Automated Screening Method and System for Automotive Fastener Generation

Abstract

This invention relates to the field of automatic sorting technology, specifically to an automated screening method and system for automotive fasteners, comprising the following steps: determining end face orientation through bi-angle reflection, detecting outer diameter elliptical offset using radial radius difference, determining thread flattening by combining multi-angle valley reflection attenuation sequence, determining scratch direction based on surface multi-angle peak offset trajectory, and configuring rejection trajectory by combining conveying timing and spray compensation. In this invention, end face orientation is classified and directional identifiers are established through bi-angle reflection differences, elliptical offset features are extracted using radial radius difference distribution, flattened sections are located using multi-incident angle valley reflection attenuation sequence, surface reflection peaks form continuous directional trajectories in the angle sequence and scratch coordinates are output, defect judgment and motion time calculation are linked, timing compensation is performed by combining conveying speed and nozzle lag, and multi-level pressure trajectories are matched according to quality grouping, enhancing posture adaptability and detectability of minor defects, improving the consistency of full-circumference detection and the stability of classification and rejection.

Description

Technical Field

[0001] This invention relates to the field of automatic sorting technology, and in particular to an automated screening method and system for automotive fasteners. Background Technology

[0002] The field of automatic sorting technology involves technologies for the automatic identification and classification of various solid materials or components. Core aspects include methods for detecting the physical characteristics of the target object, such as size, shape, color, and material; the structural design and working mechanism of the sorting device; and the coordination and optimization of the sorting process by the control system. Applications include industrial manufacturing, agricultural processing, and resource recycling, aiming to improve sorting efficiency, reduce labor costs, and enhance product quality. Specifically, this involves using visual recognition systems for image acquisition and processing, using photoelectric sensors for positioning and feature judgment, employing pneumatic or mechanical structures to complete the sorting action, and implementing the sorting through programmable logic controllers or industrial computers. The logic control and execution management of the process, in particular, the automated screening method for traditional automotive fastener generation refers to the specific process of screening the appearance quality and dimensional consistency of fasteners such as bolts, nuts, and washers after production. It usually uses optical lenses and industrial cameras to collect multi-angle images of fasteners during the transmission process, and uses image processing algorithms to identify undesirable features such as surface defects, dimensional deviations, and thread damage. Based on the detection results, a rejection device controlled by a solenoid valve is used to automatically remove unqualified parts from the conveyor track. Combined with physical means such as drum screening, laser ranging, and weighing sensors, the size and weight are quickly judged to achieve the initial screening of fasteners.

[0003] Existing technologies rely on multi-angle images and comparison with conventional dimensions for screening. They are sensitive to changes in fastener posture, and the outer diameter and thread area are prone to deviation when the end face orientation is unstable, leading to fluctuations in judgment. They are weak in responding to small elliptical eccentricity and thread valley flattening defects. Surface scratches have insufficient contrast under single imaging conditions and are easily covered by noise. The classification threshold is often based on empirical ranges and requires frequent adjustment. Fixed delay and single-stage pressure are commonly used in the rejection process. When the conveying speed drifts or the nozzle lags, the hit deviation increases, resulting in missed rejections and false rejections, and the consistency of sorting is limited. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide an automated screening method for generating automotive fasteners, comprising the following steps:

[0005] To achieve the above objectives, the present invention adopts the following technical solution: an automated screening method for generating automotive fasteners, comprising the following steps:

[0006] S1: Call the end face double angle reflection data group, analyze the reflection data of the fastener end face under two angles of incident light, calculate the difference between the reflection peak intensity and valley reflection signal under the two incident angles, determine the orientation category according to the classification range set by the end face type, and generate end face orientation judgment information.

[0007] S2: Based on the end face orientation judgment information, calculate the intersection points of multiple radial sections and fastener contours within the outer diameter region, determine the existence of offset based on the difference in radius of adjacent sections, identify the circumferential offset direction, classify the offset trend, and generate an outer diameter elliptical offset index.

[0008] S3: Based on the outer diameter ellipse offset index, call the thread valley bottom reflection data group, analyze the reflection intensity change under multiple incident angles, calculate the reflection signal attenuation difference and generate an attenuation difference sequence, determine the continuity of attenuation change, identify the flattened area, and generate thread flattening judgment result.

[0009] S4: Based on the thread flattening determination result, call the fastener surface multi-angle reflection curve data group, analyze the distribution difference of reflection peaks in spatial coordinates, calculate the reflection peak offset, determine the continuity of the offset sequence, filter the offset trajectory direction, and generate surface scratch direction coordinate information.

[0010] S5: Based on the surface scratch direction coordinate information, calculate the movement time of the defective fastener from the detection position to the spraying position, use the conveying speed and nozzle response lag for offset correction, match the spraying control signal through quality grouping, and perform classification matching by combining the spraying pressure in the pushing and correction stages to generate the spraying trajectory control configuration.

[0011] As a further embodiment of the present invention, the end face orientation judgment information includes orientation category information, direction status identification information, and diversion channel control information; the outer diameter elliptic offset index specifically includes outer diameter ellipticity parameter, circumferential eccentricity parameter, and offset direction characteristic parameter; the thread flattening judgment result includes flattening presence mark, flattening area range, and flattening severity level; the surface scratch direction coordinate information specifically includes scratch direction vector information, scratch start and end coordinate information, and scratch trajectory set information; and the spray trajectory control configuration includes spray start timing parameter, spray pressure combination parameter, and spray trajectory position parameter.

[0012] As a further aspect of the present invention, the step of obtaining the end face orientation determination information specifically includes:

[0013] S101: Obtain the double-angle reflection data set of the end face, extract the reflection data formed by the end face under the first incident angle and the reflection data formed under the second incident angle, calculate the numerical difference between the corresponding reflection peak intensity and reflection valley signal, and generate the reflection peak valley difference value.

