Electric appliance bolt installation quality detection system based on machine vision

Abstract

The invention belongs to the technical field of electric appliance bolt installation quality detection, and particularly discloses and provides an electric appliance bolt installation quality detection system based on machine vision, which comprises the steps of randomly selecting a quality overhaul sample, detecting an apparent image and offset data when the system is not electrified, analyzing scratches, deformation and the like to obtain an installation apparent offset index; the non-electrified installation quality is judged in combination with offset data, meanwhile, position images are collected after electrification, the electrified installation quality is judged, unqualified bolts are screened according to non-electrified and electrified judgment results, and then the abnormal mechanical arm is positioned and fed back; according to the method, the non-electrified detection installation quality qualification of each bolt is judged based on the deviation data in combination with the installation apparent deviation index, the reliability of the non-electrified quality evaluation of the bolts is improved, and meanwhile, whether the electrified detection installation quality is qualified or not is judged by comparing the electrified position image with the initial position image of the electrified position image. And the limitation that only the non-electrified size is detected at present is broken through.

Description

Technical Field

[0001] The invention belongs to the technical field of electrical appliance bolt installation quality detection, and relates to an electrical appliance bolt installation quality detection system based on machine vision. Background Art

[0002] In the field of modern electrical manufacturing and operation and maintenance, bolts are the basic components for achieving electrical connections and mechanical fixation. Their installation quality plays a key role in the safe operation and stable performance of the equipment. As electrical systems develop towards high voltage, high current, and intelligence, the complexity of equipment structure and the requirements for operational reliability are simultaneously increasing, and the potential risks caused by bolt installation defects are becoming increasingly prominent. The quality of bolt installation is directly related to the structural integrity of the appliance. If the bolt preload is insufficient, the position is offset, or it is loose, during the operation of the equipment, it may be affected by factors such as vibration, thermal expansion and contraction, which may cause the component to shift or fall off, and then cause mechanical failure of the equipment. This requires inspection and analysis of the bolt installation quality.

[0003] Prior art solutions have emerged for automated bolt inspection using machine vision technology. For example, Chinese invention patent publication number CN116754575A discloses a machine vision-based screw surface defect detection system. This technical solution primarily includes: an image preprocessing module, which uses morphological image processing methods to identify and remove burrs from screw images and correct for uneven illumination to optimize image quality, paving the way for subsequent processing; and a screw surface detection module, which identifies and determines surface defects based on extracted geometric features of the screw images. This solution further incorporates statistical analysis techniques to process dimensional measurement and defect detection data, aiming to implement self-diagnosis capabilities for screw production lines, improving detection efficiency and product yield.

[0004] The above existing technologies have the following deficiencies: 1. The current quality analysis of bolts only focuses on the static geometric dimensions and surface morphology defects of the bolts themselves, without considering the state changes of the bolts when the equipment is actually powered on and running. As a result, the evaluation dimension of the bolt installation quality is single and lacks adaptability to actual operating conditions, making it difficult to fully guarantee the connection reliability of the equipment under dynamic loads.

[0005] 2. Currently, the focus is only on the quality analysis of the bolt surface, and it does not involve the detection of key installation parameters such as the positioning accuracy of the bolt installation hole, the fitting clearance between the bolt and the installation hole, and the relative position relationship when multiple bolts are installed in a coordinated manner. As a result, the connection reliability of the bolt installation cannot be guaranteed. Summary of the Invention

[0006] In view of this, in order to solve the problems raised in the above background technology, a machine vision-based electrical bolt installation quality detection system is proposed.

[0007] The purpose of the present invention can be achieved through the following technical solutions: The present invention provides an electrical bolt installation quality inspection system based on machine vision, including: randomly selecting a preset number of products from the bolts tightened by each manipulator as quality inspection samples.

[0008] When power is not supplied, the surface image and offset data of each bolt in the quality inspection sample are inspected.

[0009] The scratch influence index and deformation influence index of the bolts are analyzed based on the surface image, and the installation apparent deviation index of each bolt is confirmed accordingly.

[0010] Based on the offset data and the installation apparent offset index, each bolt is subjected to an unpowered test to determine whether the installation quality is acceptable.

