A method for measuring building damage traces based on laser point cloud difference

Abstract

The application relates to a kind of measurement methods for building damage trace based on laser point cloud difference, and relates to the field of building weapon damage trace measurement and damage effect evaluation.The application solves the problems of high operation risk, low efficiency and poor precision, fully excavates and extracts the three-dimensional information of building damage area (including small-scale damage information such as splashed and fallen stones) of building and its components, comprehensively obtains the detailed point cloud information of building, efficiently and accurately constructs the three-dimensional point cloud model of building, objectively and accurately describes the damage trace of building by using the difference, segmentation and feature description method provided in the application, and gives the corresponding description function, which can provide more reliable data basis support for quantitative analysis of building damage and damage effectiveness evaluation, and is suitable for wide promotion and application.

Description

Technical Field

[0001] This invention relates to the field of measuring weapon damage traces and assessing damage effects on buildings and structures, and particularly to a method for measuring damage traces on buildings and structures based on laser point cloud differential analysis. Specifically, it relates to a method for extracting typical damage traces on buildings and structures and describing damage features using multi-parameter parameters based on multi-platform repeated scanning laser point cloud data differential analysis technology. Background Technology

[0002] The high-precision extraction and measurement of damage traces on buildings and structures has long been a fundamental method for assessing the damage effects of buildings and structures and evaluating the destructive capabilities of weapons. Buildings and structures are various engineering facilities used to avoid or mitigate the destructive effects of weapons, such as protective buildings, bunkers, and command centers, and generally also include various urban infrastructure facilities such as residences, shopping malls, hospitals, airports, pipelines, and railways. Important buildings and structures face the risk of being hit and damaged by weapons during wartime, especially various protective engineering projects with significant military value. As important support and guarantee forces in modern warfare, they are also primary targets of weapon attacks and damage. Damage assessment is the main method for analyzing the degree of damage to buildings and structures, and can scientifically reflect the damage status of buildings and structures and the destructive power of weapon systems. It provides a comprehensive evaluation of the degree of loss of core functions of buildings and structures after damage, and can also provide necessary basic data for analyzing the damage effectiveness of weapon systems and optimizing attack strategies. Therefore, high-precision quantitative extraction and characteristic parameter analysis of damage traces are important supports for assessing damage to buildings and structures and the damage effectiveness of weapon systems.

[0003] Generally, on-site measurement, image acquisition, and expert evaluation of damaged areas of buildings and structures are currently the main methods for analyzing and identifying weapon damage to buildings and structures. The main characteristics of building and structure damage are craters and gaps caused by weapon impacts, as well as the resulting structural damage. Extracting and analyzing the grayscale information, shape distribution, size, and texture of the corresponding typical features of the damage is currently the main approach based on image processing methods to identify and analyze relevant features. This is mainly applied in the identification of weapon damage craters, such as craters on runways. By statistically analyzing the corresponding feature points, the degree of damage from different impacts can be assessed. It can also play a role in the extraction of damage traces in buildings and structures, such as identifying the central area and the extent of damage.

[0004] However, the aforementioned image-based damage identification and assessment algorithms suffer from low accuracy in extracting damage traces and rely on external light sources. They generally only identify the location and size of damage, lacking sufficient mining of three-dimensional information about the damaged area and missing volumetric information of the damage traces. Moreover, traditional field investigation methods are easily limited by complex damage scenarios and potential collapses, resulting in high operational risks, low efficiency, and poor accuracy. In addition, traditional methods relying on manual labor or image recognition often focus more on large-scale damage traces, and their measurement accuracy and reliability are insufficient. Due to the difficulty in reaching damaged areas or the presence of potential accident risks, some damaged areas cannot be observed internally, leading to the loss of key damage information and greatly affecting the accuracy of damage experts' interpretation. Furthermore, in addition to the damage to building components, the debris generated by the damage is also an important reference for judging the damage traces and size, but traditional damage identification methods tend to ignore this small-scale damage information, resulting in an inability to comprehensively and accurately characterize the distribution of damage-related features. Summary of the Invention

[0005] To overcome the shortcomings of the prior art, this invention provides a method for measuring building damage traces based on laser point cloud differential analysis. This invention utilizes a multi-platform stereoscopic monitoring method to comprehensively acquire point cloud data containing geometric structural information of buildings in the target area. It then precisely registers the three-dimensional point cloud data and segmentation model of the building before and after damage. Using the differential, segmentation, and feature description methods proposed in this invention, it mines damage-related three-dimensional geometric feature information, achieving accurate extraction and feature analysis of building damage traces, thus ensuring the scientific rigor and rationality of damage assessment.

