Search

CN-122023366-A - Intelligent bridge template deformation monitoring method integrating machine vision

CN122023366ACN 122023366 ACN122023366 ACN 122023366ACN-122023366-A

Abstract

The invention discloses an intelligent bridge template deformation monitoring method integrating machine vision, which belongs to the crossing field of computer image processing and civil engineering monitoring technology and comprises the steps of generating a theoretical edge feature map according to an acquired three-dimensional model and calibration parameters, performing distance transformation on the theoretical edge feature map to generate a two-dimensional distance potential energy field, collecting video streams, extracting current video frames to generate an actual measurement edge feature map, constructing a reverse projection objective function based on the actual measurement edge feature map and the two-dimensional distance potential energy field, solving the reverse projection objective function to generate a total displacement field, performing orthogonal decomposition, decoupling a rigid motion component, an accumulated plastic deformation component and a local elastic deformation component, and generating a deformation visualization result and a safety early warning signal based on the local elastic deformation component and the accumulated plastic deformation component. And the total displacement field is solved by adopting the reverse projection based on the three-dimensional model and the orthogonal decomposition scheme is carried out on the total displacement field, so that the rigid motion and the real deformation of the structure can be separated, and the full-field deformation monitoring of the bridge template is realized.

Inventors

  • ZHAO MANXIANG

Assignees

  • 山东中菏桥梁模板有限公司

Dates

Publication Date
20260512
Application Date
20260205

Claims (10)

