Seamless steel pipe produced by image and tofd combined detection method

Abstract

The application provides a seamless steel pipe produced by a combined detection method of images and TOFD, wherein the detection method comprises the following steps: an image detection unit is used to acquire image detection data of a seamless steel pipe to be detected; crack detection is performed based on the image detection data, so that a target position of a suspected crack and first crack detection data are obtained; the transmission speed of the seamless steel pipe to be detected is acquired; time delay compensation is performed based on the transmission speed, so that TOFD detection time delay is acquired; the target position is detected by a TOFD detection unit based on the detection time delay, so that second crack detection data are obtained; and the first crack detection data and the second crack detection data are input into a pre-trained crack detection model, so that the first crack detection data and the second crack detection data are subjected to feature extraction and weighted fusion by the crack detection model, and a crack detection result is obtained by combining corresponding parameter thresholds. According to the technical scheme, the accuracy and reliability of crack detection can be improved.

Description

Technical Field

[0001] This application relates to the field of seamless steel pipe production technology, and more particularly to a seamless steel pipe produced by a combined image and TOFD detection method. Background Technology

[0002] Seamless steel pipes are prone to surface defects such as cracks, scratches, and rolling marks after cold rolling. Related technologies for detecting surface defects in seamless steel pipes mainly include manual visual inspection, single-vision inspection, single-TOFD inspection, and eddy current testing. Manual visual inspection is inefficient, has a high rate of missed detections, and cannot be used in high-temperature environments; single-vision inspection can identify most surface defects but cannot obtain crack depth; single-TOFD inspection can accurately measure crack depth, but it is mostly offline or semi-online inspection, with slow scanning speed, making it difficult to achieve rapid screening along the entire length; eddy current testing is suitable for surface and near-surface defects, but is easily affected by material and surface condition. Summary of the Invention

[0003] This application aims to at least partially address one of the technical problems in the related art.

[0004] In a first aspect, this application proposes a seamless steel pipe inspection method combining image processing and Time-of-Flight (TOFD) detection, comprising: acquiring image detection data of the seamless steel pipe to be inspected based on the image detection unit; wherein the image detection unit is located at the exit of a cold rolling mill; performing crack detection based on the image detection data to obtain the target location of a suspected crack and first crack detection data; acquiring the transmission speed of the seamless steel pipe to be inspected; performing time delay compensation based on the transmission speed to obtain a TOFD detection delay; controlling the TOFD detection unit to detect the target location based on the detection delay to obtain second crack detection data; wherein the TOFD detection unit is located downstream of the image detection unit; inputting the first crack detection data and the second crack detection data into a pre-trained crack detection model, so that the crack detection model performs feature extraction and weighted fusion on the first crack detection data and the second crack detection data, and combines the corresponding parameter thresholds to obtain a crack detection result.

[0005] In one implementation, the step of performing crack detection based on the image detection data to obtain the target location of the suspected crack and the first crack detection data includes: preprocessing the image detection data to obtain image data to be detected; using a pre-trained target detection network to detect the image data to be detected to obtain the pixel location of the suspected crack and the first crack detection data; and performing position calibration based on the pixel location to obtain the axial position and circumferential angle as the target location.

[0006] In one alternative implementation, the circumferential angle is determined by the following steps: based on the pixel position, the offset of the crack in the circumferential direction of the steel pipe is calculated; the offset is converted based on a preset reference orientation to obtain the circumferential angle.

[0007] In one implementation, the step of performing delay compensation based on the transmission speed to obtain the TOFD detection delay includes: performing speed smoothing filtering on the transmission speed to obtain a corrected speed; and obtaining the detection delay based on the corrected speed and the distance between the image detection unit and the TOFD detection unit.

[0008] In one implementation, the crack detection model processes the first crack detection data and the second crack detection data through the following steps to obtain the crack detection result: obtaining a first confidence level corresponding to the first crack detection data and a second confidence level corresponding to the second crack detection data; performing feature extraction on the first crack detection data to obtain an image feature vector; performing feature extraction on the second crack detection data to obtain a TOFD feature vector; performing weighted fusion of the image feature vector and the TOFD feature vector based on the first confidence level and the second confidence level to obtain a fused feature vector; performing parameter parsing on the fused feature vector to obtain crack parameters; and comparing the crack parameters with the parameter thresholds to obtain the crack detection result.

[0009] In one optional implementation, obtaining the first confidence level corresponding to the first crack detection data and the second confidence level corresponding to the second crack detection data includes: obtaining image detection quality parameters of the image detection data; obtaining the first confidence level based on the image detection quality parameters; obtaining ultrasonic detection signal feature parameters corresponding to the second crack detection data; and obtaining the second confidence level based on the ultrasonic detection signal feature parameters.

