A flange bolt abnormal installation sequence identification method, device and equipment and a readable storage medium

Abstract

The application discloses a flange bolt abnormal installation sequence identification method, device, equipment and readable storage medium in the safety production technical field, and can complete all analysis work in a time much less than that required by manual review after the installation video is recorded, and automatically compares and checks the recorded actual operation sequence with the standard procedure, so that the abnormal installation sequence is detected. The application mode of the rapid verification ensures the quality and safety of the assembly process, can help the on-site managers and workers to immediately take corrective measures, and eliminates potential structural hidden dangers before the equipment is put into use.

Description

Technical Field

[0001] This application relates to the field of safety production technology, and in particular to a method, device, equipment and readable storage medium for identifying abnormal installation sequence of flange bolts. Background Technology

[0002] In oil and gas drilling and production (well site) operations, wellhead safety control is a critical aspect of ensuring the smooth operation. Blowout preventer (BOP) assemblies, as an important component of the well control system, function to control and prevent high-pressure fluids from gushing to the surface. Given the extreme pressures of the operating conditions and the hazardous nature of the downhole media, a failure or malfunction of the BOP system can directly lead to a blowout, causing incalculable casualties, enormous economic losses, and severe environmental disasters.

[0003] Therefore, the standardized installation and structural integrity of the blowout preventer (BOP) system are essential prerequisites for ensuring the safety of the entire drilling and production operation. Components of critical well control units such as the BOP are primarily connected and stacked with high strength using large flanges and flange bolts. In such high-pressure sealing applications, the quality of flange bolt tightening is crucial, and standard procedures such as cross-diagonal tightening must be strictly followed to ensure uniform stress on the flange and reliable gasket sealing.

[0004] Existing industrial safety monitoring methods analyze operator status and overall compliance with regulations. Specifically, the core objectives of these methods include assessing operator safety status (e.g., whether helmets, reflective vests, and other personal protective equipment are worn), overall behavior (e.g., falls, climbing, boundary crossings), or physiological state (e.g., fatigue, drowsiness). These methods are characterized by relatively low granularity, relying primarily on the features of the human skeleton, key points, or the overall target (person / object) for judgment. However, in practical applications, these industrial safety monitoring methods still fail to achieve effective safety monitoring and quality control, leaving safety hazards unresolved.

[0005] In conclusion, how to effectively solve problems such as ensuring the fastening quality of flange bolts is a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] The purpose of this application is to provide a method, device, equipment, and readable storage medium for identifying abnormal installation sequence of flange bolts. By detecting abnormal installation sequence of flange bolts during the installation process, the tightening quality of flange bolts can be ensured, further guaranteeing the functional stability of blowout preventers based on flange and flange bolt tightening, and reducing safety risks.

[0007] To solve the above-mentioned technical problems, this application provides the following technical solution.

[0008] A method for identifying abnormal installation sequence of flange bolts includes: acquiring installation video segments of multiple flange bolts simultaneously captured from multiple perspectives; fixing the multiple flange bolts to the same set of flanges, and capturing images of flange bolts with corresponding numbers from different perspectives; performing target detection on the installation video segments for the operating medium and flange bolts to obtain detection results; determining the inscribed circle of the detection frame corresponding to the flange bolt as the bolt circle; for flange bolts located above the flange, extending the upper half of the bolt circle upward along the longitudinal axis by a specified length to obtain a bolt capsule; for flange bolts located below the flange, extending the lower half of the bolt circle downward along the longitudinal axis by a specified length to obtain a bolt capsule; performing overlap detection on the detection frame of the bolt capsule and the operating medium, and determining the installation sequence in conjunction with the flange bolt numbers; comparing the installation sequence with standard procedures to detect abnormal installation sequences.

[0009] Preferably, the overlap detection of the bolt capsule and the detection frame of the operating medium is performed, and the installation sequence is determined in combination with the flange bolt number, including: creating a bolt capsule mask based on the bolt capsule, and creating an operating medium rectangular mask based on the detection frame of the operating medium; performing a logical AND operation on the bolt capsule mask and the operating medium rectangular mask to obtain an intersection mask; using the intersection mask to determine the overlap area of ​​the detection frame of the bolt capsule and the operating medium; determining the bolt capsule corresponding to the maximum coverage at each time step as the operating bolt according to the time sequence; and obtaining the installation sequence based on the number of the operating bolt.

[0010] Preferably, acquiring installation video segments of multiple flange bolts simultaneously captured from multiple perspectives includes: acquiring original installation videos of multiple flange bolts simultaneously captured from multiple different perspectives; extracting installation video segments of installation time periods from the multiple original installation videos; and performing frame extraction processing on the installation video segments.

[0011] Preferably, extending the upper half of the bolt circle upward along the longitudinal axis by a specified length to obtain a bolt capsule includes: extending the upper half of the bolt circle upward along the longitudinal axis by a radius length to obtain a bolt capsule.

[0012] Preferably, the installation video segment is subjected to target detection of the operating medium and flange bolts to obtain detection results, including: target detection of the hand, torque wrench and flange bolts in the installation video segment to obtain detection results.

[0013] Preferably, the target detection of flange bolts in the installation video segment includes: detecting flange bolts in the installation video segment; after detecting flange bolts located above the flange, shifting the detection frame of the flange bolt downwards by two flange widths to obtain the detection frame corresponding to the flange bolt located below the flange; wherein, the flange width is the width of the rectangular detection frame of the flange obtained by target detection of the flange.

[0014] Preferably, it also includes: outputting a reinstallation prompt message when an abnormal installation sequence is detected.

[0015] A device for identifying abnormal installation sequence of flange bolts includes: a video acquisition module for acquiring installation video segments of multiple flange bolts simultaneously captured from multiple perspectives; multiple flange bolts are fixed to the same set of flanges, and flange bolts with corresponding numbers are captured from different perspectives; a target detection module for performing target detection on the installation video segments of the operating medium and flange bolts to obtain detection results; a bolt capsule creation module for determining the inscribed circle of the detection frame corresponding to the flange bolt as the bolt circle, and for flange bolts located above the flange, extending the upper half of the bolt circle upward along the longitudinal axis by a specified length to obtain a bolt capsule, and for flange bolts located below the flange, extending the lower half of the bolt circle downward along the longitudinal axis by a specified length to obtain a bolt capsule; an overlap detection module for performing overlap detection on the bolt capsule and the detection frame of the operating medium, and determining the installation sequence in conjunction with the flange bolt numbers; and an abnormal installation sequence detection module for comparing the installation sequence with standard procedures to detect abnormal installation sequences.

[0016] An electronic device includes: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the above-described method for identifying abnormal installation sequence of flange bolts.

[0017] A readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described method for identifying abnormal installation sequence of flange bolts.

