BIM Data Acquisition Method and System Applied to Highway Engineering

Abstract

The present invention relates to the technical field of electronic digital data processing, and particularly to a BIM data acquisition method and system applied to highway engineering, including acquiring acquisition data and flight data; calculating the acquisition influence degree according to the flight data; performing regional segmentation and grouping on the point cloud data to obtain the point cloud data of the region; calculating the surface morphology complexity degree of the region according to the point cloud data of the region; calculating the acquisition data interference degree according to the surface morphology complexity degree and the acquisition influence degree of the region; performing perspective projection transformation processing on the point cloud data and the image of the region to obtain a high-dimensional image; and performing weighted fusion on the high-dimensional image according to the acquisition data interference degree of the region to obtain a three-dimensional space coordinate set corresponding to the region. The present invention evaluates the influence of the acquisition time, divides the coverage area, and evaluates the surface complexity. Based on the evaluation results, the regional interference degree is calculated, and the point cloud data is fused, considering the influence of wind force, to improve the data reliability.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic digital data processing, and particularly to a BIM data acquisition method and system applied to highway engineering. Background Art

[0002] With the rapid development of infrastructure construction in China, highway engineering, as a key part of the transportation network, has seen a continuous increase in its construction scale and complexity. To improve the design accuracy, construction quality, and operation and maintenance efficiency of highway engineering, more and more projects have started to adopt Building Information Modeling (BIM) technology. The application of BIM technology can digitalize, visualize, and integrate the management of engineering information, thus providing strong support for the whole-life cycle management of highway engineering.

[0003] However, during the BIM data acquisition process, due to the limitations of complex terrain or construction sections, on-vehicle lidar cannot be used for three-dimensional scanning. Therefore, unmanned aerial vehicle (UAV) scanning operations have become an alternative. However, when collecting road surface information along a predefined flight path in the air, UAVs are vulnerable to wind influence, resulting in deviations in the collected road data, which in turn affects the accuracy of highway BIM data. Currently, there is no effective technology to repair these data deviations. Therefore, there is an urgent need for a BIM data optimized acquisition method applicable to highway engineering. Summary of the Invention

[0004] In order to solve the above technical problems, the purpose of the present invention is to provide a BIM data acquisition method and system applied to highway engineering. To achieve the above purpose, the technical solutions adopted by the present invention are as follows:

[0005] In a first aspect, the present application provides a BIM data acquisition method for highway engineering. The method includes: obtaining the acquisition data and flight data of a drone within a preset time period, where the preset time period is the time period during which the drone flies uniformly along a target highway section at a preset speed, and the acquisition data includes point cloud data and images; calculating the acquisition influence degree at each moment based on the flight data and the images within the preset time period; performing region segmentation based on the point cloud data corresponding to each moment, and performing grouping division based on the segmentation result to obtain the point cloud data when each region is acquired multiple times; calculating the surface morphology complexity degree of each region based on the point cloud data when each region is acquired multiple times, where the surface morphology complexity degree is calculated from the point cloud density distribution and the point cloud curvature difference during each acquisition; calculating the acquisition data interference degree of each region during a single acquisition based on the surface morphology complexity degree of each region and the acquisition influence degree of the region during a single acquisition; performing perspective projection transformation processing on the point cloud data and images of each region during a single acquisition, and performing linear interpolation based on the processing result to obtain a high-dimensional image corresponding to each acquisition, where each pixel point in the high-dimensional image corresponds to a mapped three-dimensional space coordinate; performing weighted fusion based on the acquisition data interference degree and the high-dimensional image of each region during a single acquisition to obtain a three-dimensional space coordinate set corresponding to each region, and the three-dimensional space coordinate set is used to generate a BIM model of highway engineering.

[0006] In combination with the first aspect, in a possible implementation manner, calculating the acquisition influence degree at each moment based on the flight data and the images within the preset time period includes: calculating the image deviation degree corresponding to each moment based on the difference between adjacent images in time sequence; calculating the drone attitude deviation degree corresponding to each moment based on the flight data; and multiplying the image deviation degree and the drone attitude deviation degree corresponding to each moment to calculate the acquisition influence degree.

[0007] In combination with the first aspect, in a possible implementation manner, calculating the image deviation degree corresponding to each moment according to the difference between adjacent images in time sequence includes: processing all the images according to the optical flow algorithm to obtain the optical flow vector of each pixel point in each image; counting the number of optical flow vectors in each image to obtain the optical flow number corresponding to each image; calculating the first optical flow mean vector and the second optical flow mean vector corresponding to each image respectively according to the time sequence, where the first optical flow mean vector is the mean of the optical flow vectors within the first preset time period, the second optical flow mean vector is the mean of the optical flow vectors within the second preset time period, the cut-off moments of the first preset time period and the second preset time period are both the acquisition moment of this image, and the time length of the first preset time period is greater than the time length of the second preset time period; calculating the image deviation degree corresponding to each image according to the first optical flow mean vector, the second optical flow mean vector, the optical flow number corresponding to each image, and the optical flow vector of each pixel point.

