time 60μs, focal length 95mm, aperture F10" in the library, the reference feature quantity associated with this entry is directly called, such as 7; if there is no completely matching entry, the reference value is determined through interpolation calculation, for example, taking "exposure time 50μs, focal length 95mm, aperture F10" and "exposure time 70μs, focal length 95mm, aperture F10" as references, the interpolation calculation determines that the reference feature quantity corresponding to the current parameters is 7; finally, the determined reference feature quantity is temporarily stored as the standard for judging whether the image quality is qualified.

[0146] Step S402: When the number of features recognized by the workpiece image feature recognition is less than the reference feature quantity, perform environmental feature analysis to determine the reflective region on the workpiece surface;

[0147] The number of features recognized by the workpiece image feature recognition refers to the number of workpiece structure features that can be clearly distinguished extracted from the current collected image by an image recognition algorithm, for example, the number of main blade edge contours, secondary blade edge lines, and corner numbers recognized from the MIM electric clip blade image. If the number is lower than the reference value determined in step S401, it indicates that there is a problem with the image quality, and the cause needs to be further analyzed. The environmental feature analysis refers to the process of quantitatively analyzing the brightness distribution, gray scale gradient, and highlight area of the current image. By calculating the gray scale values of each pixel point in the image, the area with a gray scale value exceeding the preset highlight threshold is selected. These areas are usually over-bright areas caused by light reflection and are the main cause of feature blurring. The workpiece surface reflection area refers to the high-brightness area formed by the high-reflectivity of metal materials such as silver S440C stainless steel blades. The features in this area, such as the blade edge, are covered by light overflow, making it impossible to be recognized, or the gray scale value of the 3-5mm segment on the right side of the blade edge reaches 245, forming a reflection band and causing the secondary blade edge feature to be missing.

[0148] The implementation process of this step is as follows: first, the number of features that can be clearly recognized in the current workpiece image is counted by an image recognition algorithm, for example, 1 main blade edge, 2 corners, and a total of 3 features are recognized, while the reference feature number determined in step S401 is 7, indicating that the number of features is insufficient; then, the environmental feature analysis is started, and the pixel area with a gray scale value exceeding 230 is marked; through region connectivity analysis, the dispersed highlight pixel points are merged into continuous reflection areas, and the coordinate range of each reflection area in the workpiece image is recorded

[0149] Step S403: Based on the surface reflection area and the preset light compensation sub-strategy, the auxiliary light source adjustment parameters, including the light source irradiation angle and the light source brightness, are analyzed to determine the surface reflection area.

[0150] The surface reflection area is the highlight area marked in step S402 that causes feature blurring, and its position, area, and reflection intensity directly determine the direction and amplitude of light compensation.

[0151] The preset light compensation sub-strategy is a method for calculating auxiliary light source adjustment parameters. This strategy determines the light source irradiation angle based on the position of the reflection area and determines the light source brightness adjustment amplitude based on the gray scale value of the reflection area. By adjusting the angle, the light is uniformly irradiated on the workpiece surface, reducing the reflection intensity.

[0152] The auxiliary light source adjustment parameter refers to the key parameter for controlling the compensation light unit, and the light source irradiation angle refers to the included angle between the auxiliary light source and the workpiece surface. If the angle is too large, it is easy to directly reflect the workpiece and form a reflection, and if the angle is too small, the light is insufficient.

[0153] The light source brightness refers to the output light intensity of the auxiliary light source, which needs to be dynamically adjusted according to the reflection intensity. The stronger the reflection, the greater the brightness adjustment range.

[0154] Step S404: adjusting the preset compensation light unit based on the auxiliary light source adjustment parameter.

[0155] The surface reflection area is the highlight area marked in step S402 that causes feature blurring, and its position and area directly determine the direction and amplitude of light compensation. The preset light compensation sub-strategy is a method for calculating the auxiliary light source adjustment parameter, which combines the position of the reflection area. The light source irradiation angle is determined, and the light source brightness adjustment range is determined in combination with the gray value of the reflection area.

