Strain gauge defect detection device and method

Abstract

The invention discloses a strain gauge defect detection device and method.The device comprises a multi-stage motion platform, industrial cameras, an image processing unit, a defect marking module and a main control unit, the multi-stage motion platform is used for moving a strain gauge array to be detected, and the industrial cameras comprise the first industrial camera and the second industrial camera; the first industrial camera is used for scanning the strain gauge array and acquiring images of strain gauges in the strain gauge array, and the second industrial camera is used for scanning sensitive grids of the strain gauges and acquiring images of the sensitive grids of the strain gauges; the image processing unit is used for carrying out image recognition processing on the images acquired by the industrial camera and determining the strain gauges with macroscopic defects or microscopic defects; and the defect marking module is used for physically marking the strain gauges with macroscopic defects or microdefects in the strain gauge array. According to the invention, two-stage detection of macroscopic defects and microscopic defects and physical marking can be respectively carried out on the strain gauge, and the detection efficiency and the detection precision of the defects of the strain gauge can be effectively improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of strain gauges, and in particular relates to a strain gauge defect detection device and method. Background Art

[0002] Strain gauges are devices that can convert their deformation due to force into changes in resistance. They are widely used in aviation, aerospace, and marine equipment testing, as well as material testing and high-speed rail collision testing. The sensitive grid is the core component of the strain gauge. In order to ensure the detection accuracy and reliability of the strain gauge, the sensitive grid of the strain gauge needs to be strictly tested for morphology and electrical characteristics during its production process. Among them, the morphology defect detection of the strain gauge sensitive grid is the primary link in quality inspection and plays a decisive role in the quality of the strain gauge.

[0003] Existing strain gauge array defect detection generally uses manual visual inspection and simple contact measurement tools for detection. These methods have the disadvantages of low detection efficiency, strong subjectivity, and limited ability to detect tiny defects. Summary of the invention

[0004] The purpose of the present invention is to provide a strain gauge defect detection device and method to address the defects existing in the prior art, which can perform two-level detection of macro defects and micro defects on the strain gauge and perform physical marking, thereby effectively improving the detection efficiency and detection accuracy of strain gauge defects.

[0005] In order to achieve the above-mentioned object, one aspect of the present invention provides a strain gauge defect detection device, comprising a multi-stage motion platform, an industrial camera, an image processing unit, a defect marking module and a main control unit, wherein the main control unit is respectively connected to the multi-stage motion platform, the industrial camera, the image processing unit and the defect marking module; The multi-stage motion platform is used to move the strain gauge array to be tested under the control of the main control unit, and includes a first motion platform and a second motion platform, wherein the first motion platform is arranged below the second motion platform, and a loading platform is arranged on the second motion platform for carrying the strain gauge array to be tested, and a positioning mark is arranged on the loading platform for determining the position information of each strain gauge in the strain gauge array on the loading platform; The industrial camera is arranged above the multi-stage motion platform, and is used to collect images of the strain gauge array under the control of the main control unit and transmit the collected images to the image processing unit. The industrial camera includes a first industrial camera and a second industrial camera. The first industrial camera is a low-magnification large-field-of-view camera, and the second industrial camera is a high-magnification small-field-of-view camera. The first industrial camera is used to scan the strain gauge array and collect images of the strain gauges in the strain gauge array. The second industrial camera is used to scan the sensitive grid of the strain gauge and collect images of the sensitive grid of the strain gauge. The image processing unit is used to perform image recognition processing on the images collected by the industrial camera under the control of the main control unit, and determine the strain gauges with macroscopic defects or microscopic defects. The defect marking module is used to physically mark the strain gauges with macroscopic defects or microscopic defects in the strain gauge array under the control of the main control unit.

[0006] Preferably, the device further includes a drive control unit, which is respectively connected to the first motion platform, the second motion platform and the main control unit, and is used to drive the first motion platform and the second motion platform according to the control instructions of the main control unit.

[0007] Preferably, the device further includes a first camera bracket and a second camera bracket, which are respectively used to support the first industrial camera and the second industrial camera. The first camera bracket is arranged on the first side of the multi-stage motion platform, the second camera bracket is arranged on the second side adjacent to the first side, and the defect marking module is arranged on the third side of the multi-stage motion platform.

