Building settlement observation system and method based on image recognition

Abstract

The invention relates to the technical field of image recognition, and discloses a building settlement observation system and method based on image recognition. According to the system and the method, a high-definition camera is used for carrying out image acquisition and obtaining high-quality images of a building in real time, an image preprocessing module is used for eliminating image noise, enhancing image quality and carrying out necessary cutting and zooming, and in the feature extraction stage, the edge, angular point and straight line features of the building are extracted; the relative height features of the building are accurately calculated, the monitoring of the building settlement is more accurate and comprehensive due to the features, and the building settlement detection module can highly analyze the displacement of feature points by adopting an image matching algorithm, so that the settlement condition of the building is quickly judged, and the accuracy of building settlement monitoring is improved. And the data comprehensive report module comprehensively analyzes the detection result, and a generated building settlement report can be displayed on a large building settlement observation screen in real time, so that related personnel can make timely decisions and responses.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and in particular to a building settlement observation system and method based on image recognition. Background Art

[0002] Building settlement refers to the vertical displacement of a building on its foundation, usually caused by factors such as soil compression, groundwater level changes, or geological condition changes. Timely monitoring and judgment of the building settlement situation are of great significance, which can effectively prevent potential structural safety hazards caused by settlement, ensure the use safety of the building and the life and property safety of residents. In addition, through the dynamic monitoring of building settlement, relevant departments can take timely measures for repair, extend the service life of the building, reduce maintenance costs, and thus improve the overall management level of construction projects.

[0003] Traditional building settlement monitoring methods often rely on manual measurement and traditional instrument equipment, which are not only time-consuming and laborious, but also have poor accuracy and real-time performance. With the acceleration of urbanization, the density of buildings increases, and traditional monitoring methods are difficult to meet the high-frequency and large-scale settlement monitoring requirements. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides a building settlement observation system and method based on image recognition. By using a high-definition camera for image acquisition, high-quality images of the building are obtained in real time, avoiding errors that may be brought by manual measurement, and improving the accuracy and reliability of data. The image preprocessing module is used to eliminate image noise and enhance image quality, and perform necessary cropping and scaling, so as to ensure the effectiveness of subsequent feature extraction. In the feature extraction stage, not only the edge, corner, and straight line features of the building are extracted, but also the relative height feature of the building is accurately calculated. These features make the monitoring of building settlement more accurate and comprehensive. By using an image matching algorithm, the building settlement detection module can highly analyze the displacement of feature points, so as to quickly judge the building settlement situation. The data comprehensive report module comprehensively analyzes the detection results, generates a building settlement report, and can be displayed in real time on the building settlement observation large screen, facilitating timely decision-making and response by relevant personnel.

[0005] To achieve the above object, the present invention provides the following technical solution: A building settlement observation system based on image recognition, including an image acquisition module, an image preprocessing module, a feature extraction module, a building settlement detection module, and a data comprehensive report module;

[0006] The image acquisition module acquires images of the building through a high-definition camera and sends them to the image preprocessing module;

[0007] The image preprocessing module preprocesses the image of the building, including image denoising, image enhancement, image cropping, and image scaling. The preprocessed building image is sent to the feature extraction module;

[0008] The feature extraction module extracts features based on the preprocessed building image, including building edge features, building corner features, building straight line features, and the relative height feature of the building;

[0009] The building settlement detection module analyzes the displacement of feature points by using an image matching algorithm according to the building features extracted by the feature extraction module, judges the settlement situation of the building, and sends the judgment result to the data comprehensive report module;

[0010] The data comprehensive report module comprehensively analyzes the judgment result output by the building settlement detection module, generates a report on the building settlement situation, and sends it to the building settlement observation large screen.

[0011] Preferably, the formula for image denoising is as follows:

[0012]

[0013] In the formula, I filt (x, y) represents the pixel value of the denoised building image at the position (x, y), and I(x ′ , y ′ ) represents the pixel value of the original building image at the position (x, y). σ represents the standard deviation of the Gaussian kernel, which controls the degree of blurring, and (x, y) represents the position of the target pixel.

