Viaduct disease detection method based on Beidou and laser scanning technology

Abstract

The invention discloses an elevated bridge disease detection method based on the Beidou and laser scanning technology, and the method comprises the steps: obtaining the three-dimensional point cloud data of the surface of a bridge based on the Beidou and laser scanning technology, and carrying out the data processing, and obtaining target data; constructing a three-dimensional model of the bridge surface according to the target data, and performing crack detection and deflection analysis on the three-dimensional model of the bridge surface through a deep learning algorithm to obtain an analysis result of bridge structure deformation; constructing a deep learning model, and performing deformation prediction on the bridge based on the target deep learning model and the analysis result of the bridge structure deformation to obtain a prediction result; and judging whether the prediction result exceeds a preset threshold value or not, and when the prediction result exceeds the preset threshold value, triggering an early warning mechanism to notify related personnel to check. According to the invention, high-precision positioning is provided through the Beidou system, bridge surface data is obtained by using a laser scanning technology, and rapid and accurate detection of elevated bridge diseases is realized by combining an image processing technology and a data analysis technology.

Description

Technical Field

[0001] The invention belongs to the field of bridge defect detection, and in particular relates to an elevated bridge defect detection method based on Beidou and laser scanning technology. Background Art

[0002] With the acceleration of urbanization, urban expressways, as an important part of urban transportation, bear increasing traffic pressure. In order to ensure the safe and stable operation of urban expressways, it is particularly important to monitor their deformation. Compared with the single-point three-dimensional data collection methods such as total stations and GPS, the laser scanning monitoring method in the existing technology does not require the setting of reflective prisms and non-contact measurement, and has obvious advantages in dangerous areas that are difficult for people to reach; it breaks through the single-point measurement method and obtains massive point cloud data of scanned objects with high density and high resolution, with detailed description of the target and high sampling rate, and can perform an overall analysis of various structural deformations of expressways. However, at present, there are also some problems with the laser scanning monitoring method, including the rapid decrease in accuracy caused by the accumulation of errors in multi-station splicing in large-scale scenes, low efficiency in processing massive point clouds, and deformation analysis being greatly affected by noise.

[0003] Therefore, the existing expressway deformation monitoring technologies each have their own advantages and disadvantages. If two or more technologies can be selected for fusion processing, thereby achieving the effect of suppressing weaknesses and maximizing strengths, it will be of great significance to solving the problem of urban expressway deformation monitoring. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides a method for detecting elevated bridge defects based on Beidou and laser scanning technology, comprising:

[0005] Acquire three-dimensional point cloud data of the bridge surface based on Beidou and laser scanning technology, and process the three-dimensional point cloud data to obtain target data;

[0006] Constructing a three-dimensional model of the bridge surface according to the target data, performing crack detection and deflection analysis on the three-dimensional model of the bridge surface through a deep learning algorithm, and obtaining an analysis result of the deformation of the bridge structure;

[0007] Constructing a deep learning model, and predicting the deformation of the bridge based on the target deep learning model and the analysis results of the deformation of the bridge structure to obtain a prediction result;

[0008] Determine whether the prediction result exceeds the preset threshold. When the prediction result exceeds the preset threshold, trigger the early warning mechanism and notify relevant personnel to conduct inspection.

[0009] Preferably, the process of acquiring three-dimensional point cloud data of the bridge surface based on Beidou and laser scanning technology includes:

[0010] Select appropriate points as monitoring points, set up Beidou signal receivers and laser scanning equipment, scan the bridge surface with the laser scanning equipment, and obtain laser point cloud data of the bridge surface;

[0011] Collecting corresponding position information of the laser point cloud image through the Beidou signal receiver;

[0012] The laser point cloud data and the corresponding position information of the laser point cloud image are fused to obtain three-dimensional point cloud data.

