An inspection system for precision fasteners

Abstract

The present application relates to the technical field of data analysis, and particularly relates to a kind of inspection system for precision fastener, comprising: construction module, for uploading component standard structure parameter, application component standard structure parameter constructs component standard three-dimensional model in three-dimensional scene, synchronous storage is carried out to the component standard three-dimensional model of construction;Capture module, for collecting the image data of component to be detected, the image data of component to be detected is segmented, and the image of component to be detected in the image data of component to be detected is captured;The present application constructs the three-dimensional model of component to be detected by the high-precision restoration of component multi-party image acquisition, and uploads component standard structure parameter, and component standard three-dimensional model is constructed synchronously, to compare component standard three-dimensional model with the three-dimensional model of component to be detected, to provide effective quality determination for component to be detected.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, and more specifically to an inspection system for precision fasteners. Background Technology

[0002] Precision fasteners are components in fields such as mechanical manufacturing that require high precision in dimensions, shape, and surface quality. They are typically manufactured using advanced processing technologies and processes with strict tolerance control, and are widely used in industries such as aerospace, automotive, and electronics. They are key elements in ensuring the performance and reliability of high-end equipment and products.

[0003] Patent application No. 202411093566.3 discloses a non-destructive testing system based on the contour reconstruction of a metal component, including a robotic arm, an eddy current sensor, and a computing unit. The eddy current sensor is installed on an actuator at the end of the robotic arm and is used to detect the electromagnetic properties of the surface of the metal component and collect eddy current signals at various measurement points on the surface of the metal component. The computing unit is communicatively connected to the robotic arm and is used to obtain the pose information of the actuator at the end of the robotic arm based on the eddy current signals. By processing the pose information, the pose adjustment feedback information of the actuator and the scanning trajectory of the eddy current sensor are obtained, and the pose adjustment feedback information is fed back to the robotic arm. Finally, the surface contour of the metal component is reconstructed based on the scanning trajectory. The robotic arm is used to adjust the pose of the actuator according to the pose adjustment feedback information so that the axis of the eddy current sensor is always perpendicular to the surface of the metal component being measured and maintains a fixed lift-off distance.

[0004] The application aims to address the problem that "in the quality control process of industrial products, the non-destructive testing of complex metal components depends on the accurate measurement of the surface contour of the object being tested. Currently, surface contour information is mainly obtained through design specifications or by measuring with specialized instruments, which requires prior knowledge of relevant information or preliminary measurements during testing, thus reducing testing efficiency."

[0005] However, for precision fasteners such as bolts, the key lies in controlling the precision of the threads and slots. Currently, there are technologies that can perform quality inspection on such precision fasteners, but these technologies have not been innovated for a long time.

[0006] Therefore, we propose a brand-new precision fastener inspection system. Summary of the Invention

[0007] In view of the above-mentioned shortcomings of the prior art, the present invention provides an inspection system for precision fasteners, which solves the technical problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] An inspection system for precision fasteners, comprising:

[0010] The construction module uploads standard structural parameters of components, applies these parameters to build a standard 3D model of the component in a 3D scene, and simultaneously stores the constructed standard 3D model. The capture module acquires image data of the component to be inspected, segments the image data, and captures the image of the component to be inspected within the image data. The solid modeling module obtains the structural parameters of the component to be inspected from the captured images in the capture module and creates a 3D model of the component based on these parameters. The reference point picking module receives the standard 3D model of the component built by the construction module and the 3D model of the component to be inspected created by the solid modeling module, and picks reference points on both 3D models. The demonstration module receives the reference points picked by the reference point picking module and determines the two 3D models based on these reference points. The system comprises a line model, which controls the line model to move at a constant speed along the thread in its corresponding 3D model until it reaches the end of the thread, thus obtaining a demonstration animation for thread quality inspection; an analysis module, which receives demonstration animations from the standard 3D model of the component and the 3D model of the component to be inspected, and captures frames at the same time in the two demonstration animations based on a preset frame capture logic, and analyzes the degree of dimensional standardization of the component to be inspected based on the similarity evaluation of the frames captured at the same time in the two demonstration animations; a judgment module, which sets the component qualification judgment threshold, compares the consistency between the component qualification judgment threshold and the degree of dimensional standardization of the component to be inspected obtained from the analysis module, and determines whether the component is qualified; and a sniffing module, which monitors the qualification judgment result of the judgment module, and sniffs the location of thread defects on the surface of the component when the judgment result is negative.