[0014] Preferably, the formula for image enhancement is as follows:

[0015] I enhanced (x, y) = CDF(I filt (x, y))

[0016] In the formula, I enhanced (x, y) represents the pixel value of the enhanced building image at the position (x, y), and CDF(I filt (x, y)) represents the cumulative distribution function of the pixel value I filt (x, y).

[0017] Preferably, the formula for image cropping is as follows:

[0018] I cropped = I(x1:x2, y1:y2)

[0019] In the formula, I croppedIt represents the cropped building image. I(x1:x2,y1:y2) represents the cropped area from the original building image, where (x1,y1) is the upper left corner coordinate and (x2,y2) is the lower right corner coordinate.

[0020] Preferably, the formula for image scaling is as follows:

[0021]

[0022] In the formula, I scal (x,y) represents the pixel value of the scaled building image at the position (x,y), s represents the scaling factor, represents rounding down, and i, j represent adjacent pixel positions.

[0023] Preferably, the formula for extracting building edge features is as follows:

[0024]

[0025] In the formula, G represents the edge intensity image, G x represents the gradient of the building image in the x direction, and G y represents the gradient of the building image in the y direction.

[0026] Preferably, the formula for extracting building corner features is as follows:

[0027] R = det(M) - k * (trace(M)) 2

[0028] In the formula, R represents the corner response value, M represents the building image feature matrix, k represents the empirical parameter with a value range from 0.04 to 0.06, det(M) represents the determinant of matrix M, and trace(M) represents the sum of the diagonal elements of matrix M.

[0029] Preferably, the formula for extracting building line features is as follows:

[0030] ρ = xcos(θ) + ysin(θ)

[0031] In the formula, ρ represents the minimum distance from the coordinate origin to the line, θ represents the angle between the line and the x-axis, and x, y represent the pixel coordinates in the image.

[0032] Preferably, the formula for extracting the relative height feature of the building is as follows:

[0033] h = h1 - h2

[0034] In the formula, h represents the relative height of the building, h1 represents the height of one part of the building obtained from the 3D reconstruction of image recognition, and h2 represents the height of the building reference point.

[0035] A building settlement observation method based on image recognition includes the following steps:

[0036] S1. Collect images of the building through a high-definition camera;

[0037] S2. Preprocess the images of the building;

[0038] S3. Extract features according to the preprocessed building images;

[0039] S4. According to the extracted building features, use an image matching algorithm to analyze the displacement of feature points and judge the settlement situation of the building;

[0040] S5. Comprehensively analyze the judgment results of the building settlement situation, generate a building settlement situation report and send it to the building settlement observation large screen.

[0041] Compared with the prior art, the present invention provides a building settlement observation system and method based on image recognition, having the following beneficial effects:

[0042] By using a high-definition camera for image acquisition, the present invention can obtain high-quality images of the building in real time, avoiding errors that may be brought by manual measurement, improving the accuracy and reliability of data. The image preprocessing module is used to eliminate image noise, enhance image quality, and perform necessary cropping and scaling, so as to ensure the effectiveness of subsequent feature extraction. In the feature extraction stage, not only the edge, corner and straight line features of the building are extracted, but also the relative height feature of the building is accurately calculated. These features make the monitoring of building settlement more accurate and comprehensive. By using an image matching algorithm, the building settlement detection module can highly analyze the displacement of feature points, thereby quickly judging the settlement situation of the building. The data comprehensive report module comprehensively analyzes the detection results, generates a building settlement report that can be displayed on the building settlement observation large screen in real time, facilitating relevant personnel to make timely decisions and responses. This system not only improves the efficiency and accuracy of monitoring, but also greatly reduces the labor cost and enhances the scientific and intelligent level of building safety management. Description of the Drawings

[0043] Figure 1 It is a schematic diagram of the system flow of the present invention;

[0044] Figure 2 It is a schematic diagram of the method steps of the present invention. Specific Embodiments

[0045] Next, in combination with the accompanying drawings in the embodiments of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