[0013] Preferably, the process of collecting the corresponding position information of the laser point cloud image by the Beidou signal receiver includes:

[0014] Beidou signal receivers are placed at appropriate locations on both sides of the open-air track to continuously observe the locations of monitoring points and collect coordinate data;

[0015] The accuracy of the coordinate data is evaluated, and the coordinate change values ​​at different times are used as the point deformation values. The coordinate difference at different positions at the same time is used as the distance value of the two monitoring points, and the change value of the distance at different times is calculated as the deformation value of the distance. Deformation analysis and deformation prediction are performed based on the results.

[0016] Preferably, the process of fusing the laser point cloud image and the corresponding position information of the laser point cloud image to obtain three-dimensional point cloud data includes:

[0017] Before scanning each station, the Beidou signal receiver is used to obtain the initial position and attitude information of the point to be monitored;

[0018] During the scanning process at each station, the laser scanner collects laser point cloud data on the bridge surface, records Beidou positioning data before scanning at each station, performs multi-station scanning at different positions of the bridge, and uses the Beidou signal receiver to align and constrain the point cloud scanned at each station. Through data fusion, seamless stitching of multi-station scanning data is achieved to obtain three-dimensional point cloud data.

[0019] Preferably, the process of processing the three-dimensional point cloud data includes:

[0020] A bilateral filter with feature preservation characteristics in image processing is introduced, and the normal vector angle is used as a control operator for feature preservation to filter and smooth the random noise of the three-dimensional point cloud data;

[0021] Construct a curvature-based point cloud compression algorithm, simplify the core content of compressed point cloud data based on curvature, establish a strategy for searching data point areas in scattered point cloud data without topological relationships, calculate the curvature of data points in the search data point area, and simplify the compressed point cloud data according to the curvature threshold;

[0022] When calculating the point cloud deviation using the neighborhood weighted centroid deviation calculation method, the three-dimensional point cloud description operator is defined, the point cloud matching relationship is established, the neighborhood KNN of the point is obtained, and the points in the neighborhood are weighted based on the distance to obtain the weighted centroid. The deviation between the weighted centroid and the model is then used as the final deviation between the measured point and the model to match point cloud models of different periods.

[0023] Preferably, the process of filtering and smoothing the random noise of the three-dimensional point cloud data includes:

[0024] The bilateral filtering technology in image processing is introduced, combined with the characteristics of point cloud data, the filtering algorithm is improved, the key features in the point cloud data are maintained, over-smoothing is avoided, and the spatial relationship and feature information of the point cloud are considered at the same time, and denoising is performed through kernel function weighting;

[0025] Taking advantage of the characteristics of bilateral filtering, in the filtering process, the spatial relationship between a point and its neighbors is used for filtering, while considering the feature information of each point; for areas with drastic feature changes, the effect of the filtering algorithm is weakened to maintain the geometric features of the point cloud data;

[0026] Fitting a trend surface using the neighboring points of the measurement point, and then adjusting the trend surface along the normal vector direction of the measurement point to make the trend of the measurement point consistent with that of the neighborhood;

[0027] The spatial relationship between the point and its nearest neighbor KNN is used for filtering, so that the point is adjusted from the spatial distribution to the center of KNN.

[0028] Preferably, the process of constructing a three-dimensional model of the bridge surface according to the target data includes:

[0029] The point cloud obtained from a single scan is defined in the current coordinate system of the laser scanner, and the point clouds from different perspectives are merged into a unified coordinate system through coordinate transformation to obtain a three-dimensional point cloud model of the bridge surface.

[0030] Preferably, the process of performing crack detection on the three-dimensional model of the bridge surface by a deep learning algorithm includes: analyzing and processing point cloud data to detect whether there are cracks in the structure, and determining the location, shape and size of the cracks.

[0031] Preferably, the process of detecting whether there are cracks in the structure and determining the location, shape and size of the cracks by analyzing and processing the point cloud data includes:

[0032] Use a 3D laser scanner to obtain point cloud data of the structure surface, then process and analyze the point cloud data to identify areas where cracks may exist;

[0033] Through point cloud filtering, point cloud segmentation, and feature extraction technology, noise and redundant data are removed, and the characteristic information of cracks is extracted. At the same time, a camera is used to acquire images, and cracks are detected and quantified based on a two- and three-dimensional combined crack detection method.