[0006] To achieve the aforementioned inventive objectives, the present invention employs the following technical solution:

[0007] A method for measuring damage traces on buildings based on laser point cloud difference, the method specifically includes the following steps:

[0008] The first step is to construct a three-dimensional lidar platform and select appropriate base station locations based on the geometric structure, size, and surrounding environment of the buildings and structures, thereby enabling multi-platform three-dimensional monitoring of the buildings and structures and acquiring high-precision point cloud data of the buildings and structures in the target area.

[0009] The second step is to perform data preprocessing on the 3D point clouds acquired by LiDAR from different platforms, specifically including outlier removal, noise point removal, 3D point cloud adaptive filtering, point cloud downsampling, and feature line and surface extraction.

[0010] The third step is to perform high-precision registration of point cloud data from multiple platforms based on overlapping regions.

[0011] The fourth step is to reconstruct the three-dimensional model of the building based on the fused complete three-dimensional point cloud data, and to use the geometric morphology information and regular repetitive features of the reconstructed model to segment and assign weights to the main structure of the building to characterize the importance of each segmented component.

[0012] Step 5: Obtain the 3D point cloud model and segmentation model of the building before damage. Repeat steps 1 to 4 above to complete the 3D point cloud model and segmentation results after damage. Then, based on the line and surface features (reference points) of the undamaged area and the point cloud distance minimization optimization method, register the 3D point cloud data and segmentation model of the building before and after damage with high precision.

[0013] Step 6: Based on the registered 3D point cloud and segmented model of the building before and after damage, perform differential processing of the 3D point cloud and segmented model under the same 3D coordinate system. Quantitatively calculate the volume change of the 3D model of the building's internal and external structures before and after damage based on the differential results. By setting the minimum effective usable damage volume threshold, obtain the 3D model difference volume after damage, and segment the difference volume. Determine whether the volume change trend indicates an increase or decrease in volume.

[0014] Step 7: Based on the three-dimensional spatial distribution and volume information of the differential bodies, the characteristic parameters of the subtractive and augmenting bodies are described by using the cuboid envelope and Weil distribution function, respectively. The maximum likelihood optimization estimation method is then used to obtain the cuboid parameters corresponding to the augmenting body and the asymmetric Gaussian function parameters corresponding to the subtractive body.

[0015] Step 8: Based on the cuboid parameters (subtracted volume) and Weil distribution (increased volume) function and their characteristic parameters, and combined with the weighted volume of the subtracted and increased volumes, the corresponding damage to the building is calculated in a comprehensive manner, thereby accurately reflecting the damage effect caused by the explosion.

[0016] The method for measuring damage traces of buildings based on laser point cloud difference, in the first step of constructing a three-dimensional lidar platform, selects an unmanned aerial vehicle lidar platform P1, a ground lidar platform P2, and a mobile lidar platform P3 according to the internal and external geometric characteristics of the current building, respectively to acquire point cloud datasets D1 for the upper part of the building, D2 for the exterior facade, and D3 for the internal structure. Each point cloud dataset consists of one or more subsets, and each point cloud subset corresponds to one observation result.

[0017] In the method for measuring building damage traces based on laser point cloud differential, when selecting the corresponding base station location in the first step, in order to reduce the workload of subsequent point cloud data registration, it is necessary to select the corresponding observation station according to the geometric characteristics of the building structure, thereby reducing the number of station relocations. At the same time, it is necessary to ensure that the point cloud dataset has a certain degree of repeatability during each observation, thereby improving the accuracy and reliability of point cloud dataset registration.

[0018] The method for measuring building damage traces based on laser point cloud differential measurement, in the first step of multi-platform three-dimensional monitoring of buildings, forms a multi-platform three-dimensional monitoring method that combines indoor and outdoor monitoring of buildings, and is further assisted by the UAV lidar platform P1. This method comprehensively acquires point cloud data of the geometric structure information of buildings in the target area, while ensuring that the point cloud data of different platforms have a certain degree of repetition, so as to provide a foundation for subsequent stitching of complete point cloud data of the target area.

[0019] In the aforementioned method for measuring building damage traces based on laser point cloud differential analysis, during the second step of preprocessing the three-dimensional laser point cloud data, it is important to select appropriate parameters based on the differences in point cloud data density and accuracy acquired by different platform lidars to achieve better filtering effects.