[0008] Preferably, the distance between the first industrial camera and the multi-stage motion platform is greater than the distance between the second industrial camera and the multi-stage motion platform. The magnification of the first industrial camera is 20 - 60 times, and the field of view size is greater than 1 cm × 1 cm. The magnification of the second industrial camera is 200 - 500 times, and the field of view size is less than 2.5 mm × 2.5 mm.

[0009] Preferably, a cavity is arranged between the second motion platform and the first motion platform for placing a vacuum adsorption device, and through holes are arranged on the carrier platform for cooperating with the vacuum adsorption device to adsorb the strain gauge array on the upper surface of the carrier platform.

[0010] Another aspect of the present invention provides a method for detecting strain gauge defects, which uses the above device to detect strain gauge defects. The method includes: Step S11, controlling the first motion platform to move into the field of view of the first industrial camera, and collecting images of each strain gauge in the strain gauge array; Step S12, using the image processing unit to perform image recognition processing on the images of the strain gauges, and determining the strain gauges with macroscopic defects according to the image recognition results of the strain gauges; Step S13, controlling the multi-stage motion platform to move into the field of view of the second industrial camera, and collecting images of the sensitive grids of the strain gauges without macroscopic defects in the strain gauge array; Step S14, using the image processing unit to perform image recognition processing on the images of the sensitive grids, and determining the strain gauges with microscopic defects according to the image recognition results of the sensitive grids; Step S15: Control the defect marking module to physically mark the strain gauges with macroscopic or microscopic defects.

[0011] Preferably, before step S11, the method further includes: determining the position information of each strain gauge in the strain gauge array to be measured on the loading platform according to the positioning marks on the loading platform, and configuring strain gauge identifiers for each strain gauge according to the position information of each strain gauge.

[0012] Preferably, in step S12, first identify each strain gauge in the strain gauge image, record the position information and strain gauge identifier of each strain gauge; then perform macroscopic defect identification on the strain gauge image, identify the strain gauges with macroscopic defects, and output the corresponding strain gauge identifiers and corresponding position information; When performing macroscopic defect identification on the strain gauge image, extract the gray distribution feature information of the image of a single preprocessed strain gauge, input the extracted feature information into a trained CNN model, and the output result is the confidence level of a certain macroscopic defect of the strain gauge.

[0013] Preferably, in step S14, perform microscopic defect identification on the collected image of the sensitive grid, identify the strain gauges with microscopic defects, and output the strain gauge identifiers and position information of the strain gauges with microscopic defects; When performing microscopic defect identification on the image of the sensitive grid, perform feature extraction on the preprocessed image of the sensitive grid, input the extracted feature information into a trained CNN model, and the output result is the confidence level of a certain microscopic defect.

[0014] Preferably, the types of macroscopic defects of the strain gauge include at least one of large-area adhesion defects and large-area corrosion defects; The types of microscopic defects of the strain gauge include any one or any combination of the following: sensitive grid adhesion, sensitive grid breakage, sensitive grid morphology distortion, and abnormal sensitive grid spacing.

[0015] According to the strain gauge defect detection device and method in the above aspects of the present invention, by controlling the cooperation of the industrial camera and the multi-stage moving platform, two-level detection of macroscopic defects and microscopic defects of the strain gauge is carried out respectively and physical marking is performed, which can effectively improve the detection efficiency and detection accuracy of strain gauge defects. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained according to these drawings: Figure 1It is a schematic diagram of the principle of a strain gauge defect detection device according to an embodiment of the present invention; Figure 2 It is a schematic structural diagram of a strain gauge defect detection device according to an embodiment of the present invention; Figure 3 It is a schematic diagram of a reference mark on a loading platform according to an embodiment of the present invention; Figure 4 It is a schematic diagram of configuring a strain gauge identifier for a strain gauge according to an embodiment of the present invention; Figure 5 It is a flowchart of a strain gauge defect detection method according to an embodiment of the present invention; Figure 6 It is a schematic diagram of a large - area corrosion defect according to an embodiment of the present invention; Figure 7 It is a schematic diagram of a large - area adhesion defect according to an embodiment of the present invention; Figure 8 It is a schematic diagram of a sensitive grid pitch anomaly defect according to an embodiment of the present invention; Figure 9 It is a schematic diagram of a sensitive grid break defect according to an embodiment of the present invention; Figure 10 It is a schematic diagram of a sensitive grid morphology distortion defect according to an embodiment of the present invention; Figure 11 It is a schematic diagram of a sensitive grid adhesion defect according to an embodiment of the present invention. Detailed implementation manners

[0017] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0018] An embodiment of the present invention provides a strain gauge defect detection device. Referring to Figure 1 as shown, the device includes a main control unit, and a multi - stage motion platform, an industrial camera, an image processing unit, and a defect marking module that are respectively connected to the main control unit. The main control unit is used to send control instructions to each module and control the working states of each module.