[0156] Referring to Figure 6 , the light compensation sub-strategy includes:

[0157] Step S4031: performing image reflection feature analysis on the reflection feature to determine the exposure parameter and the shadow parameter of the reflection feature, and performing image division on the reflection feature to obtain a reflection analysis image;

[0158] The reflection feature refers to the feature presented by the highlight area formed by light reflection in the workpiece image, including the shape, area, and gray value distribution of the reflection area. These features directly reflect the intensity and distribution of the reflection. Image reflection feature analysis refers to the process of quantitatively extracting and analyzing the above-mentioned features of the reflection area. By calculating the average gray value, gray standard deviation, and edge gradient of the reflection area, the severity of the reflection and the interference range of the image feature are determined.

[0159] The exposure parameter refers to the parameter that affects the exposure effect of the image. In this step, it specifically refers to the local exposure time and exposure intensity related to the reflection area, for example, the exposure time of the reflection area needs to be shorter than that of the normal area to avoid overexposure.

[0160] The shadow parameter refers to the feature parameter of the dark area formed around the reflection area due to light obstruction, including the gray value and area ratio of the shadow, which needs to be adjusted in coordination with the reflection parameter to balance the overall brightness of the image.

[0161] The reflection analysis image refers to a sub-image obtained by separately segmenting the area containing the reflection feature in the original workpiece image. The reflection area and the surrounding associated area, such as the reflection edge transition zone, are retained through cropping, masking, and other processing, facilitating subsequent targeted analysis and adjustment.

[0162] Step S4032: adjusting the exposure parameter and the shadow parameter of the reflection analysis image based on the preset image noise reduction parameter to obtain an optimized reflection image, and performing analysis to determine the effective image noise reduction parameter when the number of features of the workpiece image is within the preset reference number range.

[0163] The preset image denoising parameter refers to a parameter set for reducing image noise, including denoising intensity, such as three levels of low, medium and high, and filter radius, such as 1-5 pixels, and different parameter combinations correspond to different denoising effects. Too high intensity may blur the real features, and insufficient intensity cannot eliminate the noise interference in the reflection area. The optimized reflection image refers to an image obtained after adjusting the exposure parameter, shadow parameter and denoising processing of the reflection analysis image. The reflection area gray value is reduced to a reasonable range, the shadow area gray value is improved to a normal level, and the edge details of the workpiece features are retained.

[0164] The effective image denoising parameter refers to a denoising parameter combination that can make the number of workpiece features in the optimized reflection image reach a preset reference range, for example, when the parameter of "medium intensity denoising, filter radius 3 pixels" is used, the number of identifiable features in the image is increased from 3 to 7, and the set reference range is 6-8, so the parameter is effective.

[0165] Step S4033: matching the corresponding auxiliary light source adjustment parameter in the preset light source database based on the effective image denoising parameter.

[0166] The effective image denoising parameter is the denoising parameter combination determined in step S4032, which can make the image feature quantity meet the standard. This parameter indirectly reflects the processing needs of the reflection area, such as higher denoising intensity, which means that the noise caused by reflection is more serious.

[0167] The preset light source database is a database constructed in advance to store the corresponding relationship between the effective image denoising parameter and the auxiliary light source adjustment parameter, for example, medium intensity denoising, filter radius 3 pixels correspond to an irradiation angle of 32° and a brightness of 500 lux. These corresponding relationships are obtained based on a large number of lighting experiments and parameter calibration, ensuring that light source adjustment can reduce reflection from the physical layer and form a synergistic effect with image denoising.

[0168] The auxiliary light source adjustment parameter refers to the parameter used to control the compensation lighting unit. The parameter obtained through database matching can specifically solve the reflection problem and avoid the risk of feature distortion caused by image algorithm processing.

[0169] Reference Figure 7 Based on the same inventive concept, the embodiment of the present application provides a MIM electric clip blade flatness detection device, which comprises:

[0170] The workpiece conveying module is configured with a detection process recognition strategy, which is used to analyze the detection process of the workpiece and issue a transmission instruction to alternately convey the detected workpiece and the workpiece to be detected.