[0020] The method for measuring building damage traces based on laser point cloud differential is described above. In the third step, when performing high-precision registration of multi-platform point cloud data based on overlapping regions, the combination of multiple platforms can leverage the advantages of different platforms, overcoming the shortcomings of traditional single platforms that only acquire point clouds of the exterior or interior of buildings. In order to unify the point cloud data of different platforms into the same coordinate system, the point cloud features of overlapping parts of different platforms and optimal methods such as weighted least squares are used to achieve high-precision registration and coordinate system unification of point cloud data from different platforms, thereby obtaining globally fused point cloud data that reflects the internal and external geometric features of buildings.

[0021] The method for measuring building damage traces based on laser point cloud difference, in the fourth step, the segmented components include walls, floors, beams, support beams or internal objects, with the load-bearing components related to structural stability having the highest weight.

[0022] In the method for measuring building damage traces based on laser point cloud difference, in the sixth step, the total volume of the increased volume and the volume of the decreased volume should be basically the same, but their spatial distribution is different. The increased volume is relatively discrete, while the decreased volume is relatively concentrated.

[0023] The method for measuring building damage traces based on laser point cloud difference, in the seventh step, mainly includes the descriptive parameters of the outer envelope cuboid of the subtracted body, which consists of the maximum and minimum values ​​of the subtracted body on the XYZ axes, thus forming a corresponding cuboid. Taking a wall damage gap perpendicular to the X-axis as an example, the corresponding cuboid of the subtracted body is the yellow area. At the point perpendicular to the X-axis and intersecting at point Xi, its affected area is the envelope rectangle (Zmin, Zmax, Ymin, Ymax) in the ZY plane, and the thickness of the wall is the thickness (H) of the corresponding cuboid, thus forming a corresponding... The subtractive outer envelope cuboid is represented by (Xi,Ymin,Zmin)(Xi,Ymin,Zmax)(Xi,Ymax,Zmin)(Xi,Ymax,Zmax)(Xi-H,Ymin,Zmin)(Xi-H,Ymin,Zmax)(Xi-H,Ymax,Zmin)(Xi-H,Ymax,Zmax). Other corresponding subtractive outer envelope cuboids are obtained using the same method. However, for the description of the augmenting body, due to its significant asymmetry, the Weil distribution is used to describe the augmenting body, and its corresponding formula function is as follows.

[0024]

[0025] Where β is the shape parameter of the volume distribution of the augment along the distance direction, η is the corresponding scaling factor, and γ is the position parameter, which is generally taken as 0. It can be seen that, with the Xc and Yc coordinates of the damage center as the center, the distance of the augment from that point on the horizontal axis, and the volume of the augment as the vertical axis, the volume distribution of the augment in a certain area can be statistically analyzed. The Weil distribution function and its corresponding parameters can be used to describe its distribution characteristics.

[0026] The method for measuring building damage traces based on laser point cloud difference, in the eighth step, for the cuboid parameters (subtraction), the total damage volume Vneg is calculated by weighting the weights and volumes of the components to which it belongs, and the volume Vs of the cuboid can be calculated by using the outer envelope of the cuboid; for the Weil distribution (addition), the cumulative value of the added volume Vpos of the corresponding function is calculated based on the corresponding Weil distribution parameters, and the total added volume Vz can also be calculated based on the added volume segmentation results. Based on these parameters and information, high-precision and refined measurement and analysis of damage can be performed.

[0027] By employing the technical solution described above, the present invention has the following advantages:

[0028] This invention solves the problems of high operational risk, low efficiency, and poor accuracy. It fully mines and extracts the three-dimensional information of the damaged areas of buildings and their components (including small-scale damage information such as splashed gravel), comprehensively obtains detailed point cloud information of buildings and structures, and efficiently and accurately constructs three-dimensional point cloud models of buildings and structures. Using the difference, segmentation, and feature description methods proposed in this invention, it objectively and accurately describes the damage traces of buildings and structures and provides corresponding description functions. It can provide a more reliable data foundation for quantitative analysis of the degree of damage to buildings and structures, and is suitable for wide-ranging promotion and application. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of multi-platform point cloud acquisition and model construction of buildings and structures in this invention;

[0030] Figure 2 This is a schematic diagram of damage-related subtraction and augmentation in this invention;

[0031] Figure 3 This is the coordinate system for the statistical distribution of damage-related augmentation in this invention;

[0032] Figure 4 This is a schematic diagram showing the statistical range distribution of the outer envelope and the volume of the damaged tissue in this invention;

[0033] Figure 5 This is the statistical distribution of volumetric weights (distance and volume coordinate system) in this invention. Detailed Implementation

[0034] The present invention can be explained in more detail through the following embodiments, but the present invention is not limited to the following embodiments;

[0035] In this invention, Figure 1 This diagram illustrates the multi-platform point cloud acquisition and model construction for buildings. ① represents various 3D LiDAR platforms, which are typically selected based on the geometric distribution of the target building. Commonly used platforms include UAV platforms, fixed ground platforms, and mobile platforms, as shown in the diagram. ② primarily illustrates walls, using a simplified wall model to represent the 3D point cloud and the model built upon it. Since the registered point cloud resides in a unified coordinate system, the wall's thickness information can be depicted through the point cloud model. ③ illustrates ground point clouds. ④ represents the unified coordinate system after multi-platform point cloud registration.

