CN116309418B - Intelligent monitoring method and device for deformation of main beam in cantilever construction of bridge - Google Patents

Abstract

The application provides an intelligent monitoring method and device for deformation of a bridge cantilever construction girder, and the method comprises the following steps: acquiring a reference image of a bridge cantilever construction girder with a specific mark in a reference state and an image to be detected in a state to be detected; the reference image and the image to be detected are acquired under the approximate image acquisition condition; image segmentation is carried out on the reference image and the image to be detected so as to obtain a segmented reference image and a segmented image to be detected respectively; acquiring a reference relative position of a reference specific mark image in the segmented reference image and a to-be-detected relative position of a to-be-detected specific mark image in the segmented to-be-detected image; and determining the deformation of the bridge cantilever construction girder according to the reference relative position and the relative position to be detected. According to the method, the deformation of the bridge cantilever construction girder is determined by analyzing images of the bridge cantilever construction girder in different states, so that the detection accuracy is improved, and equipment required in implementation is simplified.

Description

Intelligent monitoring method and device for deformation of main beam in cantilever construction of bridge

Technical Field

The present application relates to the technical field of intelligent construction monitoring in bridge engineering, and specifically, to an intelligent monitoring method and device for deformation of a main beam of a cantilever bridge construction.

Background technique

Deformation detection can be divided into contact deformation detection and non-contact deformation detection. Contact deformation detection mainly uses micrometers, wire displacement meters, vibrating wire displacement meters, etc. to measure the main beam of the bridge cantilever construction, and determines the deformation of the bridge cantilever construction main beam based on the measurement results; non-contact deformation detection mainly uses non-contact monitoring equipment to detect the main beam of the bridge cantilever construction.

However, contact deformation detection uses measuring tools for mechanical measurement and requires the use of various embedded sensors, which results in low detection accuracy and low monitoring efficiency. Non-contact detection uses a sensor system to detect the deformation of the cantilever construction girder of the bridge. Compared with contact deformation detection, although it improves measurement efficiency, its measurement accuracy still needs to be further improved.

Especially in the process of constructing prestressed rigid frame bridges using the cantilever construction method of hanging basket support, this construction method will make the force and line shape of the structure complicated and unable to be consistent with the expected design values, thereby threatening the long-term operation safety of the bridge and even causing problems such as large deflection of the main beam and concrete cracking. In order to ensure the safety of the bridge during construction and achieve the expected structural internal force and line shape during the construction period and the long-term operation of the structure, it is necessary to implement scientific and reasonable construction monitoring of the bridge through the implementation of bridge deformation detection.

In the prior art, traditional non-contact deformation detection is mostly used to detect bridge deformation, such as theodolite, total station, optical plumb line, laser instrument, GPS positioning system, etc. However, the accuracy of bridge deformation detection using such sensor systems still needs to be further improved.

Summary of the invention

The purpose of the embodiment of the present application is to provide an intelligent monitoring method and device for the deformation of the main beam of a bridge cantilever construction, by acquiring images of the main beam of a bridge cantilever construction with specific marks in different states, and determining the deformation of the main beam of the bridge cantilever construction by processing and analyzing the images, so as to solve the problem that the deformation detection accuracy in the prior art is still not high enough.

In the first aspect, an embodiment of the present application provides an intelligent monitoring method for the deformation of a bridge cantilever construction main beam, comprising: obtaining a reference image of a bridge cantilever construction main beam with a specific mark in a reference state, and an image to be detected in a state to be detected; wherein the reference image and the image to be detected are acquired under similar image acquisition conditions; performing image segmentation on the reference image and the image to be detected to obtain a segmented reference image and a segmented image to be detected, respectively; obtaining a reference relative position of a reference specific mark image in the segmented reference image in the segmented reference image, and a relative position to be detected of a reference specific mark image in the segmented image to be detected in the segmented image to be detected; and determining the deformation of the bridge cantilever construction main beam according to the reference relative position and the relative position to be detected.

The above-mentioned intelligent monitoring method for deformation of the main beam of the cantilever construction of the bridge, by obtaining images of the cantilever construction main beam of the bridge with specific marks in different states under approximate conditions, and by processing the image, the position change of the specific mark image in the image is known, and the specific deformation of the cantilever construction main beam of the bridge is determined according to the position change, which improves the accuracy of detection compared with the prior art. In addition, the deformation detection of the cantilever construction main beam of the bridge by the method in the above-mentioned embodiment is simpler to implement and the required equipment is also simpler than using theodolite, total station, optical plumb line, laser instrument, GPS positioning system and the like. In addition, by scientifically and reasonably monitoring the process of building a prestressed rigid frame bridge by the cantilever construction method of the hanging basket, the construction personnel can comprehensively and accurately grasp the linear shape of the bridge and the structural internal force it is subjected to during the construction process, and finally ensure the construction quality of the bridge and the safety of the construction personnel.

In combination with the first aspect, optionally, the performing image segmentation on the reference image and the image to be detected to obtain the segmented reference image and the segmented image to be detected, respectively, includes: performing semantic segmentation on the reference image and the image to be detected, respectively, using a neural network, to obtain the segmented reference image and the segmented image to be detected, respectively; wherein the neural network includes a SegNet fully convolutional neural network.

The above-mentioned intelligent monitoring method for deformation of the main beam of the cantilever construction of the bridge uses a neural network to perform semantic segmentation on the collected images to distinguish the image with specific marks in the image from its background image, which just meets the requirements of image processing in the above-mentioned embodiment. Compared with other panoramic segmentation and instance segmentation, which make more detailed classifications of pixels in the image, the semantic segmentation algorithm is relatively simple, which makes the efficiency of executing the method steps relatively high. In addition, the SegNet fully convolutional neural network used in the above-mentioned embodiment, compared with other neural networks such as U-Net, DenseNetsE-Net and Link-Net, the SegNet fully convolutional neural network achieves a good balance between memory and precision involved in segmentation performance, and can maintain the integrity of high-frequency details in segmentation.

In combination with the first aspect, optionally, the obtaining of the benchmark relative position of the benchmark marked image in the segmented benchmark image in the segmented image to be detected, and the relative position to be detected of the specific marked image to be detected in the segmented image to be detected, includes: extracting the benchmark feature point set of the benchmark specific marked image in the segmented benchmark image, and the feature point set to be detected of the specific marked image to be detected in the segmented image to be detected; establishing a mapping relationship between the benchmark feature point set and the feature point set to be detected by similarity comparison; selecting the benchmark feature points in the benchmark feature point set and the feature points to be detected from the feature points to be detected that have the mapping relationship with the benchmark feature points; and obtaining the coordinate parameters of the benchmark feature points and the feature points to be detected, and using them as the benchmark relative position and the relative position to be detected, respectively.