[0014] S102: Based on the difference value of the reflection peak and valley, compare it with the classification range set by the end face type, make a judgment on the orientation category, form the classification information of the nut end face direction, and obtain the end face category range information;

[0015] S103: Based on the end face category range information, establish a direction status identifier for each nut and output the control information of the corresponding diversion channel to obtain end face orientation judgment information.

[0016] As a further aspect of the present invention, the step of obtaining the outer diameter ellipse offset index specifically includes:

[0017] S201: Based on the end face orientation judgment information, calculate the coordinates of the intersection points of multiple radial sections and the fastener contour within the outer diameter detection area, and extract the radius data corresponding to each radial section to generate a radial radius data group;

[0018] S202: Based on the radial radius data set, perform difference calculations on the radius values ​​between each pair of adjacent cross-sections, and combine the circumferential arrangement order of the cross-sections to identify the offset direction in the circumferential order, thereby obtaining the circumferential offset distribution parameters.

[0019] S203: Call the circumferential offset distribution parameters, classify and analyze the offset trend of the sectional line, perform stability merging analysis on the offset distribution structure within the entire circumference, and obtain the outer diameter ellipse offset index.

[0020] As a further aspect of the present invention, the step of obtaining the thread flattening determination result is specifically as follows:

[0021] S301: Based on the outer diameter ellipse offset index, obtain the fastener thread valley bottom multi-angle reflection data group, and calculate the intensity attenuation value between adjacent incident angle reflection signals based on the thread valley bottom reflection intensity under multiple incident angles to obtain the thread reflection attenuation difference sequence.

[0022] S302: Call the thread reflection attenuation difference sequence, make a continuous judgment on the attenuation change trend of each point in the sequence, analyze the fluctuation concentration characteristics of the reflection difference in multiple incident angles, and obtain the thread attenuation fluctuation range parameters.

[0023] S303: Based on the thread attenuation fluctuation range parameters and combined with the range segment of fluctuation concentration, identify the spatial distribution of thread valley flattening, establish the regional characteristics and grade parameters of thread flattening, and obtain the thread flattening judgment result.

[0024] As a further aspect of the present invention, the step of obtaining the surface scratch direction coordinate information specifically includes:

[0025] S401: Based on the thread flattening determination result, call the fastener surface multi-angle reflection curve data group, extract the distribution position of reflection peaks in spatial coordinates under multiple illumination angles, calculate the spatial coordinate offset value between adjacent angle reflection peaks, and obtain the peak position coordinate offset sequence.

[0026] S402: Call the peak position coordinate offset sequence, determine the continuous extension characteristics of the peak position in the spatial direction, filter the offset trajectory structure according to the direction consistency, and generate a set of scratch offset trajectories.

[0027] S403: Based on the set of scratch offset trajectories, merge the direction vector parameters of the offset trajectories, integrate the scratch direction vector and start and end coordinate information, and obtain the surface scratch direction coordinate information.

[0028] As a further aspect of the present invention, the process of selecting the offset trajectory structure based on directional consistency specifically includes:

[0029] Based on the peak position offset sequence, calculate the spatial coordinate difference between adjacent angular reflection peaks, obtain the coordinate offset vector direction angle corresponding to each spatial coordinate difference, and arrange them in angular order to form a direction angle sequence.

[0030] The absolute difference between adjacent direction angles in the direction angle sequence is used to form a direction angle difference sequence, and the 90th decimal place of the direction angle difference sequence is used to set the direction consistency threshold.

[0031] Elements in the direction angle difference sequence that are less than or equal to the direction consistency threshold and whose angle indices are consecutive are grouped into offset trajectory candidate units, and the offset trajectory structure is obtained by filtering the offset trajectory candidate units with no less than three elements.

[0032] The trajectory direction vector information is determined by the average direction angle of all coordinate offset vectors within the offset trajectory structure, and the index segment corresponding to the trajectory direction vector information is output as the scratch offset trajectory set.

[0033] As a further aspect of the present invention, the step of obtaining the injection trajectory control configuration specifically includes:

[0034] S501: Based on the surface scratch direction coordinate information, combined with the spatial coordinates of the detection position and the spraying position, and using the conveying speed data and nozzle response hysteresis, perform timing calculation and offset correction on the movement path of the fastener to obtain the fastener movement time parameters.

[0035] S502: Call the fastener movement time parameters, match the injection start control signal according to the mass grouping parameters, and combine the injection pressure levels of the pushing stage and the correction stage to obtain the injection control combination parameters;

[0036] S503: Based on the injection control combination parameters, integrate motion state data and injection behavior characteristics, configure multiple timing and pressure parameters of the injection trajectory, and obtain the injection trajectory control configuration.

[0037] An automated screening system for automotive fastener generation includes:

[0038] The end face orientation recognition module calls the end face double angle reflection data group, analyzes the reflection data of the fastener end face under two angles of incident light, calculates the difference between the reflection peak intensity and valley reflection signal under the two incident angles, judges the orientation category according to the classification range set by the end face type, and generates end face orientation judgment information.

[0039] The outer diameter offset evaluation module calculates the intersection points of multiple radial sections and fastener contours within the outer diameter region based on the end face orientation judgment information, determines the existence of offset based on the difference in radius of adjacent sections, identifies the circumferential offset direction, classifies the offset trend, and generates an outer diameter elliptical offset index.

[0040] The thread flattening detection module calls the thread valley bottom reflection data group according to the outer diameter ellipse offset index, analyzes the reflection intensity change under multiple incident angles, calculates the reflection signal attenuation difference and generates an attenuation difference sequence, judges the continuity of attenuation change, identifies the flattening area, and generates thread flattening judgment result.