[0018] Using the method provided in the embodiments of this application, installation video segments of multiple flange bolts are simultaneously captured from multiple perspectives; multiple flange bolts are fixed to the same set of flanges, and flange bolts with corresponding numbers are captured from different perspectives; target detection of the operating medium and flange bolts is performed on the installation video segments to obtain the detection results; the inscribed circle of the detection frame corresponding to the flange bolt is determined as the bolt circle; for flange bolts located above the flange, the upper half of the bolt circle is extended upward along the longitudinal axis by a specified length to obtain a bolt capsule; for flange bolts located below the flange, the lower half of the bolt circle is extended downward along the longitudinal axis by a specified length to obtain a bolt capsule; overlap detection is performed on the detection frame of the bolt capsule and the operating medium, and the installation sequence is determined in combination with the flange bolt numbers; the installation sequence is compared with the standard procedure to detect abnormal installation sequences.

[0019] This application takes into account that any deviation in the bolt tightening sequence may lead to uneven distribution of preload, thereby affecting the sealing performance of the flange connection. Given the high-risk nature of blowout preventer systems and the extremely high requirements for assembly quality, computer vision-based industrial safety monitoring typically focuses on identifying personnel's overall posture, movement trajectory, or the wearing of personal protective equipment for safety supervision, which is no longer sufficient to meet actual safety supervision needs.

[0020] Therefore, this application uses the entire process of on-site workers installing flange bolts as input. The actual bolt tightening sequence is automatically analyzed and recorded. By comparing this sequence with standard procedures, abnormal installation sequences can be detected. This avoids traditional manual inspections or time-consuming video playback, efficiently delegating the tedious sequence comparison work to computer equipment.

[0021] Specifically, firstly, multiple installation video segments of flange bolts are acquired simultaneously from multiple perspectives. These flange bolts are used to fix the same set of flanges, and the flange bolts with corresponding numbers are captured from different perspectives. This means that the global number of the detected flange bolts can be uniquely determined based on the video's perspective. Then, target detection of the operating medium and flange bolts is performed on the installation video segments to obtain the detection results. The inscribed circle of the detection frame corresponding to the flange bolt is defined as the bolt circle. For flange bolts located above the flange, the upper half of the bolt circle is extended upwards along the longitudinal axis by a specified length to obtain a bolt capsule. For flange bolts located below the flange, the lower half of the bolt circle is extended downwards along the longitudinal axis by a specified length to obtain a bolt capsule. Considering that the center position of the bolt circle and the rectangular detection frame represents the position where the bolt is already installed, in actual operation, operators must use a torque wrench or their hands to apply force at a position far from the bolt circle's center. This results in a necessary structural positional offset between the detection frame of the operating medium and the bolt center. This leads to a generally low overlap between the tool boundary frame and the bolt circle, severely affecting the accuracy of bolt detection and causing a large number of missed detections. To address this, by deforming the inspection frame for flange bolts, especially by extending the bolt circle in different directions in conjunction with the inscribed circle and the specific installation characteristics corresponding to its location, it becomes easier to detect the flange bolts currently being operated. Specifically, the bolt capsule and the inspection frame of the operating medium are overlapped for inspection, and the installation sequence is determined by combining the flange bolt number. Finally, the installation sequence is compared with the standard procedure to detect any abnormal installation sequences.

[0022] As can be seen, this application can complete all analysis work in far less time than manual review after the installation video is recorded, and automatically compare and verify the recorded actual operation sequence with standard procedures, thereby detecting abnormal installation sequences. This rapid verification application mode ensures the quality and safety of the assembly process, and can help on-site managers and workers take immediate corrective measures to eliminate potential structural hazards before the equipment is put into use.

[0023] Accordingly, embodiments of this application also provide a flange bolt abnormal installation sequence identification device, equipment, and readable storage medium corresponding to the above-described flange bolt abnormal installation sequence identification method, which have the above-described technical effects, and will not be elaborated further here. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a flowchart illustrating the implementation of a method for identifying abnormal installation sequence of flange bolts in an embodiment of this application.

[0026] Figure 2 This is a schematic diagram showing the initial connection of a flange in one of the embodiments of this application;

[0027] Figure 3 This is a schematic diagram of a flange after it has been tightened, according to an embodiment of this application.

[0028] Figure 4 This is a top-view schematic diagram of a flange in an embodiment of this application;

[0029] Figure 5 This is a schematic diagram illustrating the prediction of the bolt position below the flange in an embodiment of this application;

[0030] Figure 6 This is a front view of a flange after installation, as described in one embodiment of this application.

[0031] Figure 7 This is a schematic diagram of a bolt target detection using a rectangular detection frame in an embodiment of this application;

[0032] Figure 8 This is a schematic diagram showing the correspondence between a bolt circle and a rectangular detection frame of the bolt in an embodiment of this application;

[0033] Figure 9 This is a schematic diagram of bolt target detection in a circular bolt shape according to an embodiment of this application;

[0034] Figure 10 This is a schematic diagram showing the overlap between a bolt circle and an operating medium detection frame in an embodiment of this application;

[0035] Figure 11 This is a schematic diagram of single bolt target detection in a bolt circle shape according to an embodiment of this application;

[0036] Figure 12 This is a schematic diagram of a single bolt target detection in the form of a bolt capsule, as described in an embodiment of this application.

[0037] Figure 13 This is a schematic diagram showing the overlap between a bolt capsule and an operating medium detection frame in an embodiment of this application;

[0038] Figure 14 This is a schematic diagram of an overlap detection method in an embodiment of this application;

[0039] Figure 15 This is a schematic diagram of a standard flange bolt manufacturing process in an embodiment of this application;

[0040] Figure 16 This is a schematic diagram of flange bolt numbering in an embodiment of this application;

[0041] Figure 17 This is a schematic diagram of the structure of a flange bolt abnormal installation sequence identification device in an embodiment of this application;

[0042] Figure 18 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application;

[0043] Figure 19 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0044] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0045] Please refer to Figure 1 , Figure 1 This is a flowchart of a method for identifying abnormal installation sequence of flange bolts according to an embodiment of this application. The method includes the following steps.

[0046] S101. Acquire video clips of the installation of multiple flange bolts simultaneously captured from multiple perspectives.

[0047] In this process, multiple flange bolts are used to fix the same set of flanges, and the flange bolts with corresponding numbers are photographed from different angles.

[0048] Please refer to Figure 2 and Figure 3 A flange is a disc-shaped structure used to connect pipes, valves, and other equipment. Two flanges are fastened together by multiple flange bolts, and a gasket is placed between the two flanges to achieve a seal. Please refer to [reference needed]. Figure 4 The flange bolts are symmetrically distributed on the flange.

[0049] In the actual installation and fastening of flanges, for effective safety monitoring, multiple fixed-angle image acquisition devices can be used to simultaneously capture the installation process of multiple flange bolts. It is important to note that different angles can be used to capture flange bolts with corresponding numbers. For example, if there are 20 flange bolts, four fixed-angle cameras (e.g., cameras placed at equal distances and orthogonal) can be used. Within the frame of each camera, the flange bolts with the same number are captured. For instance, in camera 1, the flange bolts numbered 1, 2, 3, 4, and 5 are selected to be captured from the center of the field of view to the left and right (in practice, the numbers of adjacent flange bolts can be non-consecutive, but the bolt numbers on the same flange must be unique; also, the actual number of flange bolts that can be captured by camera 1 is greater than 5; here, for effective recording, only the bolts in the middle position with the best field of view are recorded). In other words, there are multiple bolts in the field of view of camera 1. Using the center line of the field of view as a reference, the bolts on the center line, plus the two bolts on the left and two bolts on the right near the center line, these five bolts are the bolts recorded from this angle. Thus, given the camera's angle of view and the position of the flange bolt within the camera, the bolt number can be uniquely determined.