[0034] Preferably, the process of performing deflection analysis on the three-dimensional model of the bridge surface by a deep learning algorithm comprises:

[0035] Annotate the point cloud data and mark the characteristic points of deflection changes;

[0036] A convolutional neural network model is constructed, and the convolutional neural network model is used to automatically identify and analyze the deflection changes on the bridge surface, and the characteristic points and change values ​​of the deflection changes are extracted from the model output to obtain the deformation of the bridge structure.

[0037] Compared with the prior art, the present invention has the following advantages and technical effects:

[0038] The method of the present invention can realize rapid and accurate detection of elevated bridge defects. The high-precision positioning of the Beidou system ensures the accuracy of data collection, the laser scanning technology provides comprehensive surface data, and the image processing and data analysis technology improve the accuracy and efficiency of defect identification. This method not only improves the detection efficiency, but also significantly improves the reliability of the detection results, providing strong support for bridge maintenance and safety management.

[0039] The method of the present invention can significantly improve the overall stitching accuracy of multi-station laser scanning point clouds and reduce error accumulation. The high-precision positioning data of the Beidou positioning system provides a strong constraint for point cloud alignment, and advanced data fusion technology and intelligent error correction algorithms further improve the stitching effect.

[0040] The method of the present invention can realize rapid and accurate detection of elevated bridge defects. The high-precision positioning data of the Beidou system provides a strong constraint for point cloud alignment, and the advanced data fusion technology and intelligent error correction algorithm further improve the splicing effect. This method not only improves the detection efficiency, but also significantly improves the reliability of the detection results, providing strong support for bridge maintenance and safety management. The multi-station splicing method based on deep learning and Beidou positioning constraints effectively solves the error accumulation problem in the multi-station scanning splicing process, improves the splicing accuracy, and ensures the accuracy and reliability of the detection data.

[0041] The method of the present invention can effectively reduce the noise in the point cloud data, while maintaining the key features of the point cloud data and improving the accuracy of disease detection. The improved image bilateral filtering algorithm significantly improves the smoothness and analyzability of the point cloud data while maintaining the features, providing a high-quality data foundation for subsequent disease identification and analysis.

[0042] The method of the present invention can realize the automatic and high-precision analysis of the deflection change of the elevated bridge. The application of deep learning algorithm significantly improves the efficiency and accuracy of deflection analysis, reduces manual intervention, and provides strong technical support for bridge maintenance and safety management. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0044] Figure 1 The figure is a schematic diagram of a method flow of an embodiment of the present invention. DETAILED DESCRIPTION

[0045] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0046] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0047] like Figure 1 As shown, this embodiment provides a method for detecting elevated bridge defects based on Beidou and laser scanning technology, including:

[0048] Based on Beidou and laser scanning technology, the three-dimensional point cloud data of the bridge surface is obtained, and the three-dimensional point cloud data is processed to obtain the target data;

[0049] A three-dimensional model of the bridge surface is constructed based on the target data. The three-dimensional model of the bridge surface is used for crack detection and deflection analysis through a deep learning algorithm to obtain the analysis results of the bridge structure deformation.

[0050] Construct a deep learning model, predict the deformation of the bridge based on the target deep learning model and the analysis results of the bridge structure deformation, and obtain the prediction results;

[0051] Determine whether the prediction result exceeds the preset threshold. When the prediction result exceeds the preset threshold, the early warning mechanism is triggered and relevant personnel are notified to conduct inspections.

[0052] Furthermore, the process of obtaining three-dimensional point cloud data of the bridge surface based on Beidou and laser scanning technology includes:

[0053] Select appropriate points as monitoring points, set up Beidou signal receivers and laser scanning equipment, scan the bridge surface with the laser scanning equipment, and obtain laser point cloud data of the bridge surface;

[0054] Collect the corresponding position information of the laser point cloud image through the Beidou signal receiver;

[0055] The three-dimensional point cloud data is obtained by fusing the corresponding position information of the laser point cloud data and the laser point cloud image.