The above-mentioned intelligent monitoring method for the deformation of the main beam of the bridge cantilever construction, by respectively obtaining the benchmark feature point set of the benchmark specific marked image in the segmented benchmark image and the feature point set to be detected of the specific marked image to be detected in the segmented image to be detected, and after establishing a mapping relationship between the benchmark feature point set and the feature point set to be detected, selects at least one group of benchmark feature points and feature points to be detected with a mapping relationship, and determines the deformation mode of the main beam of the bridge cantilever construction based on the change in the relative position between the benchmark feature point and the feature point to be detected, so that the method can be applied to marked images of any shape, thereby expanding the scope of use.

In combination with the first aspect, optionally, the specific mark includes a circular optical mark; the obtaining of the reference relative position of the reference specific mark image in the segmented reference image, and the relative position to be detected of the specific mark image to be detected in the segmented image to be detected, includes: using a gradient-based Hough transform to identify the reference center coordinates of the reference circular image in the segmented reference image, and the coordinates of the center of the circular image to be detected in the segmented image to be detected; and using the reference center coordinates and the coordinates of the center of the circle to be detected as the reference relative position and the relative position to be detected, respectively.

The above-mentioned intelligent monitoring method for deformation of the main beam of the bridge cantilever construction adopts a circular optical marker as a specific marker, and combines it with the gradient-based Hough transform to identify the coordinates of its center of the circle, which can more efficiently determine the position change of the image to be detected and the reference image, so as to determine the deformation of the main beam of the bridge cantilever construction, thereby improving the efficiency of deformation detection. In addition, the optical marker image makes the specific marker image in the image collected by the image acquisition module such as the industrial camera more obvious, reducing the pressure of subsequent image processing and analysis, thereby further improving the efficiency of deformation detection.

In combination with the first aspect, optionally, before using the gradient-based Hough transform to identify the reference center coordinates of the circular image in the segmented reference image and the center coordinates of the circular image to be detected in the segmented image to be detected, the method also includes: performing boundary identification on the segmented reference image and the segmented image to be detected, respectively, to obtain reference boundary points of the segmented reference image and boundary points to be detected of the segmented image to be detected; fitting the reference boundary points and the boundary points to be detected, respectively, to obtain a reference fitting ellipse of the segmented reference image and a fitting ellipse to be detected of the segmented image to be detected; and performing ellipse correction on the reference fitting ellipse and the fitting ellipse to be detected, respectively, to obtain a reference circular image in the segmented reference image and a circular image to be detected in the segmented image to be detected.

The above-mentioned intelligent monitoring method for deformation of the bridge cantilever construction main beam, when various factors cause the circular optical marker to be elliptical in the collected reference image and the image to be detected during image acquisition, respectively fits the reference boundary point and the to-be-detected boundary point into a reference fitting ellipse and a to-be-detected fitting ellipse through ellipse fitting, and further corrects the reference fitting ellipse and the to-be-detected fitting ellipse into a reference circular image and a to-be-detected circular image according to an ellipse correction algorithm, so as to facilitate the subsequent determination of the reference circle center coordinates and the to-be-detected circle center coordinates, and finally determines the deformation of the bridge cantilever construction main beam according to the reference circle center coordinates and the to-be-detected circle center coordinates, avoiding the need for strict acquisition conditions in the image acquisition process to make the circular optical marker circular in the collected image, so as to determine the deformation of the bridge cantilever construction main beam through a circular recognition algorithm. That is, the scope of use of the intelligent monitoring method for deformation of the bridge cantilever construction main beam provided in the embodiment of the present application is expanded, and the convenience of implementing the method is improved.

In combination with the first aspect, optionally, the performing elliptical correction on the reference fitting ellipse and the fitted ellipse to be detected respectively to obtain a reference circular image in the segmented reference image and a circular image to be detected in the segmented image, includes: performing elliptical correction on the reference fitting ellipse and the fitted ellipse to be detected respectively to obtain a reference corrected circle of the segmented reference image and a corrected circle to be detected of the fitted ellipse to be detected; and performing circular normalization on the reference correction circle and the corrected circle to be detected respectively to obtain a reference circular image in the segmented reference image and a circular image to be detected in the segmented image.

The above-mentioned intelligent monitoring method for deformation of the main beam of the cantilever construction of the bridge simplifies the subsequent calculation process of determining the coordinates of the center of the reference circle and the coordinates of the center of the circle to be detected by circular normalization, thereby improving the efficiency of deformation detection.

In combination with the first aspect, optionally, the circular optical mark includes a first circular optical mark and a second circular optical mark; the reference center coordinates include a first reference coordinate corresponding to the first circular optical mark, and a second reference coordinate corresponding to the second circular optical mark; the center coordinates to be detected include a first coordinate to be detected corresponding to the first circular optical mark, and a second coordinate to be detected corresponding to the second circular optical mark; the deformation of the bridge cantilever construction main beam is determined according to the reference relative position and the relative position to be detected, including: determining the first deformation information of the bridge cantilever construction main beam according to the first reference coordinate and the first coordinate to be detected, and determining the second deformation information of the bridge cantilever construction main beam according to the second reference coordinate and the second coordinate to be detected; determining the deformation of the bridge cantilever construction main beam according to the first deformation information and the second deformation information.

The intelligent monitoring method for deformation of a bridge cantilever construction main beam improves the accuracy of the final determined result by comprehensively determining the deformation of the bridge cantilever construction main beam based on at least one circular optical marker.

In the second aspect, an embodiment of the present application also provides an intelligent monitoring device for the deformation of a bridge cantilever construction main beam, comprising: an acquisition module, an image segmentation module and a determination module; wherein the acquisition module is used to acquire a reference image of a bridge cantilever construction main beam with a specific mark in a reference state, and an image to be detected in a state to be detected; wherein the reference image and the image to be detected are acquired under similar image acquisition conditions; the image segmentation module is used to perform image segmentation on the reference image and the image to be detected, so as to obtain a segmented reference image and a segmented image to be detected, respectively; the acquisition module is also used to acquire a reference relative position of a reference specific mark image in the segmented reference image in the segmented reference image, and a relative position to be detected of a specific mark image to be detected in the segmented image to be detected; the determination module is used to determine the deformation of the bridge cantilever construction main beam according to the reference relative position and the relative position to be detected.