[0008] In combination with the first aspect, in a possible implementation manner, the flight data includes acceleration and angular velocity, and calculating the UAV attitude deviation degree corresponding to each moment according to the flight data includes: calculating the mean values of acceleration and angular velocity respectively according to the acceleration and angular velocity at each moment within the preset time period to obtain the acceleration mean value and the angular velocity mean value; calculating the UAV attitude deviation degree corresponding to each moment according to the difference between the acceleration at each moment and the acceleration mean value and the difference between the angular velocity and the angular velocity mean value.

[0009] In combination with the first aspect, in a possible implementation manner, calculating the surface morphology complexity degree of each area according to the point cloud data when the area is collected multiple times includes: calculating the point cloud density and the point cloud curvature variance at the time of single collection respectively according to the point cloud data at the time of single collection; calculating the mean values of the point cloud density and the point cloud curvature variance during multiple collections to obtain the density mean value and the curvature variance mean value; multiplying the density mean value by the curvature variance mean value to obtain the surface morphology complexity degree.

[0010] In combination with the first aspect, in a possible implementation manner, calculating the point cloud density and the point cloud curvature variance at the time of single collection according to the point cloud data at the time of single collection includes: performing density division processing on the point cloud data at the time of single collection to obtain the number of point cloud cluster classes, and using the number of point cloud cluster classes as the point cloud density; performing local surface fitting on each point at the time of single collection to obtain the curvature of each point; calculating the variance of the curvatures of all points to obtain the curvature variance.

[0011] In combination with the first aspect, in a possible implementation manner, the number of point cloud cluster classes is obtained by processing with the DBSCAN algorithm.

[0012] In a second aspect, the present application further provides a BIM data acquisition system for highway engineering, including: a data acquisition module, configured to acquire the acquisition data and flight data of the unmanned aerial vehicle within a preset time period, where the preset time period is the time period during which the unmanned aerial vehicle flies uniformly along the target highway section at a preset speed, and the acquisition data includes point cloud data and images; an influence factor calculation module, configured to calculate the acquisition influence degree at each moment according to the flight data and the images within the preset time period; a grouping module, configured to perform regional segmentation according to the point cloud data corresponding to each moment, and perform grouping division based on the segmentation result to obtain the point cloud data when each region is collected multiple times; a morphological factor calculation module, configured to calculate the surface morphological complexity of each region according to the point cloud data when each region is collected multiple times, where the surface morphological complexity is calculated from the point cloud density distribution and the point cloud curvature difference during each collection; a disturbance factor calculation module, configured to calculate the acquisition data disturbance degree of each region during a single collection according to the surface morphological complexity of each region and the acquisition influence degree of the region during a single collection; a transformation processing module, configured to perform perspective projection transformation processing on the point cloud data and images of each region during a single collection, and perform linear interpolation based on the processing result to obtain a high-dimensional image corresponding to each collection, where each pixel point in the high-dimensional image corresponds to a mapped three-dimensional space coordinate; a fusion calculation is performed to perform weighted fusion based on the acquisition data disturbance degree and the high-dimensional image of each region during a single collection to obtain a three-dimensional space coordinate set corresponding to each region, and the three-dimensional space coordinate set is used to generate a BIM model of highway engineering.

[0013] In combination with the second aspect, in a possible implementation manner, the influence factor calculation module includes: an image deviation calculation module, configured to calculate the image deviation degree corresponding to each moment according to the difference between adjacent images in time series; an attitude deviation calculation module, configured to calculate the attitude deviation degree of the unmanned aerial vehicle corresponding to each moment according to the flight data; a first calculation module, configured to multiply the image deviation degree and the attitude deviation degree corresponding to each moment to calculate the acquisition influence degree.

[0014] In combination with the second aspect, in a possible implementation manner, the morphological factor calculation module includes: a point cloud calculation module, configured to calculate the point cloud density and the point cloud curvature variance during a single collection according to the point cloud data during a single collection respectively; an average value calculation module, configured to calculate the density average value and the curvature variance average value according to the point cloud density and the point cloud curvature variance during multiple collections; a second calculation module, configured to multiply the density average value by the curvature variance average value to obtain the surface morphological complexity.

[0015] The present invention has the following beneficial effects:

[0016] In the present invention, first, the influence caused by the flight postures at different times is estimated to obtain the acquisition influence degree. Subsequently, the coverage area during the acquisition task of the UAV is divided into regions, and the complexity of the surface morphology of each region is evaluated to obtain the surface morphology complexity. Based on these evaluation results, the data interference degree of each region during the acquisition process at each acquisition moment is comprehensively calculated. Then, in the present invention, the point cloud data and image data obtained by the acquisition are subjected to mapping processing, and these data are effectively fused according to the interference degree at each acquisition. Through the above processing, a three-dimensional space coordinate set is obtained, which can reduce the weight of the point cloud data when the wind force is affected, thereby significantly improving the reliability and accuracy of the BIM acquisition data for highway engineering. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work.

[0018] Figure 1 It is a schematic flowchart of a BIM data acquisition method applied to highway engineering provided by Embodiment 1 of the present invention;

[0019] Figure 2 It is a schematic flowchart of step S2 provided by Embodiment 1 of the present invention;

[0020] Figure 3 It is a schematic flowchart of step S4 provided by Embodiment 1 of the present invention;

[0021] Figure 4 It is a schematic structural diagram of a BIM data acquisition system applied to highway engineering described in Embodiment 2 of the present invention;

[0022] Figure 5 It is a schematic structural diagram of an influence factor calculation module described in Embodiment 2 of the present invention;

[0023] Figure 6 It is a schematic structural diagram of a morphology factor calculation module described in Embodiment 2 of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0024] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation manner, structure, features and effects of a BIM data acquisition method and system applied to highway engineering proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this invention belongs.