  1. 1. The intelligent bridge template deformation monitoring method integrating machine vision is characterized by comprising the following steps of: acquiring a three-dimensional model for describing the geometric form of the bridge template and calibration parameters for describing the imaging relation of a camera, generating a theoretical edge feature map according to the three-dimensional model and the calibration parameters, performing distance transformation on the theoretical edge feature map, and generating a two-dimensional distance potential energy field; extracting a current video frame from a video stream acquired by a field camera, and extracting an actually measured edge feature map from the current video frame; Constructing a reverse projection objective function based on the actually measured edge feature map and the two-dimensional distance potential energy field, and solving to generate a total displacement field; Performing orthogonal decomposition on the total displacement field, and decoupling a rigid motion component, an accumulated plastic deformation component and a local elastic deformation component; and generating a deformation visualization result and a safety early warning signal based on the local elastic deformation component and the accumulated plastic deformation component.
  2. 2. The intelligent bridge template deformation monitoring method based on machine vision integration according to claim 1, wherein the generating a theoretical edge feature map comprises: acquiring a deformation initial field for simulating micro deformation of the bridge template in an initial stress state; applying the deformation initiation field to the three-dimensional model to generate a set of candidate three-dimensional models; Rendering each candidate three-dimensional model in the candidate three-dimensional model set according to the calibration parameters to generate a candidate reference feature image set; Extracting a first frame image of the video stream, and processing the first frame image to generate a first frame edge feature map; Calculating the image quality confidence coefficient weight of each region of the first frame edge feature map, and generating comprehensive matching degree according to the matching degree of the image quality confidence coefficient weight and the first frame edge feature map and the candidate reference feature map set; and selecting a feature map with the highest matching degree from the candidate reference feature map set according to the comprehensive matching degree as a theoretical edge feature map.
  3. 3. The intelligent bridge formwork deformation monitoring method based on machine vision integration according to claim 2, wherein the generating the total displacement field comprises: determining an initial state of the three-dimensional model according to the theoretical edge feature diagram, and initializing vertex displacement; Forward projection is carried out according to the current vertex displacement and the calibration parameters, and a calculated edge feature map is generated; Comparing the calculated edge feature map with the actually measured edge feature map to generate a residual map; Analyzing a spatial distribution mode of the residual map, dynamically adjusting the regional weight of the two-dimensional distance potential energy field according to the spatial distribution mode, and generating a self-adaptive distance potential energy field; and iteratively optimizing the vertex displacement according to the smooth physical constraint of the self-adaptive distance potential energy field and the adjacent vertex displacement, and generating a total displacement field.
  4. 4. The intelligent bridge formwork deformation monitoring method based on machine vision integration according to claim 3, wherein the generating the self-adaptive distance potential energy field comprises: identifying the distribution form of the pixel points in the residual image and classifying the distribution form into a random noise mode, a directional stripe mode or a structural continuous mode; Acquiring an image area corresponding to the random noise pattern and the directional stripe pattern, and generating a weight factor for reducing the influence of the image area in the two-dimensional distance potential energy field; and applying the weight factors to the two-dimensional distance potential energy field to generate an adaptive distance potential energy field.
  5. 5. The intelligent bridge formwork deformation monitoring method based on machine vision integration according to claim 3, wherein the method further comprises: Acquiring the convergence speed of the iterative optimization vertex displacement process, combining the spatial distribution mode of the residual error map and the image quality confidence weight, and comprehensively calculating to generate the optimal result confidence of the current frame; And correlating the confidence coefficient of the optimized result with the total displacement field to generate displacement data with a confidence coefficient label.
  6. 6. The intelligent bridge template deformation monitoring method based on machine vision integration according to claim 1, wherein the performing orthogonal decomposition on the total displacement field comprises: collecting a plurality of total displacement fields generated by continuous video frames, and constructing a displacement time sequence for each vertex of the three-dimensional model; performing time-frequency analysis on the displacement time sequence, and decomposing the displacement time sequence into a high-frequency displacement component, a middle-low frequency displacement component and an accumulated plastic deformation component; acquiring a vibration event time point generated by external construction equipment, and filtering out an instantaneous interference component by correlating the vibration event time point with the occurrence time of a high-frequency displacement component; and analyzing the spatial distribution consistency of the medium-low frequency displacement components, and separating out the rigid displacement components and the local elastic deformation components.
  7. 7. The intelligent bridge formwork deformation monitoring method based on machine vision integration according to claim 1, wherein the generating of the deformation visualization result and the safety pre-warning signal comprises: Vector synthesis is carried out on the basis of the local elastic deformation component and the accumulated plastic deformation component, and a net deformation field is generated; acquiring a safe deformation threshold value for defining a deformation risk level; And comparing the size of the net deformation field with a safety deformation threshold value, triggering a safety early warning signal, rendering the net deformation field to the surface of the three-dimensional model, and generating the deformation visualization result.
  8. 8. The intelligent bridge formwork deformation monitoring method based on machine vision integration according to claim 7, wherein the method further comprises: Evaluating the integrated credibility of the net deformation field; And acquiring a confidence threshold value for judging whether the result is reliable, reversely fusing the accumulated plastic deformation component to the three-dimensional model when the comprehensive reliability is higher than the confidence threshold value, generating an updated three-dimensional reference model, and executing a subsequent monitoring process by adopting the updated three-dimensional reference model.
  9. 9. The intelligent bridge template deformation monitoring method based on machine vision integration according to claim 8, wherein the generating the updated three-dimensional reference model comprises: Acquiring a proportionality coefficient used for a smoothing model updating process; attenuating the accumulated plastic deformation component according to a proportionality coefficient to generate attenuated node displacement; and applying the attenuated node displacement to the corresponding vertex of the three-dimensional model to update the geometric data of the three-dimensional model, and generating an updated three-dimensional reference model.
  10. 10. A bridge form deformation intelligent monitoring system integrating machine vision, which is applied to a bridge form deformation intelligent monitoring method integrating machine vision as claimed in any one of claims 1 to 9, and is characterized in that the system comprises: The reference construction module is used for acquiring a three-dimensional model for describing the geometric form of the bridge template and calibration parameters for describing the imaging relation of the camera, generating a theoretical edge feature map according to the three-dimensional model and the calibration parameters, performing distance transformation on the theoretical edge feature map and generating a two-dimensional distance potential energy field; the feature extraction module is used for extracting a current video frame from a video stream acquired by the field camera and extracting an actually measured edge feature map from the current video frame; the deformation calculation module is used for constructing a reverse projection objective function based on the actually measured edge feature map and the two-dimensional distance potential energy field and solving the reverse projection objective function to generate a total displacement field; The decoupling analysis module is used for performing orthogonal decomposition on the total displacement field and decoupling a rigid motion component, an accumulated plastic deformation component and a local elastic deformation component; And the output early warning module is used for generating a deformation visualization result and a safety early warning signal based on the local elastic deformation component and the accumulated plastic deformation component.