[0041] The surface scratch positioning module calls the fastener surface multi-angle reflection curve data group according to the thread flattening judgment result, analyzes the distribution difference of reflection peaks in spatial coordinates, calculates the reflection peak offset, judges the continuity of the offset sequence, filters the offset trajectory direction, and generates surface scratch direction coordinate information.

[0042] The spray sorting control module calculates the movement time of the defective fastener from the detection position to the spray position based on the surface scratch direction coordinate information, uses the conveying speed and nozzle response lag for offset correction, matches the spray control signal by quality grouping, and performs classification matching by combining the spray pressure in the pushing and correction stages to generate the spray trajectory control configuration.

[0043] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0044] In this invention, the orientation of the end face is classified and a direction identifier is established by the difference in reflection at two angles. The radial radius difference distribution is used to extract the elliptical offset feature. The valley reflection attenuation sequence of multiple incident angles is used to locate the flattened section. The surface reflection peak position forms a continuous directional trajectory in the angle sequence and outputs the scratch coordinates. The judgment of each defect is linked with the calculation of the movement time. The timing compensation is combined with the conveying speed and nozzle lag and multi-level pressure trajectory is matched according to the quality group. This enhances the attitude adaptability and the detectability of minor defects, and improves the consistency of full-circumference detection and the stability of classification and rejection. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is a schematic diagram of the steps of the present invention;

[0047] Figure 2 This is a detailed schematic diagram of S1 of the present invention;

[0048] Figure 3 This is a detailed schematic diagram of S2 of the present invention;

[0049] Figure 4 This is a detailed schematic diagram of S3 of the present invention;

[0050] Figure 5 This is a detailed schematic diagram of S4 of the present invention;

[0051] Figure 6 This is a detailed schematic diagram of S5 of the present invention;

[0052] Figure 7 This is a system module diagram of the present invention. Detailed Implementation

[0053] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0054] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0055] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent.

[0056] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0057] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0058] Please see Figure 1 This invention provides an automated screening method for generating automotive fasteners, comprising the following steps:

[0059] S1: Call the end face double angle reflection data group, analyze the reflection data of the fastener end face under two angles of incident light, calculate the difference between the reflection peak intensity and valley reflection signal under the two incident angles, determine the orientation category according to the classification range set by the end face type, and generate end face orientation judgment information.

[0060] S2: Based on the end face orientation information, calculate the intersection of multiple radial sections and fastener contours within the outer diameter region, determine the existence of offset based on the difference in radius of adjacent sections, identify the circumferential offset direction, classify the offset trend, and generate the outer diameter elliptical offset index.

[0061] S3: Based on the outer diameter ellipse offset index, call the thread valley bottom reflection data group, analyze the reflection intensity change under multiple incident angles, calculate the reflection signal attenuation difference and generate an attenuation difference sequence, determine the continuity of attenuation change, identify the flattened area, and generate thread flattening judgment result.

[0062] S4: Based on the thread flattening judgment result, call the fastener surface multi-angle reflection curve data group, analyze the distribution difference of reflection peaks in spatial coordinates, calculate the reflection peak offset, judge the continuity of the offset sequence, filter the offset trajectory direction, and generate surface scratch direction coordinate information.

[0063] S5: Based on the coordinate information of the surface scratch direction, calculate the movement time of the defective fastener from the detection position to the spraying position, use the conveying speed and nozzle response lag to correct the offset, match the spraying control signal through quality grouping, and combine the spraying pressure in the pushing and correction stages to classify and match, and generate the spraying trajectory control configuration.

[0064] The end face orientation judgment information includes orientation category information, direction status identification information, and diversion channel control information. The outer diameter elliptic offset index specifically includes the outer diameter ellipticity parameter, circumferential eccentricity parameter, and offset direction characteristic parameter. The thread flattening judgment result includes the flattening presence mark, flattening area range, and flattening severity level. The surface scratch direction coordinate information specifically includes scratch direction vector information, scratch start and end coordinate information, and scratch trajectory set information. The injection trajectory control configuration includes injection start timing parameters, injection pressure combination parameters, and injection trajectory position parameters.

[0065] Please see Figure 2 The specific steps for obtaining end face orientation information are as follows:

[0066] S101: Obtain the double-angle reflection data set of the end face, extract the reflection data formed by the end face under the first incident angle and the reflection data formed under the second incident angle, calculate the numerical difference between the corresponding reflection peak intensity and reflection valley signal, and generate the reflection peak valley difference value.

[0067] At the image acquisition station of the automated inspection production line, the system controls an industrial camera to capture images of M12 flange nuts on the conveyor belt. First, a first ring-shaped LED light source is illuminated at a first incident angle of 30 degrees, and the camera exposure time is set to 150 microseconds to acquire the first grayscale image. Then, after a 5-millisecond interval, a second ring-shaped LED light source is illuminated at a second incident angle of 60 degrees, and the camera acquires the second grayscale image with the same exposure parameters, thus obtaining the end-face bi-angle reflection data set. The system locks the circular region of interest (ROI) containing the fastener end face in the image and extracts the grayscale values ​​of all pixels within that region. For the first incident angle image, the average intensity of the bright pixels with a grayscale value greater than 220 within the ROI is calculated and denoted as the reflection peak intensity. The average intensity of dark pixels with gray values ​​less than 50 is calculated and denoted as the reflection valley signal. Similarly, for the image at the second incident angle, the reflection peak intensity is obtained. With reflection valley signal The system performs numerical difference calculations to determine the contrast at the first angle. and contrast at the second angle Based on this, the contrast difference between the two angles is used as the final reflection peak-valley difference value. That is, to perform operations For example, when inspecting a specific nut, the measured value is... , ,but ; measured , ,but The final generated reflection peak-valley difference value This value quantifies the difference in response of the end face morphology to light at different angles. For example, the chamfered structure on the flange face will produce significant highlights under low-angle light, while the reflection will be weakened at high angles. The planar structure is the opposite. This difference value directly reflects the geometric characteristics of the end face.