[0011] Furthermore, when storing the standard 3D model of the component, the construction module generates a data packet using the structural parameters applied in the construction stage of the standard 3D model of the component, and binds and stores the data packet with the corresponding constructed standard 3D model of the component.

[0012] When acquiring image data of the component to be detected, the capture module uses the following viewing angles: top, bottom, front, and back. The background is a solid-color curtain preset by the system user, distinct from the surface color of the component. The capture module has sub-modules, including:

[0013] The segmentation unit is used to receive the image data of the component to be detected collected by the capture module, segment the image data of the component to be detected, and output the image of the component to be detected.

[0014] The segmentation unit is equipped with image segmentation processing logic, and the segmentation unit performs segmentation processing operations on the image data of the component to be detected based on the image segmentation processing logic.

[0015] Furthermore, the image segmentation processing logic in the segmentation unit is as follows:

[0016] Based on the background of the image data of the component to be detected, a threshold for determining the target pixel value of pixel segmentation is set. The target pixel value of pixel segmentation is applied to perform a traversal operation in the image data of the component to be detected, and pixels that meet the target pixel value of pixel segmentation are picked up. The picked pixels are used as the target pixels for segmentation and the segmentation operation is performed in the image data of the component to be detected to obtain the image of the component to be detected.

[0017] The edge pixels of the image of the component to be detected are captured, and each edge pixel is used as a retrieval target to retrieve the pixel with the highest similarity in the image of the component to be detected.

[0018] The search result is copied to a pixel, and the corresponding search target is iterated with the copied pixel. After all the edge pixels of the image of the component to be detected have been iterated, the output operation of the image of the component to be detected is performed.

[0019] ;

[0020] In the formula: For pixel brightness, saturation, and hue; The value of the pixel in the three color channels; These represent the maximum and minimum values ​​of the pixel in the three color channels, respectively.

[0021] Calculate each pixel based on the above formula. Then the similarity between the two pixels is:

[0022] ;

[0023] In the formula: To adjust the index;

[0024] Among them, the adjustment index It is 1 or -1. The numerator of the fraction is less than or equal to the denominator. =1, The numerator of the fraction is greater than the denominator. =-1.

[0025] Furthermore, during the operation phase of the entity modeling module, when the structural parameters of the component to be detected are acquired, the image of the component to be detected is placed in a two-dimensional coordinate system. Based on the two-dimensional coordinate system, the coordinates of each corner point on the image of each component to be detected are identified. The structural parameters of the component to be detected are determined by applying the coordinates of each corner point. Based on the structural parameters of the component to be detected, a three-dimensional model of the component to be detected is created, so that each thread and thread groove on the three-dimensional model of the component to be detected is constructed based on independent structural parameters.

[0026] The structural parameters of the top and bottom of the three-dimensional model of the component to be tested are derived from the images of the component to be tested in four directions, and the structural parameters of the thread and thread groove of the three-dimensional model of the component to be tested are derived from the images of the component to be tested in the front and back directions.

[0027] Furthermore, when the reference point picking module picks reference points on two 3D models, it follows the following rules:

[0028] Identify the connection surface between the component head and the screw part on the component head in the 3D model, and use the corner point of the connection surface as a reference point;

[0029] A sub-module is provided between the reference point picking module and the demonstration module, including:

[0030] The connection unit is used to connect all reference points on the 3D model of the component to be tested to the thread start point on the 3D model of the component to be tested, and to connect all reference points on the standard 3D model of the component to the thread start point on the standard 3D model of the component.

[0031] After the connection unit finishes running, a set of three-dimensional models composed of multiple line segments are obtained in the three-dimensional model of the component to be tested and the standard three-dimensional model of the component, respectively. These are denoted as line model one and line model two, with one end intersecting at a point and the other end having different positions. Model one and line model two correspond to the three-dimensional model of the component to be tested and the standard three-dimensional model of the component, respectively.

[0032] Furthermore, the control logic for the control line models in the demonstration module to move at a constant speed along the thread in their respective corresponding 3D models is as follows:

[0033] The intersection of each line segment in the line model is used as the moving end. The system and user terminals define the movement speed, so that the intersection ends move along the thread on the surface of the component based on the set movement speed. During the movement, the line segments in the line model are stretched and offset based on the movement of the intersection ends to form a demonstration animation.

[0034] When the user-defined motion speed of the line model on the system side is set, it follows the logic that the tighter the thread on the component surface and the higher the precision requirement of the component, the slower the motion speed, and vice versa.