[0046] Aiming at the problem that traditional building settlement monitoring methods often rely on manual measurement and traditional instrument equipment, which are not only time-consuming and laborious, but also have poor accuracy and real-time performance, a building settlement observation system based on image recognition is proposed. Please refer to Figure 1 . The system includes an image acquisition module, an image preprocessing module, a feature extraction module, a building settlement detection module, and a data comprehensive report module;

[0047] The image acquisition module uses a high-performance high-definition camera with a resolution of up to 4K and a wide-angle lens, and can work stably under different lighting conditions, such as low-light environments and strong light reflections. The module sends the captured building images to the image preprocessing module in real time through an efficient image transmission interface (such as Ethernet or USB 3.0). During the data transmission process, the module also has real-time compression and encryption functions to ensure the security and effectiveness of the image data. At the same time, the camera is equipped with intelligent autofocus and image stabilization technology to ensure that the captured images are clear and free of blur, thereby providing high-quality basic data for subsequent image analysis and processing;

[0048] The image preprocessing module performs a series of important preprocessing operations on the building images to improve the accuracy and efficiency of subsequent feature extraction. First, Gaussian filtering algorithm is used for image denoising, which is implemented through the following formula:

[0049]

[0050] This step can effectively reduce the random noise in the image, improve the clarity of the image, and create better conditions for subsequent operations. By using denoising algorithms such as Gaussian filtering, important edges and details can be retained, pixel errors can be reduced. After denoising, a clearer image can be obtained, which helps subsequent feature extraction and analysis, improves the recognition accuracy, reduces the false positive and false negative rates, and ensures that the system can detect the tiny settlement changes of the building in time. For example, for the monitoring of high-rise buildings, tiny changes may affect the safety of the overall structure. Therefore, clear images can provide reliable data display to help engineers make more accurate analysis and judgment;

[0051] Next, the image enhancement operation improves the image contrast through histogram equalization, making the details more prominent, and thus making the building edges more obvious. The formula is:

[0052] I enhanced (x,y) = CDF(I filt (x,y))

[0053] Through image enhancement, the accuracy of feature extraction algorithms (such as Hough transform or Canny edge detection) can be improved, ensuring that building edges and other important features can be accurately identified in building settlement monitoring. In practical applications, image enhancement technology can help engineers quickly discover potential structural problems, reducing the time and cost required for subsequent reviews and repairs, providing a basis for taking timely measures to ensure building safety;

[0054] Image cropping selects the key parts according to a specific area, using the formula:

[0055] I cropped = I(x1:x2,y1:y2)

[0056] Image cropping is the process of extracting the region of interest (ROI) from the complete image. Selecting an appropriate area for cropping can effectively reduce the computational complexity of subsequent processing because only the parts of interest are processed, avoiding unnecessary redundant data. In addition, cropping can focus on the key features of the building, such as the foundation part, window frames, or important nodes, without being interfered by background information. Through the cropped image, the feature extraction and analysis process has higher efficiency and accuracy, which can better help engineers evaluate the settlement status of the building, provide important data support and decision-making basis, and take necessary maintenance and reinforcement measures in a timely manner to ensure the safety and service life of the building;

[0057] Finally, image scaling is performed through bilinear interpolation technology to meet the input requirements of the feature extraction module:

[0058]

[0059] Image scaling is an important step in adjusting the image size to meet the requirements of subsequent analysis modules. Through appropriate scaling techniques, it can be ensured that the images input to the feature extraction algorithm are consistent in size, thereby improving the processing efficiency. Scaling can avoid errors caused by inconsistent image sizes. Commonly used bilinear interpolation method or bicubic interpolation method can adjust the image size while maintaining image clarity. At the same time, specific scaling ratios can help the system better identify minor settlement changes, especially when monitoring large buildings such as high-rise buildings, and can more accurately capture the subtle movements and deformations of the building;

[0060] Through these preprocessing steps, the images finally transmitted to the feature extraction module not only remove unnecessary interferences but also enhance important features, making the subsequent straight-line feature extraction and building contour analysis more accurate and effective, thus improving the overall performance and reliability of building settlement monitoring;