[0010] In one implementation, the parameter threshold is determined through the following steps: obtaining the target steel pipe specification and production environment parameters of the seamless steel pipe to be tested; obtaining a basic parameter threshold from a pre-configured threshold library based on the target steel pipe specification and the production environment parameters; wherein the threshold library stores multiple sets of parameter thresholds, each set of parameter thresholds corresponding to a set of steel pipe specifications and production environment parameters; obtaining first historical testing data of qualified seamless steel pipes within a preset historical period; wherein the steel pipe specification of the qualified seamless steel pipe is the same as the target steel pipe specification; adjusting the basic parameter threshold based on the first historical testing data to obtain the parameter threshold.

[0011] In one optional implementation, the method further includes: acquiring second historical detection data according to a pre-configured detection volume or detection cycle; performing threshold recalibration based on the second historical detection data to obtain an updated threshold; and updating the threshold library using the updated threshold.

[0012] In one implementation, the first crack detection data includes at least one of the following: crack length, crack width, morphological parameters, and texture parameters; the second crack detection data includes at least one of the following: crack depth and crack height.

[0013] Secondly, this application proposes a seamless steel pipe, which is produced by the testing method described in the first aspect.

[0014] Thirdly, this application proposes a seamless steel pipe inspection system combining image processing and Time-of-Flight (TOFD) detection. The system includes an image detection unit, a TOFD detection unit, and a processing unit. Along the seamless steel pipe conveying direction, the image detection unit is located at the cold rolling mill exit, and the TOFD detection unit is sequentially located downstream of the image detection unit. The processing unit is used to: acquire image detection data of the seamless steel pipe to be inspected based on the image detection unit; perform crack detection based on the image detection data to obtain the target location of a suspected crack and first crack detection data; acquire the transmission speed of the seamless steel pipe to be inspected; perform time delay compensation based on the transmission speed to obtain the TOFD detection delay; control the TOFD detection unit to detect the target location based on the detection delay to obtain second crack detection data; input the first crack detection data and the second crack detection data into a pre-trained crack detection model, so that the crack detection model performs feature extraction and weighted fusion on the first crack detection data and the second crack detection data, and combines the corresponding parameter thresholds to obtain the crack detection result.

[0015] In one implementation, the processing unit can be used to: preprocess the image detection data to obtain image data to be detected; use a pre-trained target detection network to detect the image data to be detected to obtain the pixel position of the suspected crack and the first crack detection data; and perform position calibration based on the pixel position to obtain the axial position and circumferential angle as the target position.

[0016] In one alternative implementation, the processing unit may be used to determine the circumferential angle by the following steps: calculating the offset of the crack in the circumferential direction of the steel pipe based on the pixel position; and converting the offset based on a preset reference orientation to obtain the circumferential angle.

[0017] In one implementation, the processing unit can be used to: perform speed smoothing filtering on the transmission speed to obtain a corrected speed; and based on the corrected speed, combine the distance between the image detection unit and the TOFD detection unit to obtain the detection delay.

[0018] In one implementation, the crack detection model processes the first crack detection data and the second crack detection data through the following steps to obtain the crack detection result: obtaining a first confidence level corresponding to the first crack detection data and a second confidence level corresponding to the second crack detection data; performing feature extraction on the first crack detection data to obtain an image feature vector; performing feature extraction on the second crack detection data to obtain a TOFD feature vector; performing weighted fusion of the image feature vector and the TOFD feature vector based on the first confidence level and the second confidence level to obtain a fused feature vector; performing parameter parsing on the fused feature vector to obtain crack parameters; and comparing the crack parameters with the parameter thresholds to obtain the crack detection result.

[0019] In one optional implementation, the processing unit may be used to: acquire image detection quality parameters of the image detection data; acquire a first confidence level based on the image detection quality parameters; acquire ultrasonic detection signal feature parameters corresponding to the second crack detection data; and acquire a second confidence level based on the ultrasonic detection signal feature parameters.

[0020] In one implementation, the processing unit can be used to: acquire the target steel pipe specifications and production environment parameters of the seamless steel pipe to be tested; obtain a basic parameter threshold from a pre-configured threshold library based on the target steel pipe specifications and the production environment parameters; wherein the threshold library stores multiple sets of parameter thresholds, each set of parameter thresholds corresponding to a set of steel pipe specifications and production environment parameters; acquire first historical test data of qualified seamless steel pipes within a preset historical period; wherein the steel pipe specifications of the qualified seamless steel pipes are the same as the target steel pipe specifications; and adjust the basic parameter threshold based on the first historical test data to obtain the parameter threshold.

