Automatic detecting and sorting system for thread protector

Abstract

The invention provides an automatic detection and sorting system for thread protectors, and relates to the technical field of detection and sorting of petroleum and natural gas pipeline accessories, and the system comprises an acquisition device which carries out image acquisition, size measurement and spectrum detection on the thread protectors to form a to-be-detected data set; the to-be-detected data set comprises appearance image data, size parameter data and material spectrum data. According to the automatic detection and sorting system for the thread protectors, the thread protectors are comprehensively detected through multiple means such as image acquisition, size measurement and spectrum detection, the appearance defects, geometric parameters and material classification characteristics of the thread protectors are accurately recognized, and the detection precision is remarkably improved. And through application of intelligent algorithms such as the convolutional neural network, the system can automatically judge the detection result, so that automatic sorting is realized, manual intervention is effectively reduced, and the sorting efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of oil and gas pipeline accessory inspection and sorting technology, specifically an automatic inspection and sorting system for thread protectors. Background Technology

[0002] Thread protectors are commonly used protective components at the ends of oil casing and tubing. They prevent thread damage caused by collisions or friction during transportation, handling, and storage, ensuring the sealing performance and service life of the tubing string. Existing products are mainly divided into all-plastic protectors and steel-plastic composite protectors. All-plastic protectors are typically made of polyethylene, which is lightweight and easy to process, but is prone to dimensional instability and impact cracking under extreme temperature conditions. To improve their performance, the industry has developed modified materials composed of polypropylene, glass fiber reinforcement, and toughening agents, which remain stable in high and low temperature environments and meet the testing requirements of API 5CT Appendix I. Steel-plastic composite protectors combine steel strips with polyethylene using a snap-fit ​​structure, avoiding stress concentration and delamination problems caused by traditional nailing methods, thus improving both appearance and reliability. With the expansion of oil development, the production volume of protectors is rapidly increasing, making inspection and sorting crucial processes for ensuring quality. Currently, the industry generally uses manual visual inspection and semi-automatic sorting, relying on the experience and judgment of operators. While simple, these methods have limited efficiency.

[0003] The core problem with existing inspection methods lies in their reliance on manual observation, resulting in poor result stability and a high risk of missed detections or misjudgments. Defective protectors entering the operational phase pose a safety hazard to the tubing. In production environments, the sheer number of protectors, with their often simultaneous presence of appearance defects, dimensional deviations, and material variations, makes efficient and accurate sorting by manual labor insufficient. This method fails to meet the efficiency and quality requirements of modern production, and the defects become even more pronounced in large-scale operations and extreme application scenarios. Therefore, there is an urgent need for a system capable of automated inspection and intelligent sorting to improve inspection accuracy and operational efficiency. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an automatic detection and sorting system for thread protectors. The technical problem this invention aims to solve is: how to achieve automatic detection and intelligent sorting of thread protectors by employing image acquisition, dimensional measurement, and spectral detection technologies, combined with convolutional neural networks for feature extraction and judgment processing of data, thereby improving detection accuracy and production efficiency.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an automatic detection and sorting system for thread protectors, comprising: a data acquisition device, wherein the data acquisition device performs image acquisition, size measurement and spectral detection on the thread protectors to form a dataset to be inspected, wherein the dataset to be inspected includes appearance image data, size parameter data and material spectral data.

[0006] The feature extraction device performs feature extraction processing on the dataset to be inspected to form a feature set, which includes defect crack features, geometric parameter features, material classification features, and snap-fit ​​structure integrity index features.

[0007] The determination device performs determination processing on the feature set to form a detection result, which includes qualified thread protectors and unqualified thread protectors.

[0008] The sorting device sorts the thread protectors, and the sorting process includes conveying the qualified thread protectors to the qualified product channel and the unqualified thread protectors to the rejection channel.

[0009] A traceability device is used to trace the test results and form a quality record.

[0010] Preferably, the image acquisition uses an industrial camera to form the appearance image data, the size measurement uses a laser rangefinder to form the size parameter data, and the spectral detection uses a spectral sensor to form the material spectral data.

[0011] Preferably, the feature extraction process uses a convolutional neural network to perform defect and crack detection processing on the appearance image data to form the defect and crack features; the feature extraction process uses an edge contour analysis method to perform geometric feature modeling processing on the size parameter data to form the geometric parameter features; and the feature extraction process uses a spectral feature analysis method to perform classification modeling on the material spectral data to form the material classification features. This achieves automated and accurate extraction and identification of multiple features of the thread protector's appearance defects, geometric dimensions, and material composition.