[0061] The feature extraction module deeply analyzes the preprocessed building images to extract key features for building settlement monitoring. First, the building edge features are extracted using the Canny edge detection algorithm, which is achieved based on the gradient intensity and direction through the following formula:

[0062]

[0063] Using this formula, the gradient intensity of each pixel can be calculated. The edge features can reflect the overall visual form of the building, including important information such as corners and wall boundaries. When the building undergoes settlement or deformation, its edges will also change slightly;

[0064] The building corner features adopt the Harris corner detection algorithm, and its core formula is:

[0065] R = det(M) - k * (trace(M)) 2

[0066] Among them, M is the image feature matrix. By identifying the corners in the image, the key positions of the building can be accurately located, which helps analyze its structural stability. During the use of the building, any displacement of the corners may imply potential structural problems. By real-time monitoring the position changes of the corners, engineers can timely discover and handle possible safety hazards;

[0067] The building straight-line features are extracted through the Hough transform, and its basic equation is:

[0068] ρ = xcos(θ) + ysin(θ)

[0069] Extracting the building straight-line features can provide a reliable basis for monitoring the straight-line deformation of the building structure. Many structural parts of the building, such as beams, columns, and walls, usually appear in the form of straight lines, and the geometric relationships of these straight lines are crucial for the overall stability. Using methods such as the Hough transform to obtain the straight-line features can clearly mark these structural lines in the image. When the building tilts or cracks, the original straight-line relationships will change;

[0070] The calculation of the relative height features of the building is achieved through the height differences of each key point, and the formula is:

[0071] h = h1 - h2

[0072] High-precision feature extraction can provide detailed data for settlement monitoring, enabling the system to identify minute structural changes, enhancing the accuracy and timeliness of building safety assessments. Through the acquisition of these precise features, engineers can quickly judge the structural health of buildings and take necessary maintenance and reinforcement measures in a timely manner, thus effectively preventing potential safety hazards;

[0073] The building settlement detection module deeply analyzes the building features extracted by the feature extraction module and uses efficient image matching algorithms (such as feature point-based matching algorithms) to evaluate the displacement of feature points. These algorithms usually adopt SIFT (Scale-Invariant Feature Transform) or ORB (Oriented FAST and Rotated BRIEF) techniques, which can accurately match key points in images under different lighting conditions and perspectives. By calculating the coordinate changes of feature points in images at different times, the settlement detection module can quantify the settlement degree of the building. If the displacement between feature points is detected to exceed the set threshold, the system will determine it as a settlement phenomenon and send the result to the data comprehensive report module through preset logical conditions;

[0074] In the data comprehensive report module, the system will conduct a comprehensive analysis based on the judgment results output by the settlement detection module. The module integrates multiple image matching results, combines historical data and statistical models, and deeply evaluates the settlement situation of the building. The comprehensive report not only includes the absolute value change of settlement but also analyzes the settlement rate, distribution, and potential influencing factors, forming a detailed report on the building settlement situation. This report is then sent to the building settlement observation large screen for relevant engineers and managers to view and monitor in real time. Through such a comprehensive feedback mechanism, the system can effectively support the long-term health management of buildings, enhance safety, and provide data basis for subsequent maintenance and decision-making. The whole process is efficient and intelligent, ensuring that building safety hazards can be quickly identified and processed, bringing significant convenience to project management.

[0075] Please refer to Figure 2 , an image recognition-based building settlement observation method, comprising the following steps:

[0076] S1. Collect images of the building through a high-definition camera;

[0077] S2. Preprocess the images of the building;

[0078] S3. Extract features based on the preprocessed images of the building;

[0079] S4. According to the extracted building features, adopt an image matching algorithm to analyze the displacement of feature points and judge the settlement situation of the building;

[0080] S5. Comprehensively analyze the judgment results of the building settlement situation, generate a building settlement situation report, and send it to the building settlement observation large screen.

[0081] Through the comprehensive application of the above system and method, not only the monitoring efficiency and accuracy are improved, but also the labor cost is greatly reduced, and the scientific and intelligent level of building safety management is enhanced.