The above-mentioned embodiment provides an intelligent monitoring device for deformation of a main beam of a bridge cantilever construction, which has similar beneficial effects to the intelligent monitoring method for deformation of a main beam of a bridge cantilever construction provided by the above-mentioned first aspect, or any optional implementation method of the first aspect, and will not be elaborated here.

In summary, the intelligent monitoring method and device for deformation of the bridge cantilever construction main beam provided by the present application obtains images of the bridge cantilever construction main beam with specific marks in different states, and processes and analyzes the images to obtain the position change of the specific mark compared with the image background, and determines the deformation of the bridge cantilever construction main beam according to the position change. Compared with the prior art, the accuracy of detection is improved, the implementation is simpler, and the required equipment is also simpler. And it enables construction personnel to fully and accurately grasp the linear shape of the bridge and the structural internal forces received during the construction process, thereby ensuring the construction quality of the bridge and the safety of the construction personnel. In addition, the neural network is used to perform semantic segmentation on the collected image to distinguish the image of the specific mark in the image from its background image, which just meets the requirements of the above-mentioned embodiment for image processing, so that the efficiency of executing the method steps is relatively high. In addition, by obtaining the feature points of the specific mark in the collected image, the relative displacement of the specific mark in the image background is determined based on the feature, so that the method can be applied to marked images of any shape, expanding the scope of use. In addition, through fitting, ellipse correction and circular normalization algorithms combined with gradient-based Hough transform, the circular optical marker corresponding to the center of the circle in the collected image is identified, and the deformation of the main beam of the bridge cantilever construction is determined according to the position of the center of the circle, which helps to improve the efficiency of deformation detection.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

FIG1 is a flow chart of an intelligent monitoring method for deformation of a bridge cantilever construction main beam provided in an embodiment of the present application;

FIG2 is a schematic diagram of an application of deformation monitoring of a bridge cantilever girder provided in an embodiment of the present application;

FIG3 is an example diagram of a specific deformation corresponding to a processed image provided in an embodiment of the present application;

FIG4 is a first detailed flow chart of step S160 in the intelligent monitoring method for deformation of a bridge cantilever construction main beam provided in an embodiment of the present application;

FIG5 is a second detailed flow chart of step S160 in the intelligent monitoring method for deformation of a bridge cantilever construction main beam provided in an embodiment of the present application;

FIG6 is a detailed flow chart of step S167 in the intelligent monitoring method for deformation of a bridge cantilever construction main beam provided in an embodiment of the present application;

FIG. 7 is a functional module diagram of an intelligent monitoring device for deformation of a main beam in cantilever construction of a bridge provided in an embodiment of the present application.

Detailed ways

The following embodiments of the technical solution of the present application will be described in detail in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present application, and are therefore only used as examples, and cannot be used to limit the scope of protection of the present application.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application.

In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is more than two, unless otherwise clearly and specifically defined.

The applicant has found in research that if machine vision technology is applied to deformation detection, the accuracy of deformation detection can be further improved compared to the above-mentioned background technologies such as: using theodolites, total stations, optical plummets, laser instruments, GPS positioning systems, etc. for deformation detection. Therefore, the present application provides an intelligent monitoring method and device for deformation of the main beam of a bridge cantilever construction, so as to further improve the accuracy of deformation detection. Specifically, please refer to the embodiments and drawings provided in the present application.

Please refer to Figure 1, which is a flow chart of an intelligent monitoring method for deformation of a bridge cantilever construction main beam provided in an embodiment of the present application. The intelligent monitoring method for deformation of a bridge cantilever construction main beam provided in an embodiment of the present application includes:

Step S120: Acquire a reference image of the cantilever construction main beam of the bridge with a specific mark in a reference state and an image to be detected in a state to be detected, wherein the reference image and the image to be detected are acquired under similar image acquisition conditions.

In the above step S120, please refer to Figures 2 and 3. Figure 2 is an application schematic diagram of the deformation monitoring of the bridge cantilever main beam provided in the embodiment of the present application; Figure 3 is an example diagram of the specific deformation corresponding to the processed image provided in the embodiment of the present application. The above embodiment can be specifically implemented as follows: first, two industrial cameras are respectively connected to two tripods and two computers, and are respectively placed in an open and bright place on the shore at the corresponding position of the main beam cantilever end. Then, circular optical markers are respectively pasted on the two ends of the main beam cantilever to complete the construction of the visual imaging system.

In order to achieve full monitoring of the bridge construction process, an industrial camera can continuously capture the two ends of the main beam cantilever with specific circular marks.

The specific mark may be a built-in bridge cantilever construction beam, or it may be added to the bridge cantilever construction beam for the purpose of detecting the bridge cantilever construction beam. The specific mark may be a triangular figure, a square figure, a circular figure, etc. The reference state is used as a reference. According to the changes of the bridge cantilever construction beam under the state to be detected relative to the bridge cantilever construction beam under the reference state, the specific deformation that has occurred can be determined. The reference image and the image to be detected can be obtained by an image acquisition module, such as an industrial camera. For the accuracy of subsequent results, it is necessary to acquire the reference image and the image to be detected under similar image acquisition conditions. The similar image acquisition conditions may be that when acquiring the image, the image acquisition module such as an industrial camera acquires the same distance and angle of the image, or the image acquisition module acquires the reference image and the image to be detected at the same position and the same posture, thereby ensuring that the whole or part of the bridge cantilever construction beam acquired in the reference image and the image to be detected has the same posture and size, while the temperature, weather, and time when the image is acquired may not be the same.

It should be understood that in order to achieve continuous detection of the cantilever construction main beam of the bridge, a video image of the cantilever construction main beam of the bridge can be collected within a certain period of time, and a reference image and at least two images to be detected can be determined from several frames of the video image. The specific process of deformation of the cantilever construction main beam of the bridge within the period of time can be obtained by subsequent processing and analysis of the reference image and the images to be detected.

Step S140: performing image segmentation on the reference image and the image to be detected to obtain a segmented reference image and a segmented image to be detected, respectively.

In the above step S140, the purpose of performing graphic segmentation on the reference image and the image to be detected is to separate the specific marker image in the image from the background image, so as to facilitate the subsequent processing and analysis of the specific markers in the reference image and the image to be detected. The image segmentation method can be semantic segmentation, example segmentation, and panorama segmentation.

Step S160: obtaining the reference relative position of the reference specific marker image in the segmented reference image and the relative position of the specific marker image to be detected in the segmented image to be detected.