[0019] The multi - stage motion platform is used to carry the strain gauge array to be measured. The multi - stage motion platform includes a first motion platform and a second motion platform, and the first motion platform is arranged below the second motion platform. The first motion platform and the second motion platform can move along the X and Y directions of the horizontal plane under the control of the main control unit, so as to move the carried strain gauge array.

[0020] The industrial camera is arranged above the multi-stage motion platform, and the industrial camera is electrically connected to the main control unit. The industrial camera includes a first industrial camera and a second industrial camera, and the first industrial camera and the second industrial camera are fixedly arranged in parallel.

[0021] The first industrial camera is a low-magnification large-field-of-view camera, and the second industrial camera is a high-magnification small-field-of-view camera; the first industrial camera is used to scan the strain gauge array, collect images of the strain gauges in the strain gauge array, and identify macroscopic defects existing in the strain gauges; the second industrial camera is used to scan the sensitive grid of the strain gauge, collect images of the sensitive grid of the strain gauge, and identify microscopic defects existing in the strain gauge.

[0022] The image processing unit is electrically connected to the industrial camera, and is used to perform defect identification on the images collected by the industrial camera to determine the strain gauges with macroscopic defects or microscopic defects; the defect marking module is used to physically mark the strain gauges with macroscopic defects or microscopic defects in the strain gauge array.

[0023] The main control unit is respectively connected to the multi-stage motion platform, the industrial camera, the image processing unit, and the defect marking module, and is used to control the working states of each module. Specifically, the control unit is used to control the movement of the motion platform according to the position information of the motion platform and the strain gauge to realize the scanning of the strain gauge array and the sensitive grid of the strain gauge; control the industrial camera to collect images of the strain gauge array and transfer the image information to the image processing unit; control the image processing unit to perform defect identification and transfer the defect identification result to the main control unit; control the defect marking module to physically mark the strain gauges with macroscopic or microscopic defects according to the identification result and the position information of the defective strain gauges.

[0024] Exemplarily, the device further includes a drive control unit, which is respectively connected to the first motion platform, the second motion platform, and the main control unit, and is used to drive the first motion platform and the second motion platform according to the control instructions of the main control unit.

[0025] In a specific embodiment, referring to Figure 2 As shown, a working platform 1 can be provided, and the defect detection device is arranged on the working platform 1. A slide rail 7 is arranged on the working platform 1, and a base 6 is arranged at the bottom of the first motion platform 8. The base 6 is provided with a groove matching the slide rail 7. The slide rail 7 is arranged along the X direction, so that the first motion platform 8 can move along the X direction.

[0026] A track 9 is arranged on the upper part of the first motion platform 8, and a sliding groove matching the track is arranged at the bottom of the second motion platform 10. The track 9 is arranged along the Y direction, so that the second motion platform 10 can move along the Y direction.

[0027] A load platform 11 for carrying the strain gauge array to be measured is provided on the second moving platform 10. Among them, the load platform 11 is provided with a smooth upper surface for adsorbing the strain gauge array on the upper surface of the load platform 11.

[0028] The upper surface of the load platform 11 is provided with positioning marks for determining the position information corresponding to each strain gauge of the strain gauge array. The positioning marks are, for example, Figure 3 the reference marks 13 as shown. A plurality of reference marks 13 can be evenly distributed on the load platform 11, and the coordinates of each reference mark 13 are pre-configured. After the strain gauge array is placed on the load platform 11, the position information corresponding to each single strain gauge can be determined according to the position information of each reference mark 13.

[0029] Alternatively, a strain gauge position mark 17 can also be provided as a positioning mark, and the corresponding position information is pre-configured. When the strain gauge array is placed on the load platform 11, each strain gauge in the strain gauge array can be aligned with one or more position marks 17. Then, referring to Figure 4 as shown, each strain gauge can be numbered and the corresponding strain gauge identifier can be configured.

