CN-122023398-A - Method and system for detecting damage of wall inside historical building based on image vision

CN122023398ACN 122023398 ACN122023398 ACN 122023398ACN-122023398-A

Abstract

The application provides a method and a system for detecting wall damage in a historical building based on image vision, and belongs to the field of image vision. The method comprises the steps of building an enhanced image according to an original image of a wall body in a historical building, executing structural semantic partition processing according to the enhanced image to generate a structural semantic region diagram, executing residual form extraction processing of the enhanced image in each wall body structural region by utilizing the structural semantic region diagram to construct a wall body residual form feature set, generating a residual form response field according to the structural semantic region diagram and the wall body residual form feature set, forming a wall body residual propagation feature diagram by combining the structural semantic region, identifying a propagation path, calculating a residual propagation intensity index, and generating a wall body residual detection result by utilizing the residual propagation intensity index and the structural semantic region diagram. The comprehensive and accurate identification of the wall damage in the historical building is realized through image visual analysis, and the evolution trend of the damage is effectively reflected.

Inventors

LIN XIAHUA
Lei Bangbo
MA YUANFANG
DONG KESONG
LUO YANAN
ZHANG KUN
GAO QIANG
ZHANG KEHUI
ZHU HANGYI
WANG YA

Assignees

中建国际城市建设有限公司

Dates

Publication Date: 20260512
Application Date: 20260410

Claims (10)