[0021] In an optional implementation, the processing unit may further be used to: acquire second historical detection data according to a pre-configured detection quantity or detection cycle; perform threshold recalibration based on the second historical detection data to obtain an updated threshold; and update the threshold library using the updated threshold.

[0022] In one implementation, the first crack detection data includes at least one of the following: crack length, crack width, morphological parameters, and texture parameters.

[0023] In one implementation, the second crack detection data includes at least one of the following: crack depth and crack height.

[0024] Fourthly, this application provides an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the detection method as described in the first aspect.

[0025] Fifthly, this application proposes a storage medium storing instructions that, when executed on an electronic device, cause the electronic device to perform the detection method as described in the first aspect.

[0026] In a sixth aspect, this application proposes a program product comprising at least one of a program and instructions, wherein when the program and instructions are executed by an electronic device, they implement the steps of the detection method described in the first aspect.

[0027] The seamless steel pipe inspection method, apparatus, equipment, and storage medium combining image processing and Time-of-Flight (TOFD) provided in this application can acquire image detection data of the seamless steel pipe to be inspected through an image detection unit. Based on the image detection data, crack detection is performed to obtain the target location of suspected cracks and the first crack detection data. Then, the steel pipe transmission speed is acquired and time delay compensation is performed to obtain the TOFD detection delay. Based on this delay, the TOFD detection unit is controlled to detect the target location to obtain the second crack detection data. Finally, the two types of detection data are input into a pre-trained crack detection model. The model performs feature extraction and weighted fusion, and combines the results with parameter thresholds to obtain the crack detection result. This allows for the coordinated operation of vision and TOFD detection, effectively improving the accuracy and reliability of crack detection. Furthermore, the inspection method of this application adopts a real-time detection mode, which does not require interruption of the normal operation of the seamless steel pipe production line, nor does it require stopping the steel pipe to be inspected. Relevant equipment can be directly installed on the existing seamless steel pipe conveying and production line to achieve tracking and inspection of seamless steel pipes in motion, adapting to continuous production line processes.

[0028] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0029] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein,

[0030] Figure 1 This is a flowchart illustrating the first embodiment provided in this application. Figure 2This is a flowchart illustrating the second embodiment provided in this application. Figure 3 This is a flowchart illustrating the third embodiment provided in this application. Figure 4 This is a schematic diagram of the structure of the fourth embodiment provided in this application; Figure 5 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0031] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0032] The following description, with reference to the accompanying drawings, describes a seamless steel pipe inspection method and system combining image and TOFD techniques according to embodiments of this application.

[0033] It should be noted that the seamless steel pipe inspection method combining image and TOFD provided in this application embodiment can be applied to a seamless steel pipe inspection system combining image and TOFD. The system includes an image detection unit, a TOFD detection unit and a processing unit. Along the seamless steel pipe conveying direction, the image detection unit is set at the cold rolling mill exit, and the TOFD detection unit is sequentially set downstream of the image detection unit.

[0034] It should be noted that, in the embodiments of this application, the above-mentioned image detection unit may include multiple image sensors (e.g., cameras), which are arranged in a ring to cover the entire circumference of the seamless steel pipe, thereby enabling full-coverage scanning of the steel pipe produced by the cold rolling mill and avoiding missed detections.

[0035] For example, a ring-shaped inspection frame is set up on the travel path of the seamless steel pipe at the cold rolling exit. Multiple industrial cameras are evenly arranged circumferentially on the ring-shaped inspection frame. Each camera is equipped with an LED (Light Emitting Diode) light source. The cameras continuously collect surface images of the seamless steel pipe during the conveying process from the cold rolling mill to the next process.

[0036] It should be noted that the TOFD detection unit may include at least one pair of TOFD probes. The probes are held at a fixed distance from the surface of the steel pipe by a clamping mechanism, and a stable coupling agent is provided by a configured coupling medium supply module.

[0037] Figure 1 This is a flowchart illustrating a first embodiment provided in this application. For example... Figure 1As shown, the method may include, but is not limited to, the following steps: S110: Obtain image data of the seamless steel pipe to be inspected based on the image detection unit.

[0038] For example, the image detection unit is controlled to detect the surface of the newly rolled seamless steel pipe to be inspected at the exit of the cold rolling mill in order to obtain the corresponding image data.

[0039] S120: Crack detection is performed based on image data to obtain the target location of the suspected crack and the first crack detection data.