[0068] S102: Based on the difference between reflection peaks and valleys, compare the classification range set by the end face type, determine the orientation category, form the classification information of the nut end face direction, and obtain the end face category range information;

[0069] The system pre-sets a classification range for end face types. This range is based on statistical analysis of experimental data from 1000 standard qualified samples (including both upright and reverse orientations). The experimentally measured mean of the reflection peak-valley difference in the "upright (flange face up)" state was 105, with a standard deviation of 5. Based on the principle of three times the standard deviation, the classification range for upright orientation is set as follows: The sample mean for the "reverse placement (welded side up)" state was -40, and the standard deviation was 8. The classification interval for reverse placement was set as follows: The difference between reflection peaks and valleys calculated in system call S101. The value is then compared with the aforementioned preset interval. The judgment logic is as follows: If... If it is, then it is determined to be the "upright placement category"; if If the value falls within any of the above ranges, it is classified as "reverse playback category"; if the value does not fall within any of the above ranges, it is classified as "unknown undetermined category". In this embodiment, Falling Within the specified range, the system determines the end face orientation of the nut to be "upright". This step, through strict matching of numerical ranges, completes the digital classification of physical orientation, forming clear end face category range information. This provides a basic logical basis for subsequent defect detection targeting only specific faces, avoiding invalid calculations on non-detection faces.

[0070] S103: Based on the end face category range information, establish a direction status identifier for each nut and output the control information of the corresponding diversion channel to obtain end face orientation judgment information;

[0071] Based on the end-face category range information, the system establishes a unique orientation status identifier ID for the currently detected nut, assigning it the value "DIR_UP_01". This identifier follows the nut's data flow to subsequent processing units. Simultaneously, the system queries the diversion channel control logic table: identifier "DIR_UP_01" corresponds to "Main Detection Channel A", identifier "DIR_DOWN_02" corresponds to "Flip Channel B", and identifier "DIR_UNKNOWN" corresponds to "Rejection Channel C". Since the current identifier is "DIR_UP_01", the system generates an enable signal for Main Detection Channel A. This signal is a 32-bit binary instruction, sent to the diversion PLC via industrial Ethernet. After parsing the instruction, the PLC controls the pneumatic lever to remain stationary, allowing the nut to continue its linear movement into the subsequent visual inspection area. At this point, the system outputs end-face orientation judgment information containing physical position attributes (upright) and logical control attributes (channel A). This process ensures that all fasteners entering the next stage of high-precision testing maintain a uniform test posture, eliminating the risk of algorithm misjudgment caused by posture randomness.

[0072] Please see Figure 3The specific steps for obtaining the outer diameter ellipse offset index are as follows:

[0073] S201: Based on the end face orientation information, calculate the coordinates of the intersection points of multiple radial sections and the fastener profile within the outer diameter detection area, and extract the radius data corresponding to each radial section to generate a radial radius data set;

[0074] After confirming that the fastener is in a detectable posture based on the end face orientation information, the system establishes a polar coordinate system in the image with the centroid of the fastener as the origin. The system then... to Within the outer diameter detection area, with 360 virtual radial sections are generated at intervals. For each radial section, the system uses a sub-pixel edge extraction algorithm to scan the gray-level gradient changes radiating outward from the center, locating the point of maximum gradient magnitude as the intersection of the fastener contour and the section. The system reads the pixel coordinates of this intersection point. And based on the camera calibration parameters (e.g., 1 pixel = 0.05 mm), calculate the Euclidean distance from the point to the center of the circle, i.e., the radius data. The system iterates through all 360 intercepts, storing the obtained radius values ​​in angular order into a one-dimensional array to generate a radial radius dataset. .

[0075] Table 1 Sampling table of radial radius data sets

[0076]

[0077] As shown in Table 1, the radius measurement data at some angles are listed (only a portion is shown for illustration). This data set fully describes the geometric dimension distribution of the fastener's outer perimeter.

[0078] S202: Based on the radial radius data set, perform difference calculations on the radius values ​​between each pair of adjacent cross-sections, and combine the circumferential arrangement order of the cross-sections to identify the offset direction in the circumferential order and obtain the circumferential offset distribution parameters.

[0079] Call the radial radius data set and perform the difference operation between adjacent intercepts. Set the index. Iterate from 1 to 359, and calculate... And calculate the difference between the first and last closed loops. Based on the data example in Table 1, calculate... mm, mm. The system identifies the offset direction based on the sign of the difference: a positive value indicates outward expansion of the contour, and a negative value indicates inward contraction. The system will... The numerical sequence is mapped to the direction sequence ,in Marked as +1, Marked as -1, and marked as 0 if the absolute value is less than or equal to 0.001. For example, if a sequence appears consecutively... This indicates that there is a local deformation in the region that first bulges and then sinks. This step, through difference and symbolization processing, eliminates the interference of absolute size and focuses on the relative rate of change of the contour shape, obtaining the circumferential offset distribution parameters, and accurately capturing the tiny shape abrupt changes or trend changes in the circumferential sequence.

[0080] S203: Call the circumferential offset distribution parameters, classify and analyze the offset trend of the sectional line, perform stability merging analysis on the offset distribution structure within the entire circumference, and obtain the outer diameter ellipse offset index.