[0036] Figure 2The diagram illustrates the reduction and addition of structures related to damage. It is mainly used to show the gaps in the wall caused by the explosion, as well as the scattered fragments caused by the wall damage. The part missing from the original building target is called the reduction, as shown in Figure ⑤. The part added to the original geometric structure by the explosion is called the addition, as shown in Figure ⑥. The additions caused by the explosion on the inner and outer sides of the wall are not connected in space and are divided by the wall.

[0037] Figure 3 The coordinate system for the statistical distribution of damage-related growths is mainly based on one side of the wall to illustrate the constructed statistical distribution of growths. As shown in Figure ⑦, the statistical coordinate system of growths is based on the projection point O' of the damage center on the XY plane as the origin, the distance of the growth from the origin as the horizontal axis, and the volume of the growth as the vertical axis. The growths on both sides of the wall are statistically analyzed, and the dashed semicircles are the corresponding equidistant lines.

[0038] Figure 4 The diagram shows the statistical distribution of the outer envelope and volume of the damaged wall, where the outer envelope cuboid of the damaged wall is shown in Figure ⑧.

[0039] Figure 5 The figure shows the statistical distribution of the increased volume (distance and volume coordinate system). In the figure, ⑨ represents the distribution of increased volume on both sides of the wall, and the vertical axis represents the cumulative volume at the same distance.

[0040] 3D laser scanning technology, characterized by its speed, accuracy, and high degree of automation, is an efficient and accurate method for acquiring 3D point cloud models of buildings and structures. It provides crucial methodological and technical support for the accurate and efficient acquisition of 3D building models. Because this technology relies on the laser emitted by the lidar itself reflecting off the target object to acquire 3D coordinate data, it does not depend on external light sources and can perform high-precision 3D point cloud observation and modeling in areas without light illumination. Compared to traditional image-based damage assessment methods, lidar offers advantages such as high point cloud density, high measurement accuracy, and multi-platform compatibility. It eliminates the need for personnel to enter the damaged interior of buildings and structures, allowing for comprehensive acquisition of detailed point cloud information. It enables all-day measurement of damaged areas, significantly improving the efficiency and reliability of damage feature extraction. Based on the building point cloud data, high-precision 3D models of buildings and structures in different scenarios can be created, accurately depicting the different structural components of buildings. Compared to traditional methods, point cloud-based building feature depiction capabilities are stronger, ensuring the scientific rigor and rationality of damage assessment. Therefore, by mining the three-dimensional geometric feature information related to the damage based on the three-dimensional point cloud model of the building before and after the damage, and constructing a reasonable evaluation model, we can provide more accurate data support for the assessment of building damage and the evaluation of weapon damage effectiveness, and further enrich the methodology for the extraction and application of building damage features.

[0041] Leveraging the advantages of lidar scanning—large scanning area, long working distance, high data acquisition accuracy, and high point cloud density—this invention utilizes point cloud filtering and smoothing, point cloud segmentation, and point cloud-based 3D modeling to establish a high-precision 3D model of a building before and after damage. Based on this model, the proposed differential, segmentation, and feature description methods enable accurate extraction and feature analysis of building damage traces. Compared to traditional expert-led on-site damage assessment, this method is more efficient, accurate, and reliable. Furthermore, the accurate 3D information provides a more comprehensive and layered analysis than traditional image-based methods, making the content more measurable. The 3D spatial distribution characteristics of the damage features and corresponding volume measurements objectively and accurately describe the damage traces and provide corresponding descriptive functions, offering a more reliable data foundation for quantitative analysis of the extent of building damage.

[0042] Combined with appendix Figures 1-5 The present invention discloses a method for measuring damage traces of buildings and structures based on laser point cloud difference. The measurement method specifically includes the following steps:

[0043] Step 1, such as Figure 1 As shown, firstly, based on the geometric structure, size, and surrounding environment of the buildings, a three-dimensional lidar platform is constructed and corresponding base station locations are selected to form a multi-platform three-dimensional monitoring of the buildings and obtain high-precision point cloud data of the buildings in the target area.