[0030] In addition, through holes 12 are provided on the load platform 11 for cooperating with the vacuum adsorption device 16 to adsorb the strain gauge array on the upper surface of the load platform 11. The through holes 12 can be randomly distributed or evenly distributed on the load platform 11. Each through hole is connected to the vacuum adsorption device 16 through a pipeline. For example, a cavity is formed between the second moving platform 10 and the first moving platform 8, and the vacuum adsorption device 16 can be arranged in the cavity.

[0031] The first industrial camera 3 is arranged above the multi-stage moving platform for collecting images of the strain gauges in the strain gauge array; the second industrial camera 4 is arranged above the multi-stage moving platform for collecting images of the sensitive grids of the strain gauges.

[0032] Exemplarily, in the Z direction, the distance between the first industrial camera 3 and the multi-stage moving platform is greater than the distance between the second industrial camera 4 and the multi-stage moving platform. This is because the first industrial camera 3 is a low-magnification camera and the second industrial camera 4 is a high-magnification camera, and the focal length of the low-magnification camera is greater than the focal length of the high-magnification camera.

[0033] For example, referring to Figure 2As shown, the first industrial camera 3 and the second industrial camera 4 can be arranged above the multi-stage motion platform through the first camera support 14 and the second camera support 15 respectively. For example, the camera support can be an inverted L-shaped structure or a portal-shaped structure. The base of the camera support can be fixed on the working platform 1, and the industrial camera can be set at the front end of the camera support, so that the camera is set above the multi-stage motion platform.

[0034] Among them, the first camera support 14 corresponding to the first industrial camera 3 is arranged on the first side of the multi-stage motion platform, and the second camera support 15 corresponding to the second industrial camera 4 is arranged on the adjacent second side. And the defect marking module 5 can be arranged on the third side of the multi-stage motion platform. For example, the defect marking module 5 can be a laser beam type component or a robotic arm component.

[0035] Exemplarily, the magnification of the first industrial camera 3 is 20 - 60 times, and the field of view size is greater than 1 cm × 1 cm; the magnification of the second industrial camera 4 is 200 - 500 times, and the field of view size is less than 2.5 mm × 2.5 mm.

[0036] Correspondingly, the height of the first camera support 14 can be greater than the height of the second camera support 15, so that the distance between the first industrial camera 3 and the multi-stage motion platform is greater than the distance between the second industrial camera 4 and the multi-stage motion platform. Or, the first camera support 14 and the second camera support 15 can also use the same height, and a threaded rod can be built into the camera support to adjust the focal length of the industrial camera.

[0037] Exemplarily, an intelligent terminal device 2 can be provided, and the main control unit, the image processing unit, and the drive control unit can be set in the intelligent terminal device 2 to implement corresponding functions. For example, the user can send control instructions to the industrial camera and the multi-stage motion platform on the terminal device 2, control the industrial camera to collect images and upload them to the terminal device 2, then perform recognition processing on the collected images, and control the robotic arm component to mark the single strain gauge with defects.

[0038] Exemplarily, after placing the strain gauge array on the loading platform 11, a defect detection task for the current strain gauge array can be created on the intelligent terminal device 2. First, the strain gauge array can be positioned according to the reference marks 13 on the loading platform 11, and based on the position information of the reference marks 13, the position information corresponding to each strain gauge can be calculated. A strain gauge identifier can be configured for each strain gauge, and the position information and the strain gauge identifier can be recorded. Then, the position of the first moving platform 8 is driven into the field of view of the first industrial camera 3, and the first industrial camera 3 is used to collect an image of the strain gauge. This image can be a low-precision image for macroscopic defect identification of the strain gauge. Then, the first moving platform 8 and the second moving platform 10 can be driven to move in coordination and displaced into the field of view of the second industrial camera 4, and the second industrial camera 4 is used to scan the sensitive grid of the strain gauge and collect the corresponding image; this image can be a high-resolution and high-precision image and can be used for microscopic defect identification of the strain gauge. For example, first, a low-precision image of the first strain gauge can be collected; then, a high-precision image can be collected; according to the strain gauge identifier in sequence, two image collections are performed on each strain gauge in turn.

[0039] After completing the two image collections of a strain gauge, the corresponding image recognition task is triggered to perform macroscopic defect identification and microscopic defect identification on the strain gauge. If there are macroscopic defects or microscopic defects in the current strain gauge, the strain gauge identifier and the position information corresponding to the strain gauge are recorded. After completing the defect detection of the current strain gauge, image collection and defect identification can be performed on each strain gauge in sequence according to the strain gauge identifier.