[0040] For example, the image is first preprocessed by filtering, enhancement, and distortion correction, and then deep learning object detection networks such as YOLO are used to identify the image data to obtain the target location of possible cracks on the surface of the seamless steel pipe to be detected and the first crack detection data.

[0041] In the embodiments of this application, the aforementioned first crack detection data includes at least one of the following: crack length, crack width, morphological parameters, and texture orientation parameters. The crack morphological parameters include at least one of the following: crack area, crack aspect ratio, crack perimeter, and crack opening degree; the crack texture orientation parameters include at least one of the following: crack circumferential angle, crack axial dip angle, and crack extension direction.

[0042] In one implementation, the axial position and circumferential angle of the suspected crack can be obtained as the target position mentioned above. For an example, please refer to... Figure 2 , Figure 2 This is a flowchart illustrating the second embodiment provided in this application. Figure 2 As shown, step S120 may include the following steps: S1201: Preprocess the image data to obtain the image data to be detected.

[0043] For example, image data is filtered, enhanced, and distortion corrected to obtain the image data to be detected.

[0044] S1202: A pre-trained target detection network is used to detect the image data to be detected, and the pixel positions of the suspected cracks and the first crack detection data are obtained.

[0045] For example, a pre-trained target detection model is used to identify cracks in the image data to be detected, thereby obtaining the pixel locations of suspected cracks and the first crack detection data.

[0046] S1203: Perform position calibration based on pixel position, and obtain the axial position and circumferential angle as the target position.

[0047] For example, firstly, the conversion relationship between image pixels and actual physical size is established through pixel equivalent calibration. Then, the actual axial position of the crack from the end of the seamless steel pipe to be inspected is calculated by combining the pixel position of the crack with the pixel equivalent. The circumferential angle is calculated based on the crack's relative reference position, and the axial position and axial angle are used together as the target position of the crack.

[0048] For example, the following formula can be used for pixel equivalent calibration: k = L actual / L pixel Where k represents pixel equivalent, L actual L represents the actual length of the calibration object (i.e., the steel pipe to be tested). pixel This represents the number of pixels occupied by the calibration object in the image. The actual length and width of the crack are then calculated using the following formula: L crack = N pixel × k Among them, L crack N represents the actual length or width of the crack. pixel The number of pixels in the image representing the length or width of the crack.

[0049] In one alternative implementation, the circumferential angle can be determined by the following steps: calculating the offset of the crack in the circumferential direction of the steel pipe based on the pixel position; and converting the offset based on a preset reference orientation to obtain the circumferential angle.

[0050] For example, the circumferential angle can be calculated using the following formula: θ = (x offset / W image ) × 360 Where θ represents the circumferential angle of the crack, x offset W represents the pixel offset of the crack from the reference position. image This represents the total width of the circumferentially unfolded image of the seamless steel pipe.

[0051] S130: Obtain the transmission speed of the seamless steel pipe to be inspected.

[0052] Specifically, the transmission speed of the seamless steel pipe to be inspected on the production line is obtained.

[0053] S140: Perform delay compensation based on transmission speed to obtain TOFD detection delay.

[0054] For example, the TOFD detection delay is obtained by dividing the distance between the image detection unit and the TOFD detection unit by the transmission speed and then adding it to the signal processing time.

[0055] In one implementation, the above-mentioned delay compensation based on transmission speed to obtain the TOFD detection delay may include the following steps: performing speed smoothing filtering on the transmission speed to obtain a corrected speed; and obtaining the detection delay based on the corrected speed and the distance between the image detection unit and the TOFD detection unit.

[0056] For example, the transmission speed of the seamless steel pipe to be inspected is acquired in real time, and the acquired transmission speed is smoothed and filtered to obtain a corrected speed. Then, the detection delay is calculated based on the fixed distance between the image detection unit and the TOFD detection unit and the corrected speed.

[0057] S150: The TOFD detection unit, based on detection delay control, detects the target position and obtains the second crack detection data.

[0058] For example, the time when the location of the suspected crack reaches the detection position of the TOFD probe is added to the detection delay and used as the TOFD detection trigger time. At this trigger time, the TOFD detection unit is triggered to detect the target location and obtain the second crack detection data.

[0059] In the embodiments of this application, the second crack detection data includes at least one of the following: crack depth, crack height, crack tip position, crack extension length, and diffraction time difference between the upper and lower endpoints.

