Bridge detection method and system based on unmanned aerial vehicle

Abstract

The invention provides a bridge detection method and system based on an unmanned aerial vehicle, and the method comprises the steps: building a three-dimensional live-action model through the analysis of data obtained after a bridge is detected through a laser radar carried by the unmanned aerial vehicle; and analyzing the three-dimensional live-action model to obtain a first detection result. According to the invention, low-cost, effective, rapid and accurate bridge detection is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of bridge detection, and in particular to a bridge detection method and system based on an unmanned aerial vehicle. Background Art

[0002] As an important part of the transportation system, the safety and reliability of bridges are directly related to the overall operation of the transportation system. In addition, during the long-term use of bridges, they will be affected by various factors such as the natural environment and vehicle loads, resulting in structural damage and performance degradation. Through bridge health monitoring, abnormal conditions and potential risks of bridges can be discovered in time, bridge accidents can be avoided, thereby ensuring traffic safety. In addition, damage and aging of bridges can be discovered in time, and corresponding maintenance measures can be taken, thereby extending the service life of the bridge and reducing maintenance costs.

[0003] Existing monitoring is performed remotely by placing sensors on bridges. Due to the cost of sensors, not every bridge can be equipped with them and they are not fully deployed. How to achieve low-cost and effective bridge detection is a technical problem that needs to be solved urgently. Summary of the invention

[0004] One of the purposes of the present invention is to provide a bridge inspection method and system based on drones to achieve low-cost, effective, fast and accurate bridge inspection.

[0005] An embodiment of the present invention provides a bridge detection method based on a drone, comprising: By analyzing the data obtained from bridge inspection using a laser radar carried by a drone, a three-dimensional real-scene model is constructed; The three-dimensional real scene model is analyzed to obtain a first detection result.

[0006] Preferably, the analysis steps of the three-dimensional real scene model are as follows: Perform point sampling on the three-dimensional real scene model to obtain multiple sampling points and sampling data of each sampling point; Determine the correlation analysis result according to the sampling data of the associated sampling point, the stored historical sampling data of the sampling point and the preset corresponding analysis library; The correlation analysis results are statistically analyzed to obtain the first detection result.

[0007] Preferably, the bridge detection method based on drone also includes: The second detection result is determined by analyzing the image of the bridge taken by the image acquisition device carried by the unmanned aerial vehicle.

[0008] Preferably, the image analysis steps are as follows: Matching the image with a standard abnormal image in a pre-configured abnormal database, and when a match exists, obtaining the annotation data associated with the standard abnormal image as a second detection result; and / or, The image is matched with a standard normal image in a pre-configured normal database, and when a match exists, the annotation data associated with the standard normal image is obtained as the second detection result.

[0009] Preferably, before the drone photographs the bridge, it is determined that the image to be photographed corresponds to the position of the bridge; The steps for determining the location are as follows: Obtain structural parameter data and usage data of the bridge; The structural parameter data and the usage data are analyzed to determine the center position points of the multiple images.

[0010] The present invention also provides a bridge detection system based on an unmanned aerial vehicle, comprising: a construction module and a first analysis module; wherein the construction module constructs a three-dimensional real-scene model by analyzing data obtained after detecting the bridge using a laser radar carried by the unmanned aerial vehicle; the first analysis module analyzes the three-dimensional real-scene model to obtain a first detection result.

[0011] Preferably, the first analysis module analyzes the three-dimensional real scene model in the following steps: Perform point sampling on the three-dimensional real scene model to obtain multiple sampling points and sampling data of each sampling point; Determine the correlation analysis result according to the sampling data of the associated sampling point, the stored historical sampling data of the sampling point and the preset corresponding analysis library; The correlation analysis results are statistically analyzed to obtain the first detection result.

[0012] Preferably, the UAV-based bridge detection system further includes: a second analysis module; wherein the second analysis module determines the second detection result by analyzing the image of the bridge taken by the image acquisition device carried by the UAV.

[0013] Preferably, the second analysis module analyzes the image in the following steps: Matching the image with a standard abnormal image in a pre-configured abnormal database, and when a match exists, obtaining the annotation data associated with the standard abnormal image as a second detection result; and / or, The image is matched with a standard normal image in a pre-configured normal database, and when a match exists, the annotation data associated with the standard normal image is obtained as the second detection result.

