Search

CN-122015716-A - Curtain wall structure adhesive operation and maintenance safety monitoring system and method based on multispectral-visual morphology bimodal fusion

CN122015716ACN 122015716 ACN122015716 ACN 122015716ACN-122015716-A

Abstract

The embodiment of the application provides a curtain wall structure adhesive operation and maintenance safety monitoring system and method based on multi-spectrum-visual morphology bimodal fusion, and belongs to the technical field of building curtain wall safety monitoring. The system comprises an optical detection module, a visual detection module, an environment sensor assembly and a control and fusion decision module. The method comprises the steps of obtaining multiband reflection spectrum data and surface morphology images of the structural adhesive in parallel, performing temperature compensation and illumination normalization processing on the spectrum data based on real-time environment data, extracting optical features and visual morphology features, calculating respective aging grades and confidence degrees, dynamically calculating fusion weights according to the bimodal confidence degrees, carrying out weighted fusion on the aging grades of two modes, and outputting a comprehensive aging evaluation result. According to the application, the detection stability is improved through environment self-adaptive compensation, and the comprehensive, accurate and reliable evaluation of structural adhesive from internal chemical aging to external physical damage is realized by utilizing an intelligent fusion mechanism based on confidence coefficient.

Inventors

  • LI ZHIYUAN
  • Qiu longxiang
  • LIU JUNMING

Assignees

  • 华南理工大学
  • 广东晶启激光科技有限公司

Dates

Publication Date
20260512
Application Date
20251225

Claims (10)

  1. 1. Curtain wall construction adhesive operation and maintenance safety monitoring system based on multispectral-visual morphology bimodal fusion is characterized by comprising: the optical detection module comprises a wide-spectrum light source, an optical assembly, a multiband filter rotating wheel, a photoelectric detector and an analog-to-digital converter and is used for acquiring multiband reflection spectrum data of the structural adhesive sample; The visual detection module comprises an illumination light source and an industrial camera and is used for acquiring a surface morphology image of the structural adhesive sample; The environment sensor assembly is used for acquiring temperature and illuminance data of a detection environment in real time; The control and fusion decision module is respectively in communication connection with the optical detection module, the visual detection module and the environment sensor assembly; wherein the control and fusion decision module is configured to: Based on the data of the environment sensor assembly, performing temperature compensation and illumination normalization processing on the multiband reflection spectrum data to eliminate environment interference; extracting optical features from the processed multiband reflectance spectrum data, and extracting visual morphology features from the surface morphology images; respectively calculating to obtain an optical modal aging grade and the confidence coefficient thereof, and a visual modal aging grade and the confidence coefficient thereof according to the optical characteristics and the visual morphology characteristics; Dynamically calculating fusion weights of the two modes based on the optical mode confidence coefficient and the visual mode confidence coefficient; And carrying out weighted fusion on the optical mode aging grade and the visual mode aging grade according to the fusion weight, and outputting a comprehensive aging grade evaluation result.
  2. 2. The monitoring system of claim 1, wherein the multi-band filter wheel comprises at least three filters of different wavelength bands covering a spectral range of 320nm to 1100nm.
  3. 3. The monitoring system of claim 1, wherein the control and fusion decision module comprises: an environment compensation unit for performing the temperature compensation and illumination normalization processing; An optical characteristic extraction unit for extracting broadband spectral distribution curve characteristics and characteristic peak parameters related to chemical bond changes from the compensated spectral data; And the visual characteristic extraction unit is used for extracting at least one morphological characteristic parameter of crack density, surface roughness and peeling area proportion from the surface morphology image by adopting a multiscale morphological gradient and self-adaptive region growth algorithm.
  4. 4. The monitoring system according to claim 3, wherein the optical feature extraction unit is configured to use a partial least square method to reduce the dimension of the high-dimensional spectral feature, and input the feature after the dimension reduction into a support vector machine classification model to obtain the optical modal aging level and the classification probability.
  5. 5. The monitoring system of claim 1, wherein the control and fusion decision module is configured to dynamically calculate the fusion weights according to the following formula: Wherein, the And Respectively fusion weights of the optical mode and the visual mode, And The calculated optical mode confidence and the calculated visual mode confidence are respectively obtained.
  6. 6. The monitoring system of claim 5, wherein the optical modality confidence level And visual modality confidence And calculating at least one factor of the signal-to-noise ratio, the classification probability and the feature consistency based on the corresponding mode.
  7. 7. The monitoring system of claim 1, wherein the control and fusion decision module further comprises a collision detection unit configured to: Calculating a difference value between the optical modal ageing grade and the visual modal ageing grade; And triggering a historical data cross-validation or secondary checking process when the difference value exceeds a first preset threshold value.
  8. 8. A method of operation based on the monitoring system of any one of claims 1 to 7, comprising the steps of: optical detection and visual detection are executed in parallel, namely multiband reflection spectrum data of a structural adhesive sample are obtained through the optical detection module, and meanwhile, a surface topography image of the sample is obtained through the visual detection module; Acquiring environmental data, namely acquiring environmental temperature and illuminance in real time through an environmental sensor assembly; performing environment self-adaptive compensation, namely performing temperature compensation and illumination normalization processing on the multiband reflection spectrum data based on the environment temperature and the illumination; extracting optical features from the compensated spectrum data and visual morphology features from the surface morphology images; The confidence evaluation, namely respectively calculating an optical modal aging level and the confidence thereof, and a visual modal aging level and the confidence thereof based on the optical features and the visual morphology features; and dynamically fusing the decision, namely dynamically calculating the fusion weight according to the optical mode confidence coefficient and the visual mode confidence coefficient, and carrying out weighted fusion on the ageing grades of the two modes by utilizing the weight to output the comprehensive ageing grade.
  9. 9. The method of claim 8, wherein the step of extracting visual topographical features comprises: processing the surface morphology image by adopting a multi-scale morphological gradient algorithm to obtain a multi-scale gradient map and fusing the multi-scale gradient map; dividing the fused gradient map based on a local self-adaptive threshold value to obtain candidate crack pixels; and growing and verifying the candidate crack pixels by adopting an adaptive region growing algorithm so as to quantify the crack density.
  10. 10. The method of operation of claim 8, further comprising the step of conflict handling: comparing the optical modality aging level with the visual modality aging level; if the difference between the two is beyond the preset range, at least one operation of starting the historical detection data for cross verification, triggering the system for secondary detection or generating prompt information needing manual review is selected according to the difference degree.