[0044] In specific implementation, when constructing a 3D lidar platform, based on the internal and external geometric characteristics of the current building, a UAV lidar platform P1, a ground lidar platform P2, and a mobile lidar platform P3 are selected to acquire point cloud datasets D1 (upper part of the building), D2 (exterior facade), and D3 (internal structure), respectively. Each point cloud dataset consists of one or more subsets, and each point cloud subset corresponds to one observation result.

[0045] Furthermore, when selecting appropriate base station locations, in order to reduce the workload of subsequent point cloud data registration, it is necessary to select appropriate observation stations based on the geometric characteristics of the building structure, thereby reducing the number of station relocations. At the same time, it is necessary to ensure that the point cloud dataset has a certain degree of repeatability during each observation, thereby improving the accuracy and reliability of point cloud dataset registration.

[0046] Furthermore, when conducting multi-platform three-dimensional monitoring of buildings and structures, a multi-platform three-dimensional monitoring method is developed that combines indoor and outdoor monitoring of buildings and structures, and is further assisted by the UAV lidar platform P1. This method comprehensively acquires point cloud data of the geometric structure information of buildings and structures in the target area. At the same time, it is necessary to ensure that the point cloud data of different platforms have a certain degree of redundancy, so as to lay the foundation for subsequent stitching of complete point cloud data of the target area.

[0047] The second step is to perform data preprocessing on the 3D point clouds acquired by LiDAR from different platforms. This includes outlier removal, noise removal, adaptive filtering of 3D point clouds, point cloud downsampling, feature line and surface extraction, to reduce the impact of environmental noise and system noise and lay a high-quality data foundation for the subsequent registration of point cloud data from different platforms.

[0048] In practice, when preprocessing 3D laser point cloud data, since the density and accuracy of point cloud data acquired by different platform lidars are different, attention should be paid to selecting appropriate parameters in a differentiated manner during point cloud data preprocessing in order to achieve better filtering effect.

[0049] The third step is to perform high-precision registration of point cloud data from multiple platforms based on overlapping regions.

[0050] In practice, when performing high-precision registration of point cloud data from multiple platforms based on overlapping regions, the collaboration of multiple platforms can leverage the advantages of different platforms and overcome the shortcomings of traditional single platforms that only acquire point clouds of the exterior or interior of buildings. In order to unify the point cloud data from different platforms into the same coordinate system, the point cloud features of overlapping parts of different platforms and optimal methods such as weighted least squares are used to achieve high-precision registration and coordinate system unification of point cloud data from different platforms, thereby obtaining globally fused point cloud data that reflects the internal and external geometric features of buildings.

[0051] During implementation, since the external and internal structures of buildings are generally quite complex, it is difficult to obtain complete point cloud information of the corresponding target by relying on a single observation station. Moreover, the initial reference of the three-dimensional point cloud data from multiple platforms and observation stations is different. In order for the point cloud data to accurately reflect the geometric features of the target, it is necessary to obtain the corresponding rotation matrix Rij and offset matrix Tij based on the point cloud data of the repeated areas observed each time. i and j represent the point cloud data with common parts obtained at stations i and j. In order to further improve the accuracy and reliability of registration, the constraints of the corresponding point cloud line and surface features are also added to the final objective function min||Rij*Di-Tij-Dj||+line or surface feature coincidence constraint to obtain the coordinate transformation parameters between different point clouds, thereby realizing the transformation registration of point clouds and obtaining a complete point cloud that reflects the geometric features of the building.

[0052] The fourth step is to reconstruct the three-dimensional model of the building based on the fused complete three-dimensional point cloud data, and to use the geometric morphology information and regular repetitive features of the reconstructed model to segment and assign weights to the main structure of the building to characterize the importance of each segmented component.

[0053] In practice, the components to be divided include walls, floors, beams, support beams or internal objects, with load-bearing components that are related to structural stability having the highest weight.

[0054] Step 5: Obtain the 3D point cloud model and segmentation model of the building before damage. Repeat steps 1 to 4 above to complete the 3D point cloud model and segmentation results after damage. Then, based on the line and surface features (reference points) of the undamaged area and the point cloud distance minimization optimization method, register the 3D point cloud data and segmentation model of the building before and after damage with high precision.

[0055] Step 6, as Figure 2 As shown, based on the registered three-dimensional point cloud and segmented model of the building before and after damage, differential processing of the three-dimensional point cloud and segmented model is carried out under the same three-dimensional coordinate system. Based on the differential results, the volume change of the three-dimensional model of the internal and external structure of the building before and after damage is quantitatively calculated. By setting the minimum effective usable damage volume threshold, the difference volume of the three-dimensional model after damage is obtained, and the difference volume is segmented. Based on the volume change trend, its attribute is determined to be either an increase or a decrease.