In the above step S160, the reference relative position of the reference specific marker image in the segmented reference image in the segmented reference image may be the coordinate position of a designated point or center, centroid, etc. of the reference specific marker image in the coordinate system established in the segmented reference image. Similarly, the relative position to be detected may be the coordinate position of a corresponding point (designated point or center, centroid, etc.) in the specific marker image to be detected and the reference image in the same coordinate system established in the segmented image to be detected.

Step S180: Determine the deformation of the cantilever construction main beam of the bridge according to the reference relative position and the relative position to be detected.

In the above step S180, taking the horizontally placed crossbar as an example, when the coordinates of the relative position to be detected are located directly above or obliquely above the coordinates of the reference relative position, it means that the crossbar has bent upward; when the coordinates of the relative position to be detected are located directly below or obliquely below the coordinates of the reference relative position, it means that the crossbar has bent downward; when the coordinates of the relative position to be detected are the same as the ordinates of the reference relative position coordinates, and the coordinates of the detected relative position are located on the side close to the center of the reference relative position coordinates, it means that the crossbar has contracted, otherwise it means that it has stretched. When the coordinates of the relative position to be detected do not show any changes compared to the coordinates of the reference relative position, it means that the crossbar has not undergone any deformation during the time period.

In addition, regarding the occurrence of stretching and contraction, the specific length of the stretching or contraction can be determined based on the difference between the coordinates of the relative position to be detected and the horizontal coordinate parameters of the reference relative position coordinates. Similarly, the specific arc, curvature, etc. of the bending can also be determined based on the relationship between the coordinates of the relative position to be detected and the reference relative position coordinates.

In the above implementation process, by obtaining images of the cantilever construction main beam of the bridge with specific marks in different states under approximate conditions, and by processing the image, the position change of the specific mark image in the image is known, and the specific deformation of the cantilever construction main beam of the bridge is determined according to the position change, which improves the accuracy of detection compared with the prior art. In addition, the deformation detection of the cantilever construction main beam of the bridge by the method in the above embodiment is simpler to implement and the required equipment is simpler than using theodolite, total station, optical plumb line, laser instrument, GPS positioning system and the like. In addition, by scientifically and reasonably monitoring the process of building a prestressed rigid frame bridge by the cantilever construction method of the hanging basket, the construction personnel can comprehensively and accurately grasp the linear shape of the bridge and the structural internal force it is subjected to during the construction process, and finally ensure the construction quality of the bridge and the safety of the construction personnel.

In an optional implementation, the above step S140 includes:

Step S141: semantically segment the reference image and the image to be detected using a neural network to obtain a segmented reference image and a segmented image to be detected, respectively. The neural network includes a SegNet fully convolutional neural network.

In the above step S141, the neural network used can be a trained neural network, for example: SegNet fully convolutional neural network. Taking the SegNet fully convolutional neural network as an example, the semantic segmentation performed can be specifically implemented as follows: first, using the first 13 convolutional layers of VGG16, each convolutional layer contains convolution, batch normalization and ReLU nonlinear activation function operations, the image data can be compiled to obtain image features. Next, upsampling is performed using the index of the maximum pooling in the encoding network, and then a convolution operation is performed on it. The low-resolution feature map can be mapped to the same size as the original image, and finally sent to the softmax classifier for pixel-by-pixel classification to achieve the distinction between the reference specific marker image and its background image in the reference image, and the distinction between the specific marker image to be detected and its background image in the image to be detected.

In the above implementation process, a neural network is used to perform semantic segmentation on the collected images to distinguish the image with specific marks in the image from its background image, which just meets the requirements of image processing in the above embodiment. Compared with other panoramic segmentation and instance segmentation, which make more detailed classifications of pixels in the image, the semantic segmentation algorithm is relatively simple, which makes the efficiency of executing the method steps relatively high. In addition, the SegNet fully convolutional neural network used in the above embodiment, compared with other neural networks such as U-Net, DenseNetsE-Net and Link-Net, the SegNet fully convolutional neural network achieves a good balance between memory and precision involved in segmentation performance, and can maintain the integrity of high-frequency details in segmentation.

Please refer to FIG. 4 , which is a first detailed flow chart of step S160 in the intelligent monitoring method for deformation of a bridge cantilever construction main beam provided in an embodiment of the present application. The above step S160 includes:

Step S161: extracting a reference feature point set of a reference specific marker image in the segmented reference image and a to-be-detected feature point set of a to-be-detected specific marker image in the segmented to-be-detected image.

In the above step S161, the reference feature point set includes corner points, edge points and intersection points in the reference specific marker image; similarly, the feature point set to be detected includes corner points, edge points and intersection points in the specific marker image to be detected.

Step S162: establishing a mapping relationship between the reference feature point set and the feature point set to be detected by similarity comparison.

In the above step S162, the mapping relationship established may be a group of reference feature points and feature points to be detected having a mapping relationship, and the relationship between the reference feature points and the surrounding feature points, the position relationship in the reference specific marker image and the specific marker image to be detected, the invariant moment, the angle and other feature parameters are the same. Taking an isosceles triangle as an example, the vertex point of the isosceles triangle in the reference specific marker image corresponds to the vertex point in the specific marker image to be detected, the first base point of the isosceles triangle in the reference specific marker image corresponds to the first base point in the specific marker image to be detected, and the second base point of the isosceles triangle in the reference specific marker image corresponds to the second base point in the specific marker image to be detected.

Step S163: selecting reference feature points in the reference feature point set and feature points to be detected that have a mapping relationship with the reference feature points from among the feature points to be detected.

In the above step S163, the selected reference feature points with a mapping relationship and the feature points to be detected may be one group, two groups, three groups, etc. Continuing to take the above isosceles triangle as an example, the selected reference feature points with a mapping relationship and the feature points to be detected include: at least one of the vertex point of the isosceles triangle in the reference specific marker image and the vertex point in the specific marker image to be detected, the first base point of the isosceles triangle in the reference specific marker image and the first base point in the specific marker image to be detected, and the second base point of the isosceles triangle in the reference specific marker image and the second base point in the specific marker image to be detected.

Step S164: Obtain coordinate parameters of the reference feature point and the feature point to be detected, and use them as the reference relative position and the relative position to be detected, respectively.

In the above step S164, by obtaining the coordinate parameters of the benchmark feature points and the feature points to be detected with a mapping relationship, the relative position of the entire specific marker image to be detected in the image to be detected after segmentation is obtained, and the change in the relative position of the benchmark specific marker image in the benchmark image after segmentation is compared to obtain the deformation of the main beam of the bridge cantilever construction.