[0081] The system calls upon the circumferential offset distribution parameters to categorize the offset trends across the entire circumference. The system sets the sliding window size to 10 data points, within the direction sequence. Move upwards to count the percentage of same-direction symbols (all +1 or all -1) within the window. If the percentage of same-direction symbols exceeds 80% in a continuous segment and the duration exceeds [a certain value], [further action is taken]. If the condition is not met, the segment is determined to be a "monotonically deformable zone". The system counts the number and distribution interval of "monotonically deformable zones" throughout the entire circumference. For a standard circle, there should be no significant monotonically deformable zones throughout the entire circumference; for an ellipse, the entire circumference should exhibit an alternating structure of "expansion-contraction-expansion-contraction", i.e., there should be four main deformable zones. The system calculates the standard deviation of the entire circumference radius data. And combined with the number of deformation zones Perform stability merge analysis. Define the decision logic: if... and If the value is in mm, then an elliptic trend is determined. For example, the standard deviation of 360 points over the entire circumference is calculated. mm, and identified two expansion regions and two contraction regions symmetrically distributed. The system then calculates the standard deviation value. The output is an index of outer diameter ellipse offset. This index is not simply the difference between the maximum and minimum diameters, but a statistical measure that incorporates the circumferential distribution pattern, which can effectively distinguish between local burrs (high-frequency noise) and overall elliptical deformation (low-frequency structural deviation).

[0082] Please see Figure 4 The specific steps for obtaining the thread flattening determination result are as follows:

[0083] S301: Based on the outer diameter ellipse offset index, obtain the multi-angle reflection data set of the fastener thread valley bottom, calculate the intensity attenuation value between the reflection signals of adjacent incident angles based on the reflection intensity of the thread valley bottom under multiple incident angles, and obtain the thread reflection attenuation difference sequence.

[0084] The system receives the outer diameter elliptic offset index. If this index is less than 0.1mm (indicating that the outer diameter positioning datum is reliable), the thread inspection program is initiated. The system controls multi-angle light sources to sequentially... The spiral valley area was illuminated from five incident angles, and the camera captured five frames for each frame. The system located the same pixel position at the bottom of the spiral valley in each frame, extracted the reflected grayscale intensity at that position, and constructed an intensity array. Perform attenuation calculations for adjacent angles, using the following formula: ,in Assume the collected intensity data is The calculated decay sequence is: , , , This sequence reveals the extent to which light is blocked by the thread crest as the angle of incidence changes. A normal thread structure produces a smooth or regular attenuation, while flattened areas cause reflected light to abruptly change or remain bright at certain angles, thus manifesting as numerical anomalies in the attenuation sequence.

[0085] S302: Call the thread reflection attenuation difference sequence, make a continuous judgment on the attenuation change trend of each point in the sequence, analyze the fluctuation concentration characteristics of the reflection difference in multiple incident angles, and obtain the thread attenuation fluctuation range parameters.

[0086] Call the thread reflection attenuation difference sequence The continuity characteristics are analyzed. The system calculates the first difference of the sequence. And calculate the sequence variance The fluctuation threshold is set to 50. If any element exists in the sequence... If the deviation from the mean exceeds a threshold, or the maximum absolute value of the first difference exceeds a threshold, then concentrated fluctuations are considered to exist. In the above example, the maximum value of the first difference, 80, exceeds the threshold of 50, indicating that... to A dramatic reflection abrupt change occurred during the lighting switching (the light was suddenly reflected by the flattened metal surface, rather than being blocked by the tooth profile). The system records the index position of the abrupt change (i.e., the 3rd attenuation interval) and calculates the span of this fluctuation within the multi-angle reflection group. If the attenuation value of multiple consecutive angles (e.g., more than 3) is close to 0 (i.e., the intensity does not change and remains bright), it indicates that the area is a planar reflection. The system encapsulates the identified fluctuation amplitude (80), the starting angle index (3) of the fluctuation, and the duration (1) into a threaded attenuation fluctuation interval parameter, quantifying the instability of the reflection characteristics in the angular domain.

[0087] S303: Based on the thread attenuation fluctuation range parameter and combined with the range segment of fluctuation concentration, identify the spatial distribution of thread valley flattening, establish the regional characteristics and grade parameters of thread flattening, and obtain the thread flattening judgment result.

[0088] Based on the thread attenuation fluctuation range parameters, the system identifies the flattening region by combining spatial mapping relationships. It is known that the fluctuation occurs... to Within the incident angle range, the system maps this angle range to the physical depth position of the thread profile. If the fluctuation exhibits "high reflection intensity with no attenuation," it is determined that the tooth crest has been flattened and filled into the valley; if it exhibits "abnormal dark area at a specific angle," it is determined that foreign object obstruction has occurred. System setting level parameter table: Fluctuation amplitude For being without defects, It is a slight indentation. This indicates severe flattening. In this example, the fluctuation amplitude is 80, falling within the severe flattening range. The system further examines the area where the fluctuation is concentrated; if this abnormal feature extends continuously in the thread circumference direction for more than [a certain period], [further action will be taken]. (Obtained through image circumferential unfolding analysis), thus confirming the presence of physical damage. The system ultimately generates a thread flattening determination result of "NG_CRITICAL_FLAT", along with the defect center coordinates and coverage area parameters. This result not only qualitatively confirms the existence of the defect but also quantitatively determines its level, providing a basis for subsequent classification and removal decisions.

[0089] Please see Figure 5 The specific steps for obtaining the surface scratch direction coordinate information are as follows:

[0090] S401: Based on the thread flattening determination result, call the fastener surface multi-angle reflection curve data group, extract the spatial coordinate distribution position of the reflection peaks under multiple illumination angles, calculate the spatial coordinate offset value between adjacent angle reflection peaks, and obtain the peak position coordinate offset sequence.