[0171] The workpiece recognition module is configured with a key region division strategy, and performs image feature recognition on the workpiece to determine the workpiece blade structure feature and divide the workpiece detection region into different detection priority workpieces.

[0172] The workpiece detection module is configured with a scanning parameter adaptation strategy, and the sampling scanning parameters are adapted according to the detection priority of the workpiece detection region, and the planeness detection data is obtained by the planeness detection of the laser scanning unit according to the sampling scanning parameters.

[0173] In addition, a detection feedback module connected by a wireless signal is also provided, the detection feedback module is built-in in the workpiece recognition module as a data analysis chip, and is configured with an identification feedback strategy, which is used to analyze whether the feature quantity of the workpiece image acquisition meets the preset reference feature quantity, and to feedback and adjust the acquisition parameters of the image acquisition unit.

[0174] The embodiment of the application provides a computer readable storage medium, which stores a computer program capable of being loaded by a processor and executing the MIM electric clip blade planeness detection method.

[0175] The computer storage medium includes, for example, a U disk, a mobile hard disk, a read-only memory (Read-Only Memory, ROM), a random access memory (Random Access Memory, RAM), a magnetic disk or an optical disk and various storage program codes.

[0176] Based on the same inventive concept, the embodiment of the application provides an intelligent terminal, which comprises a memory and a processor, and the memory stores a computer program capable of being loaded by the processor and executing the MIM electric clip blade planeness detection method.

[0177] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional modules is taken as an example for illustration, and in actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the above-described system, device and unit can refer to the corresponding process in the foregoing method embodiments, which will not be repeated here.

[0178] The above are preferred embodiments of the application, and are not intended to limit the protection scope of the application, any feature disclosed in the specification (including the abstract and the drawings) can be replaced by other equivalent or similar features unless specifically described. That is, unless specifically described, each feature is only an example of a series of equivalent or similar features.

Claims

1. A MIM electric clip blade flatness detection method, configured with a flatness detection device, the flatness detection device is provided with a laser scanning unit, an image acquisition unit and a workpiece conveying unit, characterized in that, The method comprises the following steps: a workpiece conveying step configured with a detection process identification strategy for analyzing the detection process of the workpiece and issuing a transmission instruction to alternate the detected workpiece and the workpiece to be detected; a workpiece identification step configured with a key area division strategy for identifying the image features of the workpiece to determine the structure features of the workpiece blade and divide the workpiece detection area into different detection priorities; a workpiece detection step configured with a scanning parameter adaptation strategy for adapting the sampling scanning parameters according to the detection priority of the workpiece detection area and instructing the laser scanning unit to perform flatness detection according to the sampling scanning parameters to obtain flatness detection data; and a flatness error correction strategy, comprising: identifying the surface roughness features from the workpiece image to determine the area roughness of the workpiece rough area and analyzing the workpiece rough area to determine the abnormal rough area; based on the abnormal rough area, identifying the area roughness, and when the area roughness exceeds the preset allowable roughness, calculating the actual flatness by combining the flatness detection data through the preset flatness correction model; a detection feedback step configured with an identification feedback strategy for analyzing whether the number of features collected from the workpiece image meets the preset reference feature quantity and adjusting the collection parameters of the image collection unit to keep the number of features of the workpiece image within the preset reference quantity range, the identification feedback strategy comprising: matching the corresponding reference feature quantity in the preset standard feature quantity library according to the collection parameters of the image collection unit; when the number of features identified from the workpiece image is less than the reference feature quantity, analyzing the environmental features to determine the surface reflective area of the workpiece; based on the surface reflective area and the preset light compensation sub-strategy, analyzing to determine the auxiliary light source adjustment parameters of the surface reflective area, including the light source irradiation angle and the light source brightness; adjusting the preset compensation light unit based on the auxiliary light source adjustment parameters; a freely movable adjacent laser scanning unit is also provided for forming two scanning units for simultaneously detecting the flatness of the workpiece, configured with a batch detection and viewing angle optimization strategy: analyzing the features from the workpiece image to determine the current number of workpieces being detected and adjusting through the preset viewing angle sharing strategy to determine the minimum detection angle of the adjacent laser scanning unit; based on the minimum detection angle, adjusting the laser scanning of the adjacent laser scanning unit and analyzing through the preset adjustment gain model to obtain the multi-workpiece synchronous scanning gain; when the workpiece surface flatness detection time corresponding to the multi-workpiece synchronous scanning gain is greater than the preset optimal time gain, triggering the minimum detection angle adjustment instruction to reduce the total number of synchronous detection workpieces and the minimum detection angle.