[0035] Furthermore, the frame capture logic in the analysis module is represented as follows:

[0036] The duration of two demonstration animations is identified. The intersection of the durations of the two demonstration animations is used as the time domain for frame capture. The frame capture time interval is set. The frame at the beginning of the time domain is captured for the first time. Based on the timestamp of the first captured frame and the frame capture time interval, the frame is continuously captured in the two demonstration animations. Each captured frame is marked with its capture timestamp.

[0037] In the setting stage of the frame capture time interval, the logic is that the tighter the thread on the component surface and the higher the precision requirement of the component, the shorter the frame capture time interval, and vice versa.

[0038] Furthermore, the analysis module has sub-modules at its lower level, including:

[0039] The matching unit is used to identify the capture timestamps marked on each captured frame in the analysis module, and to match the frames marked with the same capture timestamps with each other, which are then used as analysis targets by the analysis module.

[0040] The analysis logic for the degree of standardization of the dimensions of the component to be detected in the analysis module (6) is expressed as follows:

[0041] ;

[0042] In the formula: Let the midline model morphology similarity between two images at the i-th capture timestamp; It is the set of endpoints of the corresponding corner points of each line segment in the line model; Let be the area of ​​the triangular region formed by the intersection of the i-th endpoint and the (i+1)-th endpoint and the line segments to which they belong in line model one; Let be the area of ​​the triangular region formed by the intersection of the i-th endpoint and the (i+1)-th endpoint and the line segments to which they belong in line model 2; The degree of standardization of the dimensions of the component to be inspected; A collection of timestamps captured from video frames;

[0043] in, Indicates to The operation of finding the mean. Similarly.

[0044] Furthermore, the component qualification threshold in the judgment module is defined by the system user. When the standardization degree of the component size to be detected is greater than or equal to the component qualification threshold, the component is judged as qualified; otherwise, the component is judged as unqualified.

[0045] During the operation of the sniffing module, respectively, it acquires... The value, in Each item is compared with the component's qualification threshold, and items that do not meet the qualification threshold are captured. The location of the thread defect on the component's surface is recorded as the position on the 3D model of the component to be inspected, which is the intersection of the line models corresponding to the non-compliant items.

[0046] Furthermore, the construction module interacts with the capture module via a wireless network, the capture module interacts with the segmentation unit via a wireless network, the capture module interacts with the entity modeling module via a wireless network, the entity modeling module interacts with the segmentation unit via a wireless network, the entity modeling module interacts with the reference point picking module and the demonstration module via a wireless network, the reference point picking module and the demonstration module interact with the connection unit via a wireless network, the demonstration module interacts with the analysis module via a wireless network, the analysis module interacts with the matching unit via a wireless network, and the analysis module interacts with the judgment module and the sniffing module via a wireless network.

[0047] Compared with known public technologies, the technical solution provided by this invention has the following beneficial effects:

[0048] (1) The present invention constructs a three-dimensional model of the component to be tested by constructing a high-precision reconstruction of multi-source image acquisition, and uploads the standard structural parameters of the component, and simultaneously constructs a standard three-dimensional model of the component. The standard three-dimensional model of the component is compared with the three-dimensional model of the component to be tested to provide an effective quality judgment for the component to be tested.

[0049] (2) When comparing the standard three-dimensional model of the component with the three-dimensional model of the component to be tested, the present invention picks up the feature points on the model surface, creates line models for the standard three-dimensional model of the component and the three-dimensional model of the component to be tested respectively, and generates a demonstration animation of the line model. The frame of the screen in the demonstration animation is used as the comparison data, which greatly improves the accuracy of the quality judgment of the component to be tested.

[0050] (3) After the quality judgment of the component to be tested, the present invention can also detect the location of the surface defect of the component when the quality judgment of the component to be tested is unqualified, thereby using this as data support to assist the production user in reverse inspection of the hidden defects of the production equipment based on the defect location. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0052] Figure 1 This is a schematic diagram of a testing system for precision fasteners;

[0053] Figure 2 This is a schematic diagram of the morphology of the midline model in different frames in this invention. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0055] The present invention will be further described below with reference to embodiments.

[0056] Example:

[0057] This embodiment provides an inspection system for precision fasteners, such as... Figure 1 As shown, it includes:

[0058] Module 1 is used to upload the standard structural parameters of the component, apply the standard structural parameters of the component to build a standard 3D model of the component in the 3D scene, and simultaneously store the built standard 3D model of the component.