[0040] After completing the defect identification of all strain gauges, a marking task can be triggered on the intelligent terminal device, the multi-stage moving platform is moved into the working range of the robotic arm assembly, and according to the position information of the defective strain gauges, the robotic arm assembly is controlled to add physical marks, such as scratches, to the defective individual strain gauges.

[0041] Alternatively, it can also be that after completing all the low-precision image collections, the image recognition task is executed to recognize the images collected by the first industrial camera 3, and the strain gauge identifiers and the corresponding position information of the individual strain gauges with macroscopic defects are output; at the same time, the images collected by the second industrial camera 4 can be recognized, and the strain gauge identifiers and the corresponding position information of the individual strain gauges with microscopic defects are output. Alternatively, it can also be that after completing the collection of an image, the corresponding image recognition task is triggered, so that image collection and image recognition can be synchronized.

[0042] Alternatively, in some embodiments, it may also be to first drive the first moving platform 8 to move into the field of view of the first industrial camera 3, collect the corresponding image of a strain gauge, and perform macro defect identification, determine whether there are macro defects, and output the defect identification result. Then, collect the corresponding low-precision image of the next strain gauge and perform macro defect identification until all strain gauges complete macro defect identification. At this time, the strain gauge identification and position information of the strain gauges with macro defects can be output; and the strain gauge identification and position information of the strain gauges without macro defects.

[0043] After that, the first moving platform 8 and the second moving platform 10 can be driven to move collaboratively and displace into the field of view of the second industrial camera 4, and the second industrial camera 4 is used to scan the sensitive grid of a strain gauge without macro defects to collect the corresponding image; this image can be a high-resolution and high-precision image. And perform micro defect identification on the current image and output the identification result. Then, collect the corresponding high-precision image of the sensitive grid of the next strain gauge and perform micro defect identification until all strain gauges without macro defects complete micro defect identification. According to the micro defect identification results of each strain gauge, output the strain gauge identification and position information of the strain gauges with micro defects. Then, when a marking task is triggered on the intelligent terminal device, move the multi-stage moving platform into the working range of the robotic arm assembly, and control the robotic arm assembly to add physical marks, such as scratches, to the defective strain gauges according to the position information of the strain gauges with macro or micro defects.

[0044] Alternatively, it may also be to trigger the execution of the image recognition task after all image collections are completed, recognize the images collected by the first industrial camera 3, and output the strain gauge identification and the corresponding position information of the strain gauges with macro defects; at the same time, the images collected by the second industrial camera 4 can be recognized to output the strain gauge identification and the corresponding position information of the strain gauges with micro defects, and then physical marking is performed.

[0045] An embodiment of the present invention also provides a method for detecting strain gauge defects, which uses the strain gauge defect detection device of the above embodiment to detect strain gauge defects. Refer to Figure 5 As shown, the method includes: Step S11, control the multi-stage moving platform to move into the field of view of the first industrial camera 3 and collect the image (low-precision image) of the strain gauges in the strain gauge array; Step S12, use the image processing unit to perform image recognition processing on the image of the strain gauge to obtain the macro defect recognition result, and determine the strain gauges with macro defects according to the recognition result of the strain gauge; Step S13: Control the multi-level moving platform to move into the field of view of the second industrial camera 4, and collect images (high-precision images) of the sensitive grids of the strain gauges in the strain gauge array that have no macroscopic defects. Step S14: Use the image processing unit to perform image recognition processing on the images of the sensitive grids of the strain gauges that have no macroscopic defects, and determine the strain gauges with microscopic defects based on the image recognition results of the sensitive grids. Step S15: Control the defect marking module to physically mark the individual strain gauges with defects.

[0046] Exemplarily, before step S11, the method further includes: determining the position information of the strain gauge array on the load platform 11 according to the reference mark 13 on the load platform 11 of the multi-level moving platform; based on the position information of the strain gauge array to be measured on the load platform 11, determining the corresponding position information of each strain gauge, and configuring the corresponding strain gauge identifier for each strain gauge.

[0047] For example, the strain gauge array can be first placed on the load platform 11, and the initial position of the strain gauge array is determined by using the strain gauge reference mark 13 on the load platform 11. Then, corresponding identifiers are configured for each strain gauge, and based on the initial position of the strain gauge array and the strain gauge reference mark 13, the position coordinates of each strain gauge are calculated. The user can turn on the vacuum adsorption device through the intelligent terminal device, so that the strain gauge array is stabilized on the load platform 11.