[0091] Based on the "NG_CRITICAL_FLAT" determination result, the surface finish of the fastener is inspected. (Call...) From illumination images in four directions, extract the spatial coordinates of surface reflection peaks (centroids of connected components with grayscale values ​​greater than 230) at each angle. ,in The coordinates are obtained by photometric stereo method; if only a two-dimensional image is available, then... The system calculates the Euclidean distance shift of the reflection peak under adjacent illumination angles. Assume... Peak position under illumination , Peak position under illumination Then the coordinate offset value Pixels. The system sequentially calculates the offset of all adjacent angles to generate a peak position coordinate offset sequence. For smooth surfaces, as the illumination angle changes, the highlight point will undergo a smooth, large displacement due to the surface curvature; however, for scratches (V-groove structures), the highlight point is often confined within the scratch groove, with a smaller displacement and directional movement. This sequence quantifies the dynamic characteristics of the highlight point's movement with illumination.

[0092] S402: Call the peak position coordinate offset sequence, determine the continuous extension characteristics of the peak position in the spatial direction, filter the offset trajectory structure according to the consistency of direction, and generate a set of scratch offset trajectories.

[0093] The peak position coordinate offset sequence is invoked to perform directional consistency filtering. First, the coordinate difference between adjacent angular reflection peaks is calculated. Find the direction angle of the offset vector. For example, the resulting direction angle sequence is: Calculate the sequence of absolute differences between adjacent direction angles. The system calculates the 90th tenth of the difference sequence. Assume the result is... Set the directional consistency threshold to The system iterates through the difference sequence, filtering for values ​​less than or equal to... And elements with consecutive indices. The first two items of the sequence. The conditions are met, and the corresponding original angle indices are 1, 2, 3, with a number of elements of 3 (not less than 3), satisfying the filtering criteria. This indicates that under the first three illumination angles, the reflection peaks are along the same direction (approximately...). The system moves along the optical characteristics of linear scratches. It merges this set of continuous offset vectors into an offset trajectory structure, generating a scratch offset trajectory set, thus eliminating the random jumps (noise) that occur at index 4.

[0094] Table 2. Offset Trajectory Direction Angle Analysis Data Table

[0095]

[0096] See Table 2 for the data on the direction angle sequence processing.

[0097] S403: Based on the scratch offset trajectory set, merge the direction vector parameters of the offset trajectory, integrate the scratch direction vector and start and end coordinate information, and obtain the surface scratch direction coordinate information;

[0098] Based on the set of scratch offset trajectories, the system extracts the direction angles of all coordinate offset vectors that are determined to be continuous trajectories. Calculate its average value This determines the direction of the scratch's extension. Simultaneously, the system extracts the starting pixel coordinates corresponding to this trajectory segment. and termination pixel coordinates The pixel coordinates are converted to physical coordinates (millimeters) by combining the camera calibration coefficients. The system integrates the above information to generate the surface scratch direction coordinate information as follows: {Type: Linear scratch, Direction vector:} The starting point is (2.1mm, 4.0mm), the ending point is (2.24mm, 4.2mm), and the length is 0.24mm. This result indicates the presence of a tiny linear scratch on the fastener surface, clearly defining its spatial location and orientation. This high-dimensional geometric description is far more accurate than a simple "defect present / absent" judgment, sufficient to support subsequent precise blasting removal at specific locations.

[0099] Please see Figure 6 The specific steps for obtaining the injection trajectory control configuration are as follows:

[0100] S501: Based on the surface scratch direction coordinate information, combined with the spatial coordinates of the detection position and the spraying position, and using the conveying speed data and nozzle response hysteresis, the movement path of the fastener is calculated and offset corrected in a timing manner to obtain the fastener movement time parameters.

[0101] Based on the coordinate information of the surface scratch direction, the current detection position coordinates of the defective fastener on the conveyor belt are determined. mm. The physical coordinates of the downstream pneumatic nozzle are known. mm, real-time monitored conveyor belt encoder speed data mm / s. The system first calculates the theoretical motion time. s=500ms. Query the system configuration file to obtain the response hysteresis of the nozzle solenoid valve. ms (including coil excitation time and air path transmission time). To ensure that the airflow ejected from the nozzle accurately hits the center of the moving fastener, the system performs timing calculations and corrections: ms. This means the system needs to issue a trigger signal 488 milliseconds after detecting a defect. The calculated fastener movement time parameter at this point is 488ms. This corrected calculation eliminates the spatiotemporal misalignment caused by hardware physical delays, ensuring a high rejection hit rate under high-speed movement.

[0102] S502: Call the fastener motion time parameters, match the injection start control signal according to the mass grouping parameters, and combine the injection pressure levels of the pushing stage and the correction stage to obtain the injection control combination parameters;

[0103] The system calls the fastener movement time parameter (488ms) and combines it with the defect level information generated in the previous steps. The system presets a quality grouping parameter table, classifying defects into "Class A: Severe dimensional deviation / flattening," "Class B: Surface micro-scratches," and "Class C: Exterior oil stains." For different categories, the system matches different injection pressures to balance rejection success rate and energy consumption. Based on the judgment information "severe thread flattening" and "micro-scratches," the system classifies the workpiece into Class A according to the "priority of severe defects" principle. The pressure level table (as shown in Table 3) is consulted; the high-pressure injection mode corresponding to Class A is: 0.6MPa during the pushing stage and 0.4MPa during the correction stage (airflow tail section). Class B corresponds to a low-pressure mode (0.3MPa). Based on this, the system generates the injection control combination parameters: {Trigger delay: 488ms, On-time: 20ms, Pressure setting: 0.6MPa}.

[0104] Table 3 Injection Pressure Level Configuration Table

[0105]

[0106] Referring to Table 3, through this classification and matching, the system can use strong airflow to remove large or severely jammed waste products, and use weak airflow to remove slightly defective products, preventing excessive airflow from causing waste products to rebound or accidentally damaging good products.

[0107] S503: Based on the injection control combination parameters, it integrates motion state data and injection behavior characteristics, configures multiple timing and pressure parameters of the injection trajectory, and obtains the injection trajectory control configuration.