[0050] In one specific embodiment of this application, acquiring installation video segments of multiple flange bolts simultaneously captured from multiple perspectives includes: acquiring original installation videos of multiple flange bolts simultaneously captured from multiple different perspectives; extracting installation video segments of installation time periods from the multiple original installation videos; and performing frame extraction processing on the installation video segments.

[0051] Specifically, since the blowout preventer is composed of multiple large flanges stacked together, there are many obstacles overhead when shooting from above. Therefore, at the installation site, multiple cameras can be evenly fixed around the flanges to simultaneously film the flange bolt installation process from multiple angles. This approach allows each camera to focus only on the few bolts with the clearest and most obvious spacing within its field of view, significantly reducing the false bolt detection rate.

[0052] In addition, the start time of video recording under multiple perspectives is synchronized, and keyframes are extracted at equal time intervals to ensure time synchronization under different perspectives. Specifically, since the start time of cameras with different fixed perspectives in a construction site usually varies, if the video is analyzed directly, the images from different perspectives will correspond to different actual time points, making it impossible to compare the multi-view detection results at the same moment. Therefore, the camera that started last can be used as the time reference, and the segments in the videos of other cameras that are earlier than that time point can be discarded, so that all videos have a unified start time on the timeline. Then, frames are extracted from each video at the same time interval to ensure that the images from different perspectives correspond one-to-one at the same moment, thereby achieving time synchronization of multi-view videos and providing a reliable temporal consistency basis for subsequent joint analysis of bolt operation behavior. Of course, the camera that shut down first can also be used as the end time, and the start time can be determined backward from the preset installation duration to extract the installation video segment corresponding to the installation time period.

[0053] S102. Perform target detection on the operating medium and flange bolts of the installation video segment to obtain the detection results.

[0054] Once the installation video segment is obtained, the operating medium and flange bolts appearing in the installation video segment can be targeted for detection, thereby obtaining the corresponding detection results.

[0055] Target detection is an image processing and computer vision technique that automatically identifies the presence and precise spatial location of specific objects in an image (e.g., bolts and operating media in this embodiment, such as hands and torque wrenches). The core output of this process is a traditional rectangular bounding box (bbox), which is a minimum axis-aligned rectangle used to roughly define the edges and contours of each identified target.

[0056] In one specific embodiment of this application, target detection of the operating medium and flange bolts is performed on the installation video segment to obtain detection results, including: target detection of the hand, torque wrench and flange bolts on the installation video segment to obtain detection results.

[0057] Specifically, flange bolts are usually tightened by hand first, and then tightened again with a torque wrench. Therefore, the operating medium can be hand or torque wrench.

[0058] A torque wrench is a specialized tool that can be set and precisely controlled to tighten bolts. It will make a sound or slip when the specified torque value is reached.

[0059] During flange installation, for a single frame image, to determine which bolt the worker is tightening, it is necessary to detect the position of the operating medium and the bolt. If the rectangular detection frame of the operating medium has a large overlap with the area of ​​a bolt with a certain number, it is assumed that the worker is tightening that bolt. Based on the extracted keyframes, a target detection model is used to obtain the corresponding detection frames and position coordinates.

[0060] In one specific embodiment of this application, target detection of flange bolts in an installation video segment includes: detecting flange bolts in the installation video segment; after detecting flange bolts located above the flange, shifting the detection frame of the flange bolt downwards by two flange widths to obtain the detection frame corresponding to the flange bolt located below the flange; wherein, the flange width is the width of the rectangular detection frame of the flange obtained by target detection of the flange.

[0061] Given that the bolts below the flange were not installed during the initial installation, if Figure 5 As shown. Therefore, target detection cannot draw the corresponding rectangular detection box. The width of the rectangular detection box for the flange can be obtained by performing target detection on the flange. In this embodiment, a strategy of shifting the upper detection box downwards by two flange widths is adopted to predict the position of the lower bolts, thereby detecting the positions of all flange bolts, i.e., the detection boxes corresponding to the flange bolts. For example... Figure 6 As shown, Figure 6 This is a front view of the flange after installation, showing bolts on both the top and bottom.

[0062] S103. Determine the inscribed circle of the detection frame corresponding to the flange bolt as the bolt circle. For the flange bolt located above the flange, extend the upper half of the bolt circle upward along the longitudinal axis for a specified length to obtain the bolt capsule. For the flange bolt located below the flange, extend the lower half of the bolt circle downward along the longitudinal axis for a specified length to obtain the bolt capsule.

[0063] The bolt circle is a geometric representation method used to optimize the accuracy of bolt feature localization and recognition, aiming to overcome the shortcomings of rectangular detection boxes in handling closely arranged bolt targets. It is defined by replacing the rectangular detection boxes of bolt features obtained from multi-frame detection with the inscribed circle of those boxes. The fundamental purpose of this substitution is to address the technical challenge that rectangular detection boxes easily interfere with subsequent detection processes and cause a large number of false detections when adjacent bolts are too close together. By transforming the rectangular detection boxes into more compact circles, this method effectively increases the geometric distance between adjacent bolt features, thereby achieving clear isolation between targets and robustness in subsequent processing.

[0064] Among them, the bolt capsule is a geometric model of a non-standard operating area designed for optimized posture during industrial assembly operations. It aims to overcome the structural omissions inherent in bolt positioning methods during refined operational scenarios and to assist in accurate calculation of overlap. For example... Figure 7 , Figure 8 and Figure 9 As shown, the center of the bolt circle and rectangular inspection frame represents the position where the bolt has been installed. However, in actual operation, operators must use a torque wrench or their hands to apply force away from the bolt's center, resulting in a necessary structural positional offset between the inspection frame and the bolt's center. For example... Figure 10 (Where M1 is the operation medium detection frame, such as a hand or torque wrench, M4 is the bolt circle, and M3 is the overlapping area) as shown, so that the tool boundary frame (such as...) Figure 10 The overlap between the M1 bolt and the bolt circle is generally low, which seriously affects the accuracy of the judgment of the operating bolt and causes a large number of missed inspections.

[0065] Considering that the center of the bolt circle and the rectangular detection frame represents the position where the bolt is already installed, in actual operation, operators must use a torque wrench or their hands to apply force away from the bolt's center. This results in a structural positional offset between the detection frame and the bolt's center. Consequently, the overlap between the tool boundary frame and the bolt circle is generally low, severely affecting the accuracy of bolt detection and causing numerous missed detections. To address this, deforming the flange bolt detection frame, especially by extending the bolt circle in different directions in conjunction with the inscribed circle and its specific installation characteristics, can better facilitate the detection of the flange bolts currently being operated on.