[0048] Exemplarily, in step S11, the user can initiate a control instruction for the first moving platform 8 in the lower layer of the multi-level moving platform on the intelligent terminal device, control the multi-level moving platform to move into the field of view of the first industrial camera 3, and then send a control instruction to the first industrial camera 3 to control the first industrial camera 3 to collect images of the strain gauges.

[0049] Exemplarily, in step S12, after the first image acquisition task is completed by using the first industrial camera 3, the first-stage image recognition processing can be performed on the currently acquired strain gauge images. First, identify each strain gauge in the strain gauge array in the strain gauge image, record the position information of each strain gauge, and configure the strain gauge identifier for each strain gauge. Among them, the strain gauge identifier can be a virtual identifier. For example, the intelligent terminal device can construct a matrix corresponding to the strain gauge array, and mark the virtual identifier and position of the strain gauge monomer in the matrix.

[0050] Secondly, the strain gauge image can be input into the macroscopic defect recognition model to identify the strain gauges with macroscopic defects, and output the corresponding strain gauge identifiers and corresponding position information.

[0051] Among them, a coordinate system can be established based on the reference mark 13 of the load platform 11, so as to determine the coordinates of each strain gauge in this coordinate system.

[0052] Exemplarily, in step S13, after the image recognition task of the first stage is completed, a control instruction can be issued to the multi-stage motion platform to control the second motion platform 10 to move into the field of view of the first industrial camera 3; or, synchronously control the first motion platform 8 and the second motion platform 10 to move collaboratively to move the load platform 11 into the field of view of the second industrial camera 4, and use the second industrial camera 4 to collect the image of the strain gauge grid.

[0053] In some exemplary embodiments, it may be after the recognition task of the image of the strain gauge is completed and the image recognition result is output, and then control the multi-stage motion platform to move into the field of view of the second industrial camera 4.

[0054] Or, it may also be after the image acquisition of the strain gauge is completed, first control the first motion platform 8 to move to perform preprocessing on the displacement of the multi-stage motion platform. The displacement processing of the multi-stage motion platform here means that after the strain gauge is scanned, first drive the first motion platform to move into the field of view of the first industrial camera; then drive the second motion platform to perform high-precision movement within the field of view of the second industrial camera to realize the scanning of the strain gauge grid; at the same time, after obtaining the recognition result of the strain gauge image, control the second motion platform 10 to perform a small range of displacement to realize the precise displacement control of the load platform 11.

[0055] Exemplarily, in step S14, after the intelligent terminal device obtains the current image of the strain gauge grid, it can trigger the second image recognition task, and based on the recognition result of the single strain gauge with macroscopic defects, only perform the recognition and judgment of microscopic defects on the strain gauges without macroscopic defects. For example, the collected image of the strain gauge grid can be input into the microscopic defect recognition model to realize the scanning of the strain gauge grid and output the defect recognition result of the grid, as well as the strain gauge identification and position information of the strain gauge with microscopic defects, so as to realize the defect recognition of the second stage of the strain gauge.

[0056] Exemplarily, in step S15, after the microscopic defect recognition of the strain gauge is completed, the intelligent terminal device can control the first motion platform 8 or the second motion platform 10 of the multi-stage motion platform to move into the operation range of the defect marking module. Then, according to the position information of each defective strain gauge, control the defect marking module to sequentially add physical marks to each defective strain gauge. For example, add scratches on the defective strain gauge.

[0057] For example, referring to Figure 6 and Figure 7 shown, the macroscopic defects of the strain gauge include at least one of large-area adhesion defects and large-area corrosion defects.

[0058] Reference Figures 8 - 11 As shown, the microscopic defects of the strain gauge include any one or any combination of the following: sensitive grid adhesion, sensitive grid open circuit, sensitive grid morphology distortion, and abnormal sensitive grid spacing.

[0059] Exemplarily, the first industrial camera 3 can collect the image information of the strain gauge; the second industrial camera 4 scans the sensitive grid of the strain gauge to obtain the sensitive grid image information.

[0060] When identifying defects in the image, the image can be preprocessed first, including: first converting the image into a grayscale image, using the Gaussian filtering algorithm to eliminate the random noise in image acquisition, then enhancing the visual recognition of the micron-level sensitive grid through the histogram equalization method, and then applying the Canny edge detection algorithm to strengthen the grid line contour features.