In the above implementation process, by respectively obtaining the benchmark feature point set of the benchmark specific marking image in the segmented benchmark image and the feature point set to be detected of the specific marking image to be detected in the segmented image to be detected, and after establishing a mapping relationship between the benchmark feature point set and the feature point set to be detected, at least one group of benchmark feature points and feature points to be detected with a mapping relationship is selected, and the deformation mode of the target to be detected is determined by the change in the relative position between the benchmark feature point and the feature point to be detected, so that the method can be applied to marking images of any shape, thereby expanding the scope of use.

Please refer to Figure 5, which is a second detailed flow chart of step S160 in the intelligent monitoring method for deformation of a bridge cantilever construction main beam provided in an embodiment of the present application. In an optional embodiment, the above-mentioned specific mark includes a circular optical mark.

Accordingly, the above step S160 includes:

Step S168: using the gradient-based Hough transform to identify the reference center coordinates of the reference circular image in the segmented reference image and the center coordinates of the to-be-detected circle of the to-be-detected circular image in the segmented to-be-detected image.

Step S169: using the reference circle center coordinates and the to-be-detected circle center coordinates as the reference relative position and the to-be-detected relative position respectively.

In the above steps, when the specific mark is a circular optical mark, the image acquisition module such as an industrial camera is used to acquire the segmented reference image with the reference circular image and the segmented image to be detected with the circular image to be detected. The corresponding reference circle center coordinates and the circle center coordinates to be detected are respectively identified using the gradient-based Hough transform. This is equivalent to the reference circle center coordinates and the circle center coordinates to be detected with a mapping relationship in the aforementioned embodiment. The deformation of the bridge cantilever construction main beam is determined based on the difference between the reference circle center coordinates and the circle center coordinates to be detected.

In the above implementation process, a circular optical marker is used as a specific marker, and the coordinates of its center are identified in combination with the gradient-based Hough transform, which can more efficiently determine the position change of the image to be detected and the reference image, so as to determine the deformation of the main beam of the bridge cantilever construction, thereby improving the efficiency of deformation detection. In addition, the optical marker image makes the specific marker image in the image collected by the image acquisition module such as the industrial camera more obvious, reducing the pressure of subsequent image processing and analysis, thereby further improving the efficiency of deformation detection.

Please refer to FIG. 6 , which is a detailed flow chart of step S167 in the intelligent monitoring method for deformation of a bridge cantilever construction main beam provided in an embodiment of the present application. In an optional embodiment, before the above step S168, the intelligent monitoring method for deformation of a bridge cantilever construction main beam provided in an embodiment of the present application further includes:

Step S165: performing boundary recognition on the segmented reference image and the segmented image to be detected respectively, so as to obtain reference boundary points of the segmented reference image and boundary points to be detected of the segmented image to be detected.

In the above step S165, after the image segmentation algorithm, the reference fixed mark image in the segmented reference image has been clearly distinguished, and the specific mark to be detected and the specific mark to be detected image in the segmented image to be detected have also been clearly separated. Therefore, the pixel points at the intersection of the images are the reference boundary points and the boundary points to be detected, respectively.

Step S166: fitting the reference boundary points and the boundary points to be detected respectively to obtain the reference fitting ellipse of the segmented reference image and the fitting ellipse to be detected of the segmented image to be detected.

In the above step S166, when the specific mark is a circular optical mark, and the image acquisition module is not directly facing the bridge cantilever construction main beam for image acquisition, the circular optical mark will be in an elliptical shape in the acquired reference image and the image to be detected, and then the reference specific mark image in the segmented reference image and the image to be detected in the segmented image to be detected will also be elliptical. Of course, the factors that cause the circular optical mark to be in an elliptical shape in the acquired reference image and the image to be detected are not limited to the fact that the image acquisition module is not directly facing the bridge cantilever construction main beam for image acquisition.

However, although the reference boundary points of the reference image after segmentation and the boundary points to be detected of the image to be detected after segmentation can be extracted after boundary recognition, the figure formed by all the reference boundary points is not an ellipse formed by a perfect smooth curve, and the figure formed by all the boundary points to be detected is not an ellipse in the strict sense. Therefore, it is necessary to fit the reference boundary points and the boundary points to be detected respectively. Specifically, the ellipse fitting method can be used to fit the reference boundary points and the boundary points to be detected into the reference fitting ellipse and the fitting ellipse to be detected respectively.

Step S167: performing ellipse correction on the reference fitting ellipse and the fitting ellipse to be detected respectively, so as to obtain a reference circular image in the segmented reference image and a circular image to be detected in the segmented image to be detected.

In the above step S167, the reference fitting ellipse and the fitting ellipse to be detected are respectively corrected into a reference circular image and a circular image to be detected by an ellipse correction algorithm, so as to obtain the reference circle center coordinates and the circle center coordinates to be detected having a corresponding relationship.

In the above implementation process, when the circular optical marker is in an elliptical shape in the captured reference image and the image to be detected due to various factors during the image acquisition process, the reference boundary points and the boundary points to be detected are respectively fitted into reference fitting ellipses and fitting ellipses to be detected by ellipse fitting, and further the reference fitting ellipses and fitting ellipses to be detected are respectively corrected into reference circular images and circular images to be detected according to the ellipse correction algorithm, so as to facilitate the subsequent determination of the reference circle center coordinates and the circle center coordinates to be detected, and finally the deformation of the bridge cantilever construction main beam is determined according to the reference circle center coordinates and the circle center coordinates to be detected, avoiding the need for strict acquisition conditions in the image acquisition process to make the circular optical marker appear circular in the captured image, so as to determine the deformation of the bridge cantilever construction main beam through the circle recognition algorithm. That is, the scope of use of the intelligent monitoring method for the deformation of the bridge cantilever construction main beam provided in the embodiment of the present application is expanded, and the convenience of the implementation of the method is improved.

Please refer to FIG. 4 , which shows that the above step S167 includes:

Step S1671: performing ellipse correction on the reference fitting ellipse and the fitting ellipse to be detected respectively to obtain a reference correction circle of the segmented reference image and a correction circle to be detected of the fitting ellipse to be detected.

Step S1672: performing circular normalization on the reference correction circle and the correction circle to be detected respectively, so as to obtain a reference circular image in the segmented reference image and a circular image to be detected in the segmented image to be detected.

In the above step S1672, the reference correction circle and the correction circle to be detected respectively obtained after elliptical correction are normalized to simplify the subsequent calculation process of identifying the reference center coordinates of the quasi-circular image and the center coordinates of the circular image to be detected by using the gradient-based Hough transform.