[0014] Preferably, the drone-based bridge detection system further includes: a position determination module, which is used to determine the position of the bridge corresponding to the image to be captured before the drone captures the bridge; The location determination module performs the following operations: Obtain structural parameter data and usage data of the bridge; The structural parameter data and the usage data are analyzed to determine the center position points of the multiple images.

[0015] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0016] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 Schematic diagram of a bridge detection method based on drone in an embodiment of the present invention; Figure 2 Schematic diagram of a bridge detection system based on a drone in an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0019] The embodiment of the present invention provides a bridge detection method based on a drone, such as Figure 1 As shown, including: Step 1: Build a 3D real scene model by analyzing the data obtained from bridge inspection using a laser radar carried by a drone; When a bridge needs to be inspected, the laser radar carried by the drone is used to survey the bridge, which can quickly obtain the required data, especially for bridges without fixed monitoring sensors installed. The three-dimensional point cloud data obtained during surveying is used to build a three-dimensional real-scene model. The main steps of building a three-dimensional real-scene model include: data preprocessing, modeling, and post-processing; data preprocessing includes: denoising, simplification, registration, etc.; modeling can rely on Rhino, BIM, 3Dmax, ContextCapture, and post-processing includes adjustment and optimization. Building a three-dimensional real-scene model through three-dimensional point cloud data is an existing mature technology, which will not be explained here. Step 2: Analyze the three-dimensional real scene model to obtain a first detection result; For the detection of the bridge, the analysis steps of the three-dimensional real scene model are as follows: sampling points on the three-dimensional real scene model to obtain multiple sampling points and sampling data of each sampling point (the sampling data may be the coordinate data of the sampling point in the three-dimensional real scene model); determining the correlation analysis result according to the sampling data of the associated sampling point, the stored historical sampling data of the sampling point and the preset corresponding analysis library; and statistically analyzing the correlation analysis result to obtain the first detection result; Among them, the analysis library corresponds to the associated association rules, that is, one association rule corresponds to one analysis library; that is, the sampling data and historical sampling data of the sampling points involved in the association rules are analyzed by the analysis library to obtain an association analysis result; when sampling the points of the three-dimensional real scene model, all the points involved in the association rules are sampled, and then the association groups are performed according to the association rules, and each group is analyzed separately, and then the first detection result is obtained after statistics of the association analysis results; the point sampling rule can be specifically to construct a sampling library in advance, and the sampling method of the sampling points corresponding to each component of the bridge is stored in the sampling library. For example, the sampling rule of the bridge body can be to set at least one sampling point at each end, and the sampling points at both ends correspond to each other to form at least one group, and at least one sampling point is set in the middle of the bridge body. The analysis of the bridge body is to associate the sampling points of the bridge body; the specific analysis includes: analyzing the deformation through the position change between the sampling points; The specific sampling process can be broken down into the following steps: decomposing the 3D real scene model to obtain the component models of the bridge, matching each component model with the standard model in the sampling library, determining the sampling points on each component model based on the sampling points on the standard model, and then extracting the sampling data of the sampling points from the 3D real scene model; In order to achieve accurate analysis, the specific operation of sampling has high requirements. Taking the scenario of sampling one point at each end of the bridge body and at least one point in the middle as an example, segmentation is first performed. The central axis of the bridge body is segmented in both sides in turn by equidistant segmentation. A sampling point is determined on the central axis, and a line segment parallel to the central axis is constructed at the points at both ends of the bridge body. A sampling point is determined on the parallel line segment, and a sampling point is determined at the center of each segment; wherein, the sampling point on the central axis is at the same distance from the two boundaries of the bridge body; the sampling point on the parallel line segment is the midpoint of the line segment; the distance of the segment can be determined by querying the pre-configured and stored segment distance table through the total length of the bridge body; In order to improve the accuracy of the analysis, when performing the association analysis, the historical sampling data of each associated sampling point is also extracted. As the name implies, the historical sampling data is obtained and stored after sampling the three-dimensional real-life model constructed after the last or several times of detection by the laser radar carried by the drone; the specific analysis is to arrange the sampling data of the sampling points in order to form an analysis data set; each row of the analysis data set represents the sampling data of the same batch of sampling points; the analysis data set is matched with the standard data set corresponding to each association analysis result in the analysis library, and the association analysis results associated with the standard data set matching the analysis data set are extracted; the association analysis results may specifically include: offset direction, offset, etc.; when the historical sampling data is not stored, the data obtained by sampling the three-dimensional real-life model constructed by sampling the bridge design data can be used as the historical sampling data; Furthermore, the components of the bridge are numbered, and the associated analysis results of the components corresponding to the numbers are arranged in the order of the numbers to form a comprehensive analysis data set; the comprehensive analysis results are retrieved from a pre-configured comprehensive analysis library through the comprehensive analysis data set; wherein the comprehensive analysis library is pre-configured, and the comprehensive analysis data set and the comprehensive analysis results are associated one by one in the library; the comprehensive analysis results include: whether the bearing position is incorrect or damaged, causing tilt, misalignment and displacement, whether the component offset exceeds the preset offset, etc.; Of course, the detection method of this embodiment can also be used to detect bridges installed with fixed monitoring sensors. On the one hand, the bridge can be further comprehensively monitored, that is, the positions on the bridge where no monitoring sensors are installed can be monitored; on the other hand, the monitoring of the fixed monitoring sensors can be verified.

