Indoor-outdoor fusion modeling method, device and medium based on orthogonal trihedral corner target

Abstract

The application discloses an indoor and outdoor fusion modeling method and device based on a quadrantal trihedral corner target and a medium, and belongs to the field of multi-source data fusion and three-dimensional modeling. The method comprises the following steps: after the quadrantal trihedral corner target is arranged at a specified position of a target building, obtaining the unmanned aerial vehicle outdoor oblique photo data and indoor point cloud original data of the target building respectively; for the indoor point cloud original data, filtering is performed on the high point cloud density area, outlier deletion is performed on the low point cloud density area, and secondary processing is performed to obtain indoor point cloud optimization data with uniform point cloud density; the unmanned aerial vehicle outdoor oblique photo data and the indoor point cloud optimization data of the indoor and outdoor overlapping area are fused and registered by using a feature-region two-stage registration method based on the target, multi-source fusion data is obtained, and an indoor and outdoor integrated model of the target building is generated by reverse modeling. The application improves the registration accuracy, stability and adaptability of indoor and outdoor fusion modeling.

Description

Technical Field

[0001] This invention belongs to the field of multi-source data fusion and 3D modeling, and more specifically, relates to an indoor and outdoor fusion modeling method, device and medium based on orthogonal trihedral targets. Background Technology

[0002] In 3D building modeling, high-quality digital fusion modeling relies on high-precision raw data and efficient modeling procedures. Traditional modeling methods typically involve collecting data step-by-step through outdoor UAV oblique photography and indoor static laser scanning, followed by manual rough selection of corresponding points for ground features for data registration, and finally modeling in 3D modeling software based on computer-aided design (CAD) drawings. This approach suffers from poor fusion accuracy, low modeling efficiency, and an inability to organically combine the two processes into a single procedure.

[0003] In indoor-outdoor collaborative monitoring systems, various monitoring sensors belong to different observation platforms and environments. When collecting data independently, the lack of a unified and universal geometric constraint benchmark easily leads to fusion problems such as spatial misalignment of multi-source point clouds and insufficient registration accuracy. Although traditional target spheres have certain versatility, they have core defects: First, registration relies on spherical fitting, which is significantly affected by point cloud density, occlusion, and viewing angle, resulting in poor accuracy stability. Second, during indoor-outdoor collaborative acquisition and fusion, the ground target sphere point cloud is sparse due to the limited altitude of the UAV, making it impossible to form complete spherical features and failing to meet the convergence requirement of the Iterative Closest Point (ICP) algorithm for sufficiently overlapping feature points, causing the existing ICP registration method to fail. Third, in the face of indoor and underground feature degradation scenarios, three non-collinear target spheres are required to solve spatial transformations, and the deployment of a large number of turning points increases the workload, making it difficult and costly. Fourth, the target spheres are small and the edges are not sharp enough, resulting in poor recognition as underground features and insufficient adaptability for real-time localization and mapping (SLAM) in underground long-distance corridor scenarios.

[0004] In addition, during indoor static laser scanning, the actual point cloud data acquired may suffer from uneven density distribution due to factors such as scanning distance, scanning angle, laser incident angle, and the reflectivity of the building's interior surfaces. This issue can affect the accuracy and reliability of subsequent registration. Existing point cloud preprocessing methods based on statistical filtering or voxel downsampling struggle to balance denoising and density equalization requirements, resulting in insufficient accuracy and poor stability in indoor-outdoor fusion modeling. Summary of the Invention

[0005] The main objective of this invention is to provide an indoor-outdoor fusion modeling method, device, and medium based on orthogonal trihedral targets, so as to overcome the shortcomings of the prior art.