1. The method for detecting the damage of the wall body inside the historical building based on the image vision is characterized by comprising the following steps of: After an original image of an inner wall of a historical building is obtained, performing illumination balancing processing and texture enhancement processing on the original image, and establishing an enhanced image; identifying structural boundary lines and material texture distribution characteristics of the wall body in the enhanced image, and executing structural semantic partition processing to generate a structural semantic region diagram; Performing the damaged form extraction processing of the enhanced image in each wall structure area by utilizing the structural semantic area diagram, identifying crack textures, peeling boundaries and chromaticity degradation areas on the wall surface, and constructing a wall damaged form feature set according to the trend continuity of the crack textures, the form contour of the peeling boundaries and the gradient change of the chromaticity degradation areas; Generating a damaged form response field according to the structural semantic region graph and the wall damaged form feature set, and calculating damaged propagation potential values by combining the adjacency relation between the structural semantic regions, the directional consistency of crack textures, the neighborhood expansion trend of the peeling boundary and the gradient diffusion direction of the chromaticity degradation region to form a wall damaged propagation feature graph; and identifying a propagation path by using the wall damage propagation characteristic map, calculating damage propagation intensity indexes, and generating a wall damage detection result by using the damage propagation intensity indexes and the structural semantic region map.
2. The method for detecting wall damage in a historic building based on image vision according to claim 1, wherein the processing for extracting the damage form of the enhanced image is performed in each wall structure region by using the structural semantic region map, comprising: determining a pixel set of each wall structure area in the enhanced image according to the structure semantic area diagram, and taking the pixel set as a damage detection area; Performing direction gradient calculation on the enhanced image in the damaged detection area, identifying a pixel sequence with linear continuity characteristics, and extracting crack texture characteristics according to the direction consistency and space connectivity of the pixel sequence; performing edge detection and region segmentation processing on the enhanced image in the damaged detection region, identifying an irregular defect region on the surface of the wall body, and extracting peeling boundary features according to the contour curvature of the boundary of the irregular defect region; Performing a chrominance component gradient analysis on the enhanced image within the impairment detection region, identifying a pixel region having continuous chrominance decay features, and extracting chrominance decay features from the spatial distribution of the chrominance gradient; And constructing a wall damaged morphological feature set according to the crack texture features, the peeling boundary features and the chromaticity degradation features.
3. The method for detecting wall damage in a historic building based on image vision according to claim 2, wherein extracting crack texture features comprises: establishing a pixel direction distribution map according to a direction gradient calculation result, and generating a plurality of candidate crack pixel sets by performing direction clustering on pixels with similar direction gradients; Connectivity analysis is carried out on each candidate crack pixel set to form a crack communication segment, and effective crack segments are screened according to the extension length and the directional stability of the crack communication segment; and calculating a crack direction consistency index according to the relation of the included angle between the extension direction of the effective crack segment and the texture direction of the adjacent wall masonry, and generating crack texture features according to the crack direction consistency index.
4. The method for detecting wall damage in a historic building based on image vision according to claim 2, wherein extracting the peeling boundary feature comprises: performing region growing processing on edge pixels in the damage detection region to form an irregular defect region; performing boundary contour curve extraction of the irregular defect area, and calculating a curvature change sequence of the boundary contour curve; Identifying boundary turning points with abrupt curvature characteristics according to the curvature change sequence, and calculating boundary irregularity of the defect area according to the spatial distribution of the boundary turning points; an exfoliation boundary feature is generated based on the boundary irregularities and the irregular defect region area.
5. The method for detecting wall damage in a historic building based on image vision according to claim 2, wherein extracting the chromaticity degradation features comprises: Converting the enhanced image into a chrominance component space and extracting a chrominance channel image; performing local gradient calculation on the chrominance channel image in a damage detection area to generate a chrominance gradient distribution map; identifying a continuous chromaticity attenuation region according to the chromaticity gradient distribution map, and determining a chromaticity degradation region range through region expansion; and generating chromaticity degradation characteristics according to the chromaticity gradient change amplitude and the space diffusion range in the chromaticity degradation area.
6. The method for detecting wall damage in a historic building based on image vision according to claim 1, wherein forming a wall damage propagation feature map comprises: Forming a damaged morphology response field according to the spatial distribution of the wall damaged morphology feature set in the enhanced image, wherein the damaged morphology response field is used for representing damaged morphology response intensity of each pixel position; Determining the adjacent relation between the wall structure areas according to the structure semantic area diagram, and establishing the adjacent propagation relation of the structure areas in the damaged form response field; Calculating crack propagation direction weight according to the propagation direction of the crack texture features, calculating peeling propagation weight according to the boundary propagation direction of peeling boundary features, and calculating chromaticity diffusion weight according to the gradient diffusion direction of chromaticity degradation signs; under the constraint of adjacent propagation relationship, carrying out directional weighted propagation calculation on the damage form response field based on crack propagation direction weight, peeling propagation weight and chromaticity diffusion weight, generating damage propagation potential value distribution, and forming a wall damage propagation characteristic diagram according to the damage propagation potential value.
7. The method for detecting wall damage in a historic building based on image vision according to claim 1, wherein the step of identifying a propagation path by using a wall damage propagation feature map and calculating a damage propagation intensity index comprises: Identifying pixel areas with continuously enhanced propagation potential values in the wall damage propagation feature map, and performing connected domain analysis on the pixel areas to construct a plurality of candidate damage propagation areas; Performing path skeleton extraction processing on each candidate damage propagation region to generate propagation paths representing damage propagation trends; And calculating corresponding damage propagation intensity indexes according to the extension length, the branch number and the path direction stability of the propagation path.
8. The method for detecting wall damage in a historic building based on image vision according to claim 7, wherein generating a wall damage detection result using the damage propagation intensity index and the structural semantic region map comprises: mapping the damage propagation intensity index to a corresponding wall structure area in the structure semantic area diagram; Determining the regional damage degree of the structural region according to the mapping result; and generating a wall damage detection result by using the regional damage degree.
9. The method for detecting the wall damage in the historic building based on the image vision according to claim 1, wherein a visual damage identifier is configured according to the wall damage detection result, and early warning processing is performed by using the visual damage identifier.
10. An image vision-based historical building interior wall damage detection system, for implementing the image vision-based historical building interior wall damage detection method according to any one of claims 1 to 9, comprising: The image enhancement module is used for performing illumination equalization processing and texture enhancement processing on the original image after acquiring the original image of the wall body inside the historical building, and establishing an enhanced image; the structural semantic partitioning module is used for identifying structural boundary lines and material texture distribution characteristics of the wall body in the enhanced image, executing structural semantic partitioning processing and generating a structural semantic region diagram; The damaged form extraction module is used for executing damaged form extraction processing of the enhanced image in each wall structure area by utilizing the structural semantic area diagram, identifying crack textures, peeling boundaries and chromaticity degradation areas on the wall surface, and constructing a wall damaged form feature set according to the trend continuity of the crack textures, the form contour of the peeling boundaries and the gradient change of the chromaticity degradation areas; The propagation feature construction module is used for generating a damaged form response field according to the structural semantic region graph and the wall damaged form feature set, and calculating a damaged propagation potential value by combining the adjacency relation between the structural semantic regions, the directional consistency of crack textures, the neighborhood expansion trend of the peeling boundary and the gradient diffusion direction of the chromaticity degradation region to form a wall damaged propagation feature graph; The damage detection output module is used for identifying a propagation path by using the wall damage propagation characteristic diagram, calculating damage propagation intensity indexes and generating a wall damage detection result by using the damage propagation intensity indexes and the structural semantic region diagram.