2. The MIM electric clip blade flatness detection method according to claim 1, wherein, The scanning parameter adaptation strategy comprises: matching the corresponding reference scanning parameters in the preset scanning parameter database according to the detection priority, the reference scanning parameters reflecting the accuracy of the workpiece scanning; analyzing the feature clarity during the image feature identification process of the workpiece image, and when the feature clarity is lower than the preset reliable clarity, calculating and updating the detection priority through the preset priority analysis model.

3. The MIM electric clip blade flatness detection method according to claim 2, wherein, The priority analysis model is calculated by the following formula: ； wherein, is an updated detection priority, is a reference detection priority of an initial match, determined according to a preset scanning parameter database, is a priority adjustment coefficient, dynamically adapted according to a hardness of a workpiece material, is a preset reliable definition threshold, is an actual recognized feature definition, determined through pixel analysis of an image.

4. The MIM electric clip blade flatness detection method of claim 1, wherein, The flatness correction model is calculated by the following formula: ； wherein, is the corrected actual flatness, is the original flatness detection data, directly obtained by the laser scanning unit, k is the roughness correction coefficient, is determined according to the hardness of the blade material, is the actual area roughness of the abnormal rough area, is a preset allowable roughness threshold, and is the area of the abnormal rough area, is the total area of the workpiece detection area.

5. The MIM electric clip blade flatness detection method of claim 1, wherein, The preset adjustment gain model includes: The multi-workpiece synchronous scanning gain is determined according to the scanning gain time consumption and the number of multi-workpiece synchronous scanning; The adjustment gain model is calculated by the following formula: ； Wherein, G is the time-consuming gain of multi-workpiece synchronous scanning, N is the total number of workpieces detected synchronously, T is the reference time-consuming of a single laser scanning unit detecting 1 workpiece, T is the actual total time-consuming of two laser scanning units synchronously detecting N workpieces, is the preset angle influence coefficient, is the additional time-consuming generated by the minimum detection angle adjustment, The detection priority is the current minimum detection angle, is the set optimal reference detection angle.

6. The MIM electric clip blade flatness detection method of claim 1, wherein, The light compensation sub-strategy includes: According to the reflection feature, the exposure parameter and the shadow parameter of the reflection feature are determined by analyzing the image reflection feature, and the reflection analysis image is obtained by dividing the reflection feature; Based on the preset image denoising parameter, the optimized reflection image is obtained by adjusting the exposure parameter and the shadow parameter of the reflection analysis image, and the effective image denoising parameter is determined when the number of workpiece image features is within the preset reference number range; The effective image denoising parameter is matched with the corresponding auxiliary light source adjustment parameter in the preset light source database.

7. A MIM electric clip blade flatness detection device loaded with a MIM electric clip blade flatness detection method according to any one of claims 1 to 6, characterized in that, Including: The workpiece conveying module is configured with a detection process identification strategy, which is used to analyze the detection process of the workpiece, and issues a transmission instruction to alternately convey the detected workpiece and the workpiece to be detected; The workpiece recognition module is configured with a key area division strategy, which is used to identify the image features of the workpiece to determine the workpiece blade structure features, and divide the workpiece detection area with different detection priorities; The workpiece detection module is configured with a scanning parameter adaptation strategy, which is used to adapt the sampling scanning parameter according to the detection priority of the workpiece detection area, and perform flatness detection according to the sampling scanning parameter to obtain flatness detection data; The detection feedback module is configured with an identification feedback strategy, which is used to analyze whether the number of image features collected by the workpiece image acquisition unit meets the preset reference feature number, and to feedback and adjust the acquisition parameter of the image acquisition unit.

Images