[0012] Preferably, the convolutional neural network is a convolutional neural network with a kernel size of 3×3 and a stride of 1. The spectral analysis processing uses a maximum-minimum normalization method to map the spectral intensity values ​​of the material spectral data to 0-1. The material classification features include distinguishing between all-plastic modified protectors and steel-plastic composite protectors.

[0013] Preferably, the determination process includes comparing the defect crack features with a preset threshold, comparing the geometric parameters with the allowable dimensional deviation range, and determining the connection reliability of the thread protector by the snap-fit ​​structure integrity index, thereby realizing a multi-dimensional, quantitative, and traceable automated and accurate determination of the quality of the thread protector.

[0014] Preferably, the preset threshold is set according to the standard of extreme working conditions simulated in Appendix I of API 5CT. The extreme working conditions include low temperature working conditions and high temperature working conditions. The temperature of the low temperature working condition is -46°C and the allowable dimensional deviation range of the low temperature working condition is ±0.5mm. The temperature of the high temperature working condition is 66°C and the allowable dimensional deviation range of the high temperature working condition is ±0.8mm.

[0015] Preferably, the traceability device includes operator information binding processing, which generates operation records, and includes binding employee ID and inspection shift number.

[0016] Preferably, the quality records are stored in JSON format, and the quality records are visualized to form traceability reports. The visualization process uses a combination of time series line charts and statistical pie charts.

[0017] This invention provides an automatic detection and sorting system for thread protectors. It has the following advantages:

[0018] This invention utilizes multiple methods, including image acquisition, dimensional measurement, and spectral detection, to comprehensively inspect thread protectors, accurately identifying features such as appearance defects, geometric parameters, and material classification, significantly improving inspection accuracy. Through the application of intelligent algorithms such as convolutional neural networks, the system can automatically determine the inspection results, thereby achieving automated sorting, effectively reducing manual intervention and improving sorting efficiency.

[0019] This automatic thread protector detection and sorting system, through its automatic traceability function, saves the detection results in the form of quality records and generates traceability reports through visualization. This not only ensures the traceability of product quality but also facilitates production monitoring and quality analysis for operators, thereby further enhancing the reliability and transparency of the production process. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the automatic detection and sorting system for thread protectors;

[0021] Figure 2 This is a flowchart of the feature extraction process;

[0022] Figure 3 This is a flowchart of the decision-making process;

[0023] Figure 4 This is a flowchart of the traceability process;

[0024] Figure 5 This is a flowchart of data acquisition and testing. Detailed Implementation

[0025] 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 embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Example 1

[0027] like Figure 1-5 As shown, this embodiment of the invention provides an automatic inspection and sorting system for thread protectors, including a data acquisition device. The data acquisition device performs image acquisition, dimensional measurement, and spectral detection on the thread protectors to form a dataset to be inspected. The dataset includes appearance image data, dimensional parameter data, and material spectral data. Image acquisition uses an industrial camera to generate appearance image data, dimensional measurement uses a laser rangefinder to generate dimensional parameter data, and spectral detection uses a spectral sensor to generate material spectral data.

[0028] The image acquisition section utilizes an industrial camera to capture appearance image data. The industrial camera boasts a 50-megapixel resolution, enabling it to clearly capture surface images of the thread protector. To ensure inspection accuracy under varying angles and lighting conditions, the system is equipped with an automatically adjusting light source and lens, allowing for precise detection of minute cracks, stains, and other appearance defects on the thread protector surface. The acquired appearance image data includes the surface features of the thread protector, including potential cracks, dents, and other morphological defects.

[0029] Dimensional measurement utilizes a laser rangefinder for precise data acquisition of dimensional parameters. The laser rangefinder boasts extremely high accuracy, capable of measuring the length, width, thickness, and other critical dimensions of the thread protector. The system records dimensional data in real-time from all directions through multiple laser scans, generating a detailed dataset of dimensional parameters. With an accuracy of 0.01mm, the laser rangefinder provides timely feedback even with minimal dimensional changes, ensuring the thread protector's dimensions meet standard requirements.