[0026] Embodiment 1:

[0027] The following specifically describes the specific solution of a BIM data acquisition method applied to highway engineering provided by the present invention in conjunction with the accompanying drawings.

[0028] Please refer to Figure 1 , which shows a schematic flow chart of a BIM data acquisition method applied to highway engineering provided by an embodiment of the present invention. Among them, in this embodiment, this method altogether includes steps S1 to S7.

[0029] S1. Obtain the acquisition data and flight data of the unmanned aerial vehicle within a preset time period. The preset time period is the time period during which the unmanned aerial vehicle flies uniformly along the target highway section at a preset speed. The acquisition data includes point cloud data and images;

[0030] In this step, the unmanned aerial vehicle will fly along the center line of the target highway section and perform vertical downward image shooting. This unmanned aerial vehicle is equipped with a Velodyne lidar device, whose scanning accuracy exceeds 200 points per square meter, and supports multi-echo technology, capable of penetrating the vegetation layer to obtain accurate surface data. In addition, the unmanned aerial vehicle is also equipped with a high-definition camera with a resolution of 4K, which supplements the deficiencies in details of the lidar scan point cloud data by collecting high-resolution image data. At the same time, in this embodiment, this camera has an optical image stabilization function to ensure that clear road surface images can be captured during flight. In addition, the unmanned aerial vehicle is also equipped with a positioning system to ensure the spatial consistency of the lidar point cloud data and the image data, and record the real-time position and flight trajectory of the unmanned aerial vehicle. And the communication module can transmit the collected external data to the ground end in real time for construction and modeling analysis. Further, regarding the lidar device and the camera mentioned in this embodiment, those skilled in the art can also replace them with lidar devices and cameras of other specifications according to actual needs, and no specific limitations are made in this embodiment.

[0031] Secondly, in this embodiment, the target highway segments refer to those areas where 3D scanning using vehicle-mounted lidar is not feasible due to terrain complexity or construction restrictions. These areas may involve special terrains such as mountains, tunnels, bridges, or areas where normal scanning is not possible due to construction activities. Additionally, this embodiment can also be used for BIM data collection of specific highway segments according to actual needs. Regarding the preset speed mentioned in this step, its value can be 1 meter per second or 0.5 meter per second, or adjusted according to actual needs. This embodiment does not impose specific limitations on the preset speed. The determining factors for the preset time include the preset speed and the length of the target highway segment. It can be understood that if the preset speed is lower, the camera shooting duration and the lidar device scanning times for a certain area in the target highway segment will increase accordingly, thereby improving the accuracy of BIM data.

[0032] Meanwhile, in this embodiment, the images mentioned can be obtained by synchronously shooting and collecting during the scanning process of the lidar device; or they can be video data captured during the flight of the drone. Subsequently, according to the scanning nodes of the lidar device, corresponding frames are extracted from the video data as the corresponding images. Additionally, the scanning speed of the lidar device and the frame rate of the captured video can be set to be the same to ensure that the point cloud data obtained from each scan can correspond one-to-one with the video frames. In other words, in this embodiment, in order to maximize the utilization of all the collected point cloud data, the moments mentioned in the subsequent steps are all the moments of each point cloud scan, and each such moment corresponds to an image, or can be referred to as the moment of each point cloud data collection. The specific implementation method can be selected by those skilled in the art according to the actual situation, and this embodiment does not make specific limitations on this.

[0033] In this embodiment, considering that the drone may be affected by environmental factors such as external wind during flight, which may cause the flight attitude to deviate, thereby affecting the accuracy of the collected highway point cloud data. To solve the above problems, in this embodiment, in-depth analysis is carried out on the road surface point cloud data and its flight data collected by the drone during flight. By evaluating the interference of external environmental changes on data collection and the complexity within the collection area, the error of the collected point data can be evaluated, and based on this, the actual collected data can be adjusted and corrected to ensure the acquisition of accurate highway BIM data. Therefore, the specific implementation steps are detailed in steps S2 to S4.

[0034] S2. Calculate the acquisition influence degree at each moment based on the flight data and the images within the preset time period;

[0035] Specifically, in this embodiment, considering the characteristics of the image, that is, as a kind of planar data, compared with point cloud data, the image lacks a spatial dimension. This characteristic leads to larger errors being more likely to occur in the image data during the later processing, especially during linear interpolation. Therefore, in order to improve the accuracy of data processing, this embodiment proposes a new method, which involves combining the attitude changes of the unmanned aerial vehicle at different flight times and the corresponding image changes captured at those times to calculate the influence degree of the images collected at each time. To illustrate this process more clearly, reference can be made to Figure 2 the flowchart shown. In Figure 2 , the further refinement of step S2 is shown in detail, which includes step S21, step S22, and step S23. These steps together constitute the calculation process of the influence degree of image acquisition.