[0061] When identifying macroscopic defects in the image, the gray distribution feature information can be extracted from the image information of a single preprocessed strain gauge. The extracted feature information is input into a trained CNN model, and the output result is the confidence level that the strain gauge is of a certain defect type. Among them, the CNN model includes a global pooling layer, a 128-dimensional Dense layer - the final classification layer, which are set in sequence.

[0062] When identifying microscopic defects in the image, feature extraction can be performed on the preprocessed sensitive grid image information, including: gray distribution features, sensitive grid contour features, adjacent grid line edge contour features, and sensitive grid edge continuity features. Each feature information is used as an input parameter of the model and input into a trained CNN model, and the output result is the confidence level that the single strain gauge is of a certain defect type. Among them, the CNN model includes a global pooling layer, a 128-dimensional Dense layer - the final classification layer, which are set in sequence.

[0063] A confidence threshold can be preset, and classification can be determined according to the model output result. For example, the confidence threshold can be configured as 80%; if it is greater than 80%, it is considered to be of this defect type. If all four defect types are lower than 80%, it is considered that the single strain gauge has no defects. Among them, the open circuit confidence can be configured as 0.95, the adhesion confidence as 0.8, the morphology distortion confidence as 0.7, and the spacing anomaly confidence as 0.6.

[0064] In summary, the strain gauge defect detection device and method of the embodiments of the present invention have the following beneficial effects: (1) Two-stage detection of the strain gauge array: First, a first industrial camera with a low magnification quickly screens for macroscopic defects and marks the positions with macroscopic defects; the strain gauges without macroscopic defects enter the microscopic detection, reducing the ineffective image processing. Then, a second industrial camera with a high magnification scans the sensitive grids of individual strain gauges without macroscopic defects one by one to identify the microscopic defects of the strain gauge sensitive grids.

[0065] (2) Detection of the multi-stage motion platform and the dual industrial cameras: The large-range motion of the first motion platform (covering the strain gauge array in the X / Y directions), and the precise small-range motion of the second motion platform (error < 10 μm) achieve fully automatic high-precision detection. Cooperating with two industrial cameras with different magnifications can perform staged detection on the macroscopic and microscopic defects of the strain gauge array.

[0066] According to the strain gauge defect detection device and method of the embodiments of the present invention, by controlling the cooperation of the industrial camera and the multi-stage motion platform, two-level judgments of macroscopic defects and microscopic defects are respectively performed on the strain gauges to determine the defect types, indicate the positions of the defective strain gauges, and perform physical markings, which can solve the problems of low efficiency and insufficient defect recognition ability in the prior art of strain gauge morphology detection technology.

[0067] Specifically, the strain gauge defect detection device and method of the embodiments of the present invention can accurately identify the strain gauges with macroscopic defects in the strain gauge array and record them by using the first industrial camera to collect the images of the strain gauges in the strain gauge array and performing recognition processing on the images of the strain gauges. Moreover, in the subsequent recognition processing, only the strain gauges without macroscopic defects are subjected to microscopic defect recognition processing, thereby reducing the calculation amount; the images of the sensitive grids of the strain gauges without macroscopic defects are collected by using the second industrial camera to obtain the microscopic defect recognition results of the strain gauges; the two-level detection of the strain gauges is realized by using two different images to perform macroscopic and microscopic defect detections on the strain gauges in sequence; and thus the detection efficiency of the strain gauge defects is effectively improved.

[0068] Only some exemplary embodiments of the present invention have been described by way of illustration above. Without doubt, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the protection scope of the claims of the present invention.

Claims

1. A strain gauge defect detection device, characterized in that, It includes a multi-level motion platform, an industrial camera, an image processing unit, a defect marking module and a main control unit. The main control unit is respectively connected to the multi-level motion platform, the industrial camera, the image processing unit and the defect marking module; The multi-level motion platform is used to move the strain gauge array to be measured under the control of the main control unit. It includes a first motion platform and a second motion platform. The first motion platform is arranged below the second motion platform. A loading platform is arranged on the second motion platform for carrying the strain gauge array to be measured. Positioning marks are arranged on the loading platform for determining the position information of each strain gauge in the strain gauge array on the loading platform; The industrial camera is arranged above the multi-level motion platform and is used to collect images of the strain gauge array under the control of the main control unit and transfer the collected images to the image processing unit. The industrial camera includes a first industrial camera and a second industrial camera. The first industrial camera is a low-magnification large-field-of-view camera, and the second industrial camera is a high-magnification small-field-of-view camera. The first industrial camera is used to scan the strain gauge array and collect images of the strain gauges in the strain gauge array. The second industrial camera is used to scan the sensitive grid of the strain gauge and collect images of the sensitive grid of the strain gauge; The image processing unit is used to perform image recognition processing on the images collected by the industrial camera under the control of the main control unit to determine the strain gauges with macroscopic defects or microscopic defects; The defect marking module is used to physically mark the strain gauges with macroscopic defects or microscopic defects in the strain gauge array under the control of the main control unit.