[0030] Spectroscopic detection acquires spectral data of the material using a spectral sensor. Operating in the visible and near-infrared bands, the sensor interacts with the surface material of the thread protector to identify the spectral characteristics of different materials. The sensor scans the material surface and acquires spectral data. This spectral data contains information about the material's molecular structure and physical properties, helping the system accurately identify different types of thread protector materials. The spectral sensor has a resolution of 0.1 nm, effectively distinguishing subtle material differences.

[0031] The feature extraction device processes the dataset to be inspected to form a feature set, which includes defect and crack features, geometric parameter features, material classification features, and snap-fit ​​structure integrity index features. The feature extraction process employs a convolutional neural network to detect defects and cracks in the appearance image data, generating defect and crack features. It also uses edge contour analysis to model the dimensional parameter data, generating geometric parameter features, and spectral feature analysis to classify the material spectral data, generating material classification features.

[0032] The convolutional neural network is used with a kernel size of 3×3 and a stride of 1. The spectral analysis process uses the maximum-minimum normalization method to map the spectral intensity values ​​of the material spectral data to 0-1. The material classification features include distinguishing between all-plastic modified protectors and steel-plastic composite protectors.

[0033] The judgment device processes the feature set to generate a test result, which includes qualified and unqualified thread protectors. The judgment process includes comparing defect crack features with preset thresholds, comparing geometric parameters with allowable dimensional deviation ranges, and determining the connection reliability of the thread protector based on the integrity index of the snap-fit ​​structure.

[0034] The preset thresholds are set according to the extreme operating conditions simulated in Appendix I of API 5CT. The extreme operating conditions include low temperature conditions and high temperature conditions. The temperature of the low temperature condition is -46℃, and the allowable dimensional deviation range of the low temperature condition is ±0.5mm. The temperature of the high temperature condition is 66℃, and the allowable dimensional deviation range of the high temperature condition is ±0.8mm.

[0035] The sorting device sorts the thread protectors, including conveying qualified thread protectors to the qualified product channel and unqualified thread protectors to the rejection channel.

[0036] Qualified thread protectors are conveyed to the qualified product channel via an automated sorting system. A moving conveyor belt on the qualified product channel transports the protectors to the packaging or further processing area. A sensor system continuously monitors the protectors' passage through the channel, ensuring that qualified thread protectors are correctly positioned to prevent misplacement or confusion. The sensors operate using vision-based systems or infrared detection, identifying the status and position of each thread protector to ensure that qualified products are delivered to the qualified product channel.

[0037] For defective thread protectors, the sorting device transports them to a rejection channel. This channel is equipped with pneumatic actuators; based on sensor feedback, the system automatically rejects defective products using pneumatic actuators. By controlling airflow, the pneumatic actuators precisely push defective protectors to designated waste or recycling areas, ensuring that defective products do not enter the production chain or are misused, thus guaranteeing product quality stability.

[0038] The control system adjusts the sorting actions in real time based on the detection results, ensuring that each thread protector is accurately assigned to the appropriate channel. The control system works closely with the sensor system and actuators to ensure high efficiency and precision in the sorting process. The operating speed of the entire sorting system can be adjusted according to production needs to adapt to different production environments and thread protector sorting requirements.

[0039] The traceability device processes test results to create quality records. This device includes operator information binding, which generates operation records. The operator information binding process includes binding employee ID and shift number. Quality records are stored in JSON format and visualized to create traceability reports. The visualization uses a combination of time-series line charts and statistical pie charts.

[0040] Operator Information Binding: The traceability device uses operator information binding to associate each detection operation with the operator's identity information, ensuring that each detection result can be traced back to a specific operator. The specific steps of operator information binding are as follows:

[0041] Employee ID Binding: Each operator has a unique employee ID in the system, which is automatically recorded in the operation log. For example, the operator's employee ID is "G1243".

[0042] Binding Inspection Shift Number: To further track the time period of inspection operations, the system will bind the operation record with the corresponding inspection shift number. For example, the inspection operation occurred in shift number "B20250929".

[0043] Example: Operator ID: G1243, Inspection Shift Number: B20250929, Operation Time: 08:30:00 on September 29, 2025.

[0044] Quality record storage and management: The traceability device stores each inspection result, operator information, and other relevant data in JSON format. JSON format is easy to store and transmit, and has good compatibility and scalability. Quality record content includes: Inspection result: whether it is qualified, e.g., "qualified" or "unqualified"; Operator information: including employee number and shift number; Inspection time: the specific time the operation occurred; Other information: such as the defect type of the thread protector, dimensional deviation, and snap-fit ​​structure integrity, etc.