[0011] The quality inspection sample is powered on, and a position image after the power-on is collected.

[0012] The position image after power-on is compared with the initial position image to determine whether the power-on detection and installation quality of each bolt is qualified.

[0013] Based on the installation quality judgment results of the non-power-on detection and the power-on detection, the bolts with unqualified installation quality are screened, the abnormal manipulator is located and corresponding feedback is provided.

[0014] Compared with the existing technology, the beneficial effects of the present invention are as follows: (1) The present invention realizes full life cycle quality monitoring from assembly accuracy to working condition reliability by combining static detection without power supply with dynamic detection with power supply. It not only makes up for the defect of the existing technology that only focuses on static dimensions, but also significantly improves the comprehensiveness, accuracy and engineering practicality of detection through multi-parameter independent judgment and robot tracing mechanism.

[0015] (2) The present invention performs non-energized detection and installation quality assessment of each bolt based on the offset data combined with the installation apparent offset index, breaking through the current limitation of focusing only on surface quality or single position offset, and improving the reliability of bolt non-energized quality assessment.

[0016] (3) The present invention compares the position image after power-on with the image of its initial position, analyzes the horizontal and vertical offsets, deformation area ratios, and gap deviation values, and then determines whether the installation quality of each bolt is qualified after power-on inspection. This breaks through the current limitation of only detecting the dimensions before power-on, and realizes the dynamic performance evaluation of the bolts under actual power-on operation.

[0017] (4) The present invention adopts independent threshold judgment logic of position deviation index, deformation deviation index and looseness deviation index to determine whether the installation quality of each bolt is qualified during power-on detection, thereby avoiding misjudgment of a single dimension, realizing accurate identification of power-on operation defects, and improving the comprehensiveness and reliability of quality analysis.

[0018] (5) The present invention analyzes the installation gap offset index and the installation axis offset index, and combines the installation apparent offset index to perform non-power detection and installation quality judgment, thereby avoiding abnormal bolt installation quality caused by gap offset, axis offset and apparent defects, and ensuring the reliability of bolt installation from three aspects: position accuracy, assembly gap and surface defects. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0020] Figure 1 Schematic diagram of the connection of the system of the present invention.

[0021] Figure 2 This is a connection diagram of the steps for determining the quality of the installation during the power-off detection process of the present invention.

[0022] Figure 3 This is a connection diagram of the steps for determining whether the installation quality of each bolt is qualified by power-on detection according to the present invention. DETAILED DESCRIPTION

[0023] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0024] See also Figure 1 As shown, the present invention provides an electrical bolt installation quality inspection system based on machine vision, which includes: randomly selecting a preset number of products from the bolts tightened by each manipulator as quality inspection samples.

[0025] When power is not supplied, the surface image and offset data of each bolt in the quality inspection sample are inspected.

[0026] It should be supplemented that the offset data includes: the initial gap between each bolt head and the connecting piece, the axis deviation angle of each bolt, and the distance from the position where the bolt axis deviation angle offset occurs to the bottom of the connecting piece.

[0027] It should be noted that the apparent image was captured by a camera, the initial gap was detected by a laser vertical sensor, the axis deviation was detected by a laser radar, and the distance from the location where the bolt axis deviation occurred to the bottom of the connector was detected by a laser vertical sensor.

[0028] The scratch influence index and deformation influence index of the bolts are analyzed based on the surface image, and the installation apparent deviation index of each bolt is confirmed accordingly.

[0029] Exemplarily, the bolt installation apparent quality analysis specifically includes: W1, extracting the length, depth and number of scratches on each bolt from the apparent image, performing a scratch analysis on the bolt, and obtaining a scratch influence index of each bolt.

[0030] Furthermore, the bolt scratch analysis includes: W1-1, calculating the ratios of the length and depth of each scratch to preset thresholds to obtain the length ratio and depth ratio of each scratch.

[0031] It should be added that the preset thresholds are respectively a preset scratch length and a preset scratch depth, and the maximum values ​​are respectively selected from the scratch lengths and scratch depths of each bolt as the preset scratch length and the preset scratch depth.

[0032] W1-2. Calculate the average of the product of the length ratio and depth ratio of each scratch to obtain the scratch damage index of each bolt.