In the above implementation process, the circular image and the circular image to be detected are obtained by circular normalization, which simplifies the subsequent calculation process of determining the coordinates of the center of the reference circle and the coordinates of the center of the circle to be detected, thereby improving the efficiency of deformation detection.

In an optional embodiment, the circular optical mark includes a first circular optical mark and a second circular optical mark.

The reference circle center coordinates include the first reference coordinates corresponding to the first circular optical mark and the second reference coordinates corresponding to the second circular optical mark; the to-be-detected circle center coordinates include the first to-be-detected coordinates corresponding to the first circular optical mark and the second to-be-detected coordinates corresponding to the second circular optical mark.

Preferably, the first circular optical mark is close to the second circular optical mark, so that when capturing an image, the first circular optical mark and the second circular optical mark can be captured simultaneously.

It should be understood that, according to actual needs, those skilled in the art, the circular optical mark may further include a third circular optical mark, a fourth circular optical mark, a fifth circular optical mark, etc.

The above step S180 includes:

Step S181: determining first deformation information of a bridge cantilever construction main beam according to the first reference coordinates and the first coordinates to be detected, and determining second deformation information of a bridge cantilever construction main beam according to the second reference coordinates and the second coordinates to be detected.

Step S182: Determine the deformation of the cantilever construction main beam of the bridge according to the first deformation information and the second deformation information.

In the above steps, the deformation of the bridge cantilever construction main beam is comprehensively determined based on the first deformation information determined by the first circular optical marker and the second deformation information determined by the second circular optical marker. For example, according to the first circular optical marker, it is determined that the bridge cantilever construction main beam has a deflection of 5 cm, and according to the second circular optical marker, it is determined that the bridge cantilever construction main beam has a deflection of 3 cm. The average value of the two is a deflection of 4 cm, so it can be determined that the bridge cantilever construction main beam has a deflection of 4 cm.

Of course, when the circular optical mark may also include a third circular optical mark, a fourth circular optical mark, a fifth circular optical mark or even more, calculations may be performed based on the plurality of circular optical marks in combination with the "Error Theory" to obtain a more accurate determination result.

In the above implementation process, the deformation of the main beam of the bridge cantilever construction is comprehensively determined based on at least one circular optical marker, thereby improving the accuracy of the final determined result.

In an optional embodiment, the intelligent monitoring method for deformation of a bridge cantilever construction main beam provided in the present application is applied to deformation monitoring of a bridge cantilever main beam under construction, wherein a specific mark is located at the end of the bridge cantilever main beam under construction.

Please refer to Figures 5 and 6. Figure 5 is an application diagram of deformation monitoring of the bridge cantilever main beam provided in the embodiment of the present application; Figure 6 is a specific deformation example diagram corresponding to the processed image provided in the embodiment of the present application. The above embodiment can be specifically implemented as follows: first, two industrial cameras are respectively connected to two tripods and two computers, and are respectively placed in an open and bright place on the shore at the corresponding position of the main beam cantilever end. Then, circular optical markers are respectively pasted on the two ends of the main beam cantilever to complete the construction of the visual imaging system.

The circular optical marker may also include a first circular optical marker, a second circular optical marker, a third circular optical marker, and a fourth circular optical marker. As shown in FIG6 , the position change of the circular optical marker compared to the image background can determine the corresponding deformation of the cantilever girder of the bridge, and then the linear shape of the bridge during the construction process and the structural internal force it is subjected to can be obtained to ensure the construction quality of the bridge and the safety of the construction workers.

During the above-mentioned implementation process, the intelligent monitoring method for the deformation of the main beam of the bridge cantilever construction provided by the present application is adopted to scientifically and rationally monitor the process of building a prestressed rigid frame bridge by the hanging basket cantilever construction method, so that the construction personnel can comprehensively and accurately grasp the line shape of the bridge and the structural internal forces it is subjected to during the construction process, ultimately ensuring the construction quality of the bridge and the safety of the construction personnel.

Based on the same inventive concept, please refer to FIG. 7 , which is a functional module diagram of an intelligent monitoring device 700 for deformation of a bridge cantilever construction main beam provided in an embodiment of the present application. The intelligent monitoring device 700 for deformation of a bridge cantilever construction main beam provided in an embodiment of the present application includes: an acquisition module 710 , an image segmentation module 720 , and a determination module 730 .

Among them, the acquisition module 710 is used to obtain a reference image of the bridge cantilever construction main beam with a specific mark in a reference state, and an image to be detected in a state to be detected; wherein the reference image and the image to be detected are collected under similar image acquisition conditions; the image segmentation module 720 is used to perform image segmentation on the reference image and the image to be detected to obtain a segmented reference image and a segmented image to be detected, respectively; the acquisition module 710 is also used to obtain the reference relative position of the reference specific mark image in the segmented reference image in the segmented reference image, and the relative position to be detected of the specific mark image to be detected in the segmented image to be detected; the determination module 730 is used to determine the deformation of the bridge cantilever construction main beam according to the reference relative position and the relative position to be detected.

Continuing to refer to FIG. 7 , in an optional implementation, in the process of performing image segmentation on the reference image and the image to be detected to obtain the segmented reference image and the segmented image to be detected, the image segmentation module 720 is specifically used to perform semantic segmentation on the reference image and the image to be detected using a neural network to obtain the segmented reference image and the segmented image to be detected, respectively. The neural network includes a SegNet fully convolutional neural network.

Please continue to refer to Figure 7. In an optional embodiment, in the process of obtaining the benchmark relative position of the benchmark fixed mark image in the segmented benchmark image in the segmented image to be detected, and the relative position to be detected of the specific mark image to be detected in the segmented image to be detected, the above-mentioned acquisition module 710 is specifically used to: extract the benchmark feature point set of the benchmark specific mark image in the segmented benchmark image, and the feature point set to be detected of the specific mark image to be detected in the segmented image to be detected; establish a mapping relationship between the benchmark feature point set and the feature point set to be detected by similarity comparison; select the benchmark feature points in the benchmark feature point set and the feature points to be detected that have a mapping relationship with the benchmark feature points in the feature points to be detected; and obtain the coordinate parameters of the benchmark feature points and the feature points to be detected, and use them as the benchmark relative position and the relative position to be detected, respectively.

Please continue to refer to FIG. 7 , in an optional embodiment, the specific mark includes a circular optical mark.