[0060] For example, the crack depth can be calculated using the following formula: H1 = (Δt × v) / 2 Where H1 represents the crack depth, Δt represents the propagation time difference between the direct wave and the diffracted wave at the tip, and v represents the longitudinal wave velocity of the ultrasound in the steel. The crack height can be calculated using any of the following formulas: H2 = (Δt upper-lower × v) / 2 H2 = h lower - h upper Where H2 represents the height of the crack itself, Δt upper-lower The time difference between the diffracted waves at the upper and lower tips is represented by ν, where ν represents the longitudinal wave velocity of the ultrasound in the steel, and h represents the longitudinal wave velocity. lower and h upper These represent the depths of the lower and upper endpoints, respectively.

[0061] In one alternative implementation, the refraction angle of the TOFD detection unit can be calculated using the following formula: sinθ1 / v1 = sinθ2 / v2 Where θ1 represents the incident angle in the wedge, θ2 represents the refraction angle in the steel, v1 represents the sound velocity in the wedge, and v2 represents the sound velocity in the steel.

[0062] S160: Input the first crack detection data and the second crack detection data into the pre-trained crack detection model, so that the crack detection model can extract features and perform weighted fusion on the first crack detection data and the second crack detection data, and combine them with the corresponding parameter thresholds to obtain the crack detection result.

[0063] In the embodiments of this application, the crack detection model has been pre-trained to perform feature extraction and weighted fusion on the first crack detection data and the second crack detection data respectively, and output the crack detection result in combination with the configured parameter threshold.

[0064] In one implementation, the obtained crack detection data can be fused using a pre-trained crack detection model. See [link to example]. Figure 3 , Figure 3 This is a flowchart illustrating the third embodiment provided in this application. Figure 3 As shown, step S160 may include the following steps: S1601: Obtain the first confidence level corresponding to the first crack detection data and the second confidence level corresponding to the second crack detection data.

[0065] For example, the confidence level of image detection is determined by the target detection algorithm as the first confidence level, and the second confidence level is obtained by analyzing the ultrasonic diffraction signal obtained by TOFD detection.

[0066] S1602: Extract features from the first crack detection data to obtain the image feature vector.

[0067] For example, the first crack detection data is spliced ​​and normalized according to a fixed dimension, and each independent feature parameter is transformed into a continuous numerical sequence to form an image feature vector containing information on crack size, location, and morphology.

[0068] S1603: Extract features from the second crack detection data to obtain the TOFD feature vector.

[0069] For example, the signal characteristic parameters of the second crack detection data and the TOFD detection data are spliced ​​and normalized to obtain a TOFD feature vector that includes the internal size of the crack, the timing and propagation characteristics of the ultrasonic signal.

[0070] Among them, the above-mentioned signal characteristic parameters include at least one of the following: diffraction wave time difference and refraction angle.

[0071] S1604: Based on the first confidence level and the second confidence level, the image feature vector and the TOFD feature vector are weighted and fused to obtain the fused feature vector.

[0072] For example, the image feature vector is multiplied by a first confidence level, and the TOFD feature vector is multiplied by a second confidence level, and then the vectors are added together to obtain the fused feature vector.

[0073] S1605: Perform parameter analysis on the fused feature vector to obtain crack parameters.

[0074] For example, the crack detection model first decodes and removes redundancy from the fused feature vector to obtain the core effective features, then calculates the initial crack parameters based on the core effective features, and then performs error calibration on the initial parameters by combining the steel pipe specifications and equipment calibration coefficients, and outputs crack detection parameters including crack length, width, depth and circumferential angle.

[0075] S1606: Compare the crack parameters with the parameter thresholds to obtain the crack detection results.

[0076] For example, each crack parameter is compared with its corresponding preset threshold to determine the number and extent of exceeding the limit, and then the defect is classified according to its severity. For instance, if only one parameter slightly exceeds the limit, it is judged as a minor defect; if two or more parameters exceed the limit or a single parameter significantly exceeds the limit, it is judged as a moderate defect; if key parameters such as depth or width are severely exceeded or multiple parameters significantly exceed the limit, it is judged as a severe defect.

[0077] In one alternative implementation, if the defect is determined to be unqualified, an alarm is triggered, and the marking device is controlled to mark the corresponding position on the steel pipe surface according to the determined axial position and circumferential angle, while the defect information is uploaded to the quality management system.