[0082] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An image recognition-based building settlement observation system, characterized in that: It includes an image acquisition module, an image preprocessing module, a feature extraction module, a building settlement detection module, and a data comprehensive report module; The image acquisition module acquires images of the building through a high-definition camera and sends them to the image preprocessing module; The image preprocessing module preprocesses the images of the building, including image denoising, image enhancement, image cropping, and image scaling. The preprocessed building images are sent to the feature extraction module; The feature extraction module extracts features based on the preprocessed building images, including building edge features, building corner features, building straight-line features, and the relative height features of the building; The building settlement detection module analyzes the displacement of feature points using an image matching algorithm according to the building features extracted by the feature extraction module, judges the settlement situation of the building, and sends the judgment result to the data comprehensive report module; The data comprehensive report module comprehensively analyzes the judgment results output by the building settlement detection module and generates a building settlement situation report to be sent to the building settlement observation large screen.

2. The building settlement observation system based on image recognition according to claim 1, characterized in that: The formula for the image denoising is as follows: In the formula, I filt (x, y) represents the pixel value of the denoised building image at the position (x, y), and I(x ′ , y ′ ) represents the pixel value of the original building image at the position (x, y). σ represents the standard deviation of the Gaussian kernel, which controls the degree of blurring, and (x, y) represents the position of the target pixel.

3. The building settlement observation system based on image recognition according to claim 2, characterized in that: The formula for the image enhancement is as follows: I en hanced(x,y) = CDF(I filt (x,y)) In the formula, I enhanced (x, y) represents the pixel value of the enhanced building image at the position (x, y), and CDF(I filt (x, y)) represents the cumulative distribution function of the pixel value I filt (x, y).

4. The building settlement observation system based on image recognition according to claim 3, characterized in that: The formula for the image cropping is as follows: I cropped = I(x1:x2, y1:y2) In the formula, I cropped represents the cropped building image, I(x1:x2, y1:y2) represents the area cropped from the original building image, (x1, y1) is the upper left corner coordinate, and (x2, y2) is the lower right corner coordinate.

5. The building settlement observation system based on image recognition according to claim 4, characterized in that: The formula for the image scaling is as follows: In the formula, I scal (x, y) represents the pixel value of the scaled building image at the position (x, y), s represents the scaling factor, denotes rounding down, and i, j represent adjacent pixel positions.

6. The building settlement observation system based on image recognition according to claim 5, characterized in that: The formula for extracting the building edge features is as follows: In the formula, G represents the edge intensity image, and G x represents the gradient of the building image in the x direction, and G y represents the gradient of the building image in the y direction.

7. An image recognition-based building settlement observation system according to claim 6, characterized in that: The formula for extracting the building corner features is as follows: R = det(M) - k * (trace(M)) 2 In the formula, R represents the corner response value, M represents the building image feature matrix, k represents the empirical parameter, with a value range from 0.04 to 0.06, det(M) represents the determinant of matrix M, and trace(M) represents the sum of the diagonal elements of matrix M.

8. An image recognition-based building settlement observation system according to claim 7, characterized in that: The formula for extracting the building straight-line features is as follows: ρ = xcos(θ) + ysin(θ) In the formula, ρ represents the minimum distance from the coordinate origin to the straight line, θ represents the angle between the straight line and the x-axis, and x, y represent the pixel coordinates in the image.

9. The building settlement observation system based on image recognition according to claim 8, characterized in that: The formula for extracting the relative height features of the building is as follows: h = h1 - h2 In the formula, h represents the relative height of the building, h1 represents the height of one part of the building, obtained from the 3D reconstruction of image recognition, and h2 represents the height of the building reference point.

10. A building settlement observation method based on image recognition, characterized in that, It includes the following steps: S1. Acquire images of the building through a high-definition camera; S2. Preprocess the images of the building; S3. Extract features based on the preprocessed building images; S4. Analyze the displacement of feature points using an image matching algorithm according to the extracted building features to judge the settlement situation of the building; S5. Comprehensively analyze the judgment results of the building settlement situation and generate a building settlement situation report to be sent to the building settlement observation large screen.

Images