[0108] Based on the injection control combination parameters, the system ultimately integrates motion state data with injection behavior characteristics. A hardware timer is configured in the underlying FPGA controller, with a count value set to 488ms. Simultaneously, an analog output channel (DAC) is configured, setting the output voltage to correspond to a proportional valve opening of 0.6MPa. The system generates a complete injection trajectory control configuration data package: [Timer_ID:1024,Count:488ms,PWM_Duty:100%,DAC_Value:8.5V,Duration:30ms]. When the timer countdown ends, the system immediately executes the configuration, driving the nozzle to release compressed air at the correct time and with the correct pressure curve. This configuration includes not only a single switching action but also dynamic pressure adjustment (shifting and correcting) over time, ensuring the injection airflow forms a precise "air hammer" to remove defective fasteners from the high-speed production line, accurately placing them into the waste collection bin, completing a fully closed-loop automated process from visual recognition to physical rejection.

[0109] Please see Figure 7 An automated screening system for automotive fastener generation includes:

[0110] The end face orientation recognition module calls the end face double angle reflection data group, analyzes the reflection data of the fastener end face under two angles of incident light, calculates the difference between the reflection peak intensity and valley reflection signal under the two incident angles, judges the orientation category according to the classification range set by the end face type, and generates end face orientation judgment information.

[0111] The outer diameter offset assessment module calculates the intersection points of multiple radial sections and fastener contours within the outer diameter region based on the end face orientation information, determines the existence of offset based on the difference in radius of adjacent sections, identifies the circumferential offset direction, classifies the offset trend, and generates an outer diameter elliptical offset index.

[0112] The thread flattening detection module calls the thread valley bottom reflection data group based on the outer diameter ellipse offset index, analyzes the reflection intensity change under multiple incident angles, calculates the reflection signal attenuation difference and generates an attenuation difference sequence, judges the continuity of attenuation change, identifies the flattening area, and generates thread flattening judgment result.

[0113] The surface scratch location module calls the multi-angle reflection curve data group of the fastener surface based on the thread flattening judgment result, analyzes the distribution difference of the reflection peak in spatial coordinates, calculates the reflection peak offset, judges the continuity of the offset sequence, filters the offset trajectory direction, and generates surface scratch direction coordinate information.

[0114] The spray sorting control module calculates the movement time of the defective fastener from the detection position to the spray position based on the coordinate information of the surface scratch direction. It uses the conveying speed and nozzle response lag to correct the offset. It matches the spray control signal by quality grouping and combines the spray pressure in the pushing and correction stages to perform classification matching and generate the spray trajectory control configuration.

[0115] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An automated screening method for automotive fastener generation, characterized in that, Includes the following steps: S1: Call the end face double angle reflection data group, analyze the reflection data of the fastener end face under two angles of incident light, calculate the difference between the reflection peak intensity and valley reflection signal under the two incident angles, determine the orientation category according to the classification range set by the end face type, and generate end face orientation judgment information. S2: Based on the end face orientation judgment information, calculate the intersection points of multiple radial sections and fastener contours within the outer diameter region, determine the existence of offset based on the difference in radius of adjacent sections, identify the circumferential offset direction, classify the offset trend, and generate an outer diameter elliptical offset index. S3: Based on the outer diameter ellipse offset index, call the thread valley bottom reflection data group, analyze the reflection intensity change under multiple incident angles, calculate the reflection signal attenuation difference and generate an attenuation difference sequence, determine the continuity of attenuation change, identify the flattened area, and generate thread flattening judgment result. S4: Based on the thread flattening determination result, call the fastener surface multi-angle reflection curve data group, analyze the distribution difference of reflection peaks in spatial coordinates, calculate the reflection peak offset, determine the continuity of the offset sequence, filter the offset trajectory direction, and generate surface scratch direction coordinate information. The end face orientation judgment information includes orientation category information, direction status identification information, and diversion channel control information. The outer diameter ellipse offset index specifically includes outer diameter ellipticity parameter, circumferential eccentricity parameter, and offset direction characteristic parameter. The thread flattening judgment result includes flattening presence mark, flattening area range, and flattening severity level. The surface scratch direction coordinate information specifically includes scratch direction vector information, scratch start and end coordinate information, and scratch trajectory set information.

2. The automated screening method for generating automotive fasteners according to claim 1, characterized in that, The specific steps for obtaining the end face orientation determination information are as follows: S101: Obtain the double-angle reflection data set of the end face, extract the reflection data formed by the end face under the first incident angle and the reflection data formed under the second incident angle, calculate the numerical difference between the corresponding reflection peak intensity and reflection valley signal, and generate the reflection peak valley difference value. S102: Based on the difference value of the reflection peak and valley, compare it with the classification range set by the end face type, make a judgment on the orientation category, form the classification information of the nut end face direction, and obtain the end face category range information; S103: Based on the end face category range information, establish a direction status identifier for each nut and output the control information of the corresponding diversion channel to obtain end face orientation judgment information.

3. The automated screening method for generating automotive fasteners according to claim 2, characterized in that, The specific steps for obtaining the outer diameter ellipse offset index are as follows: S201: Based on the end face orientation judgment information, calculate the coordinates of the intersection points of multiple radial sections and the fastener contour within the outer diameter detection area, and extract the radius data corresponding to each radial section to generate a radial radius data group; S202: Based on the radial radius data set, perform difference calculations on the radius values ​​between each pair of adjacent cross-sections, and combine the circumferential arrangement order of the cross-sections to identify the offset direction in the circumferential order, thereby obtaining the circumferential offset distribution parameters. S203: Call the circumferential offset distribution parameters, classify and analyze the offset trend of the sectional line, perform stability merging analysis on the offset distribution structure within the entire circumference, and obtain the outer diameter ellipse offset index.