[0078] By implementing the embodiments of this application, image detection data of the seamless steel pipe to be inspected can be obtained through an image detection unit. Crack detection is performed based on the image detection data to obtain the target location of the suspected crack and the first crack detection data. Then, the transmission speed of the steel pipe is obtained and time delay compensation is performed to obtain the TOFD detection time delay. Based on the time delay, the TOFD detection unit is controlled to detect the target location to obtain the second crack detection data. Finally, the two types of detection data are input into a pre-trained crack detection model. The model completes feature extraction and weighted fusion, and then combines the parameter threshold to obtain the crack detection result. This can realize the synergistic cooperation between vision and TOFD detection, effectively improving the accuracy and reliability of crack detection. Moreover, the detection method of this application adopts a real-time monitoring mode, which does not require interruption of the normal operation of the existing seamless steel pipe production line, nor does it require stopping the steel pipe to be inspected. The relevant equipment can be directly installed on the existing conveying and production line to realize the tracking and detection of seamless steel pipes in motion. It is fully adapted to the continuous operation process of the production line and realizes online crack detection of seamless steel pipes without stopping the machine.

[0079] In one implementation, the above parameter thresholds are determined through the following steps A1-A4: A1: Obtain the target steel pipe specifications and current production environment parameters for the seamless steel pipe to be tested.

[0080] In the embodiments of this application, the steel pipe specifications include at least one of the following: outer diameter, wall thickness, and steel type; the production environment parameters include at least one of the following: temperature and humidity.

[0081] A2: Parameter thresholds are obtained from a pre-configured threshold library based on the target steel pipe specifications and current production environment parameters.

[0082] The threshold library stores multiple sets of parameter thresholds, each set corresponding to a set of steel pipe specifications and production environment parameters.

[0083] For example, defect tests are conducted on seamless steel pipes of different specifications under different production environments in advance to obtain corresponding parameter thresholds, thereby establishing a parameter threshold library. Each parameter threshold in the library corresponds to a set of steel pipe specifications and production environment parameters, and different parameter thresholds correspond to different steel pipe specifications and / or production environment parameters. This allows the target steel pipe specification to be tested to be matched with the current production environment parameters, obtaining the parameter thresholds from the parameter threshold library as the base parameter thresholds.

[0084] A3: Obtain the first historical inspection data of qualified seamless steel pipes within the preset historical time period.

[0085] Among them, the specifications of qualified seamless steel pipes are the same as those of the target steel pipes.

[0086] For example, acquire multiple first historical inspection data of qualified seamless steel pipes produced within a preset historical period (e.g., the past 3 hours) that have the same steel pipe specifications as the seamless steel pipe to be inspected.

[0087] A4: Adjust the basic parameter thresholds based on the first historical detection data to obtain the parameter thresholds.

[0088] For example, the statistical distribution of each parameter index (e.g., mean, standard deviation, etc.) is calculated based on multiple first historical detection data, and the basic parameter thresholds are adjusted according to the above statistical distribution to obtain the parameter thresholds, thereby automatically adapting to production line drift.

[0089] In an alternative implementation, the above method may further include the following steps: acquiring second historical detection data according to a pre-configured detection volume or detection cycle; recalibrating the threshold based on the second historical detection data to obtain an updated threshold; and updating the threshold library using the updated threshold.

[0090] For example, according to the pre-configured testing volume or testing cycle, the second historical testing data obtained during the seamless steel pipe production process is collected; based on the second historical testing data, the original basic parameter thresholds of the parameter threshold library are recalibrated and recalculated in combination with the actual production environment to obtain updated parameter thresholds that are adapted to the current testing requirements; the updated parameter thresholds are used to replace the original corresponding thresholds in the threshold library to complete the update of the threshold library, thereby enabling the parameter threshold library to adapt to changes in the production environment and improve testing accuracy.

[0091] This application also provides a seamless steel pipe inspection system combining image processing and TOFD (Time-of-Flight Difference). See also... Figure 4 , Figure 4 This is a schematic diagram of the structure of the fourth embodiment provided in this application. (See attached diagram.) Figure 4 As shown, the system 400 includes: an image detection unit 401, a TOFD detection unit 402, and a processing unit 403. The image detection unit is located at the cold rolling mill exit, and the TOFD detection unit is located after the image detection unit. The processing unit 403 is used for: acquiring image detection data of the seamless steel pipe to be inspected based on the image detection unit; performing crack detection based on the image detection data to obtain the target location of the suspected crack and the first crack detection data; acquiring the transmission speed of the seamless steel pipe to be inspected; performing time delay compensation based on the transmission speed to obtain the TOFD detection time delay; controlling the TOFD detection unit to detect the target location based on the detection time delay to obtain the second crack detection data; inputting the first crack detection data and the second crack detection data into a pre-trained crack detection model, so that the crack detection model can perform feature extraction and weighted fusion on the first crack detection data and the second crack detection data, and combine the corresponding parameter thresholds to obtain the crack detection result.