[0045] Example: Inspection result: qualified, operator ID: G1243, shift number: B20250929, inspection time: 2025-09-29, 08:30:00, thread protector defect: no cracks, dimensional deviation: ±0.3mm, snap-fit ​​structure integrity: qualified.

[0046] Visualization of Quality Records: The traceability system generates traceability reports based on stored quality records and displays the data through visualization techniques, facilitating analysis and monitoring of the production process. The visualization utilizes the following two chart formats:

[0047] Time series line chart: Displays the trend of quality changes over different time periods. For example, it can show the pass rate, defect rate, etc. for different dates or shifts.

[0048] Statistical pie charts: These display the distribution of different types of defects or the pass / fail status of thread protectors made of different materials. Pie charts visually reflect the proportion of various inspection results.

[0049] Example:

[0050] Time series line chart: Used to show the change in pass rate for each shift over time, helping to analyze potential quality fluctuations in the production process. X-axis: Time, sorted by shift or date; Y-axis: Pass rate, percentage.

[0051] Statistical pie chart: Used to display the proportion of different defect types, helping to analyze common defects in products. Each section shows the percentage of different defect types out of the total number of defects.

[0052] Generating and using traceability reports:

[0053] The final traceability report will be exported in PDF or HTML format for quality control personnel, operations managers, and other relevant staff to view and analyze. The report not only helps companies effectively trace product quality but also provides data support for subsequent quality improvements.

[0054] Example report content:

[0055] Pass Rate Analysis: Pass rate trend charts for each shift help managers understand the production situation of each shift. Defect Analysis: Statistical data on various defects reveal which problems are most common in the production process, allowing for targeted improvement measures.

[0056] Example 2

[0057] This embodiment utilizes image acquisition, dimensional measurement, and spectral detection technologies, combined with convolutional neural networks and other analysis methods, to achieve defect and crack detection, geometric parameter modeling, and material classification of thread protectors, thereby automatically extracting features and performing accurate judgment and sorting.

[0058] 1. Feature extraction from appearance image data

[0059] 1.1 Image Acquisition

[0060] Industrial cameras are used to acquire high-resolution images of thread protectors. Each thread protector image includes surface features such as cracks, wear, or other defects. To obtain accurate image data, the industrial camera needs to be set at a fixed angle and adequate lighting conditions must be ensured to minimize the effects of reflections or shadows.

[0061] 1.2 Feature Extraction

[0062] The acquired appearance image data is processed using a convolutional neural network. The network employs 3×3 convolutional kernels with a stride of 1 to ensure that high resolution and detail are not lost. During training, a large number of images of thread protectors with and without defects are used. By learning the features of cracks and other defects in the images, the network can efficiently extract crack features.

[0063] In the dataset, the defect crack area in the images ranges from 10px² to 500px². The CNN can determine whether a crack exists based on features such as the crack area, depth, and shape. If the network detects a crack area exceeding a preset threshold, such as 50px², it considers the thread protector to be defective.

[0064] 2. Feature extraction from dimensional parameter data

[0065] 2.1 Dimensional Data Acquisition

[0066] Dimensional measurements were taken using a laser rangefinder, which can accurately capture the physical dimensions of the thread protector, including length, width, and thickness. The measured data are as follows: length 300mm, width 120mm, and thickness 10mm.

[0067] 2.2 Feature Extraction

[0068] The dimensional data is fed into the edge profile analysis system, which uses algorithms to perform geometric modeling on the measurement data. The edge detection algorithm can analyze whether the shape of the thread protector conforms to standards, such as roundness, symmetry, and other geometric features.

[0069] For a given thread protector, the permissible dimensional deviations may be as follows: the permissible deviation for length is ±2 mm, the permissible deviation for width is ±1 mm, and the permissible deviation for thickness is ±0.5 mm.

[0070] 3. Feature extraction from material spectral data

[0071] 3.1 Spectral Data Acquisition

[0072] Spectral data is acquired through a spectral sensor, which captures the reflection spectrum of the thread protector material. For example, spectral data measured at a given moment might show a spectral intensity value of 0.65 and a normalized reflection spectrum of 0.32.

[0073] 3.2 Feature Extraction

[0074] After being normalized to its minimum and maximum values, the spectral data is fed into a material classification model. By learning the spectral characteristics of different materials, the material classification model can classify the materials used in thread protectors.