[0056] Furthermore, the process of collecting corresponding position information of the laser point cloud image by the Beidou signal receiver includes:

[0057] Beidou signal receivers are placed at appropriate locations on both sides of the open-air track to continuously observe the locations of monitoring points and collect coordinate data;

[0058] The accuracy of coordinate data is evaluated, and the coordinate change values ​​at different times are used as the point deformation values. The coordinate difference at different positions at the same time is used as the distance value of the two monitoring points. The change value of the distance at different times is calculated as the deformation value of the distance, and deformation analysis and deformation prediction are performed based on the results.

[0059] Furthermore, the process of fusing the laser point cloud image and the corresponding position information of the laser point cloud image to obtain the three-dimensional point cloud data includes:

[0060] Before scanning each station, the Beidou signal receiver is used to obtain the initial position and attitude information of the point to be monitored;

[0061] During the scanning process at each station, the laser scanner collects laser point cloud data on the bridge surface, records Beidou positioning data before scanning at each station, performs multi-station scanning at different positions of the bridge, and uses the Beidou signal receiver to align and constrain the point cloud scanned at each station. Through data fusion, seamless stitching of multi-station scanning data is achieved to obtain three-dimensional point cloud data.

[0062] Here, "each station" refers to the various points where laser scanning or Beidou positioning equipment collects data at different positions or areas of the bridge when conducting elevated bridge disease detection or deformation monitoring.

[0063] Furthermore, the process of processing the three-dimensional point cloud data includes:

[0064] The bilateral filter with feature preservation characteristics in image processing is introduced, and the normal vector angle is used as the control operator of feature preservation to filter and smooth the random noise of 3D point cloud data.

[0065] Construct a curvature-based point cloud compression algorithm, simplify the core content of compressed point cloud data based on curvature, establish a strategy for searching data point areas in scattered point cloud data without topological relationships, calculate the curvature of data points in the search data point area, and simplify the compressed point cloud data according to the curvature threshold;

[0066] When calculating the point cloud deviation using the neighborhood weighted centroid deviation calculation method, the three-dimensional point cloud description operator is defined, the point cloud matching relationship is established, the neighborhood KNN of the point is obtained, and the points in the neighborhood are weighted based on the distance to obtain the weighted centroid. The deviation between the weighted centroid and the model is then used as the final deviation between the measured point and the model to match point cloud models of different periods.

[0067] Furthermore, the process of filtering and smoothing the random noise of the three-dimensional point cloud data includes:

[0068] The bilateral filtering technology in image processing is introduced, combined with the characteristics of point cloud data, the filtering algorithm is improved, the key features in the point cloud data are maintained, over-smoothing is avoided, and the spatial relationship and feature information of the point cloud are considered at the same time, and denoising is performed through kernel function weighting;

[0069] Taking advantage of the characteristics of bilateral filtering, in the filtering process, the spatial relationship between a point and its neighbors is used for filtering, while considering the feature information of each point; for areas with drastic feature changes, the effect of the filtering algorithm is weakened to maintain the geometric features of the point cloud data;

[0070] Use the neighboring points of the measurement point to fit a trend surface, and then adjust the measurement point along the trend surface in the direction of the normal vector to make the trend of the measurement point consistent with that of the neighborhood;

[0071] The spatial relationship between the point and its nearest neighbor KNN is used for filtering, so that the point is adjusted from the spatial distribution to the center of KNN.

[0072] Furthermore, the process of constructing a three-dimensional model of the bridge surface according to the target data includes:

[0073] The point cloud obtained from a single scan is defined in the current coordinate system of the laser scanner, and the point clouds from different perspectives are merged into a unified coordinate system through coordinate transformation to obtain a three-dimensional point cloud model of the bridge surface.