[0036] S21. Calculate the image deviation degree corresponding to each of the times according to the difference between the adjacent images in time sequence;

[0037] In this step, the differences between the images mentioned can be evaluated and compared by various methods. These methods include but are not limited to comparing pixel differences, comparing structural similarities, and calculating the mean square error, etc. First, the evaluation of pixel differences can be achieved by directly comparing the values of the corresponding pixels in two adjacent images. The specific operation is to calculate the difference of each pixel point, and then calculate the average value or standard deviation of these differences to quantify the difference degree between the images. Secondly, the evaluation of structural similarity involves comparing the brightness, contrast, and structural information of two adjacent images, and judging the similarity degree of the images through the similarity of these visual features. Finally, the calculation of the mean square error is to square the pixel differences between two images and then calculate the average of these squared differences. The smaller the mean square error, the higher the similarity between the two images. In addition to the methods mentioned above, those skilled in the art can also adopt other various algorithms for calculating differences, and this embodiment does not make specific limitations on this.

[0038] In addition, in this embodiment, a method for matching adjacent images and calculating the differences between them by using the optical flow algorithm is also provided. Specifically, it includes steps S211 - S214. These sub - steps describe how to use the optical flow algorithm to track the motion information in the image sequence, so as to achieve the accurate calculation of the image differences.

[0039] S211. Process all the images according to the optical flow algorithm to obtain the optical flow vector of each pixel point in each of the images;

[0040] In this embodiment, the motion vectors of each pixel in the image are calculated using Dense Optical Flow. The Optical Flow algorithm is a method for analyzing the motion of objects in an image sequence. By estimating the motion field of pixels in the image, this method can detect and describe the movement of objects between consecutive frames. Specifically, it can be implemented through the Lucas-Kanade algorithm, or the Horn-Schunck algorithm, or the Farneback algorithm, all of which are prior arts and the specific implementation processes will not be elaborated in this embodiment.

[0041] S212. Count the number of optical flow vectors in each of the images to obtain the optical flow number corresponding to each of the images;

[0042] S213. Calculate respectively according to the time sequence to obtain the first optical flow mean vector and the second optical flow mean vector corresponding to each of the images. The first optical flow mean vector is the mean of the optical flow vectors within the first preset time period, and the second optical flow mean vector is the mean of the optical flow vectors within the second preset time period. The end moments of both the first preset time period and the second preset time period are the acquisition moment of this image, and the time length of the first preset time period is greater than the time length of the second preset time period;

[0043] It should be noted that in this embodiment, the long time span that the unmanned aerial vehicle (UAV) may experience during flight is fully considered, for example, it can be up to 30 minutes. However, the duration of the fluctuation phenomenon that occurs during the flight of the UAV is often significantly affected by instantaneous factors such as wind force. Therefore, in this embodiment, a method is adopted, that is, taking the specific moment to be calculated or the image acquisition moment corresponding to this moment as the end point, and intercepting a period of images along the time series forward as the basis for analyzing the fluctuation phenomenon. Specifically, for example, a shorter time period is set, called the first preset time period, and its time length can be 1 minute; at the same time, a shorter time period is set, called the second preset time period, and its time length can be 10 seconds. It should be noted that for the specific durations of the first preset time period and the second preset time period, those skilled in the art can flexibly select according to the actual flight situation and analysis requirements. In this embodiment, we do not strictly limit these durations to ensure the applicability and flexibility of the method.

[0044] S214. Calculate the image deviation degree corresponding to each of the images according to the first optical flow mean vector, the second optical flow mean vector, the optical flow number corresponding to each of the images, and the optical flow vector of each pixel point.

[0045] Specifically, in this embodiment, the calculation functional formula of the image deviation degree is as follows:

[0046] ;

[0047] Among them, represents the image deviation degree corresponding to the image taken at the -th moment; represents the max - min normalization function; represents the second optical flow mean vector within the second preset time period; represents the first optical flow mean vector within the first preset time period; represents the -th number of optical flows corresponding to the image taken at the moment; represents the -th optical flow vector of the -th pixel point in the image taken at the moment; represents the calculation function for taking the modulus of the vector.

[0048] In the above calculation function formula, represents analyzing the optical flow vector corresponding to the image taken at the -th moment. If the acquisition angle deviates at the current moment, it will cause differences in the magnitude and direction of the optical flow vectors acquired at the same moment; represents analyzing each pixel point in the image taken at the -th moment one by one. If the value is larger, it indicates a higher degree of deviation at the current moment. Through the above calculation formula function, the differences between adjacent images can be better described.

[0049] S22. Calculate the attitude deviation degree of the drone corresponding to each of the said moments according to the said flight data;

[0050] In this embodiment, the flight data is acquired in real - time through the inertial measurement unit (IMU) carried by the drone itself. These flight data mainly include two aspects: acceleration and angular velocity. Among them, in this embodiment, considering that when the drone encounters large attitude changes during flight, there may be large deviations in the acquisition of flight data at this time. In view of this, for the analysis and processing of the drone's flight data, calculating the attitude deviation degree of the drone is particularly important. To achieve this goal, this embodiment also particularly includes two key steps, namely step S221 and step S222.

[0051] S221. Calculate the mean values of acceleration and angular velocity respectively according to the acceleration and angular velocity at each moment within the preset time period, and obtain the mean acceleration and mean angular velocity;

[0052] S222. Calculate the attitude deviation degree of the drone corresponding to each moment according to the difference between the acceleration at each moment and the mean acceleration and the difference between the angular velocity at each moment and the mean angular velocity.