[0033] W1-3. The maximum value is selected from the number of scratches on each bolt as the reference number of scratches, and the ratio of the number of scratches on each bolt to the reference number of scratches is used as the scratch number ratio of each bolt.

[0034] W1-4. Couple the scratch damage index of each bolt with the scratch number ratio to obtain the scratch influence index of each bolt.

[0035] W2. Extract the completed installation image of each bolt from the apparent image and overlap it with the initial image to obtain the overlapping image area and the initial image area. The difference between the two is used as the deformation area, and the ratio of the deformation area to the initial area is used as the deformation influence index of each bolt.

[0036] It should be added that the initial image refers to an image before the bolt is installed.

[0037] W3. Multiply the scratch influence index and deformation influence index of each bolt to obtain the installation apparent deviation index of each bolt.

[0038] It should be added that scratches are caused by improper operation of the robot, tool wear, or part collision during the installation process of the bolt. The length, depth, and number of scratches can be directly related to the rationality of the robot's process during the installation process. Scratches will affect the mechanical properties and durability of the bolt. Deeper scratches may become stress concentration points, reducing fatigue strength and leading to fracture during long-term use. Deformation directly reflects the installation accuracy and assembly consistency. Deformation is an intuitive representation of installation offset. If the bolt is deformed after installation, it usually indicates that its axis deflection exceeds the design allowable range or that the force is uneven during assembly. Scratches alone may indicate rough operation, but may not directly affect the function. Deformation alone may reflect design or assembly errors, but may not be caused by operation. Therefore, the coupled calculation of scratches and deformation avoids the misjudgment of a single indicator.

[0039] Based on the offset data and the installation apparent offset index, each bolt is subjected to an unpowered test to determine whether the installation quality is acceptable.

[0040] See also Figure 2 As shown, exemplarily, the unpowered detection and installation quality judgment of each bolt includes: Q1, extracting the initial gap between each bolt head and the connecting part from the offset data, and calculating the relative deviation between it and the preset gap to obtain the installation gap offset index of each bolt.

[0041] It should be added that the preset gap is the design requirement gap for bolt installation, that is, the theoretical gap value between the bolt head and the connector specified in the electrical equipment assembly drawing or process standard, and the unit is millimeter (mm). This value is determined by factors such as the pre-tightening force requirements of the bolt connection, the elastic modulus of the connector material and the surface flatness. For example, the preset gap design value of a certain model of electrical bolts is 0.2mm±0.05mm.

[0042] Q2. Compare the axis deviation angle of each bolt in the offset data with the preset axis deviation angle threshold to determine the installation axis deviation index of each bolt.

[0043] It should be noted that the preset axis deviation threshold is a standard angle value used to measure whether the bolt's axis deviation is within the allowable range. When the actual axis deviation of the bolt is compared with this threshold, if it is less than or equal to the preset axis deviation threshold, it indicates that the bolt's axis deviation is within the acceptable range. If it is greater than the preset axis deviation threshold, it indicates that the bolt's axis deviation may have a negative impact on installation quality and requires further evaluation. The preset axis deviation threshold is calculated by averaging the axis deviation angles of each historical bolt extracted from historical data.

[0044] Furthermore, the determining of the installation axis deviation index of each bolt includes: Q2-1, if the axis deviation angle of the bolt is less than or equal to a preset axis deviation angle threshold, then using the preset deviation index as the installation axis deviation index of the bolt.

[0045] It should be added that the preset offset index can specifically be set to 0.

[0046] Q2-2. If the axis deviation angle of the bolt is greater than a preset axis deviation angle threshold, the axis deviation angle is matched with the axis deviation angle interval corresponding to each installation angle offset index to obtain the installation angle offset index of the bolt.