[0075] The goal of material classification is to distinguish between all-plastic modified protectors and steel-plastic composite protectors. The formulation of all-plastic modified materials is: 69% polypropylene, 5% rubber, 8% calcium carbonate, 10% glass fiber, 3% maleic anhydride grafting, and 5% toughening masterbatch. Steel-plastic protectors are composed of SPHC steel strips and high-density polyethylene.

[0076] For example, the typical spectral reflectance intensity of an all-plastic modified protector is usually between 0.2 and 0.4 in the wavelength range of 500 nm to 1000 nm, while the spectral intensity of a steel-plastic composite protector is usually higher, possibly approaching 0.6. By comparing sensor data with known classification standards, the system can automatically determine the material category.

[0077] 4. Formation and output of comprehensive feature set

[0078] After feature extraction is complete, the feature extraction device will generate a feature set containing the following:

[0079] Defect / Crack Features: Crack detection results from image data, including crack location, size, and shape information. Geometric Parameter Features: Geometric features from dimensional measurement data, including length, width, and thickness compared to allowable deviation ranges. Material Classification Features: Material classification information from spectral data, determining whether the thread protector is an all-plastic modified protector or a steel-plastic composite protector.

[0080] The feature set will serve as the basis for subsequent judgment and sorting to ensure that the quality of the thread protector meets production requirements.

[0081] Example 3

[0082] This embodiment uses technologies such as image acquisition, size measurement, spectral detection, and convolutional neural networks to achieve automatic detection and sorting of thread protectors, and ensures product quality by judging defects, cracks, geometric parameters, and the integrity of the snap-fit ​​structure.

[0083] 1. Comparison of defect crack characteristics with preset thresholds

[0084] During the image acquisition phase, an industrial camera captures high-resolution images of the thread protector surface, identifying potential defects such as cracks, scratches, and wear. The image data is then processed by a convolutional neural network to extract feature information about the cracks, including their location, size, depth, and shape.

[0085] The judgment device compares the extracted crack features with preset thresholds. These preset thresholds are set based on experience and standards, and are determined by parameters such as the area, length, and depth of the defect. For example, if the crack area is greater than 50px², it is considered a major defect and judged as unacceptable. If the crack depth exceeds 0.3mm, it is also considered unacceptable.

[0086] Specific example: During the inspection, a crack with an area of ​​60px² and a depth of 0.4mm was found on the surface of the thread protector, which meets the non-compliance standard, and the thread protector is judged to be unqualified.

[0087] 2. Comparison of geometric parameters with allowable dimensional deviation ranges

[0088] During the dimensional measurement phase, a laser rangefinder is used to accurately measure various dimensions of the thread protector, such as length, width, and thickness. The measurement data is then compared with the standard dimensional requirements and allowable deviation ranges.

[0089] Under standard operating conditions, the dimensions of the thread protector are as follows: length: 300mm ± 2mm, width: 120mm ± 1mm, thickness: 10mm ± 0.5mm.

[0090] In actual measurements, the dimensions of the thread protector are: length: 299.5mm, width: 119mm, thickness: 10.2mm.

[0091] After comparison, it was found that the length and width of the thread protector were within the allowable deviation range, but the thickness exceeded the deviation range of ±0.5mm. Therefore, the thread protector was determined to be non-compliant in terms of geometric parameters.

[0092] 3. Determination of the integrity index and connection reliability of the snap-fit ​​structure

[0093] In terms of snap-fit ​​structure integrity testing, the system analyzes whether the snap-fit ​​structure of the thread protector meets the standard requirements, and evaluates it from the perspectives of snap-fit ​​connection reliability, structural strength, durability, etc.

[0094] For example, the API 5CT standard requires that the snap-fit ​​structure of thread protectors should maintain stable connection strength under extreme temperature conditions to prevent detachment or loosening. The assessment device compares actual test data with the standard to evaluate the integrity of the snap-fit ​​structure.

[0095] If the snap-fit ​​structure fails or loosens under extreme temperatures, it is deemed unqualified. During testing, if the snap-fit ​​of the thread protector loosens at a high temperature of 66°C, failing to meet the connection reliability standard, it is deemed unqualified.

