A method for identifying traction anomalies in the main cable strands of a suspension bridge

Abstract

The present invention relates to the technical field of main cable installation for suspension bridges, and specifically to a method for identifying abnormal traction of main cable strands of suspension bridges, comprising: S1, constructing a roller and wrapping tape identification model; S2, identifying a target area of ​​a construction site, identifying rollers and wrapping tapes in an image of the target area, calculating a transformation matrix M based on the identified rollers and their actual sizes, and converting the image of the target area into a bird's-eye view according to the transformation matrix M; S3, judging whether there is a missing wrapping tape in the bird's-eye view; S4, calculating and statistically analyzing the width of the strands in the wrapping tape area to judge whether the wrapping tape is broken. The present invention constructs a roller and wrapping tape identification model, constructs an image transformation matrix based on the geometric dimensions of the rollers, obtains a bird's-eye view of the target area, uses the fixed spacing information of the wrapping tape to identify missing strands, and uses the geometric dimension information of a single strand to identify the breakage of the wrapping tape, thereby realizing the identification of abnormal traction of the main cable strands of the suspension bridge.

Description

Technical Field

[0001] The present invention relates to the technical field of main cable installation of suspension bridges, and in particular to a method for identifying traction anomalies of main cable strands of suspension bridges. Background Art

[0002] The main cable, a critical load-bearing component of a suspension bridge, is composed of multiple strands, each made of multiple galvanized high-strength steel wires. The main cable is often called the "lifeline" of a suspension bridge. The PPWS method is commonly used for cable construction. This method involves forming high-strength steel wires into parallel strands in the factory, winding them onto cable drums, and then installing the main cable using a reciprocating traction system. Cable installation involves several steps: strand traction, strand lateral movement, strand shaping, strand insertion into the saddle, strand sag adjustment, and anchor span tension adjustment. During cable installation, problems such as twisting, loosening, tape breakage, wire drumming, and strand surface wear often occur. If not addressed promptly, these problems can seriously impact the quality and progress of main cable installation.

[0003] To address these issues, existing methods typically rely on a human operator following the tractor to monitor the cable traction process and identify any anomalies. This involves manually observing and determining whether the cable traction process is normal, whether the strand being erected falls neatly into the rollers, and whether the strand is showing signs of loose strands or broken wrapping tape. If any issues are detected, an intercom alarm is used to stop the tractor, and the erection process can continue after the anomaly is resolved. This method requires a high level of experience and physical exertion, and requires dedicated personnel to monitor the process from a catwalk. This is time-consuming and labor-intensive, and requires personnel to work at height, posing safety risks. Summary of the Invention

[0004] The purpose of the present invention is to address the defects of the existing technology and provide a method for identifying the abnormal traction of the main cable strands of a suspension bridge, so as to realize the identification of the abnormal traction state of the main cable strands of the suspension bridge.

[0005] In order to solve the above technical problems, the present invention provides a method for identifying abnormal traction of main cable strands of a suspension bridge, comprising:

[0006] S1. Constructing the roller and wrapping belt identification model;

[0007] S2. Identify the target area of ​​the construction site, identify the rollers and wrapping tape in the target area image, calculate the transformation matrix M based on the identified rollers and their actual sizes, and convert the target area image into a bird's-eye view based on the transformation matrix M;

[0008] S3. Determine whether there is any missing wrapping tape in the bird's-eye view;

[0009] S4. Calculate and count the width of the strands in the wrapping tape area to determine whether the wrapping tape is broken.

[0010] Furthermore, step S1 includes: collecting image data during the cable strand traction process, constructing an image data set, and training a target detection model based on the image data set to obtain a roller and wrapping belt recognition model.

[0011] Furthermore, in step S2, the method for calculating the transformation matrix M includes: identifying the rollers in the target area, obtaining the coordinates of the four corner points of the rollers in the target area image, setting the coordinates of the four corner points of the rollers after the target area image is transformed according to the preset actual size of the rollers, and calculating the transformation matrix M according to the coordinates of the four corner points of the rollers after the target area image is transformed and the coordinates of the four corner points of the rollers in the target area image.

[0012] Furthermore, in step S2, the coordinates of the four corner points of the roller after the target area image is transformed are set according to the width b and height h of the roller, that is, the coordinates of the four corner points of the roller after the target area image is transformed are (0,0), (b,0), (b,h), and (0,h), respectively, so that the target area image is transformed into a bird's-eye view after passing through the transformation matrix M.

[0013] Furthermore, in step S2, the transformation matrix M is calculated using the rollers in the target area close to the middle of the target area.

[0014] Furthermore, step S3 includes: calculating the distance between two adjacent wrapping tapes in the bird's-eye view, and determining whether the distance between the two adjacent wrapping tapes is within a normal distance range; if not, there is a situation where the wrapping tape is missing or misidentified, triggering a wrapping tape missing alarm or misidentification alarm; otherwise, executing step S4.