[0047] It should be noted that the axis deviation intervals corresponding to each installation angle offset index are subdivided into intervals within the bolt axis deviation range, with each interval corresponding to an installation angle offset index. When the bolt axis deviation exceeds a preset axis deviation threshold, the bolt's installation angle offset index is determined by matching it with these intervals, allowing for a more accurate assessment of the impact of axis deviation on bolt installation quality. The axis deviation intervals corresponding to each installation angle offset index were obtained through experimental verification. Bolts were installed at different deviation angles, such as 0°, 3°, 6°, and 9°, with 50 bolts per group. Vibration or force was simulated to the bolts while energized, and the number of bolt failures in each group was recorded. The ratio of the number of failed bolts in each group to the total number of failed bolts was used as the installation angle offset index. For example, 0°-3° is designated as a low-risk interval with an installation angle offset index of 0.2; 3°-6° is designated as a medium-risk interval with an installation angle offset index of 0.6; and 6° and above is designated as a high-risk interval with an installation angle offset index of 0.9.

[0048] Q2-3. Obtain the distance from the location where the bolt axis deflection occurs to the bottom of the connector from the offset data, and use it as the offset depth of the bolt axis deflection. The ratio of the offset depth to the preset offset depth threshold is used as the bolt installation position offset index.

[0049] Q2-4. The product of the installation angle offset index and the installation position offset index is calculated as the installation axis offset index of the bolt to obtain the installation axis offset index of each bolt.

[0050] Q3. Ratios of the installation clearance offset index, installation axis offset index, and installation apparent offset index of each bolt to their preset values ​​are calculated, and based on these ratios, a non-powered inspection and installation quality qualification judgment is performed to obtain the non-powered inspection and installation quality qualification judgment results of each bolt.

[0051] The embodiment of the present invention analyzes the installation gap offset index and the installation axis offset index, and combines the installation apparent offset index to perform non-power detection and installation quality judgment, thereby avoiding abnormal bolt installation quality caused by gap offset, axis offset and apparent defects, and ensuring the reliability of bolt installation from three aspects: position accuracy, assembly gap and surface defects.

[0052] Furthermore, the non-powered detection installation quality qualification judgment includes: Q3-1, the installation gap offset index is less than its preset installation gap offset index as condition 1, the installation axis offset index is less than its preset installation axis offset index as condition 2, and the installation apparent offset index is less than its preset installation apparent offset index as condition 3.

[0053] Q3-2. If conditions 1, 2, and 3 are all met, the bolt's non-powered detection and installation quality is determined to be qualified; otherwise, the bolt's non-powered detection and installation quality is determined to be unqualified, thereby obtaining the qualified judgment results of the non-powered detection and installation quality of each bolt.

[0054] The embodiment of the present invention performs a non-energized detection and installation quality assessment of each bolt based on the offset data combined with the installation apparent offset index, breaking through the current limitation of only focusing on surface quality or single position offset, and improving the reliability of the bolt non-energized quality assessment.

[0055] The quality inspection sample is powered on, and a position image after the power-on is collected.

[0056] It should be added that the position image is obtained by shooting with a camera.

[0057] The position image after power-on is compared with the initial position image to determine whether the power-on detection and installation quality of each bolt is qualified.

[0058] Using power-on detection to check the installation quality of PCB bolts can avoid the situation where only static geometric parameters such as bolt clearance, axis deviation, surface scratches, and deformation can be analyzed under non-power-on detection. The fitting clearance between the bolt and the mounting hole cannot be exposed in a static state, and the state changes under actual working conditions cannot be reflected. By collecting position images after power is applied, the displacement, deformation, and looseness of the bolts under dynamic loads such as current, vibration, thermal expansion and contraction can be analyzed, breaking through the limitations of static detection and realizing dynamic performance evaluation.

[0059] See also Figure 3 As shown, exemplarily, the method of determining whether the installation quality of each bolt after power-on detection is qualified includes: Y1, overlapping and comparing the position image of each bolt after the power-on detection with its initial position image, extracting the offset of each bolt in the horizontal and vertical directions to calculate the position deviation, and obtaining the position deviation index of each bolt.

[0060] Furthermore, the calculation of the position deviation index of each bolt includes: Y1-1, taking the ratio of the horizontal offset of each bolt to its preset offset threshold as the horizontal offset ratio of each bolt.

[0061] Y1-2. Analyze the vertical offset ratio of each bolt in the same way as the horizontal offset ratio.

[0062] Y1-3. Perform weighted fusion calculation on the horizontal offset ratio and vertical offset ratio of each bolt to obtain the position deviation index of each bolt.