[0056] In practice, the total volume of the increased volume and the volume of the decreased volume should be basically the same, but their spatial distributions should be different, with the increased volume being relatively dispersed and the decreased volume being relatively concentrated.

[0057] Step 7, as Figure 3 , 4 As shown in Figure 5, based on the three-dimensional spatial distribution and volume information of the differential bodies, the cuboid envelope and Weil distribution function are used to describe the characteristic parameters of the subtractive and augmented bodies respectively. The maximum likelihood optimization estimation method is used to obtain the cuboid parameters corresponding to the augmented body and the asymmetric Gaussian function parameters corresponding to the subtractive body respectively.

[0058] In practical implementation, the parameters describing the outer envelope cuboid of the subtractive body mainly include: the maximum and minimum values ​​of the subtractive body on the XYZ axes, thus forming the corresponding cuboid. Taking a wall damage gap perpendicular to the X-axis as an example (e.g.) Figure 4 As shown), the corresponding cuboid for the subtractive volume is the yellow area. At the point perpendicular to the X-axis and intersecting at point Xi, its area of ​​influence is the envelope rectangle (Zmin, Zmax, Ymin, Ymax) in the ZY plane. The thickness of the wall is the thickness (H) of this corresponding cuboid, thus forming the corresponding subtractive envelope cuboid (Xi, Ymin, Zmin)(Xi, Ymin, Zmax)(Xi, Ymax, Zmin)(Xi, Ymax, Zmax)(Xi-H, Ymin, Zmin)(Xi-H, Ymin, Zmax)(Xi-H, Ymax, Zmax). The other corresponding subtractive outer envelope cuboids are obtained using the same method. However, for the description of the augmented volume, due to its significant asymmetry, the Weil distribution is used to describe the augmented volume (e.g., Figure 5 As shown), its corresponding formula function is as follows:

[0059]

[0060] Where β is the shape parameter of the volume distribution of the growth along the distance, η is the corresponding scaling factor, and γ is the position parameter, which is generally taken as 0. It can be seen that with the Xc and Yc coordinates of the damage center as the center, the distance of the growth from that point on the horizontal axis, and the volume of the growth on the vertical axis, the volume distribution of the growth in a certain area can be statistically analyzed. The Weil distribution function and its corresponding parameters can be used to describe its distribution characteristics.

[0061] Step 8: Based on the cuboid parameters (reduced volume) and Weil distribution (increased volume) function and their characteristic parameters, and combined with the weighted volume of the increased and reduced volumes, the corresponding damage to the building is calculated in a comprehensive manner, so as to accurately reflect the damage effect brought about by the explosion.

[0062] In practical implementation, for cuboid parameters (subtraction), the total damaged volume Vneg is calculated by weighting the weights and volumes of the components to which they belong, and the volume Vs of the cuboid can be calculated by using the outer envelope of the cuboid. For Weil distribution (addition), the cumulative value of the added volume Vpos of the corresponding function is calculated based on the corresponding Weil distribution parameters, and the total added volume Vz can also be calculated based on the added volume segmentation results. Based on these parameters and information, high-precision and refined measurement and analysis of damage can be performed.

[0063] In practical implementation, this invention can use a constructed mask file to isolate the impact of the explosion-damaged area on the registration of point clouds and models of buildings before and after the explosion. It can utilize only the point cloud and line surface features of areas less affected or unaffected by the damage, i.e., the stable features of overlapping areas, to construct an objective function similar to step 3. The rotation and offset matrices Rqh and Tqh are then optimized and calculated, thereby achieving the registration of multi-stage three-dimensional point clouds and models of buildings before and after the damage.

[0064] Furthermore, based on the registered three-dimensional models of the building before and after the damage, model difference processing is performed in a three-dimensional coordinate system to quantitatively calculate the changes in the three-dimensional models of the building's internal and external structures before and after the damage. By setting a minimum effective usable damage volume threshold to avoid the influence of measurement errors, the difference in the three-dimensional model caused by the damage is obtained. Based on the location and change trend, the difference is segmented, and the corresponding increase and decrease are determined. Generally speaking, the total volume of the two should be basically the same, but the distribution is different.

[0065] Furthermore, to further clarify the attributes of the augmented and subtracted volumes obtained after differential processing, based on the point cloud segmentation model, the geometric morphology information and repeatability characteristics are used to segment and assign weights to the main structural components of the building, characterizing the importance of each component, and supplementing the augmented and subtracted volumes with weight information of the corresponding components. The settings of the above parameters need to be adjusted in a timely manner according to the different types and uses of the building structures. Generally speaking, in applications primarily focused on identifying structural damage traces, load-bearing structural components have higher weights, floor partitions have lower weights, general walls have lower weights, and internal facilities of the building have the lowest weights. If the criteria for judging the damage effect differ, the weight parameters of the corresponding components need to be adjusted appropriately.