[0015] Furthermore, step S4 includes: processing the bird's-eye view image using a semantic segmentation algorithm, calculating the width of the cable strands in the area where the wrapping tape is located, and determining whether the width of the cable strands in the area where the wrapping tape is located is within a preset range.

[0016] Furthermore, in step S4, the width of the strands in the area where the wrapping tape is located is the maximum width of the wrapping tape.

[0017] Furthermore, in step S4, if the width of the strands in the area where the wrapping tape is located is less than the lower limit of the preset range, an alarm for abnormal wrapping tape identification is triggered.

[0018] Furthermore, in step S4, if the width of the strands in the area where the wrapping tape is located is greater than the upper limit of a preset range, a wrapping tape breakage alarm is triggered.

[0019] The beneficial effects of the present invention are:

[0020] 1. The present invention uses machine vision technology to construct a roller and wrapping tape recognition model, builds an image transformation matrix based on the roller geometric dimension information, obtains a bird's-eye view of the target area, uses the fixed spacing information of the wrapping tape to identify missing parts, and uses the geometric dimension information of a single cable strand to identify wrapping tape breaks, thus realizing the active recognition and early warning of abnormal traction status of the main cable strands of the suspension bridge.

[0021] 2. The present invention utilizes the physical dimensions of the roller to achieve geometric transformation of the target area image, and there is a proportional correspondence between the pixels in the transformed image and the actual geometric dimensions, which facilitates the subsequent identification of the wrapping tape breakage. The method is simple and efficient. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 is a flow chart of the present invention;

[0023] Figure 2 is a schematic diagram of the target area of ​​the present invention;

[0024] Figure 3 Schematic diagram of the detection results of the rollers and wrapping belts in the target area according to the present invention;

[0025] Figure 4 Schematic diagram of calculation of the transformation matrix M of the present invention;

[0026] Figure 5 Schematic diagram of the present invention transformed into a bird's-eye view through the transformation matrix M;

[0027] Figure 6 A schematic diagram of the present invention for identifying whether a wrapping tape is missing;

[0028] Figure 7 2. It is a cross-sectional dimension diagram of the cable strand of the present invention;

[0029] Figure 8 This is a schematic diagram of calculating the width of the strands in the area where the wrapping tape is located according to the present invention. DETAILED DESCRIPTION

[0030] In order to make the technical problems, technical solutions and beneficial effects to be solved by this application more clearly understood, this application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0031] like Figure 1 As shown, a method for identifying abnormal traction of main cable strands of a suspension bridge comprises:

[0032] S1. Construct the roller and wrapping belt identification model; specifically including: Figure 2 、 3As shown, a large amount of image data of the cable strand traction process is first collected. The image data includes the cable strands, rollers, and wrapping tape. Then, LabelImg is used to annotate the image data to construct a professional image dataset that can be used for model training. Based on this dataset, target detection models such as YoloV8, EfficientDet, and DETR are used for model training (YoloV8 is used in this embodiment) to obtain roller and wrapping tape recognition models.

[0033] S2. Since the cable strands are wound around the cable drum and the entire cable strand is several kilometers long, only the portion passing through the visual area is analyzed.

[0034] Identify the target area of ​​the construction site, which is the visual area, identify the rollers and wrapping tape in the target area image, calculate the transformation matrix M based on the identified rollers and their actual sizes, and convert the target area image into a bird's-eye view based on the transformation matrix M. Specifically, the process includes:

[0035] Identify the rollers in the target area and obtain the coordinates of the four corner points of the rollers in the target area image, such as Figure 4 As shown, according to the preset real size of the roller (in this embodiment, the real size of the roller is: width b = 240mm, height h = 300mm), the coordinates of the four corner points of the roller after the target area image transformation are set to (0, 0), (240, 0), (240, 300), (0, 300), respectively. According to the coordinates of the four corner points of the roller after the target area image transformation and the coordinates of the four corner points of the roller in the target area image, the transformation matrix M is calculated. The transformation matrix M is a coordinate mapping matrix, that is, the coordinates of the four corner points of the roller in the target area image can be transformed into the following four coordinates (0, 0), (240, 0), (240, 300), (0, 300) through the transformation matrix M. Similarly, other elements in the target area image can be transformed by the transformation matrix M, thereby converting the target area image into a bird's-eye view. The method used in step S2 is a perspective transformation method.

[0036] Since multiple rollers are identified in the target area, and the transformation matrix M calculated using different rollers is different, in order to reduce the error after the target area image is transformed into a bird's-eye view through the transformation matrix M, the roller near the center of the target area is used to calculate the transformation matrix M, which is beneficial to reduce the error in the spacing and size of the wrapping tapes in the bird's-eye view and reduce the possibility of misjudgment.