[0066] In one specific embodiment of this application, extending the upper half of the bolt circle upward along the longitudinal axis by a specified length to obtain a bolt capsule includes: extending the upper half of the bolt circle upward along the longitudinal axis by a radius length to obtain a bolt capsule.

[0067] Specifically, for the bolts above the flange, extend the upper semicircle of the bolt circle upwards along the longitudinal axis by a radius length; for example... Figure 11 As shown, for the bolts below the flange, the lower semicircle is extended downwards by a radius (in practical applications, this length can also be adjusted to other lengths based on experience or specific implementation), resulting in the following: Figure 12 The bolt capsule shown illustrates this directional geometric compensation strategy. This expands the effective working area of ​​the bolt, ensuring a high degree of match with the actual contact area of ​​the operating medium detection frame, thus aiding in subsequent overlap calculations.

[0068] S104. Perform overlap detection on the bolt capsule and the detection frame of the operating medium, and determine the installation sequence in conjunction with the flange bolt number.

[0069] like Figure 13 As shown (M1 is the detection frame for the operating medium, M2 is the bolt capsule, and M3 is the overlap area after processing and reasonable enlargement), after identifying the detection frames for the bolt capsule and the operating medium, overlap detection can be performed on the two. High overlap is sufficient to determine the flange bolt currently being operated. Then, the actual installation sequence is determined based on the number of the flange bolt being operated.

[0070] In one specific embodiment of this application, overlap detection is performed on the detection frames of the bolt capsule and the operating medium, and the installation sequence is determined in conjunction with the flange bolt numbers. This includes: creating a bolt capsule mask based on the bolt capsule, and creating an operating medium rectangular mask based on the detection frames of the operating medium; performing a logical AND operation on the bolt capsule mask and the operating medium rectangular mask to obtain an intersection mask; using the intersection mask to determine the overlap area of ​​the detection frames of the bolt capsule and the operating medium; determining the bolt capsules corresponding to the maximum coverage at each time point as the operating bolts according to the time sequence; and obtaining the installation sequence based on the operating bolt numbers.

[0071] Specifically, the bolt number being tightened is determined based on the degree of overlap between the hand or torque wrench and the bolt capsule area. The bolt being tightened overlaps significantly with the operating medium (hand or torque wrench), or is completely obscured by the operating medium. In this embodiment, this is used to determine the bolt number being tightened by the operator.

[0072] However, in real-world working conditions, the operating medium has a high probability of overlapping with multiple bolt capsule areas. Therefore, in this embodiment, the overlap of multiple bolts can be quantitatively calculated and sorted according to the overlap ratio. The bolt with the highest overlap ratio is selected as the bolt being tightened in this frame image. Specifically, a pixel-mask-based calculation method is used to accurately quantify the degree of overlap between the hand and the bolt determination area. This calculation process is performed on a binary mask with the same size as the image to ensure the accuracy of the determination.

[0073] First, create two independent binary masks: Bolt Capsule Mask M Capsule and the rectangular detection box mask M of the operating medium tool (That is, the rectangular mask used for manipulating the medium). All masks are initialized with background pixel values ​​set to 0. For example... Figure 14 As shown, bolt capsule mask M Capsule (Figure M2) Based on the preset bolt geometry information and partition labels, the pixel value corresponding to the capsule-shaped area is set to 255 using a geometric fill function; similarly, the media bounding box mask M is manipulated. tool (Figure M1) also sets the pixel value within the detected rectangular area to 255.

[0074] The input image size is represented as H×W. Two binary mask matrices with the same size as the image are defined: For the bolt capsule area: For the bounding box region of the operating medium: .

[0075] Subsequently, the overlapping area (as shown in the figure, M3; however, in practical applications, the shape and size of M3 will vary depending on the different overlap conditions of the bolt capsule and the bounding box of the operating medium, which will not be listed here) is calculated by performing a pixel-level logical AND operation on the two masks. This operation will generate a third mask M. intersection This refers to the intersection mask. A pixel is considered to be in the M region only if it is simultaneously located within both the bolt capsule region and the operating medium bounding box region. intersection The value in the range is 255 only if it is true; otherwise, it is 0.

[0076] The intersection mask is obtained through pixel-level logical AND operations: ;Right now: .

[0077] Furthermore, the same operating medium bounding box may overlap with different bolt capsules at the same time. When calculating the overlap area, the overlap area between the operating medium bounding box and each overlapping bolt capsule can be calculated separately. For example, if the operating medium bounding box overlaps with two bolt capsules (such as No. 4 and No. 5) at the same time, the overlap area between bolt capsule No. 4 and the operating medium bounding box can be calculated separately, and the overlap area between bolt capsule No. 5 and the operating medium bounding box can also be calculated separately.

[0078] Ultimately, the overlap area, i.e., the intersection area between the operating medium and the bolt capsule region, is obtained by counting the number of pixels with a value of 255 in the intersection mask (Mintersection). Dividing this total number of pixels by the total number of pixels (MCapsule) in the bolt capsule region yields the final overlap ratio (Coverage), which serves as the core indicator for measuring the degree of matching between the operating medium and the target bolt.

[0079] Bolt capsule area: Area of ​​the intersection region: The overlap ratio between the operating medium and the bolt capsule area is defined as the ratio of the intersection area to the bolt capsule area. Select the bolt number corresponding to the bolt area with the highest coverage as the bolt that the operator is tightening.

[0080] By integrating the recording results from all perspectives, the installation sequence of bolt numbers is output, i.e., the installation order. Specifically, since this embodiment uses multiple cameras to capture multi-view data and ensures consistency of images from all perspectives at the same time through time synchronization, integrating the recording results from all perspectives involves constructing a complete installation sequence through temporal reasoning. During the analysis of the entire video sequence, the system continuously tracks the bolt numbers identified as "being tightened" in each synchronized time frame. When a bolt number is continuously identified as the object of operation in consecutive synchronized time frames, it is confirmed as a complete bolt tightening operation. Finally, all bolt numbers confirmed as having completed tightening operations are recorded sequentially according to time order, thus outputting the actual installation sequence of the flange bolts.

[0081] S105. Compare the installation sequence with the standard procedure to detect any abnormal installation sequences.

[0082] In this embodiment, non-standard tightening behavior of flange bolts, i.e., abnormal installation sequence, can be identified. Specifically, such as... Figure 15 As shown, the standard requirement for flange bolt installation is cross-diagonal tightening. The core principle is to tighten each pair of bolts sequentially in a cross-shaped pattern to achieve symmetrical force distribution and uniform clamping on the flange. This embodiment identifies anomalies by comparing the actual bolt numbering sequence with the preset cross-diagonal tightening sequence.