Description

Intelligent bridge template deformation monitoring method integrating machine vision Technical Field The invention relates to the field of intersection of computer image processing and civil engineering monitoring technology, in particular to an intelligent bridge template deformation monitoring method integrating machine vision. Background The bridge template is a key temporary support system in bridge superstructure construction, and the stability and the form of the bridge template under the action of heavy loads such as concrete pouring and the like are directly related to the linear precision and the structural safety of a final bridge. Therefore, it is important to monitor the deformation of the bridge formwork accurately in real time during the construction process. Non-contact monitoring using machine vision technology has become an important development in this field. In the related technology, china patent publication No. CN120160585B discloses a pumped storage dam deformation monitoring method based on Beidou positioning, which comprises the steps of deploying a double-frequency Beidou receiver array and an environment sensor, fusing Beidou observation data, a database water pressure gradient, a foundation vibration spectrum and three-dimensional geological structure data, constructing an anti-difference calculation model constrained by a geological model, outputting a deformation parameter confidence interval and inhibiting abnormal values through robust estimation, matching a historical working condition library by adopting a dynamic time warping algorithm, dynamically calibrating an early warning threshold value by combining a closed loop feedback mechanism, and reversely optimizing a monitoring network. However, the above monitoring scheme based on satellite positioning or contact sensor can only obtain displacement data of sparse discrete points, and it is difficult to capture local bulge or continuous curvature change of bridge templates in stress concentration areas finely, and it is difficult to deploy in high density in scaffold systems with limited space due to large equipment volume. While existing machine vision methods have full-field monitoring potential, most rely on pasting high-contrast artificial marker points, i.e., targets, on the surface being measured or optical flow tracking using rich textures of the object surface. For a steel template with smooth surface and deficient texture characteristics and in a strong vibration environment of concrete pouring, the manual mark is extremely easy to lose efficacy due to mud splashing shielding or construction scraping, and the conventional visual algorithm is difficult to effectively distinguish self-shaking of a camera, namely rigid motion, from real deformation of the template, namely elastic deformation, so that the signal-to-noise ratio of monitoring data is low, and the requirements of high-precision and instant safety early warning on a construction site are difficult to meet. Disclosure of Invention In order to solve the problems, the invention provides an intelligent bridge template deformation monitoring method integrating machine vision, which adopts a technical scheme of solving a total displacement field based on reverse projection of a three-dimensional model and carrying out orthogonal decomposition on the total displacement field, and can effectively separate rigid motion from structure real deformation, thereby realizing high-precision full-field deformation monitoring of the bridge template. The above object can be achieved by the following scheme: A bridge template deformation intelligent monitoring method integrating machine vision comprises the steps of obtaining a three-dimensional model for describing bridge template geometric forms and calibration parameters for describing camera imaging relations, generating a theoretical edge feature map according to the three-dimensional model and the calibration parameters, performing distance transformation on the theoretical edge feature map to generate a two-dimensional distance potential energy field, extracting a current video frame from a video stream collected by a field camera, extracting an actual measurement edge feature map from the current video frame, constructing a reverse projection objective function based on the actual measurement edge feature map and the two-dimensional distance potential energy field, solving to generate a total displacement field, performing orthogonal decomposition on the total displacement field, decoupling a rigid body motion component, an accumulated plastic deformation component and a local elastic deformation component, and generating a deformation visualization result and a safety early warning signal based on the local elastic deformation component and the accumulated plastic deformation component. The method comprises the steps of obtaining a deformation initial field used for simulating small deformation of a bridge template in