[0074] Furthermore, the process of crack detection on the three-dimensional model of the bridge surface through the deep learning algorithm includes: detecting whether there are cracks in the structure by analyzing and processing the point cloud data, and determining the location, shape and size of the cracks.

[0075] Furthermore, the process of analyzing and processing the point cloud data to detect whether there are cracks in the structure and determine the location, shape and size of the cracks includes:

[0076] Use a 3D laser scanner to obtain point cloud data of the structure surface, then process and analyze the point cloud data to identify areas where cracks may exist;

[0077] Through point cloud filtering, point cloud segmentation, and feature extraction technology, noise and redundant data are removed, and the characteristic information of cracks is extracted. At the same time, a camera is used to acquire images, and cracks are detected and quantified based on a two- and three-dimensional combined crack detection method.

[0078] Furthermore, the process of performing deflection analysis on the three-dimensional model of the bridge surface through the deep learning algorithm includes:

[0079] Annotate the point cloud data and mark the characteristic points of deflection changes;

[0080] A convolutional neural network model was constructed and used to automatically identify and analyze the deflection changes on the bridge surface. The characteristic points and change values ​​of the deflection changes were extracted from the model output to obtain the deformation of the bridge structure.

[0081] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for detecting elevated bridge defects based on Beidou and laser scanning technology, characterized in that: include: Acquire three-dimensional point cloud data of the bridge surface based on Beidou and laser scanning technology, and process the three-dimensional point cloud data to obtain target data; Constructing a three-dimensional model of the bridge surface according to the target data, performing crack detection and deflection analysis on the three-dimensional model of the bridge surface through a deep learning algorithm, and obtaining an analysis result of the deformation of the bridge structure; Constructing a deep learning model, and predicting the deformation of the bridge based on the target deep learning model and the analysis results of the deformation of the bridge structure to obtain a prediction result; Determine whether the prediction result exceeds the preset threshold. When the prediction result exceeds the preset threshold, trigger the early warning mechanism and notify relevant personnel to conduct inspection.

2. The elevated bridge disease detection method based on Beidou and laser scanning technology according to claim 1 is characterized in that: The process of obtaining 3D point cloud data of the bridge surface based on Beidou and laser scanning technology includes: Select appropriate points as monitoring points, set up Beidou signal receivers and laser scanning equipment, scan the bridge surface with the laser scanning equipment, and obtain laser point cloud data of the bridge surface; Collecting corresponding position information of the laser point cloud image through the Beidou signal receiver; The laser point cloud data and the corresponding position information of the laser point cloud image are fused to obtain three-dimensional point cloud data.

3. The elevated bridge disease detection method based on Beidou and laser scanning technology according to claim 1 is characterized in that: The process of collecting the corresponding position information of the laser point cloud image by the Beidou signal receiver includes: Beidou signal receivers are placed at appropriate locations on both sides of the open-air track to continuously observe the locations of monitoring points and collect coordinate data; The accuracy of the coordinate data is evaluated, and the coordinate change values ​​at different times are used as the point deformation values. The coordinate difference at different positions at the same time is used as the distance value of the two monitoring points, and the change value of the distance at different times is calculated as the deformation value of the distance. Deformation analysis and deformation prediction are performed based on the results.

4. The elevated bridge disease detection method based on Beidou and laser scanning technology according to claim 1 is characterized in that: The process of fusing the laser point cloud image and the corresponding position information of the laser point cloud image to obtain three-dimensional point cloud data includes: Before scanning each station, the Beidou signal receiver is used to obtain the initial position and attitude information of the point to be monitored; During the scanning process at each station, the laser scanner collects laser point cloud data on the bridge surface, records Beidou positioning data before scanning at each station, performs multi-station scanning at different positions of the bridge, and uses the Beidou signal receiver to align and constrain the point cloud scanned at each station. Through data fusion, seamless stitching of multi-station scanning data is achieved to obtain three-dimensional point cloud data.