[0083] For example, during the flange bolt tightening process, the upper flange, with the bolts initially positioned, needs to be connected to the lower flange using a crane. Then, the actual installation and tightening of the bolts begins. After tightening each bolt, immediately tighten the bolts positioned opposite it on the circumference; these two opposing bolts form a diagonal bolt pair. Then, tighten another diagonal bolt pair perpendicular to the previous pair, ensuring that each pair of bolts is tightened sequentially in a cross-shaped sequence, thus achieving symmetrical force and uniform compression on the flange. Figure 15 As shown, bolts 1 and 2, and bolts 3 and 4 form a diagonal relationship, each forming a bolt pair. The two bolt pairs 1-2 and 3-4 form a cross shape, which is called an intersection.

[0084] In one specific embodiment of this application, the method further includes: outputting a reinstallation prompt message upon detecting an abnormal installation sequence. In practical applications, a standard tightening sequence model can be established first according to specifications. Subsequently, the actual recorded installation sequence is compared step by step with the standard sequence model. Once it is found that the bolt number actually tightened does not match the next expected number in the standard sequence model, an abnormal installation behavior can be identified and output, and the specific bolt number and time point of the abnormality can be indicated. This comparison mechanism can effectively identify any non-standard tightening behavior that violates the cross-diagonal tightening principle during worker operations.

[0085] It should be noted that in this application, the flange bolts located on the flange can be numbered sequentially in either a clockwise or counterclockwise direction, such as... Figure 16 The numbers are arranged in a counter-clockwise order as shown, or as follows: Figure 15 The bolts are numbered according to the diagonal tightening correspondence of the bolt tightening process. When comparing the installation sequence with the standard process, it is necessary to consider the numbering method of the flange bolts. For example, for bolts like... Figure 15 As shown in the numbering method, after tightening the flange bolts corresponding to numbers 1 and 2, tighten the flange bolts corresponding to numbers 3 and 4. However, for... Figure 16 The numbering method shown indicates that after tightening the flange bolts numbered 1 and 11, the flange bolts numbered 6 and 16 are tightened. Although the numerical values ​​of the numbers change in the installation sequence, they both conform to the standard procedure.

[0086] Using the method provided in the embodiments of this application, installation video segments of multiple flange bolts are simultaneously captured from multiple perspectives; multiple flange bolts are fixed to the same set of flanges, and flange bolts with corresponding numbers are captured from different perspectives; target detection of the operating medium and flange bolts is performed on the installation video segments to obtain the detection results; the inscribed circle of the detection frame corresponding to the flange bolt is determined as the bolt circle; for flange bolts located above the flange, the upper half of the bolt circle is extended upward along the longitudinal axis by a specified length to obtain a bolt capsule; for flange bolts located below the flange, the lower half of the bolt circle is extended downward along the longitudinal axis by a specified length to obtain a bolt capsule; overlap detection is performed on the detection frame of the bolt capsule and the operating medium, and the installation sequence is determined in combination with the flange bolt numbers; the installation sequence is compared with the standard procedure to detect abnormal installation sequences.

[0087] This application takes into account that any deviation in the bolt tightening sequence may lead to uneven distribution of preload, thereby affecting the sealing performance of the flange connection. Given the high-risk nature of blowout preventer systems and the extremely high requirements for assembly quality, computer vision-based industrial safety monitoring typically focuses on identifying personnel's overall posture, movement trajectory, or the wearing of personal protective equipment for safety supervision, which is no longer sufficient to meet actual safety supervision needs.

[0088] Therefore, this application uses the entire process of on-site workers installing flange bolts as input. The actual bolt tightening sequence is automatically analyzed and recorded. By comparing this sequence with standard procedures, abnormal installation sequences can be detected. This avoids traditional manual inspections or time-consuming video playback, efficiently delegating the tedious sequence comparison work to computer equipment.

[0089] Specifically, firstly, multiple installation video segments of flange bolts are acquired simultaneously from multiple perspectives. These flange bolts are used to fix the same set of flanges, and the flange bolts with corresponding numbers are captured from different perspectives. This means that the global number of the detected flange bolts can be uniquely determined based on the video's perspective. Then, target detection of the operating medium and flange bolts is performed on the installation video segments to obtain the detection results. The inscribed circle of the detection frame corresponding to the flange bolt is defined as the bolt circle. For flange bolts located above the flange, the upper half of the bolt circle is extended upwards along the longitudinal axis by a specified length to obtain a bolt capsule. For flange bolts located below the flange, the lower half of the bolt circle is extended downwards along the longitudinal axis by a specified length to obtain a bolt capsule. Considering that the center position of the bolt circle and the rectangular detection frame represents the position where the bolt is already installed, in actual operation, operators must use a torque wrench or their hands to apply force at a position far from the bolt circle's center. This results in a necessary structural positional offset between the detection frame of the operating medium and the bolt center. This leads to a generally low overlap between the tool boundary frame and the bolt circle, severely affecting the accuracy of bolt detection and causing a large number of missed detections. To address this, by deforming the inspection frame for flange bolts, especially by extending the bolt circle in different directions in conjunction with the inscribed circle and the specific installation characteristics corresponding to its location, it becomes easier to detect the flange bolts currently being operated. Specifically, the bolt capsule and the inspection frame of the operating medium are overlapped for inspection, and the installation sequence is determined by combining the flange bolt number. Finally, the installation sequence is compared with the standard procedure to detect any abnormal installation sequences.

[0090] As can be seen, this application can complete all analysis work in far less time than manual review after the installation video is recorded, and automatically compare and verify the recorded actual operation sequence with standard procedures, thereby detecting abnormal installation sequences. This rapid verification application mode ensures the quality and safety of the assembly process, and can help on-site managers and workers take immediate corrective measures to eliminate potential structural hazards before the equipment is put into use.

[0091] To facilitate those skilled in the art to better understand and implement the flange bolt abnormal installation sequence identification method provided in the embodiments of this application, the method will be described in detail below with reference to specific application scenarios.

[0092] The following example demonstrates a non-standard installation test using 20 flange bolts to fix the flange. In practical applications, the number of flange bolts used to fix the flange can vary depending on the actual fixing requirements, which will not be elaborated here.

[0093] The flange bolt abnormal installation sequence identification method provided in this application embodiment can identify non-standard fastening of flange bolts based on bolt capsules. The specific implementation steps include: S1 data acquisition; S2 video time synchronization and key frame extraction; S3 operation medium target detection; S4 bolt capsule area construction and bolt numbering; S5 identification of operation bolts based on overlap; S6 integration and output of fastening results; S7 standard comparison and abnormal behavior monitoring.

[0094] For S1 data acquisition, four identical industrial-grade recorders are evenly fixed around the flange to record the entire process of installing and tightening the flange. Each recorder is responsible for recording the operation of five bolts designated in the center of the field of view. Recorders one, two, three, and four are responsible for recording the operation of bolts 1-5, 6-10, 11-15, and 16-20, respectively. To effectively identify non-standard bolt tightening behavior, this embodiment deliberately simulates various bolt tightening sequences by manually guiding the operator to tighten specific bolts during data acquisition. These tightening sequences include standard tightening sequences as well as various non-standard tightening sequences (such as non-standard cross-tightening or diagonal tightening sequences) to ensure that the collected training and validation datasets can comprehensively cover the abnormal behavior patterns that this application needs to monitor.