[0063] It should be added that the calculation formula of the position deviation index is: , where is the position deviation index, and are the horizontal offset ratio and the vertical offset ratio, and are the weights of the horizontal offset ratio and the vertical offset ratio, respectively, which are used to quantify the impact of offsets in different directions on position deviation, such as , .

[0064] It should be added that the position deviation index of each bolt is calculated through weighted fusion. On the one hand, the weight distribution can reflect the actual weight of the influence of the horizontal offset ratio and the vertical offset ratio on the different dimensions of the bolt position deviation, and reflect the difference in the contribution of offsets in different directions to the degree of bolt position deviation; on the other hand, the information of the horizontal and vertical offset ratios can be directly integrated to comprehensively consider the influence of the two on the bolt position deviation.

[0065] The weights can be set based on bolt installation requirements and actual engineering experience, or obtained through test data. For example, historical data on the horizontal and vertical offsets of bolts and the corresponding impact of position deviations can be collected first, and the correlation coefficients between the horizontal offset ratio, vertical offset ratio and bolt position deviation can be calculated. The contribution of the two to the position deviation can be determined through regression analysis or logistic regression analysis. After normalization, the contribution is converted into the weights of the horizontal offset ratio and vertical offset ratio, and the total weight is 1, so as to accurately quantify the bolt position deviation index.

[0066] Y2. Based on the position image of each bolt after the power-on test and its initial position image, the deformation area and initial contour area of ​​each bolt are obtained through the edge detection algorithm, and the ratio of the two is used as the deformation deviation index of each bolt.

[0067] It should be supplemented that the deformation area refers to the area of ​​the non-overlapping portion of the bolt in the position image and its initial position image.

[0068] Y3. Obtain the current gap between each bolt head and the connecting piece from the position image of each bolt after the power-on detection is completed, perform deviation analysis based on the current gap and the initial gap, and obtain the loosening deviation index of each bolt.

[0069] Furthermore, the analysis of the loosening deviation index of each bolt includes: Y3-1, taking the difference between the current gap and the initial gap as the gap deviation value of each bolt.

[0070] Y3-2. Normalize the gap deviation values ​​of each bolt, and calculate the ratio of the absolute value of the gap deviation value to the preset maximum allowable value of the gap deviation.

[0071] Y3-3. Minimize the ratio and 1 to obtain the looseness deviation index of each bolt.

[0072] It should be noted that the preset maximum allowable clearance deviation is a safety threshold for the clearance deviation between the bolt head and the connector, expressed in millimeters (mm). This value indicates that when the clearance deviation exceeds the preset maximum allowable clearance deviation, there is a significant risk of loosening in the bolt connection, potentially leading to preload failure or vertical connection of the connector. The preset maximum allowable clearance deviation is set by reference to industry standards. For example, ISO 898-1 stipulates that the allowable clearance deviation for steel structure bolts is 0.3-0.5mm.

[0073] The embodiment of the present invention compares the position image after power-on with its initial position image, analyzes the horizontal and vertical offsets, deformation area ratios, and gap deviation values, and then determines whether the power-on detection and installation quality of each bolt is qualified. This breaks through the current limitation of only detecting the dimensions before power-on, and realizes the dynamic performance evaluation of the bolts under actual power-on operation.

[0074] Y4. Compare the position deviation index, deformation deviation index and looseness deviation index of each bolt with its preset value respectively. If any deviation index is greater than its preset value, it is judged that the power-on inspection and installation quality of the bolt is unqualified; otherwise, it is judged that the power-on inspection and installation quality of the bolt is qualified.

[0075] The embodiment of the present invention adopts independent threshold judgment logic for position deviation index, deformation deviation index, and looseness deviation index to determine whether the power-on detection installation quality of each bolt is qualified, avoiding single-dimensional misjudgment, realizing accurate identification of power-on operation defects, and improving the comprehensiveness and reliability of quality analysis.

[0076] Based on the installation quality judgment results of the non-power-on detection and the power-on detection, the bolts with unqualified installation quality are screened, the abnormal manipulator is located and corresponding feedback is provided.