[0020] When a bridge is in use, the number and type of vehicles passing through it will affect the detection results, so it is important to choose the detection time. However, since the number and type of vehicles passing through are uncertain, it is crucial to grasp the detection time in the ever-changing situation. In one embodiment, the bridge detection method based on a drone includes: Based on the historical traffic data of the bridge, the future traffic trend data of the bridge and the selection conditions of the detection time are predicted and analyzed; Comprehensively analyze future traffic data, future weather forecast data and selection conditions to determine the timing of testing; This embodiment analyzes historical driving data, makes predictions, and determines the detection timing based on the predictions, so as to perform detection at a better timing, thereby ensuring the effectiveness of bridge detection and being more conducive to the specific control of the bridge status, so as to avoid the occurrence of low detection efficiency due to artificially waiting for the timing.

[0021] Among them, predicting future traffic trend data through historical driving data can be specifically carried out by using a pre-trained neural network model, that is, screening the historical driving data to obtain the data to be analyzed, extracting features from the data to be analyzed, and inputting the extracted features into the neural network model to obtain future traffic trend data; when screening the historical driving data, the data to be analyzed is mainly extracted from the historical driving data, and the extraction rules refer to the pre-configured requirements of the neural network model. For example, the neural network model requires that the data of each time period of each day is taken as a separate sampling data, and a preset number (24 or an integer multiple of 24) of continuous sampling data is the data to be analyzed; feature extraction of the data to be analyzed is feature extraction of each sampling data in the data to be analyzed, and the extracted features include: the number of vehicles of each type, etc. In addition, it is also necessary to determine the selection conditions for the detection time. Specifically, the selection can be the sampling data with the highest traffic condition score in the data to be analyzed (representing the busiest situation), the most common traffic condition sample data, and the sampling data with the minimum score (the best is unloaded); the score of each sampling data is calculated by statistically analyzing the scores of the passing vehicles in the sampling data based on a pre-configured score table; the score table is configured with scores for each type of vehicle (for example: 1 point for a car, 3 points for a bus, 4 points for a truck, 5 points for an engineering vehicle, etc.); the sampling data for common traffic conditions is the one where the sum of the differences between the scores corresponding to the sampling data and the scores of other sampling data is the smallest; when the selection conditions for the detection time are determined, Select from the future traffic data. The future traffic data is actually composed of the forecast data for each time period in the next N days. First, compare the future weather forecast data (weather type, visibility, etc.), and eliminate some forecast data according to the pre-configured unsuitable situation table. Then calculate the scores of other forecast data, and select the data whose scores meet the selection conditions of the detection time (calculate the difference between the scores, and then sort the differences from small to large, and extract the first M as the matching data), and get three groups of candidate forecast data. Extract one from each of the three groups of candidate forecast data, and calculate the time span (the maximum value of the time difference between the two forecast data). Take the time corresponding to the forecast data extracted from the three groups of candidate forecast data with the largest time span as the detection time.

[0022] In one embodiment, the drone-based bridge inspection method further includes: The second detection result is determined by analyzing the image of the bridge taken by the image acquisition device carried by the unmanned aerial vehicle.

[0023] The image analysis steps are as follows: Matching the image with a standard abnormal image in a pre-configured abnormal database, and when a match exists, obtaining the annotation data associated with the standard abnormal image as a second detection result; and / or, The image is matched with a standard normal image in a pre-configured normal database, and when a match exists, the annotation data associated with the standard normal image is obtained as the second detection result.