[0066] Next, based on the three-dimensional spatial distribution and volume information of the differential bodies, and using the weight information obtained in the previous step, the maximum likelihood method is used to calculate the outer cuboid parameters of the subtractive body based on the cuboid and the asymmetric Gaussian function (Weil function), respectively.

[0067] Furthermore, for the augmented portion, a distribution histogram is plotted with the projection of the main subtractive center onto the plane as the center o', the distance R of the augmented portion from the center as the horizontal axis, and the weighted volume of the augmented portion as the vertical axis. If the augmented portion is divided by walls or other structures, a histogram is plotted for each segmented region using the above method, with the ground projection O' of the explosion center as the center, the distance between the augmented portion and O' as the horizontal axis, and the volume of the corresponding augmented portion as the vertical axis. Then, the Weil distribution function and the maximum likelihood method are used to estimate the corresponding model parameters.

[0068] Furthermore, based on the outer envelope cuboid (subtractive volume) and Weil distribution function (increased volume) and their characteristic parameters, high-precision monitoring of building damage traces is carried out, thereby quantitatively measuring the damage effect of explosions on buildings.

[0069] Compared with traditional damage assessment methods, this invention solves the problems of high operational risk, low efficiency, and poor accuracy. It fully mines and extracts the three-dimensional information of the damaged areas of buildings and their components (including small-scale damage information such as splashed debris), comprehensively obtains detailed point cloud information of buildings and structures, and efficiently and accurately constructs a three-dimensional point cloud model of buildings and structures. Using the difference, segmentation, and feature description methods proposed in this invention, it objectively and accurately describes the damage traces of buildings and structures and provides corresponding description functions, which can provide a more reliable data foundation for quantitative analysis of building damage and damage effectiveness assessment.

[0070] The parts of this invention not described in detail are prior art.

[0071] The embodiments selected herein for the purpose of disclosing the inventive objectives are currently considered suitable; however, it should be understood that the invention is intended to include all variations and modifications of the embodiments that fall within the scope of this concept and invention.

Claims

1. A method for measuring damage traces of buildings and structures based on laser point cloud difference, characterized in that: The measurement method specifically includes the following steps: The first step is to construct a three-dimensional lidar platform and select appropriate base station locations based on the geometric structure, size, and surrounding environment of the buildings and structures, thereby enabling multi-platform three-dimensional monitoring of the buildings and structures and acquiring high-precision point cloud data of the buildings and structures in the target area. The second step is to perform data preprocessing on the 3D point clouds acquired by LiDAR from different platforms, specifically including outlier removal, noise point removal, 3D point cloud adaptive filtering, point cloud downsampling, and feature line and surface extraction. The third step is to perform high-precision registration of point cloud data from multiple platforms based on overlapping regions. The fourth step is to reconstruct the three-dimensional model of the building based on the fused complete three-dimensional point cloud data, and to use the geometric morphology information and regular repetitive features of the reconstructed model to segment and assign weights to the main structure of the building to characterize the importance of each segmented component. Step 5: Obtain the 3D point cloud model and segmentation model of the building before damage. Repeat the above steps 1 to 4 to complete the 3D point cloud model and segmentation results after damage. Then, based on the line and surface features of the undamaged area and the point cloud distance minimization optimization method, register the 3D point cloud data and segmentation model of the building before and after damage with high precision. Step 6: Based on the registered 3D point cloud and segmented model of the building before and after damage, perform differential processing of the 3D point cloud and segmented model under the same 3D coordinate system. Quantitatively calculate the volume change of the 3D model of the building's internal and external structures before and after damage based on the differential results. By setting the minimum effective usable damage volume threshold, obtain the 3D model difference volume after damage, and segment the difference volume. Determine whether the volume change trend indicates an increase or decrease in volume. Step 7: Based on the three-dimensional spatial distribution and volume information of the differential bodies, the characteristic parameters of the subtractive and augmenting bodies are described by using the cuboid envelope and Weil distribution function, respectively. The maximum likelihood optimization estimation method is then used to obtain the cuboid parameters corresponding to the augmenting body and the asymmetric Gaussian function parameters corresponding to the subtractive body. Step 8: Based on the cuboid parameters and Weil distribution function and its characteristic parameters, and combined with the weighted volume of the added and subtracted volumes, calculate the corresponding damage to the building structure to accurately reflect the damage effect of the explosion.