Description

Curtain wall structure adhesive operation and maintenance safety monitoring system and method based on multispectral-visual morphology bimodal fusion Technical Field The application relates to the technical field of building curtain wall safety monitoring, in particular to a curtain wall structural adhesive operation and maintenance safety monitoring system and method based on multispectral-visual morphology bimodal fusion. Background The glass structural adhesive is a key bonding and sealing material of a modern building curtain wall system, and the durability of the glass structural adhesive is directly related to the overall safety and service life of the curtain wall. The structural adhesive is exposed in severe environments such as ultraviolet rays, temperature and humidity circulation, acid rain and the like for a long time, chemical degradation and physical degradation can occur, so that the adhesive force is reduced, the elasticity is lost, and serious safety accidents such as glass plate falling and the like can be possibly caused finally. Therefore, the regular and effective safety monitoring of the existing curtain wall structural adhesive is important. However, the operation and maintenance detection of the existing building curtain wall has a plurality of unique challenges, namely firstly, the initial state reference data is generally lacked, an accurate damage evolution model is difficult to build, secondly, the field environment is complex and changeable (such as illumination and temperature fluctuation), the requirements on the stability and the robustness of the detection technology are extremely high, and furthermore, the aging of the structural adhesive is a complex process comprising internal chemical structure change and external physical shape damage, and the single detection means is difficult to comprehensively characterize. At present, the common nondestructive detection method mainly has the following limitations: 1) Spectrum detection methods (such as infrared and Raman spectrum) generally reflect a single aging mechanism only aiming at specific wave bands or chemical bonds, and cannot comprehensively capture a complex aging process of multiple mechanisms of the structural adhesive. In addition, these methods are extremely sensitive to changes in ambient temperature and illumination, and lack of effective on-site compensation means, resulting in poor comparability of the detection results. 2) Visual inspection relies on visible light imaging to identify only surface defects (e.g., cracks, flaking) that have progressed to macroscopic scale, and is unable to detect early, internal chemical structural degradation, early warning hysteresis. 3) The information fusion is simple, even if research attempts are made to combine multiple technologies, a simple average or superposition mode of fixed weights is adopted, and an intelligent decision mechanism is lacked. When the detection results of different modes are contradictory, the detection results cannot be effectively identified and processed, and the misjudgment rate is high. Therefore, an intelligent operation and maintenance safety monitoring technology and system for curtain wall structural adhesive, which can adapt to complex field environments, fuse internal and external state information and realize early warning and accurate assessment, are urgently needed. Disclosure of Invention The embodiment of the application mainly aims to provide a curtain wall structural adhesive operation and maintenance safety monitoring system and method based on multi-spectrum-visual morphology bimodal fusion, which realize comprehensive, accurate and real-time evaluation of the aging state of structural adhesive through broadband multi-spectrum analysis, visual morphology detection, environment self-adaptive compensation and intelligent fusion decision based on confidence. In order to achieve the above objective, an aspect of the embodiments of the present application provides a curtain wall structural adhesive operation and maintenance safety monitoring system based on a dual-mode fusion of multispectral-visual morphology, including: the optical detection module comprises a wide-spectrum light source, an optical assembly, a multiband filter rotating wheel, a photoelectric detector and an analog-to-digital converter and is used for acquiring multiband reflection spectrum data of the structural adhesive sample; The visual detection module comprises an illumination light source and an industrial camera and is used for acquiring a surface morphology image of the structural adhesive sample; The environment sensor assembly is used for acquiring temperature and illuminance data of a detection environment in real time; The control and fusion decision module is respectively in communication connection with the optical detection module, the visual detection module and the environment sensor assembly; wherein the control and fusion decision module is configur