[0092] In one implementation, the processing unit 403 can be used to: preprocess the image detection data to obtain the image data to be detected; use a pre-trained target detection network to detect the image data to be detected to obtain the pixel position of the suspected crack and the first crack detection data; perform position calibration based on the pixel position to obtain the axial position and circumferential angle as the target position.

[0093] In one alternative implementation, the processing unit 403 can be used to determine the circumferential angle by the following steps: calculating the offset of the crack in the circumferential direction of the steel pipe based on the pixel position; and converting the offset based on a preset reference orientation to obtain the circumferential angle.

[0094] In one implementation, the processing unit 403 can be used to: perform speed smoothing filtering on the transmission speed to obtain a corrected speed; and based on the corrected speed, combine the distance between the image detection unit and the TOFD detection unit to obtain the detection delay.

[0095] In one implementation, the crack detection model processes first crack detection data and second crack detection data through the following steps to obtain crack detection results: obtaining a first confidence level corresponding to the first crack detection data and a second confidence level corresponding to the second crack detection data; extracting features from the first crack detection data to obtain an image feature vector; extracting features from the second crack detection data to obtain a TOFD feature vector; weighted fusing the image feature vector and the TOFD feature vector based on the first and second confidence levels to obtain a fused feature vector; parsing the parameters of the fused feature vector to obtain crack parameters; and comparing the crack parameters with parameter thresholds to obtain crack detection results.

[0096] In one alternative implementation, the processing unit 403 may be used to: acquire image detection quality parameters of the image detection data; acquire a first confidence level based on the image detection quality parameters; acquire ultrasonic detection signal feature parameters corresponding to the second crack detection data; and acquire a second confidence level based on the ultrasonic detection signal feature parameters.

[0097] In one implementation, the processing unit 403 can be used to: acquire the target steel pipe specifications and production environment parameters of the seamless steel pipe to be inspected; obtain basic parameter thresholds from a pre-configured threshold library based on the target steel pipe specifications and production environment parameters; wherein the threshold library stores multiple sets of parameter thresholds, each set of parameter thresholds corresponding to a set of steel pipe specifications and production environment parameters; acquire the first historical inspection data of qualified seamless steel pipes within a preset historical period; wherein the steel pipe specifications of the qualified seamless steel pipes are the same as the target steel pipe specifications; and adjust the basic parameter thresholds based on the first historical inspection data to obtain the parameter thresholds.

[0098] In an alternative implementation, the processing unit 403 may also be used to: acquire second historical detection data according to a pre-configured detection quantity or detection cycle; perform threshold recalibration based on the second historical detection data to obtain an updated threshold; and update the threshold library using the updated threshold.

[0099] In one implementation, the first crack detection data includes at least one of the following: crack length, crack width, morphological parameters, and texture parameters; the second crack detection data includes at least one of the following: crack depth and crack height.

[0100] The system of this application embodiment can acquire image detection data of the seamless steel pipe to be detected through the image detection unit, perform crack detection based on the image detection data to obtain the target location of the suspected crack and the first crack detection data, then acquire the steel pipe transmission speed and perform time delay compensation to obtain the TOFD detection delay, control the TOFD detection unit to detect the target location based on the time delay to obtain the second crack detection data, and finally input the two types of detection data into the pre-trained crack detection model. The model completes feature extraction and weighted fusion, and then combines the parameter threshold to obtain the crack detection result. It can realize the synergistic cooperation between vision and TOFD detection, effectively improving the accuracy and reliability of crack detection.

[0101] It should be noted that the foregoing explanation of the embodiment of the seamless steel pipe inspection method combining image and TOFD also applies to the image and TOFD combined inspection device of this embodiment, and will not be repeated here.

[0102] This application also proposes a seamless steel pipe produced by a seamless steel pipe inspection method or system combining images and TOFD provided in any embodiment of this application.

[0103] To implement the above embodiments, this application also proposes an electronic device. Please see [link to relevant documentation]. Figure 5 , Figure 5 This is a schematic diagram of the structure of the electronic device provided in an embodiment of this application. For example... Figure 5 As shown, the electronic device 500 includes: a processor 501 and a memory 502 communicatively connected to the processor 501; the memory 502 stores computer-executable instructions; the processor 501 executes the computer-executable instructions stored in the memory to implement the method provided in the foregoing embodiments.

[0104] To implement the above embodiments, this application also proposes a storage medium storing instructions that, when executed on an electronic device, cause the electronic device to perform the methods provided in the foregoing embodiments.

[0105] To implement the above embodiments, this application also proposes a program product, including at least one of a program and instructions, wherein when the program and instructions are executed by an electronic device, they implement the steps of the method provided in the foregoing embodiments.