[0095] S2 video time synchronization and keyframe extraction: Before the recorder is deployed, the internal clock is calibrated to a unified reference through a standard time server, and the timestamp of the file generated by the recorder is used directly as the synchronization reference.

[0096] The video files captured were automatically named by the recorder based on their recording start time. The start timestamps of the four video files were extracted to determine the latest start time (Tmax) among the four recorders. The remaining three recorders were then edited to match their start times with Tmax. Subsequently, all videos were started uniformly from their respective start times, with keyframes extracted at 5-second intervals.

[0097] S3 target detection for manipulators. It can annotate and build datasets for people, hands, torque wrenches, flanges, bolts, screws, and other objects. The datasets are divided into training and validation sets in an 8:2 ratio.

[0098] The DETR object detection model is adopted, with ResNet-50 as the backbone network and sinusoidal positional encoding. The hidden layer dimension is configured to be 256, the number of attention heads is 8, the number of encoder and decoder layers is set to 6, and the feedforward network dimension is 2048.

[0099] The trained DETR model was used to detect torque wrenches and hands in all keyframes captured by four recorders, and detection boxes were drawn.

[0100] S4 bolt capsule region construction and bolt numbering: The trained DETR model is used to detect bolts and flanges in the initial video frame. The position and height of the bolts and flanges are recorded based on the detected target bounding boxes. The height, width, and center coordinates of the bolt detection boxes are denoted as follows: The flange inspection frame has a height of H. The original inspection frame is replaced by the inscribed circle of the bolt rectangular inspection frame. Bolt circle radius formula: The origin of the coordinate system is located at the upper left corner, and the vertical axis increases downwards. The upper bolt detection frame is shifted downwards by a length of 2H along the vertical axis to construct the lower bolt position, specifically the center position of the bolt below the flange. .

[0101] Similarly, to construct the bolt circle below, for bolts above the flange, extend the upper half of the bolt circle upwards along the longitudinal axis by a radius length; for bolt circles below the flange, extend the lower half of the bolt circle downwards by a radius length, forming a new bolt action area, i.e., a bolt capsule, which is created using the following formula:

[0102] in These represent the ordinates of the centers of the top and bottom semicircular regions of the bolt capsule area, respectively. ; Bolt capsule area From the top semi-circular area and the bottom semi-circular area and the middle rectangular area composition: .

[0103] Each recorder is responsible for detecting the operation of 5 designated bolts within its field of view and numbering them counterclockwise. The four recorders obtain bolt numbers from 1 to 20. The center coordinates and radii of all bolt circles handled by the four recorders, along with the recorder number and bolt number, are saved as a global bolt position information file.

[0104] Since the viewing angle is fixed, the position of the bolt within the recorder's field of view does not change in the same video segment. Using the global bolt position information obtained above, the bolt circle or capsule is drawn on each frame of the corresponding recorder image.

[0105] S5 identifies the bolts based on overlap. Based on the bolt circle and operation medium detection box obtained from the target detection, it creates a bolt capsule mask `MCapsule` and an operation medium rectangular detection box mask `Mtool`. All masks are initialized with background pixel values ​​set to 0. The bolt capsule mask `MCapsule` uses a geometric fill function to set the pixel values ​​corresponding to the capsule-shaped region to 255, based on the global bolt position information file obtained in the previous step. Similarly, the operation medium bounding box mask `Mtool` also sets the pixel values ​​within the detected rectangular region to 255. The formula is as follows: .

[0106] Subsequently, a third intersection mask, `Mintersection`, is generated by performing a pixel-level logical AND operation on the two masks. A pixel's value in `Mintersection` is 255 only if it is simultaneously located in both the bolt capsule region and the operating medium region; otherwise, it is 0. The formula is as follows: .

[0107] The overlap area between the operating medium and the bolt capsule region is calculated by counting the number of pixels with a value of 255 in the intersection mask (Mintersection). This total number of pixels is then compared with the total number of pixels (MCapsule) in the bolt capsule region to obtain the final overlap ratio (Coverage).

[0108] Bolt capsule area: .

[0109] Area of ​​the intersection region: .

[0110] For multiple bolt capsule areas that overlap with the operating medium, the bolt number corresponding to the bolt capsule area with the highest overlap ratio exceeding a threshold of 30% is selected as the bolt being installed or tightened by the operator in that frame image, and this is recorded. All single-frame image processing results from the same recorder are saved in chronological order as a bolt tightening sequence output file. Each line in the file represents the processing result of one frame image, divided into three columns: time, overlapping bolts, and selected bolts. These columns represent the time the frame image occurred; the bolts overlapping with the operating medium; and the bolt number being tightened based on the overlap ratio. Four recorders correspond to four different output files.

[0111] S6 integrates and outputs the tightening results. For the bolt tightening sequence output files generated by each recorder in the above steps, if a single bolt number appears repeatedly within three consecutive frames, it is considered that the bolt was being tightened during that time period; otherwise, it is considered that the operator touched it arbitrarily. The bolt tightening sequence output files from the four recorders are integrated to output the actual tightening sequence of 20 bolts.

[0112] S7 specification comparison and abnormal behavior monitoring: Based on the standard installation and actual numbering of 20 bolts, the correct cross-shaped combinations are predefined: Cross-shaped 1: (bolt pair 1: (1, 11), bolt pair 2: (6, 16)). Cross-shaped 2: (bolt pair 3: (3, 13), bolt pair 4: (8, 18)). Cross-shaped 3: (bolt pair 5: (5, 15), bolt pair 6: (10, 20)). Cross-shaped 4: (bolt pair 7: (2, 12), bolt pair 8: (7, 17)). Cross-shaped 5: (bolt pair 9: (4, 14), bolt pair 10: (9, 19)).

[0113] The process begins by identifying all 10 pairs of bolts in a diagonal relationship within the actual tightening sequence (i.e., the aforementioned 10 bolt pairs). If not all 10 diagonal bolt pairs are found in the actual tightening sequence, a diagonal non-standard installation is identified. Next, all 5 complete cross-shaped combinations are identified within the actual tightening sequence. If not all 5 cross-shaped combinations are found in the actual tightening sequence, a cross-shaped non-standard installation is identified. The detected non-standard operating behaviors are then compared with deliberately simulated non-standard behaviors (i.e., non-standard behaviors in the actual operating sequence) during the data acquisition phase. The comparison is used to evaluate the bolt capsule model's performance in identifying different abnormal tightening patterns and its recall rate for detecting non-standard behaviors.

[0114] As can be seen, the detection method provided in the embodiments of this application has the following technical effects.

[0115] The robustness and accuracy of object positioning have been improved: by using a bolt circle instead of the traditional rectangular detection frame, a bolt capsule area is innovatively constructed, which precisely expands the effective working area of ​​the bolt. This ensures that even if the operating posture is off, there is still enough overlap with the extended capsule area, effectively eliminating missed detections caused by the operating posture.