[0096] 4. Overall Judgment Results

[0097] After completing the three judgments, the judgment device will synthesize the results of each judgment to determine whether the thread protector is qualified. For example, if the thread protector fails to meet the geometric parameters, but passes the snap-fit ​​structure and crack detection, the system can make a weighted judgment based on the specific circumstances. If the defects in the cracks and snap-fit ​​structure are serious, it may still be judged as unqualified, even if the deviation of the geometric parameters is small.

[0098] If all criteria meet the standards or are within the allowable range, the thread protector is deemed qualified and the system will send it to the qualified product channel; otherwise, it will be sent to the unqualified product channel for rejection.

[0099] 5. Setting the preset threshold in the determination device

[0100] Low temperature operating conditions:

[0101] In low-temperature environments, specifically -46°C, the material of thread protectors may shrink or become embrittled, necessitating a reduction in the permissible dimensional tolerances. According to API 5CT Appendix I, under low-temperature conditions, the dimensional deviation of thread protectors should be controlled within ±0.5 mm.

[0102] If the length, width, thickness, and other dimensions of the thread protector measured in a low-temperature environment exceed the allowable deviation range, it is deemed unqualified.

[0103] High-temperature operating conditions:

[0104] Under high-temperature conditions, i.e., at 66℃, thread protectors may expand or deform, and the permissible dimensional deviation range can be appropriately relaxed. According to API 5CT standard, the permissible dimensional deviation range for thread protectors under high-temperature conditions is ±0.8mm.

[0105] If the dimensional deviation of the thread protector exceeds the range under high temperature conditions, it will be deemed unqualified.

[0106] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An automatic detection and sorting system for thread protectors, characterized in that, include: The acquisition device performs image acquisition, size measurement, and spectral detection on the thread protector to form a dataset to be inspected, which includes appearance image data, size parameter data, and material spectral data. A feature extraction device performs feature extraction processing on the dataset to be inspected to form a feature set, the feature set including defect crack features, geometric parameter features, material classification features, and snap-fit ​​structure integrity index features; A determination device performs determination processing on the feature set to form a detection result, the detection result including qualified thread protectors and unqualified thread protectors; The sorting device sorts the thread protectors, and the sorting process includes conveying the qualified thread protectors to the qualified product channel and the unqualified thread protectors to the rejection channel. A traceability device is used to trace the test results and form a quality record.

2. The automatic detection and sorting system for thread protectors according to claim 1, characterized in that: The image acquisition uses an industrial camera to generate the appearance image data, the size measurement uses a laser rangefinder to generate the size parameter data, and the spectral detection uses a spectral sensor to generate the material spectral data.

3. The automatic detection and sorting system for thread protectors according to claim 1, characterized in that: The feature extraction process uses a convolutional neural network to perform defect and crack detection processing on the appearance image data to form the defect and crack features. The feature extraction process uses an edge contour analysis method to perform geometric feature modeling processing on the size parameter data to form the geometric parameter features. The feature extraction process uses a spectral feature analysis method to perform classification modeling on the material spectral data to form the material classification features.

4. The automatic detection and sorting system for thread protectors according to claim 3, characterized in that: The convolutional neural network is a 3×3 convolutional neural network with a stride of 1. The spectral analysis processing uses the maximum-minimum normalization method to map the spectral intensity values ​​of the material spectral data to 0-1. The material classification features include distinguishing between all-plastic modified protectors and steel-plastic composite protectors.

5. The automatic detection and sorting system for thread protectors according to claim 1, characterized in that: The determination process includes comparing the defect crack features with a preset threshold, comparing the geometric parameters with the allowable dimensional deviation range, and determining the connection reliability of the thread protector based on the snap-fit ​​structure integrity index.

6. The automatic detection and sorting system for thread protectors according to claim 5, characterized in that: The preset threshold is set according to the standard of extreme working conditions simulated in Appendix I of API 5CT. The extreme working conditions include low temperature conditions and high temperature conditions. The temperature of the low temperature condition is -46℃, and the allowable dimensional deviation range of the low temperature condition is ±0.5mm. The temperature of the high temperature condition is 66℃, and the allowable dimensional deviation range of the high temperature condition is ±0.8mm.

7. The automatic detection and sorting system for thread protectors according to claim 1, characterized in that: The traceability device includes an operator information binding process, which generates an operation record. The operator information binding process includes binding the employee number and the inspection shift number.

8. The automatic detection and sorting system for thread protectors according to claim 1, characterized in that: The quality records are stored in JSON format. The quality records are visualized to form traceability reports. The visualization process uses a combination of time series line charts and statistical pie charts.

Images