[0106] In the foregoing descriptions of the embodiments, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0107] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0108] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0109] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0110] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0111] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by a program instructing related hardware, and the program can be stored in a computer-readable storage medium. When executed, the program includes one or a combination of the steps of the method embodiments.

[0112] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0113] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A method for inspecting seamless steel pipes using a combination of image processing and Time-of-Flight (TOFD) analysis, characterized in that: include: Image detection data of the seamless steel pipe to be inspected is obtained based on an image detection unit; wherein, the image detection unit is set at the exit of the cold rolling mill for seamless steel pipes; Crack detection is performed based on the image detection data to obtain the target location of the suspected crack and the first crack detection data; Obtain the transmission speed of the seamless steel pipe to be tested; Based on the transmission speed, delay compensation is performed to obtain the TOFD detection delay; The TOFD detection unit controls the detection delay to detect the target position and obtain second crack detection data; wherein, the TOFD detection unit is located downstream of the image detection unit. The first crack detection data and the second crack detection data are input into a pre-trained crack detection model, which performs feature extraction and weighted fusion on the first crack detection data and the second crack detection data, and combines the corresponding parameter thresholds to obtain the crack detection result.

2. The method according to claim 1, characterized in that, The crack detection based on the image detection data, to obtain the target location of the suspected crack and the first crack detection data, includes: The image detection data is preprocessed to obtain the image data to be detected; A pre-trained target detection network is used to detect the image data to be detected, thereby obtaining the pixel position of the suspected crack and the first crack detection data; Position calibration is performed based on the pixel position to obtain the axial position and circumferential angle as the target position.

3. The method according to claim 2, characterized in that, The circumferential angle is determined through the following steps: Based on the pixel position, the offset of the crack in the circumferential direction of the steel pipe is calculated. The offset is converted based on a preset reference orientation to obtain the circumferential angle.

4. The method according to claim 1, characterized in that, The step of performing delay compensation based on the transmission speed to obtain the TOFD detection delay includes: The transmission speed is subjected to speed smoothing filtering to obtain the corrected speed; Based on the corrected speed, and combined with the distance between the image detection unit and the TOFD detection unit, the detection delay is obtained.

5. The method according to claim 1, characterized in that, The crack detection model processes the first crack detection data and the second crack detection data through the following steps to obtain the crack detection result: Obtain the first confidence level corresponding to the first crack detection data and the second confidence level corresponding to the second crack detection data; Feature extraction is performed on the first crack detection data to obtain an image feature vector; Feature extraction is performed on the second crack detection data to obtain the TOFD feature vector; Based on the first confidence level and the second confidence level, the image feature vector and the TOFD feature vector are weighted and fused to obtain a fused feature vector; The crack parameters are obtained by parsing the fused feature vector. The crack parameters are compared with the parameter thresholds to obtain the crack detection results.

6. The method according to claim 5, characterized in that, The step of obtaining the first confidence level corresponding to the first crack detection data and the second confidence level corresponding to the second crack detection data includes: Obtain the image detection quality parameters of the image detection data; The first confidence level is obtained based on the image detection quality parameters; Obtain the characteristic parameters of the ultrasonic detection signal corresponding to the second crack detection data; The second confidence level is obtained based on the characteristic parameters of the ultrasonic detection signal.

7. The method according to claim 1, characterized in that, The parameter threshold is determined through the following steps: Obtain the target steel pipe specifications and production environment parameters of the seamless steel pipe to be tested; Based on the target steel pipe specifications and the production environment parameters, basic parameter thresholds are obtained by matching from a pre-configured threshold library; wherein, the threshold library stores multiple sets of parameter thresholds, and each set of parameter thresholds corresponds to a set of steel pipe specifications and production environment parameters; Obtain the first historical test data of qualified seamless steel pipes within a preset historical period; wherein, the specifications of the qualified seamless steel pipes are the same as the specifications of the target steel pipes; The basic parameter threshold is adjusted based on the first historical detection data to obtain the parameter threshold.

8. The method according to claim 7, characterized in that, The method further includes: Obtain the second historical detection data according to the pre-configured detection volume or detection cycle; The threshold is recalibrated based on the second historical detection data to obtain the updated threshold. The threshold library is updated using the updated threshold.

9. The method according to claim 1, characterized in that, The first crack detection data includes at least one of the following: crack length, crack width, morphological parameters, and texture parameters; The second crack detection data includes at least one of the following: crack depth and crack height.

10. A seamless steel pipe, characterized in that, The seamless steel pipe is produced by the testing method as described in any one of claims 1-9.

Images