[0037] S3. Determining whether there is a missing wrapping tape in the bird's-eye view. This specifically includes calculating the distance between two adjacent wrapping tapes in the bird's-eye view and determining whether the distance between the two adjacent wrapping tapes is within a normal distance range. If not, a missing wrapping tape or misidentification has occurred, triggering a missing wrapping tape alarm or misidentification alarm. Otherwise, executing step S4. That is, when the distance between two adjacent wrapping tapes is less than the lower limit of the normal distance range, it indicates that a wrapping tape has been misidentified, and a misidentification alarm will be triggered. When the distance between two adjacent wrapping tapes is greater than the upper limit of the normal distance range, it indicates that a wrapping tape has been missing, and a missing wrapping tape alarm will be triggered.

[0038] In step S3, the normal spacing range is pre-set. Since the factory adopts mechanized processing of the cable strands, the spacing of the wrapping tapes on the cable strands is a fixed value. Therefore, the normal spacing range should be set according to the fixed value.

[0039] S4. Calculate and count the width of the strands within the wrapping tape area to determine whether the wrapping tape has broken. This specifically includes: using a semantic segmentation algorithm to process the bird's-eye view image, calculating the width of the strands within the wrapping tape area, where the strand width within the wrapping tape area is the maximum width of the wrapping tape; determining whether the strand width within the wrapping tape area is within a preset range; if the strand width within the wrapping tape area is less than the lower limit of the preset range, triggering a wrapping tape identification anomaly alarm. If the strand width within the wrapping tape area is greater than the upper limit of the preset range, triggering a wrapping tape breakage alarm. Generally, the width of the wrapping tape is smallest in the middle and largest at the ends due to the strands; therefore, the maximum width of the wrapping tape is generally at the ends.

[0040] In this embodiment, the preset range is 71 to 90 mm. If the width of the strands in the area where the wrapping tape is located is less than 71 mm, it indicates that the wrapping tape is abnormally identified. If it is greater than 90 mm, it indicates that the wrapping tape is broken.

[0041] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A method for identifying abnormal traction of main cable strands in a suspension bridge, characterized by: include: S1. Constructing the roller and wrapping belt identification model; S2. Identify the target area of ​​the construction site, identify the rollers and wrapping tape in the target area image, calculate the transformation matrix M based on the identified rollers and their actual sizes, and convert the target area image into a bird's-eye view based on the transformation matrix M; In step S2, the method for calculating the transformation matrix M includes: identifying the roller in the target area, obtaining the coordinates of the four corner points of the roller in the target area image, setting the coordinates of the four corner points of the roller after the target area image is transformed according to a preset real size of the roller, and calculating the transformation matrix M based on the set coordinates of the four corner points of the roller after the target area image is transformed and the coordinates of the four corner points of the roller in the target area image; In step S2, the coordinates of the four corner points of the roller after the target area image is transformed are set according to the width b and height h of the roller, that is, the coordinates of the four corner points of the roller after the target area image is transformed are (0, 0), (b, 0), (b, h), and (0, h), respectively, so that the target area image is transformed into a bird's-eye view after the transformation matrix M; In step S2, the transformation matrix M is calculated using the rollers in the target area close to the middle of the target area; S3. Determine whether there is any missing wrapping tape in the bird's-eye view; Step S3 includes: calculating the distance between two adjacent wrapping tapes in the bird's-eye view, and determining whether the distance between the two adjacent wrapping tapes is within a normal distance range. If not, there is a situation of missing wrapping tape or misidentification, and a wrapping tape missing alarm or misidentification alarm is triggered. Otherwise, step S4 is executed; S4. Calculate and count the width of the cable strands in the wrapping tape area to determine whether the wrapping tape is broken; Step S4 includes: processing the bird's-eye view image using a semantic segmentation algorithm, calculating the width of the cable strands in the area where the wrapping tape is located, and determining whether the width of the cable strands in the area where the wrapping tape is located is within a preset range.

2. The method for identifying abnormal traction of main cable strands of a suspension bridge according to claim 1, characterized in that: Step S1 includes: collecting image data during the cable strand traction process, constructing an image data set, training a target detection model based on the image data set, and obtaining a roller and wrapping belt recognition model.

3. The method for identifying abnormal traction of main cable strands of a suspension bridge according to claim 1, characterized in that: In step S4, the width of the strands in the area where the wrapping tape is located is the maximum width of the wrapping tape.

4. The method for identifying abnormal traction of main cable strands of a suspension bridge according to claim 3, characterized in that: In step S4, if the width of the strands in the area where the wrapping tape is located is less than the lower limit of the preset range, an alarm for abnormal wrapping tape identification is triggered.

5. The method for identifying abnormal traction of main cable strands of a suspension bridge according to claim 3, characterized in that: In step S4, if the width of the strands in the area where the wrapping tape is located is greater than the upper limit of the preset range, a wrapping tape breakage alarm is triggered.

Images