[0059] Capture module 2 is used to acquire image data of the component to be detected, segment the image data of the component to be detected, and capture the image of the component to be detected in the image data of the component to be detected;

[0060] When storing the standard 3D model of a component, Module 1 generates a data packet using the structural parameters applied during the construction phase of the standard 3D model of the component, and binds and stores the data packet with the corresponding constructed standard 3D model of the component.

[0061] When capturing image data of the component to be detected, the capture module 2 uses the following viewing angles: top, bottom, front, and back. The background is a solid-color curtain preset by the system user, distinct from the surface color of the component. The capture module has sub-modules, including:

[0062] The segmentation unit 21 is used to receive the image data of the component to be detected collected by the capture module 2, perform segmentation processing on the image data of the component to be detected, and output the image of the component to be detected.

[0063] The segmentation unit 21 is equipped with image segmentation processing logic, and the segmentation unit 21 performs segmentation processing operations on the image data of the component to be detected based on the image segmentation processing logic.

[0064] The image segmentation processing logic in segmentation unit 21 is as follows:

[0065] Based on the background of the image data of the component to be detected, a threshold for determining the target pixel value of pixel segmentation is set. The target pixel value of pixel segmentation is applied to perform a traversal operation in the image data of the component to be detected, and pixels that meet the target pixel value of pixel segmentation are picked up. The picked pixels are used as the target pixels for segmentation and the segmentation operation is performed in the image data of the component to be detected to obtain the image of the component to be detected.

[0066] The edge pixels of the image of the component to be detected are captured, and each edge pixel is used as a retrieval target to retrieve the pixel with the highest similarity in the image of the component to be detected.

[0067] The search result is copied to a pixel, and the corresponding search target is iterated with the copied pixel. After all the edge pixels of the image of the component to be detected have been iterated, the output operation of the image of the component to be detected is performed.

[0068] ;

[0069] In the formula: For pixel brightness, saturation, and hue; The value of the pixel in the three color channels; These represent the maximum and minimum values ​​of the pixel in the three color channels, respectively.

[0070] Calculate each pixel based on the above formula. Then the similarity between the two pixels is:

[0071] ;

[0072] In the formula: To adjust the index;

[0073] Among them, the adjustment index It is 1 or -1. The numerator of the fraction is less than or equal to the denominator. =1, The numerator of the fraction is greater than the denominator. =-1;

[0074] The above logical formula is used to process the image of the component to be detected, thereby improving the accuracy of the image and making the system in this embodiment more accurate in judging the quality of the component.

[0075] Solid modeling module 3 is used to obtain the structural parameters of the component to be detected from the image of the component to be detected captured by the capture module 2, and to create a three-dimensional model of the component to be detected based on the structural parameters of the component to be detected.

[0076] Reference point picking module 4 is used to receive the standard 3D model of the component built by the construction module 1 and the 3D model of the component to be inspected created by the solid modeling module 3, and pick reference points on the two 3D models.

[0077] When the reference point picking module 4 picks reference points on two 3D models, it follows the following rules:

[0078] Identify the connection surface between the component head and the screw part on the component head in the 3D model, and use the corner point of the connection surface as a reference point;

[0079] A sub-module is set between the reference point picking module 4 and the demonstration module 5, including:

[0080] The connection unit 41 is used to connect all reference points on the three-dimensional model of the component to be tested to the thread start point on the three-dimensional model of the component to be tested, and to connect all reference points on the standard three-dimensional model of the component to the thread start point on the standard three-dimensional model of the component.

[0081] After the connection unit 41 finishes running, it obtains a set of three-dimensional models composed of multiple line segments in the three-dimensional model of the component to be detected and the standard three-dimensional model of the component, respectively. These are denoted as line model one and line model two, with one end intersecting at a point and the other end having different positions. Model one and line model two correspond to the three-dimensional model of the component to be detected and the standard three-dimensional model of the component, respectively.

[0082] Demonstration module 5 is used to receive reference points picked up in reference point picking module 4, determine the line models of two three-dimensional models based on the reference points, and control the line models to move at a constant speed along the thread in their corresponding three-dimensional models until they reach the end of the thread, so as to obtain a demonstration animation for thread quality inspection.

[0083] In Demonstration Module 5, the control logic for the control line model to move at a constant speed along the thread in its respective 3D model is as follows:

[0084] The intersection of each line segment in the line model is used as the moving end. The system and user terminals define the movement speed, so that the intersection ends move along the thread on the surface of the component based on the set movement speed. During the movement, the line segments in the line model are stretched and offset based on the movement of the intersection ends to form a demonstration animation.