[0077] Exemplarily, the abnormal positioning robot includes: if the non-power detection installation quality judgment result of a bolt is that the non-power detection installation quality is qualified and the power detection installation quality judgment result is that the power detection installation quality is qualified, then the bolt installation quality is judged to be qualified; otherwise, the bolt installation quality is judged to be unqualified, and then the bolts with unqualified installation quality are obtained.

[0078] The bolts corresponding to each manipulator are extracted from the quality inspection sample, and then the bolts with unqualified installation quality are matched with the bolts corresponding to each manipulator to obtain the number of unqualified installation bolts corresponding to each manipulator, and the ratio of the number to the preset number is used as the ratio of the number of unqualified installation bolts of each manipulator.

[0079] The manipulator whose number of bolts with unqualified installation quality is greater than a preset value is regarded as an abnormal manipulator.

[0080] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0081] The above embodiments may be implemented in whole or in part through software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0082] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0083] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0084] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0085] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A machine vision-based electrical bolt installation quality inspection system, characterized by: The system includes: Randomly select a preset number of products from the bolts tightened by each robot as quality inspection samples; When power is not supplied, inspect the surface image and offset data of each bolt in the quality inspection sample; Analyze the scratch impact index and deformation impact index of the bolts based on the surface image, and confirm the installation apparent offset index of each bolt accordingly; Based on the offset data and the installation apparent offset index, each bolt is tested without power supply to determine whether the installation quality is qualified; Powering on the quality inspection sample and collecting a position image after powering on; Comparing the position image after power-on with the initial position image to determine whether the power-on detection installation quality of each bolt is qualified; Based on the installation quality judgment results of the non-power-on detection and the power-on detection, the bolts with unqualified installation quality are screened, the abnormal manipulator is located and corresponding feedback is provided.

2. The machine vision-based electrical bolt installation quality inspection system according to claim 1, characterized in that: The bolt installation apparent quality analysis specifically includes: W1. Extract the length, depth, and number of scratches on each bolt from the surface image, perform scratch analysis on the bolt, and obtain the scratch impact index of each bolt; W2. Extract the installed image of each bolt from the surface image and overlap it with the initial image to obtain the area of ​​the overlapped image and the area of ​​the initial image. The difference between the two is used as the deformation area, and the ratio of the deformation area to the initial area is used as the deformation influence index of each bolt. W3. Multiply the scratch influence index and deformation influence index of each bolt to obtain the installation apparent deviation index of each bolt.

3. The machine vision-based electrical bolt installation quality inspection system according to claim 2, characterized in that: The scratch analysis of the bolt includes: The length and depth of each scratch are calculated by comparing them with their preset thresholds to obtain the length ratio and depth ratio of each scratch; The product of the length ratio and depth ratio of each scratch is averaged to obtain the scratch damage index of each bolt; The maximum value is selected from the number of scratches on each bolt as the reference number of scratches, and the ratio of the number of scratches on each bolt to the reference number of scratches is taken as the scratch number ratio of each bolt; The scratch damage index of each bolt is coupled with the scratch number ratio to obtain the scratch influence index of each bolt.

4. The machine vision-based electrical bolt installation quality inspection system according to claim 3, characterized in that: The unpowered detection and installation quality assessment of each bolt includes: Q1. Extract the initial gap between each bolt head and the connecting piece from the offset data, and calculate the relative deviation between it and the preset gap to obtain the installation gap offset index of each bolt; Q2. Compare the axis deviation angle of each bolt in the offset data with a preset axis deviation angle threshold to determine the installation axis deviation index of each bolt; Q3. Ratios of the installation clearance offset index, installation axis offset index, and installation apparent offset index of each bolt to their preset values ​​are calculated, and based on these ratios, a non-powered inspection and installation quality qualification judgment is performed to obtain the non-powered inspection and installation quality qualification judgment results of each bolt.