2. The method for measuring building damage traces based on laser point cloud difference according to claim 1, characterized in that: In the first step, when constructing the three-dimensional lidar platform, based on the internal and external geometric features of the current building, a UAV lidar platform P1, a ground lidar platform P2, and a mobile lidar platform P3 are selected to acquire point cloud datasets D1 (upper part of the building), D2 (exterior facade), and D3 (internal structure), respectively. Each point cloud dataset consists of one or more subsets, and each point cloud subset corresponds to one observation result.

3. The method for measuring building damage traces based on laser point cloud difference according to claim 1, characterized in that: When selecting the appropriate base station location in the first step, in order to reduce the workload of subsequent point cloud data registration, it is necessary to select the appropriate observation station according to the geometric characteristics of the building structure, thereby reducing the number of station relocations. At the same time, it is necessary to ensure that the point cloud dataset has a certain degree of repeatability during each observation, thereby improving the accuracy and reliability of point cloud dataset registration.

4. The method for measuring building damage traces based on laser point cloud difference according to claim 1, characterized in that: In the first step, when conducting multi-platform three-dimensional monitoring of buildings and structures, a multi-platform three-dimensional monitoring method is formed that combines indoor and outdoor monitoring of buildings and structures, and is further assisted by the UAV lidar platform P1. This method comprehensively acquires point cloud data of the geometric structure information of buildings and structures in the target area. At the same time, it is necessary to ensure that the point cloud data of different platforms have a certain degree of redundancy, so as to provide a foundation for subsequent stitching of complete point cloud data of the target area.

5. The method for measuring building damage traces based on laser point cloud difference according to claim 1, characterized in that: In the second step, when preprocessing the 3D laser point cloud data, since the density and accuracy of the point cloud data acquired by different platform lidars are different, attention should be paid to selecting the corresponding parameters differently during the point cloud data preprocessing in order to achieve a better filtering effect.

6. The method for measuring building damage traces based on laser point cloud difference according to claim 1, characterized in that: In the third step, when performing high-precision registration of multi-platform point cloud data based on overlapping regions, the combination of multiple platforms can leverage the advantages of different platforms, overcoming the shortcomings of traditional single platforms that only acquire point clouds of the exterior or interior of buildings. In order to unify the point cloud data of different platforms into the same coordinate system, the point cloud features of overlapping parts of different platforms and the weighted least squares optimal method are used to achieve high-precision registration and coordinate system unification of point cloud data from different platforms, thereby obtaining globally fused point cloud data that reflects the internal and external geometric features of buildings.

7. The method for measuring building damage traces based on laser point cloud difference according to claim 1, characterized in that: In the fourth step, the components to be divided include walls, floors, beams, support beams, or internal objects, with load-bearing components that are related to structural stability having the highest weight.

8. The method for measuring building damage traces based on laser point cloud difference according to claim 1, characterized in that: In the sixth step, the total volume of the increased volume and the volume of the decreased volume should be basically the same, but their spatial distribution is different. The increased volume is relatively dispersed, while the decreased volume is relatively concentrated.

9. The method for measuring building damage traces based on laser point cloud difference according to claim 1, characterized in that: In the seventh step, the parameters describing the outer envelope cuboid of the subtractive body include: the subtractive body is composed of the maximum and minimum values ​​of the point cloud coordinates on the XYZ axes, thus forming the corresponding cuboid; as for the description of the augmenting body, due to its significant asymmetry, the Weil distribution is used to describe the augmenting body, and its corresponding formula function is as follows. ； in The shape parameters of the volume distribution along the distance direction of the augmentation are given. For the corresponding scaling factor, The location parameter is usually set to 0. It can be seen that, with the Xc and Yc coordinates of the damage center as the center, the distance of the augmented body from that point on the horizontal axis and the volume of the augmented body on the vertical axis, the volume distribution of the augmented body in a certain area is statistically analyzed. The Weil distribution function and its corresponding parameters are used to describe its distribution characteristics.

10. The method for measuring building damage traces based on laser point cloud difference according to claim 1, characterized in that: In the eighth step, for the cuboid parameters, the total damaged volume Vneg is calculated by weighting the weights and volumes of the components to which they belong. At the same time, the volume Vs of the cuboid can be calculated by using the outer envelope of the cuboid. For the Weil distribution, the cumulative value of the increased volume Vpos of the corresponding function is calculated based on the corresponding Weil distribution parameters. At the same time, the total increased volume Vz can be calculated based on the increased volume segmentation results. Based on these parameters and information, the damage is measured and analyzed with high precision.

Images