[0053] Specifically, the absolute value of the difference between the average acceleration during the flight of the drone and the acceleration of the drone at each moment is used as the degree of acceleration difference at each moment; the absolute value of the difference between the average angular velocity during the flight of the drone and the angular velocity of the drone at each moment is used as the degree of angular velocity difference at each moment; at each moment, normalization is performed according to the product between the corresponding acceleration difference degree and the corresponding angular velocity difference degree, and the attitude deviation degree of the drone corresponding to each moment is obtained; among them, the normalization method adopts the maximum-minimum normalization function; by measuring through the angular velocity dimension and the acceleration dimension respectively, the attitude performance of the drone during flight can be better reflected.

[0054] S23. Multiply the image deviation degree corresponding to each said moment and the attitude deviation degree of the drone to calculate the acquisition influence degree.

[0055] S3. Perform region segmentation according to the point cloud data corresponding to each moment, and perform grouping division based on the segmentation result to obtain the point cloud data collected multiple times in each region;

[0056] As described above, in this embodiment, since the drone flies uniformly along the target highway section. Therefore, each location in the target highway section will be collected multiple times within a continuous time period. If the attitude of the drone changes during this time period, there will be data deviation with time in the collected area. Therefore, in this embodiment, by performing region segmentation processing on the point cloud data corresponding to each moment in the way of a preset area, multiple regions can be segmented at a single moment and a unique code can be further assigned to each region. In this way, the point cloud data corresponding to the same region can be classified together, and thus the point cloud data collected multiple times in a single region can be obtained. At the same time, it should be noted that in this embodiment, the preset area can be a square cell of 1*1 square meter or a strip cell of 1*the width of the target highway section. The specific area and shape can be selected by those skilled in the art according to the actual situation. No specific limitation is made in this embodiment. At the same time, for those skilled in the art, in order to perform more accurate region segmentation, in this embodiment, the feature points in the point cloud data and the image can be extracted through the feature matching algorithm SIFT (Scale-Invariant Feature Transform) and matched to realize the alignment between the image and the point cloud data, so as to perform accurate preset area division.

[0057] In this embodiment, the drone performs a uniform flight task along the target highway section. Therefore, each specific position of the target highway section will be collected multiple times within consecutive time periods. If the attitude of the drone changes during the collection, the collected data will exhibit time-varying deviations. To address this issue, this embodiment adopts a method of performing regional segmentation processing on the point cloud data at each time point according to a predetermined area. In this way, the point cloud data can be segmented into multiple regions at a single time point, and a unique code is assigned to each region. The purpose of this is to classify the point cloud data of the same region, thereby obtaining a point cloud data set of a single region collected multiple times. In addition, it should be clear that in this embodiment, the predetermined area can be a square cell of 1*1 square meter or a strip cell of 1*the width of the target highway section. As for the specific area size and shape, those skilled in the art can choose according to the actual situation. This embodiment does not make specific limitations in this regard. At the same time, for those skilled in the art, in order to further improve the accuracy of regional segmentation, in this embodiment, the feature matching algorithm SIFT (Scale-Invariant Feature Transform) can be used to extract the feature points in the point cloud data and the image, and the alignment between the image and the point cloud data can be achieved by matching these feature points. In this way, the division of the predetermined area can be more accurate, thereby improving the precision of data collection.

[0058] S4. Calculate the surface morphology complexity of each of the regions based on the point cloud data when each region is collected multiple times, where the surface morphology complexity is calculated from the point cloud density distribution and the point cloud curvature difference during each collection;

[0059] In this embodiment, further considering the impact of the change in the flight attitude of the drone on the collected point cloud data when detecting any specific region. In particular, when the surface morphology complexity of the detection region is relatively high, the degree of impact of the attitude change on the data also increases accordingly. To accurately evaluate this impact, in this embodiment, the density distribution and curvature difference of the point cloud data of a single region at multiple collection moments are analyzed. The specific operation steps can be referred to Figure 3 , which details step S4 in the figure, including steps S41 to S43, and these steps together constitute an exemplary process for calculating the surface morphology complexity of a specific region.

[0060] S41. Calculate the point cloud density and the point cloud curvature variance during a single collection based on the point cloud data during the single collection;

[0061] Among them, the method for calculating the point cloud density mentioned in this step can be calculated based on the number of point clouds within an area. However, in this embodiment, considering that the change in the number of point clouds per unit area caused by the attitude of the drone at adjacent moments may not be obvious. Therefore, in this embodiment, the method for calculating the point cloud density is as follows: the point cloud data collected once is divided based on density to obtain the number of point cloud cluster classes, and the number of point cloud cluster classes is used as the point cloud density. Among them, density-based partitioning can be performed through the DBSCAN algorithm in the clustering algorithm. By using this method, the obtained point cloud cluster classes may better describe the complexity of the area. The larger the value, the more chaotic the point cloud distribution in the current area, and the higher the complexity of the surface morphology. At the same time, the point cloud cluster classes are also sensitive to the shooting angle and can better reflect a change in the attitude of the drone. Regarding the variance of the point cloud curvature, local surface fitting can be performed on each point during a single collection to obtain the curvature of each point; then, the variance of the curvature is calculated based on the curvatures of all points. Among them, regarding local surface fitting and curvature calculation, they are prior arts, and the specific process is not elaborated in this embodiment.