[0024] Before the drone takes photos of the bridge, it is determined that the image to be taken corresponds to the location of the bridge; The steps for determining the location are as follows: Obtaining the structural parameter data and usage data of the bridge; the structural parameter data are the size parameters of each component of the bridge; the usage data include the usage time and the type and quantity of vehicles passing through per hour; Analyze the structural parameter data and usage data to determine the center position of multiple images. The specific analysis steps are as follows: extract features from the structural parameter data and usage data, and arrange the extracted feature parameters in order to form a situation analysis set, determine the position retrieval result from the pre-configured position determination library based on the situation analysis set, retrieve the corresponding position from the pre-configured position retrieval model based on the position retrieval result, and use the position retrieved from the position retrieval model as the center of the captured image; the position retrieval result in the position determination library is associated with the situation analysis set in a one-to-one correspondence; The position retrieval model is essentially a three-dimensional model corresponding to the bridge, and a plurality of points are marked on the three-dimensional model. Each point is marked as a corresponding group according to the probability of an abnormality. When the position retrieval result is to retrieve the group, each point of the group is retrieved as the center of the captured image; for example, according to the probability of abnormality occurrence from low to high, the configuration groups are 1, 2, 3, and 4 respectively; when the position retrieval result is 1, the points of group 1 are retrieved; when the position retrieval result is 2, the points of groups 1 and 2 are retrieved; when the position retrieval result is 3, the points of groups 1, 2, and 3 are retrieved; when the position retrieval result is 4, the points of groups 1, 2, 3, and 4 are retrieved; The steps for determining the marking of each point on the position retrieval model are as follows: extract historical inspection data of the same or similar bridges, perform statistical analysis on the abnormal points corresponding to the historical inspection data, determine the probability of abnormal occurrence of each point on the three-dimensional model, and group and mark the points according to the pre-configured grouping rules; the probability of abnormal occurrence of a point is the ratio of the total number of abnormal data at the point divided by the total number of abnormal data; the grouping rule is: determine the difference between the minimum probability and the maximum probability, determine multiple equal intervals based on the difference, treat each equal interval as a group and number them in sequence, form groups for marking; In addition, in order to make the grouping and labeling of points more reasonable, the location of the point itself can also be considered, and the importance of the location and the probability of abnormal points can be comprehensively analyzed to perform grouping and labeling. At this time, the grouping steps are as follows: the corresponding abnormal points in the historical detection data are used as the points to be grouped and labeled; the influence area of ​​each point to be grouped and labeled is determined from the three-dimensional model (it can be a circular area with a preset radius around the point to be grouped and labeled), and an importance analysis data set is constructed based on the pre-configured important attributes of the components in the influence area; based on the importance analysis data set, the importance value of the point to be grouped and labeled is determined from the pre-configured important value analysis library; each point to be grouped and labeled is sorted for the first time according to the important value, and a first score value is assigned according to the sorting result; a second sorting is performed according to the probability of abnormal occurrence of each point to be grouped and labeled, and a second score value is assigned according to the sorting result; then, the points to be grouped and labeled are sorted for the third time according to the sum of the first score value and the second score value, and grouped according to the sorting result, and the grouping is now divided equally in order; the assignment of the first score value and the second score value is assigned through the pre-configured corresponding assignment table.

[0025] Among them, the second detection result includes normal, cracks, the length and width of cracks when cracks exist, cracks in the bridge deck pavement, subsidence in the bridge deck pavement, cracks in the bridge deck pavement, concrete peeling, etc.

[0026] The present invention also provides a bridge detection system based on a drone, such as Figure 2 As shown, it includes: a construction module 1 and a first analysis module 2; wherein the construction module 1 constructs a three-dimensional real scene model by analyzing the data obtained after the bridge is detected by using a laser radar carried by an unmanned aerial vehicle; the first analysis module 2 analyzes the three-dimensional real scene model to obtain a first detection result.

[0027] The first analysis module 2 analyzes the three-dimensional real scene model in the following steps: Perform point sampling on the three-dimensional real scene model to obtain multiple sampling points and sampling data of each sampling point; Determine the correlation analysis result according to the sampling data of the associated sampling point, the stored historical sampling data of the sampling point and the preset corresponding analysis library; The correlation analysis results are statistically analyzed to obtain the first detection result.