2. The device according to claim 1, wherein It also includes a drive control unit. The drive control unit is respectively connected to the first motion platform, the second motion platform and the main control unit and is used to drive the first motion platform and the second motion platform according to the control instructions of the main control unit.

3. The device according to claim 1 or 2, characterized in that, It also includes a first camera support and a second camera support, which are respectively used to support the first industrial camera and the second industrial camera. The first camera support is arranged on the first side of the multi-level motion platform, and the second camera support is arranged on the second side adjacent to the first side. The defect marking module is arranged on the third side of the multi-level motion platform.

4. The device according to claim 1 or 2, characterized in that, The distance between the first industrial camera and the multi-level motion platform is greater than the distance between the second industrial camera and the multi-level motion platform. The magnification of the first industrial camera is 20-60 times, and the field of view size is greater than 1 cm×1 cm. The magnification of the second industrial camera is 200-500 times, and the field of view size is less than 2.5 mm×2.5 mm.

5. The device according to claim 1 or 2, characterized in that, A cavity is arranged between the second motion platform and the first motion platform for placing a vacuum adsorption device. Through holes are arranged on the loading platform for cooperating with the vacuum adsorption device to adsorb the strain gauge array on the upper surface of the loading platform.

6. A strain gauge defect detection method, characterized in that, Using the device according to any one of claims 1-5 for strain gauge defect detection, the method includes: Step S11, controlling the first motion platform to move into the field of view of the first industrial camera and collecting images of each strain gauge in the strain gauge array; Step S12: Use the image processing unit to perform image recognition processing on the images of the strain gauges, and determine the strain gauges with macroscopic defects according to the image recognition results of the strain gauges; Step S13: Control the multi-stage moving platform to move into the field of view of the second industrial camera, and collect the images of the sensitive grids of the strain gauges in the strain gauge array that do not have macroscopic defects; Step S14: Use the image processing unit to perform image recognition processing on the images of the sensitive grids, and determine the strain gauges with microscopic defects according to the image recognition results of the sensitive grids; Step S15: Control the defect marking module to physically mark the strain gauges with macroscopic or microscopic defects.

7. The method according to claim 6, characterized in that, Before step S11, the method further includes: determining the position information of each strain gauge in the strain gauge array to be measured on the loading platform according to the positioning marks on the loading platform, and configuring strain gauge identifiers for each strain gauge according to the position information of each strain gauge.

8. The method according to claim 7, characterized in that, In step S12, first identify each strain gauge in the strain gauge image, record the position information and strain gauge identifier of each strain gauge; then perform macroscopic defect recognition on the strain gauge image, identify the strain gauges with macroscopic defects, and output the corresponding strain gauge identifier and corresponding position information; When performing macroscopic defect recognition on the strain gauge image, extract the gray distribution feature information of the image of a single preprocessed strain gauge, input the extracted feature information into the trained CNN model, and the output result is the confidence level of a certain macroscopic defect of the strain gauge.

9. The method according to claim 7 or 8, characterized in that In step S14, perform microscopic defect recognition on the collected images of the sensitive grids, identify the strain gauges with microscopic defects, and output the strain gauge identifiers and position information of the strain gauges with microscopic defects; When performing microscopic defect recognition on the images of the sensitive grids, extract features from the preprocessed images of the sensitive grids, input the extracted feature information into the trained CNN model, and the output result is the confidence level of a certain microscopic defect.

10. The method according to claim 9, wherein The types of macroscopic defects of the strain gauges include at least one of large-area adhesion defects and large-area corrosion defects; The types of microscopic defects of the strain gauges include any one or any combination of the following: sensitive grid adhesion, sensitive grid breakage, sensitive grid morphology distortion, and abnormal sensitive grid spacing.

Images