5. The elevated bridge disease detection method based on Beidou and laser scanning technology according to claim 1 is characterized in that: The process of processing the three-dimensional point cloud data includes: A bilateral filter with feature preservation characteristics in image processing is introduced, and the normal vector angle is used as a control operator for feature preservation to filter and smooth the random noise of the three-dimensional point cloud data; Construct a curvature-based point cloud compression algorithm, simplify the core content of compressed point cloud data based on curvature, establish a strategy for searching data point areas in scattered point cloud data without topological relationships, calculate the curvature of data points in the search data point area, and simplify the compressed point cloud data according to the curvature threshold; When calculating the point cloud deviation using the neighborhood weighted centroid deviation calculation method, the three-dimensional point cloud description operator is defined, the point cloud matching relationship is established, the neighborhood KNN of the point is obtained, and the points in the neighborhood are weighted based on the distance to obtain the weighted centroid. The deviation between the weighted centroid and the model is then used as the final deviation between the measured point and the model to match point cloud models of different periods.

6. The elevated bridge disease detection method based on Beidou and laser scanning technology according to claim 5 is characterized in that: The process of filtering and smoothing the random noise of the three-dimensional point cloud data includes: The bilateral filtering technology in image processing is introduced, combined with the characteristics of point cloud data, the filtering algorithm is improved, the key features in the point cloud data are maintained, over-smoothing is avoided, and the spatial relationship and feature information of the point cloud are considered at the same time, and denoising is performed through kernel function weighting; Taking advantage of the characteristics of bilateral filtering, in the filtering process, the spatial relationship between a point and its neighbors is used for filtering, while considering the feature information of each point; for areas with drastic feature changes, the effect of the filtering algorithm is weakened to maintain the geometric features of the point cloud data; Fitting a trend surface using the neighboring points of the measurement point, and then adjusting the trend surface along the normal vector direction of the measurement point to make the trend of the measurement point consistent with that of the neighborhood; The spatial relationship between the point and its nearest neighbor KNN is used for filtering, so that the point is adjusted from the spatial distribution to the center of KNN.

7. The elevated bridge disease detection method based on Beidou and laser scanning technology according to claim 1 is characterized in that: The process of constructing a three-dimensional model of the bridge surface according to the target data includes: The point cloud obtained from a single scan is defined in the current coordinate system of the laser scanner, and the point clouds from different perspectives are merged into a unified coordinate system through coordinate transformation to obtain a three-dimensional point cloud model of the bridge surface.

8. The elevated bridge disease detection method based on Beidou and laser scanning technology according to claim 1 is characterized in that: The process of crack detection on the three-dimensional model of the bridge surface through a deep learning algorithm includes: detecting whether there are cracks in the structure by analyzing and processing point cloud data, and determining the location, shape and size of the cracks.

9. The elevated bridge disease detection method based on Beidou and laser scanning technology according to claim 8 is characterized in that: The process of analyzing and processing point cloud data to detect the presence of cracks in a structure and determine the location, shape, and size of the cracks includes: Use a 3D laser scanner to obtain point cloud data of the structure surface, then process and analyze the point cloud data to identify areas where cracks may exist; Through point cloud filtering, point cloud segmentation, and feature extraction technology, noise and redundant data are removed, and the characteristic information of cracks is extracted. At the same time, a camera is used to acquire images, and cracks are detected and quantified based on a two- and three-dimensional combined crack detection method.

10. The elevated bridge disease detection method based on Beidou and laser scanning technology according to claim 1 is characterized in that: The process of performing deflection analysis on the three-dimensional model of the bridge surface by a deep learning algorithm includes: Annotate the point cloud data and mark the characteristic points of deflection changes; A convolutional neural network model is constructed, and the convolutional neural network model is used to automatically identify and analyze the deflection changes on the bridge surface, and the characteristic points and change values ​​of the deflection changes are extracted from the model output to obtain the deformation of the bridge structure.

Images