[0028] In one embodiment, the drone-based bridge detection system further includes: a second analysis module; wherein the second analysis module determines a second detection result by analyzing an image of the bridge taken by an image acquisition device carried by the drone.

[0029] The steps of analyzing the image by the second analysis module are as follows: Matching the image with a standard abnormal image in a pre-configured abnormal database, and when a match exists, obtaining the annotation data associated with the standard abnormal image as a second detection result; and / or, The image is matched with a standard normal image in a pre-configured normal database, and when a match exists, the annotation data associated with the standard normal image is obtained as the second detection result.

[0030] In one embodiment, the drone-based bridge detection system further includes: a position determination module, which is used to determine the position of the bridge corresponding to the image to be captured before the drone captures the bridge; The location determination module performs the following operations: Obtain structural parameter data and usage data of the bridge; The structural parameter data and the usage data are analyzed to determine the center position points of the multiple images.

[0031] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A bridge inspection method based on drone, characterized in that: include: By analyzing the data obtained from bridge inspection using a laser radar carried by a drone, a three-dimensional real-scene model is constructed; The three-dimensional real scene model is analyzed to obtain a first detection result.

2. The bridge detection method based on drone according to claim 1, characterized in that: The analysis steps of the 3D reality model are as follows: Perform point sampling on the three-dimensional real scene model to obtain multiple sampling points and sampling data of each sampling point; Determine the correlation analysis result according to the sampling data of the associated sampling points, the historical sampling data of the sampling points and the preset corresponding analysis library; The correlation analysis results are statistically analyzed to obtain the first detection result.

3. The bridge detection method based on drone according to claim 1, characterized in that: Also includes: The second detection result is determined by analyzing the image obtained after photographing the bridge using the image acquisition equipment carried by the unmanned aerial vehicle.

4. The method for detecting bridges based on drones as claimed in claim 3, characterized in that By analyzing the images obtained by the image acquisition device carried by the UAV after taking the bridge, the second detection result is obtained, including: Matching the image with a standard abnormal image in a pre-configured abnormal database, and when a match exists, obtaining the annotation data associated with the standard abnormal image as a second detection result; and / or, The image is matched with a standard normal image in a pre-configured normal database, and when a match exists, the annotation data associated with the standard normal image is obtained as the second detection result.

5. The bridge detection method based on drone as claimed in claim 3, characterized in that: Also includes: Before the drone takes photos of the bridge, it is determined that the image to be taken corresponds to the location of the bridge; The steps for determining the location are as follows: Obtain structural parameter data and usage data of the bridge; The structural parameter data and the usage data are analyzed to determine the center position points of the multiple images.

6. A bridge inspection system based on drones, characterized in that: include: A construction module and a first analysis module; wherein the construction module constructs a three-dimensional real-scene model by analyzing data obtained after detecting a bridge using a laser radar carried by an unmanned aerial vehicle; the first analysis module analyzes the three-dimensional real-scene model to obtain a first detection result.

7. The UAV-based bridge inspection system according to claim 6, characterized in that: The analysis steps of the first analysis module for analyzing the three-dimensional real scene model are as follows: Perform point sampling on the three-dimensional real scene model to obtain multiple sampling points and sampling data of each sampling point; Determine the correlation analysis result according to the sampling data of the associated sampling points, the historical sampling data of the sampling points and the preset corresponding analysis library; The correlation analysis results are statistically analyzed to obtain the first detection result.

8. The UAV-based bridge inspection system according to claim 6, characterized in that: Also includes: The second analysis module determines the second detection result by analyzing the image obtained after the bridge is photographed by the image acquisition device carried by the unmanned aerial vehicle.

9. The UAV-based bridge inspection system according to claim 8, characterized in that: The analysis steps of the second analysis module for analyzing the image are as follows: Matching the image with a standard abnormal image in a pre-configured abnormal database, and when a match exists, obtaining the annotation data associated with the standard abnormal image as a second detection result; and / or, The image is matched with a standard normal image in a pre-configured normal database, and when a match exists, the annotation data associated with the standard normal image is obtained as the second detection result.

10. The UAV-based bridge inspection system according to claim 8, characterized in that: Also includes: A position determination module is used to determine the position of the bridge corresponding to the image to be captured before the drone captures the bridge; The location determination module performs the following operations: Obtain structural parameter data and usage data of the bridge; The structural parameter data and the usage data are analyzed to determine the center position points of the multiple images.

Images