Correspondingly, in the process of obtaining the reference relative position of the reference specific marker image in the segmented reference image and the relative position to be detected of the specific marker image to be detected in the segmented image to be detected, the acquisition module 710 is specifically used to: identify the reference center coordinates of the reference circular image in the segmented reference image and the center coordinates of the circular image to be detected in the segmented image to be detected by using the gradient-based Hough transform; and use the reference center coordinates and the center coordinates of the circle to be detected as the reference relative position and the relative position to be detected, respectively.

Please continue to refer to Figure 7. In an optional implementation, before using the gradient-based Hough transform to identify the reference center coordinates of the circular image in the segmented reference image and the center coordinates of the circular image to be detected in the segmented image to be detected, the above-mentioned acquisition module 710 is specifically used to: perform boundary identification on the segmented reference image and the segmented image to be detected, respectively, to obtain reference boundary points of the segmented reference image and boundary points to be detected of the segmented image to be detected; fit the reference boundary points and the boundary points to be detected, respectively, to obtain a reference fitting ellipse of the segmented reference image and a fitting ellipse to be detected of the segmented image to be detected; and perform ellipse correction on the reference fitting ellipse and the fitting ellipse to be detected, respectively, to obtain a reference circular image in the segmented reference image and a circular image to be detected in the segmented image to be detected.

Please continue to refer to Figure 7. In an optional embodiment, in the process of performing elliptical correction on the baseline fitting ellipse and the fitted ellipse to be detected respectively to obtain the baseline circular image in the segmented baseline image and the circular image to be detected in the segmented image to be detected, the above-mentioned acquisition module 710 is specifically used to: perform elliptical correction on the baseline fitting ellipse and the fitted ellipse to be detected respectively to obtain the baseline corrected circle of the segmented baseline image and the corrected circle to be detected of the fitted ellipse to be detected; and perform circular normalization on the baseline corrected circle and the corrected circle to be detected respectively to obtain the baseline circular image in the segmented baseline image and the circular image to be detected in the segmented image to be detected.

Please continue to refer to FIG. 7 . In an optional implementation, the circular optical mark includes a first circular optical mark and a second circular optical mark.

The reference circle center coordinates include the first reference coordinates corresponding to the first circular optical mark and the second reference coordinates corresponding to the second circular optical mark; the circle center coordinates to be detected include the first coordinates to be detected corresponding to the first circular optical mark and the second coordinates to be detected corresponding to the second circular optical mark.

In the process of determining the deformation of the main beam of the bridge cantilever construction according to the reference relative position and the relative position to be detected, the above-mentioned determination module 730 is specifically used to: determine the first deformation information of the main beam of the bridge cantilever construction according to the first reference coordinate and the first coordinate to be detected, and determine the second deformation information of the main beam of the bridge cantilever construction according to the second reference coordinate and the second coordinate to be detected; determine the deformation of the main beam of the bridge cantilever construction according to the first deformation information and the second deformation information.

It should be understood that the device corresponds to the above-mentioned intelligent monitoring method embodiment of the deformation of the main beam of the bridge cantilever construction, and can execute the various steps involved in the above-mentioned method embodiment. The specific functions of the device can refer to the description above. To avoid repetition, the detailed description is appropriately omitted here. The device includes at least one software function module that can be stored in the memory in the form of software or firmware or solidified in the operating system (OS) of the device.

In summary, the intelligent monitoring method and device for the deformation of the bridge cantilever construction main beam provided in the embodiment of the present application, by acquiring the image of the bridge cantilever construction main beam with a specific mark in different states, and by processing and analyzing the image, the position change of the specific mark compared with the image background is known, and the deformation of the bridge cantilever construction main beam is determined according to the position change. Compared with the prior art, the accuracy of the detection is improved, the implementation is simpler, and the required equipment is also simpler, so that the construction personnel can fully and accurately grasp the linear shape of the bridge during the construction process and the structural internal force received, thereby ensuring the construction quality of the bridge and the safety of the construction personnel. In addition, the semantic segmentation of the collected image is performed by a neural network to distinguish the image of the specific mark in the image from its background image, which just meets the requirements of the above embodiment for processing the image, so that the efficiency of executing the method steps is relatively high. In addition, by acquiring the feature points of the specific mark in the collected image, the relative displacement of the specific mark in the image background is determined based on the feature, so that the method can be applied to the marked image of any shape, expanding the scope of use. In addition, through fitting, ellipse correction and circular normalization algorithms combined with gradient-based Hough transform, the circular optical marker corresponding to the center of the circle in the collected image is identified, and the deformation of the main beam of the bridge cantilever construction is determined according to the position of the center of the circle, which helps to improve the efficiency of deformation detection.

In several embodiments provided by the embodiments of the present application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely schematic. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices and methods according to the multiple embodiments of the embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of the code, and a part of the module, program segment or code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and/or flowchart, and the combination of boxes in the block diagram and/or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

In addition, the functional modules in each embodiment of the present application can be integrated together to form an independent part, or each module can exist separately, or two or more modules can be integrated to form an independent part.

The above description is only an optional implementation manner of the embodiments of the present application, but the protection scope of the embodiments of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or replacements within the technical scope disclosed in the embodiments of the present application, which should be covered within the protection scope of the embodiments of the present application.

Claims (4)

1. An intelligent monitoring method for deformation of a girder of a bridge cantilever construction is characterized by comprising the following steps:

Acquiring a reference image of a bridge cantilever construction girder with a specific mark in a reference state and an image to be detected in a state to be detected; the reference image and the image to be detected are acquired under the approximate image acquisition condition;

image segmentation is carried out on the reference image and the image to be detected so as to obtain a segmented reference image and a segmented image to be detected respectively;

acquiring a reference relative position of a reference specific mark image in the segmented reference image and a to-be-detected relative position of a to-be-detected specific mark image in the segmented to-be-detected image; and

Determining deformation of the bridge cantilever construction girder according to the reference relative position and the relative position to be detected;

The obtaining the reference relative position of the reference specific mark image in the segmented reference image in the segmented image to be detected and the relative position of the specific mark image to be detected in the segmented image to be detected includes:

Extracting a reference characteristic point set of a reference specific mark image in the segmented reference image and a to-be-detected characteristic point set of a to-be-detected specific mark image in the segmented to-be-detected image;

Establishing a mapping relation between the reference feature point set and the feature point set to be detected through similarity comparison;

Selecting reference feature points in the reference feature point set and feature points to be detected, wherein the feature points to be detected have the mapping relation with the reference feature points in the feature points to be detected; and