Description

Method and system for detecting damage of wall inside historical building based on image vision Technical Field The invention relates to the field of image vision, in particular to a method and a system for detecting wall damage in a historical building based on image vision. Background The wall inside the historical building is influenced by factors such as environmental erosion, structural aging, external force action and the like in the long-term use process, various damage phenomena are easy to generate, the damage condition of the wall is timely and accurately detected, and the method has important significance for the protection and repair and the safety evaluation of the historical building. The traditional wall damage detection mainly relies on field visual observation of professional personnel and combines experience judgment, the detection efficiency is low, the accuracy and consistency of detection results depend on the experience level of the detection personnel to a great extent, and the requirement of large-scale historical building wall damage general investigation is difficult to meet. With the development of computer vision technology, a wall damage detection method based on image analysis is gradually applied. The method can realize automatic identification of damaged areas to a certain extent by collecting and processing wall images, and improves detection efficiency. However, the existing visual inspection of images is still insufficient when actually applied to the interior wall of a historic building. On one hand, the traditional method mainly focuses on the current state identification of the damage, the detection result is difficult to reflect the evolution trend of the damage, and a reference basis of the development direction of the damage cannot be provided for subsequent repair decisions. Disclosure of Invention The invention provides a method and a system for detecting the damage of the wall inside a historical building based on image vision, aiming at the technical problems that the damage identification is not comprehensive enough and the detection result is difficult to reflect the evolution trend of the damage when the damage of the wall inside the historical building is detected in the prior art. The technical scheme for solving the technical problems is as follows: The invention provides a method for detecting wall damage in a historical building based on image vision, which comprises the steps of after an original image of the wall in the historical building is obtained, carrying out illumination balance processing and texture enhancement processing on the original image to establish an enhanced image, identifying structural boundary lines and material texture distribution characteristics of the wall in the enhanced image, carrying out structural semantic partition processing to generate a structural semantic region graph, carrying out damage form extraction processing of the enhanced image in each wall structural region by utilizing the structural semantic region graph, identifying crack textures, peeling boundaries and chromaticity degradation regions on the surface of the wall, constructing a wall damage form feature set according to trend continuity of the crack textures, form contours of the peeling boundaries and gradient changes of the chromaticity degradation regions, generating a damage form response field according to the structural semantic region graph and the wall damage form feature set, and calculating damage propagation potential values by combining the adjacent relation among the structural semantic regions, the neighborhood expansion trend of the peeling boundaries and gradient diffusion directions of the chromaticity degradation regions to form a wall damage propagation feature graph, and utilizing the wall damage propagation path identification and propagation feature graph to calculate a damage propagation strength index and a damage propagation result and generate a wall damage propagation index by utilizing the structural semantic region detection result. The invention provides an image vision-based historical building interior wall damage detection system, which comprises an image enhancement module, a structural semantic partitioning module, a damage morphology extraction module and a propagation feature construction module, wherein the image enhancement module is used for carrying out illumination balance processing and texture enhancement processing on an original image of a historical building interior wall after the original image is acquired, building an enhanced image, the structural semantic partitioning module is used for identifying structural boundary lines and material texture distribution characteristics of the wall in the enhanced image and carrying out structural semantic partitioning processing to generate a structural semantic area graph, the damage morphology extraction module is used for carrying out damage morphology extraction