[0062] S42. Calculate the density mean and the mean variance of the curvature by performing mean calculation based on the point cloud density and the variance of the point cloud curvature during multiple collections;

[0063] S43. Multiply the density mean by the mean variance of the curvature to obtain the complexity of the surface morphology.

[0064] Therefore, when the point cloud curvature of a specific area shows significant changes, and at the same time, the distribution of the point cloud cluster classes within this area also shows large differences, this usually indicates that the area has rich texture details. On the contrary, if data deviation occurs in this area during the data collection process, then due to the richness of the texture details, these deviations may be amplified during the subsequent linear interpolation process. In view of this, when using the complexity of the surface morphology to correct data in the subsequent steps of this embodiment, specifically, for the data collected under stable conditions, this embodiment will assign a greater weight, while for the data collected under non-stable conditions, a smaller weight will be assigned, so as to optimize the accuracy and reliability of the data.

[0065] S5. Calculate the interference degree of the collected data for each area during a single collection based on the complexity of the surface morphology of each area and the influence degree of the collection of this area during a single collection;

[0066] Specifically, in this embodiment, considering that the attitude change during the flight of the drone may cause changes in the shooting angle and shooting height, and the changes in the shooting angle and shooting height will directly affect the image distortion situation. Therefore, in this embodiment, it is also considered to combine and correct the shooting angle and shooting height to obtain the interference degree of the collected data.

[0067] That is, in this embodiment, the flight data further includes the shooting altitude and the shooting angle of the camera, which can be obtained by the built-in sensors of the drone respectively.

[0068] Specifically, calculate the sampling distance between the drone and the target highway section in each area at each moment; take the product of the sampling distance corresponding to each area at each moment, the acquisition influence degree of the drone at each moment, and the complexity ratio of the surface morphology in each area as the reference influence degree corresponding to each area at each moment; obtain the shooting angle of each area at each moment; normalize the ratio between the reference influence degree and the shooting angle to obtain the acquisition data interference degree of each area at each moment; among them, the normalization method uses the maximum-minimum normalization function; and when the shooting angle is 0, it is replaced with 0.001 to avoid the situation of the denominator being 0.

[0069] It should be noted that the sampling distance is calculated from the height and width of the sensors on the drone, the height and width of the image, as well as the focal length and the flight altitude, which is the prior art and will not be elaborated in this embodiment; in the process of obtaining the above acquisition data interference degree, when the drone collects data for an area, the greater the degree of interference of the drone itself at the corresponding moment, the greater the acquisition influence degree, and at the same time, the higher the complexity of the surface morphology in the corresponding acquisition area, the farther the sampling distance at the corresponding moment, and the smaller the shooting angle, the greater the influence degree of the acquisition data points in the area at the current moment.

[0070] At the same time, for those skilled in the art, it is also possible not to use these two parameters of the shooting altitude and the shooting angle, and directly multiply the complexity of the surface morphology by the acquisition influence degree and normalize it to obtain the acquisition data interference degree.

[0071] S6. Perform perspective projection transformation processing on the point cloud data and the image in each area during a single acquisition, and perform linear interpolation based on the processing result to obtain a high-dimensional image corresponding to each acquisition. Each pixel point in the high-dimensional image corresponds to a mapped three-dimensional space coordinate.

[0072] Regarding the perspective projection transformation processing, its core is the process of converting the three-dimensional coordinates of the point cloud data from the coordinate system of the lidar device to the corresponding pixel points in the camera coordinate system through the external parameter matrix and the internal parameter matrix. The specific implementation process is the prior art and will not be elaborated in this embodiment.

[0073] S7. Based on the acquisition data interference degree and the high-dimensional image in each area during a single acquisition, perform weighted fusion to obtain a set of three-dimensional space coordinates corresponding to each area. The set of three-dimensional space coordinates is used to generate the BIM model of the highway project.

[0074] After the processing of step S6, there will be multiple multi-dimensional images in each area. That is, the same pixel point may have different mapped three-dimensional space coordinates during the acquisition process. And in this step, the different mapped three-dimensional space coordinates are fused and calculated to obtain a three-dimensional space coordinate. Specifically, the functional formula for the fusion calculation of the three-dimensional space coordinates is as follows:

[0075] ;

[0076] Among them, represents the adjusted three-dimensional space coordinate of the th pixel point in the th area; represents the number of times the th area is acquired; represents the degree of interference of the acquisition data corresponding to the rd area at the th moment; represents the average value of the degree of interference of the acquisition data corresponding to the th area; represents the total number of areas; represents the mapped three-dimensional space coordinate of the th pixel point area in the image corresponding to the th area at the th moment.

[0077] In the above calculation functional formula, is obtained by calculating the average value of all the degrees of interference of the acquisition data of the th area and the number of times the th area is acquired.

[0078] In the above calculation functional formula, represents that by fusing and calculating the weights corresponding to each acquisition, when the degree of external interference received by a certain acquisition is higher, it means that the data credibility in the high-dimensional image corresponding to this acquisition is lower. Therefore, based on this, the three-dimensional space coordinates in each high-dimensional image are fused and calculated to finally obtain the three-dimensional space coordinate of the current pixel point.