[0085] Among them, when the system-side user defines the motion speed of the moving end of the line model, it follows the setting logic that the tighter the thread on the component surface and the higher the component's precision requirements, the slower the motion speed, and vice versa.

[0086] Analysis module 6 is used to receive demonstration animations from the standard 3D model of the component and the 3D model of the component to be tested. Based on the preset frame capture logic, it captures frames at the same time in the two demonstration animations. Based on the similarity evaluation of the frames captured at the same time in the two demonstration animations, it analyzes the degree of standardization of the size of the component to be tested.

[0087] The frame capture logic in analysis module 6 is represented as follows:

[0088] The duration of two demonstration animations is identified. The intersection of the durations of the two demonstration animations is used as the time domain for frame capture. The frame capture time interval is set. The frame at the beginning of the time domain is captured for the first time. Based on the timestamp of the first captured frame and the frame capture time interval, the frame is continuously captured in the two demonstration animations. Each captured frame is marked with its capture timestamp.

[0089] In the setting stage of the frame capture time interval, the setting logic is that the tighter the thread on the component surface and the higher the precision requirement of the component, the shorter the frame capture time interval, and vice versa.

[0090] Analysis module 6 has sub-modules, including:

[0091] The matching unit 61 is used to identify the capture timestamps marked on each captured frame in the analysis module 6, and to match the frames marked with the same capture timestamps with each other, which are then used as analysis targets by the analysis module 6.

[0092] The analysis logic for the degree of dimensional standardization of the component to be tested in analysis module 6 is expressed as follows:

[0093] ;

[0094] In the formula: Let the midline model morphology similarity between two images at the i-th capture timestamp; It is the set of endpoints of the corresponding corner points of each line segment in the line model; Let be the area of ​​the triangular region formed by the intersection of the i-th endpoint and the (i+1)-th endpoint and the line segments to which they belong in line model one; Let be the area of ​​the triangular region formed by the intersection of the i-th endpoint and the (i+1)-th endpoint and the line segments to which they belong in line model 2; The degree of standardization of the dimensions of the component to be inspected; A collection of timestamps captured from video frames;

[0095] in, Indicates to The operation of finding the mean. Similarly;

[0096] The above logical formula is used to calculate the degree of standardization of the dimensions of the component to be tested, providing data support for the operation of the judgment module.

[0097] The judgment module 7 is used to set the component qualification judgment threshold, and compare the consistency between the component qualification judgment threshold and the component size standardization degree obtained by the analysis module 6 to determine whether the component is qualified.

[0098] The sniffing module 8 is used to monitor the judgment result of whether the component is qualified in the judgment module 7. When the judgment result is negative, it sniffs the location of the thread defect on the surface of the component.

[0099] In the judgment module 7, the component qualification judgment threshold is defined by the system user. When the standardization degree of the component size to be inspected is greater than or equal to the component qualification judgment threshold, the component is judged as qualified; otherwise, the component is judged as unqualified.

[0100] During the 8th phase of the sniffing module's operation, the following information is obtained: The value, in Each item is compared with the component qualification threshold, and then the items that do not meet the component qualification threshold are captured. The position on the component surface thread defect is recorded as the position of the line model of the three-dimensional model of the component to be inspected, which is the intersection of the line model and the point of the line model to be inspected.

[0101] The construction module 1 interacts with the capture module 2 via a wireless network. The capture module 2's subordinates interact with the segmentation unit 21 via a wireless network. The capture module 2 interacts with the entity modeling module 3 via a wireless network. The entity modeling module 3 interacts with the segmentation unit 21 via a wireless network. The entity modeling module 3 interacts with the reference point picking module 4 and the demonstration module 5 via a wireless network. The reference point picking module 4 and the demonstration module 5 interact with the connection unit 41 via a wireless network. The demonstration module 5 interacts with the analysis module 6 via a wireless network. The analysis module 6's subordinates interact with the matching unit 61 via a wireless network. The analysis module 6 interacts with the judgment module 7 and the sniffing module 8 via a wireless network.