4. The automated screening method for generating automotive fasteners according to claim 3, characterized in that, The specific steps for obtaining the thread flattening determination result are as follows: S301: Based on the outer diameter ellipse offset index, obtain the fastener thread valley bottom multi-angle reflection data group, and calculate the intensity attenuation value between adjacent incident angle reflection signals based on the thread valley bottom reflection intensity under multiple incident angles to obtain the thread reflection attenuation difference sequence. S302: Call the thread reflection attenuation difference sequence, make a continuous judgment on the attenuation change trend of each point in the sequence, analyze the fluctuation concentration characteristics of the reflection difference in multiple incident angles, and obtain the thread attenuation fluctuation range parameters. S303: Based on the thread attenuation fluctuation range parameters and combined with the range segment of fluctuation concentration, identify the spatial distribution of thread valley flattening, establish the regional characteristics and grade parameters of thread flattening, and obtain the thread flattening judgment result.

5. The automated screening method for generating automotive fasteners according to claim 4, characterized in that, The specific steps for obtaining the surface scratch direction coordinate information are as follows: S401: Based on the thread flattening determination result, call the fastener surface multi-angle reflection curve data group, extract the distribution position of reflection peaks in spatial coordinates under multiple illumination angles, calculate the spatial coordinate offset value between adjacent angle reflection peaks, and obtain the peak position coordinate offset sequence. S402: Call the peak position coordinate offset sequence, determine the continuous extension characteristics of the peak position in the spatial direction, filter the offset trajectory structure according to the direction consistency, and generate a set of scratch offset trajectories. S403: Based on the set of scratch offset trajectories, merge the direction vector parameters of the offset trajectories, integrate the scratch direction vector and start and end coordinate information, and obtain the surface scratch direction coordinate information.

6. The automated screening method for generating automotive fasteners according to claim 5, characterized in that, The process of selecting the offset trajectory structure based on directional consistency is as follows: Based on the peak position coordinate offset sequence, calculate the spatial coordinate difference between adjacent angular reflection peaks, obtain the coordinate offset vector direction angle corresponding to each spatial coordinate difference, and arrange them in angular order to form a direction angle sequence. The absolute difference between adjacent direction angles in the direction angle sequence is used to form a direction angle difference sequence, and the 90th decimal place of the direction angle difference sequence is used to set the direction consistency threshold. Elements in the direction angle difference sequence that are less than or equal to the direction consistency threshold and whose angle indices are consecutive are grouped into offset trajectory candidate units, and the offset trajectory structure is obtained by filtering the offset trajectory candidate units with no less than three elements. The trajectory direction vector information is determined by the average direction angle of all coordinate offset vectors within the offset trajectory structure, and the index segment corresponding to the trajectory direction vector information is output as the scratch offset trajectory set.

7. The automated screening method for generating automotive fasteners according to claim 1, characterized in that, The method further includes: S5: Based on the surface scratch direction coordinate information, calculate the movement time of the defective fastener from the detection position to the spraying position, use the conveying speed and nozzle response lag for offset correction, match the spraying control signal through quality grouping, and combine the spraying pressure in the pushing and correction stages for classification matching to generate the spraying trajectory control configuration. The injection trajectory control configuration includes injection start timing parameters, injection pressure combination parameters, and injection trajectory position parameters.

8. The automated screening method for generating automotive fasteners according to claim 7, characterized in that, The specific steps for obtaining the injection trajectory control configuration are as follows: S501: Based on the surface scratch direction coordinate information, combined with the spatial coordinates of the detection position and the spraying position, and using the conveying speed data and nozzle response hysteresis, perform timing calculation and offset correction on the movement path of the fastener to obtain the fastener movement time parameters. S502: Call the fastener movement time parameters, match the injection start control signal according to the mass grouping parameters, and combine the injection pressure levels of the pushing stage and the correction stage to obtain the injection control combination parameters; S503: Based on the injection control combination parameters, integrate motion state data and injection behavior characteristics, configure multiple timing and pressure parameters of the injection trajectory, and obtain the injection trajectory control configuration.

9. An automated screening system for automotive fastener generation, characterized in that, The system is used to implement the automated screening method for generating automotive fasteners according to any one of claims 1-8, the system comprising: The end face orientation recognition module calls the end face double angle reflection data group, analyzes the reflection data of the fastener end face under two angles of incident light, calculates the difference between the reflection peak intensity and valley reflection signal under the two incident angles, judges the orientation category according to the classification range set by the end face type, and generates end face orientation judgment information. The outer diameter offset evaluation module calculates the intersection points of multiple radial sections and fastener contours within the outer diameter region based on the end face orientation judgment information, determines the existence of offset based on the difference in radius of adjacent sections, identifies the circumferential offset direction, classifies the offset trend, and generates an outer diameter elliptical offset index. The thread flattening detection module calls the thread valley bottom reflection data group according to the outer diameter ellipse offset index, analyzes the reflection intensity change under multiple incident angles, calculates the reflection signal attenuation difference and generates an attenuation difference sequence, judges the continuity of attenuation change, identifies the flattening area, and generates thread flattening judgment result. The surface scratch positioning module calls the fastener surface multi-angle reflection curve data group according to the thread flattening judgment result, analyzes the distribution difference of reflection peaks in spatial coordinates, calculates the reflection peak offset, judges the continuity of the offset sequence, filters the offset trajectory direction, and generates surface scratch direction coordinate information. The spray sorting control module calculates the movement time of the defective fastener from the detection position to the spray position based on the surface scratch direction coordinate information, uses the conveying speed and nozzle response lag for offset correction, matches the spray control signal by quality grouping, and performs classification matching by combining the spray pressure in the pushing and correction stages to generate the spray trajectory control configuration.

Images