[0116] The system achieves high stability in the monitoring process and reliable bolt positioning: it effectively solves the problem of bolts often being undetectable due to significant obstruction by the operating medium during tightening. By capturing images from a fixed perspective, bolt position information is obtained from the initial video frame, and this prior position information remains spatially consistent throughout the video sequence. The system reuses this information in all frames, thus avoiding the problem of inaccurate bolt detection due to significant obstruction by the operating medium during operation, ensuring high stability and reliability of the key target positioning data.

[0117] Serialization and high reliability verification of complex industrial process specifications: After accurately capturing the actual bolt numbering and installation sequence of each tightening action through bolt capsule and overlap calculation, the actual recorded installation sequence is compared step by step with the preset cross-diagonal tightening specification sequence model to monitor abnormal installation behavior.

[0118] Corresponding to the above method embodiments, this application also provides a flange bolt abnormal installation sequence identification device. The flange bolt abnormal installation sequence identification device described below and the flange bolt abnormal installation sequence identification method described above can be referred to each other.

[0119] See Figure 17 As shown, the device includes the following modules.

[0120] The video acquisition module 101 is used to acquire video segments of the installation of multiple flange bolts simultaneously captured from multiple perspectives; multiple flange bolts are fixed to the same set of flanges, and the flange bolts with corresponding numbers are captured from different perspectives.

[0121] The target detection module 102 is used to perform target detection on the operating medium and flange bolts in the installation video segment and obtain the detection results.

[0122] The bolt capsule creation module 103 is used to determine the inscribed circle of the detection frame corresponding to the flange bolt as the bolt circle. For flange bolts located above the flange, the upper half of the bolt circle is extended upward along the longitudinal axis by a specified length to obtain the bolt capsule. For flange bolts located below the flange, the lower half of the bolt circle is extended downward along the longitudinal axis by a specified length to obtain the bolt capsule.

[0123] The overlap detection module 104 is used to perform overlap detection on the bolt capsule and the detection frame of the operating medium, and to determine the installation sequence in conjunction with the flange bolt number.

[0124] The abnormal installation sequence detection module 105 is used to compare the installation sequence with the standard procedure to detect abnormal installation sequences.

[0125] Using the apparatus provided in this application, installation video segments of multiple flange bolts are simultaneously captured from multiple perspectives; multiple flange bolts are fixed to the same set of flanges, and flange bolts with corresponding numbers are captured from different perspectives; target detection of the operating medium and flange bolts is performed on the installation video segments to obtain detection results; the inscribed circle of the detection frame corresponding to the flange bolt is determined as the bolt circle; for flange bolts located above the flange, the upper half of the bolt circle is extended upward along the longitudinal axis by a specified length to obtain a bolt capsule; for flange bolts located below the flange, the lower half of the bolt circle is extended downward along the longitudinal axis by a specified length to obtain a bolt capsule; overlap detection is performed on the detection frame of the bolt capsule and the operating medium, and the installation sequence is determined in combination with the flange bolt numbers; the installation sequence is compared with the standard procedure to detect abnormal installation sequences.

[0126] This application takes into account that any deviation in the bolt tightening sequence may lead to uneven distribution of preload, thereby affecting the sealing performance of the flange connection. Given the high-risk nature of blowout preventer systems and the extremely high requirements for assembly quality, computer vision-based industrial safety monitoring typically focuses on identifying personnel's overall posture, movement trajectory, or the wearing of personal protective equipment for safety supervision, which is no longer sufficient to meet actual safety supervision needs.

[0127] Therefore, this application uses the entire process of on-site workers installing flange bolts as input. The actual bolt tightening sequence is automatically analyzed and recorded. By comparing this sequence with standard procedures, abnormal installation sequences can be detected. This avoids traditional manual inspections or time-consuming video playback, efficiently delegating the tedious sequence comparison work to computer equipment.

[0128] Specifically, firstly, multiple installation video segments of flange bolts are acquired simultaneously from multiple perspectives. These flange bolts are used to fix the same set of flanges, and the flange bolts with corresponding numbers are captured from different perspectives. This means that the global number of the detected flange bolts can be uniquely determined based on the video's perspective. Then, target detection of the operating medium and flange bolts is performed on the installation video segments to obtain the detection results. The inscribed circle of the detection frame corresponding to the flange bolt is defined as the bolt circle. For flange bolts located above the flange, the upper half of the bolt circle is extended upwards along the longitudinal axis by a specified length to obtain a bolt capsule. For flange bolts located below the flange, the lower half of the bolt circle is extended downwards along the longitudinal axis by a specified length to obtain a bolt capsule. Considering that the center position of the bolt circle and the rectangular detection frame represents the position where the bolt is already installed, in actual operation, operators must use a torque wrench or their hands to apply force at a position far from the bolt circle's center. This results in a necessary structural positional offset between the detection frame of the operating medium and the bolt center. This leads to a generally low overlap between the tool boundary frame and the bolt circle, severely affecting the accuracy of bolt detection and causing a large number of missed detections. To address this, by deforming the inspection frame for flange bolts, especially by extending the bolt circle in different directions in conjunction with the inscribed circle and the specific installation characteristics corresponding to its location, it becomes easier to detect the flange bolts currently being operated. Specifically, the bolt capsule and the inspection frame of the operating medium are overlapped for inspection, and the installation sequence is determined by combining the flange bolt number. Finally, the installation sequence is compared with the standard procedure to detect any abnormal installation sequences.

[0129] As can be seen, this application can complete all analysis work in far less time than manual review after the installation video is recorded, and automatically compare and verify the recorded actual operation sequence with standard procedures, thereby detecting abnormal installation sequences. This rapid verification application mode ensures the quality and safety of the assembly process, and can help on-site managers and workers take immediate corrective measures to eliminate potential structural hazards before the equipment is put into use.

[0130] In one specific embodiment of this application, the video acquisition module is specifically used to acquire original installation videos of multiple flange bolts simultaneously captured from multiple different perspectives; extract installation video segments of the installation time period from the multiple original installation videos; and perform frame extraction processing on the installation video segments.

[0131] In one specific embodiment of this application, the bolt capsule creation module is specifically used to extend the upper semicircle of the bolt circle upward along the longitudinal axis by a radius length to obtain the bolt capsule.

[0132] In one specific embodiment of this application, the overlap detection module is specifically used to create a bolt capsule mask based on the bolt capsule and an operation medium rectangular mask based on the detection frame of the operation medium; perform a logical AND operation on the bolt capsule mask and the operation medium rectangular mask to obtain an intersection mask; use the intersection mask to determine the overlap area of ​​the detection frames of the bolt capsule and the operation medium; determine the bolt capsule corresponding to the maximum coverage at each time step as the operation bolt according to the time sequence; and obtain the installation sequence based on the number of the operation bolt.

[0133] In one specific embodiment of this application, the target detection module is specifically used to perform target detection on the installation video segment for hands, torque wrenches, and flange bolts, and obtain detection results.

[0134] In one specific embodiment of this application, the target detection module is specifically used to detect flange bolts in the installation video segment. After detecting the flange bolt located above the flange, the detection frame of the flange bolt is shifted downward by two flange widths to obtain the detection frame corresponding to the flange bolt located below the flange. The flange width is the width of the rectangular detection frame of the flange obtained by performing target detection on the flange.