[0102] In this embodiment, the construction module 1 uploads the standard structural parameters of the component and uses these parameters to construct a standard 3D model of the component in a 3D scene. Simultaneously, the constructed standard 3D model is stored. The capture module 2, running subsequently, acquires image data of the component to be detected, segments this image data, and captures the image of the component to be detected within it. The segmentation unit 21 simultaneously receives the image data of the component to be detected acquired by the capture module 2, segments it, and outputs the image of the component to be detected. Then, the solid modeling module 3 obtains the structural parameters of the component to be detected from the image captured by the capture module 2 and creates a 3D model of the component to be detected based on these parameters. The reference point picking module 4 further receives the standard 3D model of the component constructed by the construction module 1 and the 3D model of the component to be detected created by the solid modeling module 3. Reference points are picked on both 3D models. The connection unit 41 simultaneously connects all reference points on the 3D model of the component to be detected to the thread start point on the 3D model of the component to be detected, and also connects all reference points on the standard 3D model of the component to the thread start point on the standard 3D model of the component. The demonstration module 5 receives reference points picked up by the reference point picking module 4, determines the line models of the two 3D models based on the reference points, and controls the line models to move at a constant speed along the thread in their respective 3D models until they reach the end of the thread, thus obtaining a demonstration animation for thread quality inspection. The analysis module 6 receives the demonstration animations from the standard 3D model of the component and the 3D model of the component to be inspected. Based on the preset frame capture logic, it captures frame frames at the same time in the two demonstration animations. Based on the similarity evaluation of the frame frames captured at the same time in the two demonstration animations, it analyzes the degree of standardization of the dimensions of the component to be inspected. The matching unit 61 identifies the capture timestamps marked on each captured frame in the analysis module 6 in real time, matches the frame frames marked with the same capture timestamps with each other, and uses them as analysis targets by the analysis module 6. Finally, the judgment module 7 sets the component qualification judgment threshold, compares the consistency between the component qualification judgment threshold and the degree of standardization of the dimensions of the component to be inspected obtained from the analysis module 6, and determines whether the component is qualified. Then, the sniffing module 8 monitors the judgment result of whether the component is qualified in the judgment module 7. When the judgment result is negative, it sniffs the location of thread defects on the surface of the component.

[0103] Through the system operation in the above embodiments, a higher precision, non-contact quality inspection technology is provided for bolts in precision fasteners, offering high-quality services for the quality control environment of precision bolt production.

[0104] See Figure 2As shown, based on the numbers 1, 2, 3, and 4 in the figure, the shape of the line model in the continuous frame is represented. In the actual application scenario of the above system, 1, 2, 3, and 4 are created for the standard three-dimensional model of the component and the three-dimensional model of the component to be tested, respectively. By comparing the corresponding two, the qualification of the component can be checked.

[0105] like Figure 1 As shown, during the operation phase of the solid modeling module 3, when the structural parameters of the component to be detected are acquired, the image of the component to be detected is placed in a two-dimensional coordinate system. Based on the two-dimensional coordinate system, the coordinates of each corner point on the image of each component to be detected are identified. The structural parameters of the component to be detected are determined by applying the coordinates of each corner point. Based on the structural parameters of the component to be detected, a three-dimensional model of the component to be detected is created, so that each thread and thread groove on the three-dimensional model of the component to be detected is constructed based on independent structural parameters.

[0106] The structural parameters of the top and bottom of the three-dimensional model of the component to be tested are derived from the images of the component to be tested in four directions, and the structural parameters of the thread and thread groove of the three-dimensional model of the component to be tested are derived from the images of the component to be tested in the front and back directions.

[0107] The above settings limit the construction logic of the 3D model of the component to be inspected, ensuring that the 3D model of the component to be inspected is stably constructed, and providing the necessary operational data support for the system operation in the above embodiments.

[0108] In summary, the system in the above embodiments constructs a high-precision reconstruction of the 3D model of the component to be inspected by acquiring multiple images and uploading the standard structural parameters of the component. Simultaneously, a standard 3D model of the component is constructed. The standard 3D model is then compared with the 3D model of the component to be inspected to provide effective quality judgment. Furthermore, during the comparison, feature points on the model surface are picked up, and line models are created for both the standard and 3D models. An animation of the line models is then generated, and frames are captured from the animation for comparison, maximizing the accuracy of the quality judgment of the component to be inspected. Additionally, after quality judgment, if the component fails the quality assessment, the system can detect the location of surface defects, providing data to assist production users in reverse-checking and identifying hidden defects in production equipment based on defect locations.