[0079] ​In summary, in this embodiment, considering that during the process of the drone collecting highway BIM data, due to being vulnerable to external environmental interference, the collected data deviates, thus affecting the accuracy of the BIM data. In response to this phenomenon, the single collection range of the drone is divided into regions, and the interference degree of the external environment during different collections and the complexity of the surface morphology in the collection area are considered, so as to determine the interference degree of a single region in the single collected data. Further, by performing mapping processing on the collected point cloud data and images, and adjusting the weights of the data according to the interference degree during each collection, combined with multiple collected data, the adjusted single-region data is obtained, further ensuring the reliability and accuracy of the highway engineering BIM collected data.

[0080] Embodiment 2:

[0081] As Figure 4 shown, this embodiment provides a BIM data collection system applied to highway engineering. The system includes:

[0082] A data acquisition module, configured to acquire the collection data and flight data of the drone within a preset time period. The preset time period is the time period during which the drone flies uniformly along the target highway section at a preset speed. The collection data includes point cloud data and images.

[0083] An influence factor calculation module, configured to calculate the collection influence degree at each moment according to the flight data and the images within the preset time period.

[0084] A grouping module, configured to perform region segmentation according to the point cloud data corresponding to each moment, and perform grouping division based on the segmentation result to obtain the point cloud data when each region is collected multiple times.

[0085] A morphology factor calculation module, configured to calculate the surface morphology complexity of each region according to the point cloud data when each region is collected multiple times. The surface morphology complexity is calculated from the point cloud density distribution and the point cloud curvature difference during each collection.

[0086] An interference factor calculation module, configured to calculate the collection data interference degree of each region during a single collection according to the surface morphology complexity of each region and the collection influence degree of the region during a single collection.

[0087] A transformation processing module, configured to perform perspective projection transformation processing on the point cloud data and images of each region during a single collection, and perform linear interpolation based on the processing result to obtain a high-dimensional image corresponding to each collection. Each pixel point in the high-dimensional image corresponds to a mapped three-dimensional space coordinate.

[0088] The fusion calculation module is used to perform weighted fusion based on the acquisition data interference degree of each region during a single acquisition and the high-dimensional image, so as to obtain a three-dimensional space coordinate set corresponding to each region, and the three-dimensional space coordinate set is used to generate a BIM model of a highway project.

[0089] Further, as Figure 5 shown, the influence factor calculation module includes:

[0090] The image deviation calculation module is used to calculate the image deviation degree corresponding to each moment according to the difference between adjacent images in time series.

[0091] The attitude deviation calculation module is used to calculate the UAV attitude deviation degree corresponding to each moment according to the flight data.

[0092] The first calculation module is used to multiply the image deviation degree and the UAV attitude deviation degree corresponding to each moment to calculate the acquisition influence degree.

[0093] Further, as Figure 6 shown, the form factor calculation module includes:

[0094] The point cloud calculation module is used to calculate the point cloud density and the point cloud curvature variance during a single acquisition respectively according to the point cloud data during a single acquisition.

[0095] The mean value calculation module is used to perform mean value calculation on the point cloud density and the point cloud curvature variance during multiple acquisitions to obtain the density mean value and the curvature variance mean value.

[0096] The second calculation module is used to multiply the density mean value by the curvature variance mean value to obtain the surface form complexity.

[0097] It should be noted that regarding the system in the above embodiments, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here in detail.

[0098] It should be noted that: the above sequence of the embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0099] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A BIM data acquisition method applied to highway engineering, characterized in that, The method includes: Obtaining the acquisition data and flight data of the unmanned aerial vehicle (UAV) within a preset time period, where the preset time period is the time period during which the UAV flies uniformly along the target highway section at a preset speed, and the acquisition data includes point cloud data and images; Calculating the acquisition influence degree at each moment based on the flight data and the images within the preset time period; Performing region segmentation based on the point cloud data corresponding to each moment, and performing grouping division based on the segmentation result to obtain the point cloud data when each region is acquired multiple times; Calculating the surface morphology complexity degree of each region based on the point cloud data when each region is acquired multiple times, where the surface morphology complexity degree is calculated from the point cloud density distribution and the point cloud curvature difference during each acquisition; Calculating the acquisition data interference degree of each region during a single acquisition based on the surface morphology complexity degree of each region and the acquisition influence degree of the region during a single acquisition; Performing perspective projection transformation processing on the point cloud data and images of each region during a single acquisition, and performing linear interpolation based on the processing result to obtain a high-dimensional image corresponding to each acquisition, where each pixel point in the high-dimensional image corresponds to a mapped three-dimensional space coordinate; Performing weighted fusion based on the acquisition data interference degree and the high-dimensional image of each region during a single acquisition to obtain a three-dimensional space coordinate set corresponding to each region, and the three-dimensional space coordinate set is used to generate a BIM model of the highway project; The process of obtaining the acquisition influence degree includes: calculating the image deviation degree corresponding to each moment according to the difference between adjacent images in time series; calculating the UAV attitude deviation degree corresponding to each moment according to the flight data; multiplying the image deviation degree and the UAV attitude deviation degree corresponding to each moment to calculate the acquisition influence degree.