Coordinate parameters of the reference feature points and the feature points to be detected are obtained and used as the reference relative position and the relative position to be detected respectively;

wherein the specific mark comprises a circular optical mark; the obtaining the reference relative position of the reference specific mark image in the segmented reference image and the to-be-detected relative position of the to-be-detected specific mark image in the segmented to-be-detected image, includes:

Identifying the reference circle center coordinates of the reference circular image in the segmented reference image and the circle center coordinates to be detected of the circular image to be detected in the segmented image by utilizing Hough transformation based on gradients; and

Taking the reference circle center coordinate and the circle center coordinate to be detected as the reference relative position and the relative position to be detected respectively;

the method further comprises the steps of:

Respectively carrying out boundary recognition on the segmented reference image and the segmented image to be detected to obtain a reference boundary point of the segmented reference image and a boundary point to be detected of the segmented image to be detected;

Fitting the reference boundary points and the boundary points to be detected respectively to obtain a reference fitted ellipse of the segmented reference image and a fitted ellipse to be detected of the segmented image to be detected; and

Respectively carrying out ellipse correction on the reference fitting ellipse and the fitting ellipse to be detected so as to obtain a reference circular image in the segmented reference image and a circular image to be detected in the segmented image to be detected;

The elliptical correction is performed on the reference fitted ellipse and the fitted ellipse to be detected respectively to obtain a reference circular image in the segmented reference image and a circular image to be detected in the segmented image to be detected, including:

Respectively carrying out ellipse correction on the reference fitting ellipse and the fitting ellipse to be detected to obtain a reference correction circle of the segmented reference image and a correction circle to be detected of the fitting ellipse to be detected; and

And respectively carrying out circular normalization on the reference correction circle and the correction circle to be detected so as to obtain a reference circular image in the segmented reference image and a circular image to be detected in the segmented image to be detected.

2. The intelligent monitoring method for deformation of the girder of the bridge cantilever construction according to claim 1, wherein the image segmentation is performed on the reference image and the image to be detected to obtain a segmented reference image and a segmented image to be detected, respectively, includes:

semantic segmentation is respectively carried out on the reference image and the image to be detected by using a neural network so as to respectively obtain the segmented reference image and the segmented image to be detected; wherein the neural network comprises SegNet full convolution neural networks.

3. The intelligent monitoring method of bridge cantilever construction girder deformation according to claim 1, wherein the circular optical markers comprise a first circular optical marker and a second circular optical marker;

The reference circle center coordinates comprise first reference coordinates corresponding to the first round optical marks and second reference coordinates corresponding to the second round optical marks; the circle center coordinates to be detected comprise first coordinates to be detected corresponding to the first circular optical mark and second coordinates to be detected corresponding to the second circular optical mark;

The deformation of the bridge cantilever construction girder is determined according to the reference relative position and the relative position to be detected, and the method comprises the following steps:

Determining first deformation information of the bridge cantilever construction girder according to the first reference coordinate and the first coordinate to be detected, and determining second deformation information of the bridge cantilever construction girder according to the second reference coordinate and the second coordinate to be detected;

and determining the deformation of the bridge cantilever construction girder according to the first deformation information and the second deformation information.

4. Intelligent monitoring device that bridge cantilever construction girder warp, its characterized in that includes: the device comprises an acquisition module, an image segmentation module and a determination module;

The acquisition module is used for acquiring a reference image of the bridge cantilever construction girder with the specific mark in a reference state and an image to be detected in a state to be detected; the reference image and the image to be detected are acquired under the approximate image acquisition condition;

the image segmentation module is used for carrying out image segmentation on the reference image and the image to be detected so as to obtain a segmented reference image and a segmented image to be detected respectively;

the acquisition module is further used for acquiring a reference relative position of a reference specific mark image in the segmented reference image and a to-be-detected relative position of a to-be-detected specific mark image in the segmented to-be-detected image;

The determining module is used for determining the deformation of the bridge cantilever construction girder according to the reference relative position and the relative position to be detected;

In the process of acquiring the reference relative position of the reference specific mark image in the segmented reference image in the segmented image to be detected and the relative position of the specific mark image to be detected in the segmented image to be detected, the acquisition module is specifically configured to: extracting a reference characteristic point set of a reference specific mark image in the segmented reference image and a to-be-detected characteristic point set of a to-be-detected specific mark image in the segmented to-be-detected image; establishing a mapping relation between the reference feature point set and the feature point set to be detected through similarity comparison; selecting reference feature points in the reference feature point set and feature points to be detected, wherein the feature points to be detected have the mapping relation with the reference feature points in the feature points to be detected; the coordinate parameters of the reference feature points and the feature points to be detected are obtained and used as the reference relative position and the relative position to be detected respectively;

wherein the specific mark comprises a circular optical mark; in the process of acquiring the reference relative position of the reference specific mark image in the segmented reference image and the to-be-detected relative position of the to-be-detected specific mark image in the segmented to-be-detected image, the acquisition module is specifically configured to: identifying the reference circle center coordinates of the reference circular image in the segmented reference image and the circle center coordinates to be detected of the circular image to be detected in the segmented image by utilizing Hough transformation based on gradients; taking the reference circle center coordinate and the circle center coordinate to be detected as the reference relative position and the relative position to be detected respectively;

Before the reference circle center coordinates of the circular image in the segmented reference image and the circle center coordinates to be detected of the circular image in the segmented image are identified by utilizing the hough transform based on gradient, the acquisition module is specifically further configured to: respectively carrying out boundary recognition on the segmented reference image and the segmented image to be detected to obtain a reference boundary point of the segmented reference image and a boundary point to be detected of the segmented image to be detected; fitting the reference boundary points and the boundary points to be detected respectively to obtain a reference fitted ellipse of the segmented reference image and a fitted ellipse to be detected of the segmented image to be detected; respectively carrying out ellipse correction on the reference fitting ellipse and the fitting ellipse to be detected so as to obtain a reference circular image in the segmented reference image and a circular image to be detected in the segmented image to be detected;

in the process of respectively carrying out ellipse correction on the reference fitting ellipse and the fitting ellipse to be detected so as to obtain a reference circular image in the segmented reference image and a circular image to be detected in the segmented image to be detected, the obtaining module is specifically configured to: respectively carrying out ellipse correction on the reference fitting ellipse and the fitting ellipse to be detected to obtain a reference correction circle of the segmented reference image and a correction circle to be detected of the fitting ellipse to be detected; and respectively carrying out circular normalization on the reference correction circle and the correction circle to be detected so as to obtain a reference circular image in the segmented reference image and a circular image to be detected in the segmented image to be detected.