[0109] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An inspection system for precision fasteners, characterized by, The method comprises the following steps: A construction module (1) is configured to upload component standard structure parameters, apply the component standard structure parameters to construct a standard three-dimensional model of a component in a three-dimensional scene, and store the constructed component standard three-dimensional model; A capturing module (2) is configured to collect image data of a component to be detected, segment the image data of the component to be detected, and capture an image of the component to be detected in the image data of the component to be detected; An entity modeling module (3) is configured to obtain structure parameters of the component to be detected from the captured image of the component to be detected, and create a three-dimensional model of the component to be detected based on the structure parameters of the component to be detected; A reference point picking module (4) is configured to receive the component standard three-dimensional model constructed by the construction module (1) and the three-dimensional model of the component to be detected created by the entity modeling module (3), and pick reference points on the two three-dimensional models; A demonstration module (5) is configured to receive the reference points picked by the reference point picking module (4), determine line models of the two three-dimensional models based on the reference points, respectively, control the line models to move uniformly along threads in the corresponding three-dimensional models until the threads reach end points, and end the movement to obtain demonstration animations for thread quality inspection; An analysis module (6) is configured to receive the demonstration animations from the component standard three-dimensional model and the three-dimensional model of the component to be detected, capture frames at the same time in the two demonstration animations based on preset frame capturing logic, and analyze a standardization degree of a size of the component to be detected based on similarity evaluation of the frames captured at the same time in the two demonstration animations; A judgment module (7) is configured to set a component qualification judgment threshold, compare the component qualification judgment threshold with the standardization degree of the size of the component to be detected obtained by the analysis module (6), and determine whether the component is qualified; A sniffing module (8) is configured to monitor a judgment result of whether the component is qualified in the judgment module (7), and sniff a thread defect position on a surface of the component when the judgment result is no.

2. An inspection system for precision fasteners as defined in claim 1, wherein, When the construction module (1) stores the component standard three-dimensional model, a data packet is generated based on structure parameters applied in a construction stage of the component standard three-dimensional model, and the data packet is stored in association with the corresponding component standard three-dimensional model; When the capturing module (2) collects the image data of the component to be detected, the collection is performed from an upper, lower, front, or rear perspective, and a background is a curtain with a pure color preset by a system end user and different from a color of a surface of the component, and the capturing module is provided with a sub-module, which comprises: A segmentation unit (21) is configured to receive the image data of the component to be detected collected by the capturing module (2), perform segmentation processing on the image data of the component to be detected, and output an image of the component to be detected. The segmentation unit (21) is provided with image segmentation processing logic, and the segmentation unit (21) performs the segmentation processing operation on the image data of the component to be detected based on the image segmentation processing logic.

3. An inspection system for precision fasteners as defined in claim 2, wherein, The image segmentation processing logic in the segmentation unit (21) is as follows: A pixel segmentation target pixel value judgment threshold is set based on the background of the component image data to be detected, a traversal operation is performed in the component image data to be detected by applying the pixel segmentation target pixel value judgment threshold, pixels meeting the pixel segmentation target pixel value judgment threshold are picked up, a segmentation operation is performed in the component image data to be detected by taking the picked-up pixels as segmentation target pixels, and the component image to be detected is obtained; An edge pixel of the component image to be detected is captured, and each edge pixel is taken as a search target to search for a pixel with the highest similarity to the edge pixel in the component image to be detected; A search result pointing pixel is copied, and the copied pixel is iterated to its corresponding search target. After the edge pixels of the component image to be detected are all iterated, an output operation of the component image to be detected is performed; ； wherein: is the pixel lightness, saturation, hue; is the pixel value in the three color channels; is the maximum and minimum of the pixel values in the three color channels, respectively. Based on the above formula, the similarity of each pixel is calculated as Then the similarity of two pixels is: ； In the formula: is an adjustment index; wherein the adjustment index is 1 or -1, the numerator of the fraction is less than or equal to the denominator, = 1, the numerator of the fraction is greater than the denominator, = -1.

4. The inspection system for precision fasteners of claim 1, wherein, In the entity modeling module (3) running stage, when the component structure parameters to be detected are acquired, the component image to be detected is placed in a two-dimensional coordinate system, the coordinates of each corner point on each component image to be detected are recognized based on the two-dimensional coordinate system, the structure parameters of the component to be detected are determined by applying the coordinates of each corner point, and a three-dimensional model of the component to be detected is created based on the structure parameters of the component to be detected, so that each thread and thread groove on the three-dimensional model of the component to be detected is constructed based on independent structure parameters. The structure parameters of the top and bottom ends of the three-dimensional model of the component to be detected are derived from the component images to be detected in four directions, and the structure parameters of the threads and thread grooves of the three-dimensional model of the component to be detected are derived from the component images to be detected in the front and rear directions.