[0135] In one specific embodiment of this application, it further includes: a warning module, used to output a reinstallation prompt message when an abnormal installation sequence is detected.

[0136] Corresponding to the above method embodiments, this application also provides an electronic device. The electronic device described below and the flange bolt abnormal installation sequence identification method described above can be referred to in correspondence.

[0137] See Figure 18 As shown, the electronic device includes: a memory 332 for storing a computer program; and a processor 322 for executing the computer program to implement the steps of the flange bolt abnormal installation sequence identification method of the above method embodiment.

[0138] For details, please refer to Figure 19 , Figure 19This is a schematic diagram of the specific structure of an electronic device provided in this embodiment. The electronic device can vary significantly due to differences in configuration or performance. It may include one or more central processing units (CPUs) (e.g., one or more processors) and a memory 332. The memory 332 stores one or more computer programs 342 or data 344. The memory 332 can be temporary or permanent storage. The program stored in the memory 332 may include one or more modules (not shown in the diagram), each module may include a series of instruction operations on the data processing device. Furthermore, the processor 322 may be configured to communicate with the memory 332 and execute the series of instruction operations stored in the memory 332 on the electronic device 301.

[0139] Electronic device 301 may also include one or more power supplies 326, one or more wired or wireless network interfaces 350, one or more input / output interfaces 358, and / or one or more operating systems 341.

[0140] The steps in the flange bolt abnormal installation sequence identification method described above can be implemented by the structure of electronic equipment.

[0141] Corresponding to the above method embodiments, this application also provides a readable storage medium. The readable storage medium described below can be referred to in conjunction with the flange bolt abnormal installation sequence identification method described above.

[0142] A readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the flange bolt abnormal installation sequence identification method described in the above method embodiments.

[0143] Specifically, the readable storage medium can be a USB flash drive, external hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, or any other readable storage medium capable of storing program code.

[0144] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0145] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0146] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0147] Finally, it should be noted that in this document, relationships such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "include," "contain," or any other variations are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus.

[0148] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for identifying abnormal installation sequence of flange bolts, characterized in that, include: Acquire installation video segments of multiple flange bolts simultaneously captured from multiple perspectives; multiple flange bolts are fixed to the same set of flanges, and flange bolts with corresponding numbers are captured from different perspectives; The installation video segment was subjected to target detection of the operating medium and flange bolts, and the detection results were obtained. The inscribed circle of the detection frame corresponding to the flange bolt is defined as the bolt circle. For the flange bolt located above the flange, the upper half of the bolt circle is extended upward along the longitudinal axis by a specified length to obtain the bolt capsule. For the flange bolt located below the flange, the lower half of the bolt circle is extended downward along the longitudinal axis by a specified length to obtain the bolt capsule. The overlap detection of the bolt capsule and the detection frame of the operating medium is performed, and the installation sequence is determined in combination with the flange bolt number; The installation sequence is compared with the standard procedure to detect abnormal installation sequences; if an abnormal installation sequence is detected, the abnormal installation behavior is output, and the specific bolt number and time point of the abnormality are indicated. Specifically, the installation video segment is subjected to target detection of the operating medium and flange bolts, and the detection results are obtained, including: The installation video segment was subjected to target detection of hands, torque wrenches, and flange bolts, and the detection results were obtained. Flange bolt detection is performed on the installation video segment. After detecting the flange bolt located above the flange, the detection frame of the flange bolt is shifted downward by two flange widths to obtain the detection frame corresponding to the flange bolt located below the flange. The flange width is the width of the rectangular detection frame of the flange obtained by target detection of the flange.

2. The method according to claim 1, characterized in that, The overlap detection of the bolt capsule and the operating medium is performed, and the installation sequence is determined by combining the flange bolt numbers, including: A bolt capsule mask is created based on the bolt capsule, and an operation medium rectangular mask is created based on the detection frame of the operation medium; Perform a logical AND operation between the bolt capsule mask and the operating medium rectangular mask to obtain the intersection mask; The overlap area between the bolt capsule and the detection frame of the operating medium is determined using the intersection mask; According to the time sequence, the bolt capsules corresponding to the maximum coverage at each time point are identified as the operating bolts; The installation sequence is determined based on the numbering of the operating bolts.

3. The method according to claim 1, characterized in that, Acquire video clips of multiple flange bolt installations simultaneously captured from multiple perspectives, including: Acquire original installation videos of multiple flange bolts simultaneously captured from multiple different perspectives; Installation video segments of the installation time period are extracted from multiple original installation videos, and frame extraction is performed on the installation video segments.

4. The method according to claim 1, characterized in that, The upper semicircle of the bolt circle is extended upwards along the longitudinal axis by a specified length to obtain the bolt capsule, including: The bolt capsule is obtained by extending the upper semicircle of the bolt circle upward along the longitudinal axis by a radius length.

5. The method according to any one of claims 1 to 4, characterized in that, Also includes: If an abnormal installation sequence is detected, a reinstallation prompt message will be output.

6. A device for identifying abnormal installation sequence of flange bolts, characterized in that, include: The video acquisition module is used to acquire video segments of the installation of multiple flange bolts simultaneously captured from multiple perspectives; the multiple flange bolts are fixed to the same set of flanges, and the flange bolts with corresponding numbers are captured from different perspectives. The target detection module is used to perform target detection on the installation video segment for the operating medium and flange bolts, and obtain the detection results; The bolt capsule creation module is used to determine the inscribed circle of the detection frame corresponding to the flange bolt as the bolt circle. For flange bolts located above the flange, the upper half of the bolt circle is extended upward along the longitudinal axis by a specified length to obtain the bolt capsule. For flange bolts located below the flange, the lower half of the bolt circle is extended downward along the longitudinal axis by a specified length to obtain the bolt capsule. The overlap detection module is used to perform overlap detection on the bolt capsule and the detection frame of the operating medium, and to determine the installation sequence in conjunction with the flange bolt number; An abnormal installation sequence detection module is used to compare the installation sequence with the standard procedure to detect abnormal installation sequences; if an abnormal installation sequence is detected, the module outputs the abnormal installation behavior and indicates the specific bolt number and time point of the abnormality. Specifically, the target detection module is used to perform target detection on the installation video segment for hands, torque wrenches, and flange bolts, and obtain detection results; for flange bolt detection on the installation video segment, after detecting the flange bolt located above the flange, the detection frame of the flange bolt is shifted downward by two flange widths to obtain the detection frame corresponding to the flange bolt located below the flange; wherein, the flange width is the width of the rectangular detection frame of the flange obtained by performing target detection on the flange.

7. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the flange bolt abnormal installation sequence identification method as described in any one of claims 1 to 5 when executing the computer program.

8. A readable storage medium, characterized in that, The readable storage medium stores a computer program that, when executed by a processor, implements the steps of the flange bolt abnormal installation sequence identification method as described in any one of claims 1 to 5.

Images