2. The BIM data acquisition method applied to highway engineering according to claim 1, wherein, Calculating the image deviation degree corresponding to each moment according to the difference between adjacent images in time series includes: Processing all the images according to the optical flow algorithm to obtain the optical flow vector of each pixel point in each image; Counting the number of optical flow vectors in each image to obtain the optical flow number corresponding to each image; Calculating the first optical flow mean vector and the second optical flow mean vector corresponding to each image respectively according to the time series, where the first optical flow mean vector is the mean of the optical flow vectors within the first preset time period, the second optical flow mean vector is the mean of the optical flow vectors within the second preset time period, the cut-off moments of the first preset time period and the second preset time period are both the acquisition moment of the image, and the time length of the first preset time period is greater than the time length of the second preset time period; Calculating the image deviation degree corresponding to each image according to the first optical flow mean vector, the second optical flow mean vector, the optical flow number corresponding to each image, and the optical flow vector of each pixel point; 3. The BIM data collection method applied to highway engineering according to claim 1, characterized in that, The flight data includes acceleration and angular velocity. Calculating the UAV attitude deviation degree corresponding to each moment according to the flight data includes: Calculate the average values of acceleration and angular velocity respectively at each moment within a preset time period to obtain the average acceleration and average angular velocity; Calculate the attitude deviation degree of the UAV corresponding to each moment according to the difference between the acceleration at each moment and the average acceleration and the difference between the angular velocity and the average angular velocity; 4. The BIM data acquisition method applied to highway engineering according to claim 1, characterized in that, Calculate the surface morphology complexity of each region based on the point cloud data when each region is collected multiple times, including: Calculate the point cloud density and point cloud curvature variance at the time of single collection respectively according to the point cloud data at the time of single collection; Calculate the average density and average curvature variance by calculating the average values of the point cloud density and point cloud curvature variance during multiple collections; Multiply the average density by the average curvature variance to obtain the surface morphology complexity; 5. The BIM data acquisition method applied to highway engineering according to claim 4, wherein Calculate the point cloud density and point cloud curvature variance at the time of single collection according to the point cloud data at the time of single collection, including: Perform density division processing on the point cloud data at the time of single collection to obtain the number of point cloud cluster classes, and use the number of point cloud cluster classes as the point cloud density; Perform local surface fitting on each point at the time of single collection to obtain the curvature of each point; Calculate the variance of the curvature based on the curvatures of all points to obtain the curvature variance; 6. The BIM data acquisition method applied to highway engineering according to claim 5, wherein, The number of point cloud cluster classes is obtained by processing with the DBSCAN algorithm; 7. A BIM data acquisition system applied to highway engineering, characterized in that, Including: A data acquisition module for acquiring the acquisition data and flight data of the UAV within a preset time period, where the preset time period is the time period when the UAV flies uniformly along the target highway section at a preset speed, and the acquisition data includes point cloud data and images; An influence factor calculation module for calculating the acquisition influence degree at each moment according to the flight data and the images within the preset time period; The process of obtaining the acquisition influence degree includes: calculating the image deviation degree corresponding to each moment according to the difference between adjacent images in time series; calculating the attitude deviation degree of the UAV corresponding to each moment according to the flight data; multiplying the image deviation degree and the attitude deviation degree of the UAV corresponding to each moment to calculate the acquisition influence degree; A grouping module for performing region segmentation according to the point cloud data corresponding to each moment, and performing grouping division based on the segmentation result to obtain the point cloud data when each region is collected multiple times; A morphology factor calculation module for calculating the surface morphology complexity of each region according to the point cloud data when each region is collected multiple times, and the surface morphology complexity is calculated from the point cloud density distribution and point cloud curvature difference at each collection; A interference factor calculation module for calculating the acquisition data interference degree of each region at the time of single collection according to the surface morphology complexity of each region and the acquisition influence degree of the region at the time of single collection; A transformation processing module for performing perspective projection transformation processing on the point cloud data and images of each region at the time of single collection, and performing linear interpolation based on the processing result to obtain a high-dimensional image corresponding to each collection, and each pixel point in the high-dimensional image corresponds to a mapped three-dimensional space coordinate; A fusion calculation module, configured to perform weighted fusion on the acquisition data interference degree and the high-dimensional image based on each of the regions during a single acquisition, so as to obtain a three-dimensional space coordinate set corresponding to each of the regions, and the three-dimensional space coordinate set is used to generate a BIM model of a highway project.

8. The BIM data acquisition system applied to highway engineering according to claim 7, wherein, The influence factor calculation module includes: An image deviation calculation module, configured to calculate an image deviation degree corresponding to each of the moments according to the difference between the temporally adjacent images; An attitude deviation calculation module, configured to calculate a UAV attitude deviation degree corresponding to each of the moments according to the flight data; A first calculation module, configured to multiply the image deviation degree and the UAV attitude deviation degree corresponding to each of the moments to calculate an acquisition influence degree.

9. The BIM data acquisition system applied to highway engineering according to claim 7, wherein The morphology factor calculation module includes: A point cloud calculation module, configured to calculate a point cloud density and a point cloud curvature variance during a single acquisition respectively according to the point cloud data during the single acquisition; An average value calculation module, configured to calculate an average density value and an average curvature variance value by performing an average value calculation on the point cloud density and the point cloud curvature variance during multiple acquisitions; A second calculation module, configured to multiply the average density value by the average curvature variance value to obtain the complexity of the surface morphology.

Images