5. The machine vision-based electrical bolt installation quality inspection system according to claim 4, characterized in that: The determination of the installation axis deviation index of each bolt includes: If the axis deviation angle of the bolt is less than or equal to the preset axis deviation angle threshold, the preset deviation index is used as the installation axis deviation index of the bolt; If the axis deflection angle of the bolt is greater than a preset axis deflection angle threshold, the axis deflection angle is matched with the axis deflection angle interval corresponding to each installation angle offset index to obtain the installation angle offset index of the bolt; The distance from the location where the bolt axis deflection occurs to the bottom of the connecting member is obtained from the offset data and used as the offset depth of the bolt axis deflection. The ratio of the offset depth to a preset offset depth threshold is used as the bolt installation position offset index. The product of the installation angle offset index and the installation position offset index is calculated as the installation axis offset index of the bolt, and the installation axis offset index of each bolt is obtained.

6. The machine vision-based electrical bolt installation quality inspection system according to claim 4, characterized in that: The non-powered detection installation quality qualification determination includes: The installation clearance offset index is smaller than the preset installation clearance offset index as condition 1, the installation axis offset index is smaller than the preset installation axis offset index as condition 2, and the installation apparent offset index is smaller than the preset installation apparent offset index as condition 3; If conditions 1, 2 and 3 are all met, the bolt's non-power detection and installation quality is judged to be qualified; otherwise, the bolt's non-power detection and installation quality is judged to be unqualified, thereby obtaining the qualified judgment results of the non-power detection and installation quality of each bolt.

7. The machine vision-based electrical bolt installation quality inspection system according to claim 1, characterized in that: The determination of whether the installation quality of each bolt is qualified by power-on detection includes: Y1. Overlap and compare the position image of each bolt after the power-on test with its initial position image, extract the offset of each bolt in the horizontal and vertical directions, calculate the position deviation, and obtain the position deviation index of each bolt; Y2. Based on the position image of each bolt after power-on detection and its initial position image, an edge detection algorithm is used to obtain the deformation area and initial contour area of ​​each bolt, and the ratio of the two is used as the deformation deviation index of each bolt; Y3. Obtaining the current gap between each bolt head and the connecting member from the position image of each bolt after the power-on test is completed, performing deviation analysis based on the current gap and the initial gap to obtain a loosening deviation index for each bolt; Y4. Compare the position deviation index, deformation deviation index and looseness deviation index of each bolt with its preset value respectively. If any deviation index is greater than its preset value, it is judged that the power-on inspection and installation quality of the bolt is unqualified; otherwise, it is judged that the power-on inspection and installation quality of the bolt is qualified.

8. The machine vision-based electrical bolt installation quality inspection system according to claim 7, characterized in that: The calculation of the position deviation index of each bolt includes: The ratio of the horizontal offset of each bolt to its preset offset threshold is used as the horizontal offset ratio of each bolt; The vertical offset ratio of each bolt is obtained by similar analysis according to the above-mentioned horizontal offset ratio analysis method; The horizontal offset ratio and vertical offset ratio of each bolt are weighted and fused to obtain the position deviation index of each bolt.

9. The machine vision-based electrical bolt installation quality inspection system according to claim 7, characterized in that: The analysis of the loosening deviation index of each bolt includes: The difference between the current gap and the initial gap is used as the gap deviation value of each bolt; Normalizing the gap deviation value of each bolt, and calculating the ratio of the absolute value of the gap deviation value to the preset maximum allowable gap deviation value; The looseness deviation index of each bolt is obtained by taking the minimum value of the ratio and 1.

10. The machine vision-based electrical bolt installation quality inspection system according to claim 1, characterized in that: The abnormal positioning manipulator includes: If the unpowered detection installation quality qualified judgment result of a bolt is that the unpowered detection installation quality is qualified and the powered detection installation quality qualified judgment result is that the powered detection installation quality is qualified, then the bolt installation quality is determined to be qualified; otherwise, the bolt installation quality is determined to be unqualified, and then the bolts with unqualified installation quality are obtained; Extract the bolts corresponding to each manipulator from the quality inspection sample, and then match the bolts with the bolts corresponding to each manipulator to obtain the number of bolts with unqualified installation quality corresponding to each manipulator, and use the ratio of the number of bolts with unqualified installation quality corresponding to each manipulator to the preset number as the ratio of the number of bolts with unqualified installation quality for each manipulator; The manipulator whose number of bolts with unqualified installation quality is greater than a preset value is regarded as an abnormal manipulator.

Images