5. The inspection system for precision fasteners of claim 1, wherein, When the reference points are picked up on the two three-dimensional models by the reference point picking module (4), the following is followed: The connecting surface between the component head and the screw rod part on the component head is recognized in the three-dimensional model, and the corner point of the connecting surface is taken as a reference point. A submodule is arranged between the reference point picking module (4) and the demonstration module (5), including: A connecting unit (41) is arranged to connect each reference point on the three-dimensional model of the component to be detected with the start point of the thread on the three-dimensional model of the component to be detected, and to connect each reference point on the three-dimensional model of the component standard with the start point of the thread on the three-dimensional model of the component standard. After the connecting unit (41) is running, a set of three-dimensional models composed of multiple line segments with one end intersecting at a point and the other end having different end point positions is obtained in the three-dimensional model of the component to be detected and the three-dimensional model of the component standard, respectively, which are denoted as line model one and line model two, and the line model one and the line model two correspond to the three-dimensional model of the component to be detected and the three-dimensional model of the component standard, respectively.

6. The inspection system for precision fasteners of claim 1, wherein, The control logic for controlling the line model to move at a constant speed along the thread in the corresponding three-dimensional model in the demonstration module (5) is as follows: The intersection end of each line segment in the line model is taken as a moving end, the system end user defines a moving speed, the intersection end moves along the thread on the component surface based on the set moving speed, and stretching and offsetting occur based on the movement of each line segment in the line model based on the intersection end during the movement to form a demonstration animation. When the system end user defines the moving speed of the moving end of the line model, the following setting logic is followed: the more compact the thread on the component surface is, the higher the component precision requirement is, and the slower the moving speed is, and vice versa.

7. The inspection system for precision fasteners of claim 1, wherein, The picture frame grabbing logic in the analysis module (6) is represented as: The time lengths of the two demonstration animations are identified, the intersection of the two demonstration animation time lengths is taken as a time domain for frame grabbing, a frame grabbing time interval is set, a first frame is grabbed at the beginning of the time domain, and frames are continuously grabbed in the two demonstration animations based on the time stamp corresponding to the first grabbed frame and the frame grabbing time interval, and each grabbed frame is marked with a grabbing time stamp; The setting of the frame grabbing time interval is subject to the following setting logic: the more compact the surface thread of the component is and the higher the precision requirement of the component is, the shorter the frame grabbing time interval is, and vice versa.

8. The inspection system for precision fasteners of claim 1, wherein, The analysis module (6) is subordinated to a sub-module, including: A matching unit (61) is configured to identify the grabbing time stamps marked on the frames grabbed in the analysis module (6), match the frames marked with the same grabbing time stamps with each other, and use the matched frames as analysis targets in the analysis module (6); The analysis logic of the standardization degree of the component size to be detected in the analysis module (6) is represented as: ； In the formula: is the shape similarity of the line model in two picture frames under the ith grabbing timestamp; is the set of endpoints of the corresponding corner points of each line segment in the line model; is the area of the triangular region composed of the intersection of the ith endpoint and the ith+1 endpoint in the line model one and the line segments to which the two endpoints belong; is the area of the triangular region composed of the intersection of the ith endpoint and the ith+1 endpoint in the line model two and the line segments to which the two endpoints belong; is the size standardization degree of the component to be detected; is the set of picture frame grabbing timestamps; wherein denotes a mean operation over same applies.​ 9. The inspection system for precision fasteners of claim 8, wherein, The component qualification threshold in the determination module (7) is defined by a system user, and when the standardization degree of the component size to be detected is greater than or equal to the component qualification threshold, the component is determined to be qualified, and vice versa, the component is determined to be unqualified; The sniffing module (8) runs a phase, respectively acquires The values of Each item is compared with the component qualification threshold value, and the items that do not meet the component qualification threshold value are captured. The positions of the non-compliant items on the surface of the three-dimensional model of the component to be detected are recorded as the positions of the thread defects on the surface of the component.

10. The inspection system for precision fasteners of claim 1, wherein, The construction module (1) is connected to the capturing module (2) through a wireless network, the capturing module (2) is connected to the segmentation unit (21) through a wireless network, the capturing module (2) is connected to the entity modeling module (3) through a wireless network, the entity modeling module (3) is connected to the segmentation unit (21) through a wireless network, the entity modeling module (3) is connected to the reference point picking module (4) and the demonstration module (5) through a wireless network, the reference point picking module (4) and the demonstration module (5) are connected to the connection unit (41) through a wireless network, the demonstration module (5) is connected to the analysis module (6) through a wireless network, the analysis module (6) is connected to the matching unit (61) through a wireless network, and the analysis module (6) is connected to the determination module (7) and the sniffing module (8) through a wireless network.

Images