[0006] To achieve the above-mentioned objectives, the present invention adopts the following technical solution: The first aspect of this invention provides an indoor-outdoor fusion modeling method based on orthogonal trihedral targets. The method includes: deploying orthogonal trihedral targets at a designated location of a target building, wherein the designated location includes at least an indoor-outdoor overlapping area, and the orthogonal trihedral targets include three mutually perpendicular planes intersecting at the same vertex; acquiring UAV outdoor oblique photograph data and indoor point cloud raw data for the target building based on the orthogonal trihedral targets; filtering regions with density higher than the uniform density range of the indoor point cloud for the indoor point cloud raw data, deleting outliers for regions with density lower than the uniform density range of the indoor point cloud, and performing secondary processing on the processed indoor point cloud raw data to obtain optimized indoor point cloud data with uniform point cloud density; fusing and registering the UAV outdoor oblique photograph data and the optimized indoor point cloud data in the indoor-outdoor overlapping area using a target-based feature-region two-level registration method to obtain multi-source fusion data; and performing reverse modeling on the multi-source fusion data to generate an integrated indoor-outdoor model of the target building.

[0007] Preferably, the process of determining the uniform density range of the indoor point cloud includes: calculating the local point cloud density of each point in the original indoor point cloud data; statistically analyzing the local point cloud density of all points to calculate the global average density and density standard deviation; and determining the uniform density range of the indoor point cloud based on the global average density and the density standard deviation.

[0008] Preferably, the uniform density range of the indoor point cloud is: ; ; in, This represents the lower limit of the range of uniform density of the indoor point cloud. This represents the upper limit of the uniform density range of the indoor point cloud. The global average density, The density standard deviation is... The first adjustment coefficient, This is the second adjustment coefficient.

[0009] Preferably, filtering is performed on regions with densities higher than the uniform density range of indoor point clouds. Specifically, this includes: for regions with densities higher than the uniform density range of indoor point clouds, determining the filtering radius and filtering threshold based on the local density of the region; filtering the region based on the filtering radius and the filtering threshold to retain points with curvature greater than the curvature threshold or normal vector changes greater than the change threshold, and removing redundant points.

[0010] Preferably, the filtering radius and the filtering threshold are respectively: ; ; in, For point clouds The filtering radius of the neighborhood. For point clouds The filtering threshold of the neighborhood For point clouds The radius of the neighborhood, For point clouds Local density of the neighborhood For point clouds Point cloud to point cloud within the neighborhood The mean distance, For point clouds Point cloud to point cloud within the neighborhood The standard deviation of the distance, It is a multiplier. This represents the lower limit of the range of uniform density of the indoor point cloud. This represents the upper limit of the uniform density range of the indoor point cloud.

[0011] Preferably, the processed indoor point cloud raw data undergoes secondary processing, specifically including: weighted sampling of the processed indoor point cloud raw data to obtain the optimized indoor point cloud data; wherein, the resampling weight is: ; in, For point clouds resampling weights, This represents the upper limit of the uniform density range of the indoor point cloud. For point clouds Local density of the neighborhood For global average density, For density standard deviation, This is the smoothing coefficient.

[0012] Preferably, the three planes of the orthogonal trihedral target are all made of square plates, and the square plates are coated with a matte coating in a checkerboard pattern.

[0013] Preferably, the fusion and registration of UAV outdoor oblique photograph data and indoor point cloud optimization data in the indoor-outdoor overlapping area specifically includes: obtaining the coordinates of multiple common points in the indoor-outdoor overlapping area where orthogonal trihedral targets are deployed; calculating the coordinate transformation parameters between the UAV outdoor oblique photograph data and the indoor point cloud optimization data based on the common point coordinates; performing coarse registration of the UAV outdoor oblique photograph data and the indoor point cloud optimization data based on the coordinate transformation parameters; and performing multi-source point cloud registration of the coarsely registered UAV outdoor oblique photograph data and the indoor point cloud optimization data with the objective of minimizing the sum of the Euclidean distances between corresponding point cloud pairs of the coarsely registered UAV outdoor oblique photograph data and the indoor point cloud optimization data.

[0014] A second aspect of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the indoor-outdoor fusion modeling method based on orthogonal trihedral targets as described above.

[0015] A third aspect of the present invention provides a non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute the indoor-outdoor fusion modeling method based on orthogonal trihedral targets as described above.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention provides an indoor and outdoor fusion modeling method, device and medium based on orthogonal trihedral targets. On the one hand, orthogonal trihedral targets are deployed at designated locations on the target building for subsequent indoor and outdoor data acquisition. Orthogonal trihedral targets avoid the inherent error of spherical fitting, and the robustness of edge features ensures the accuracy of feature localization under complex acquisition conditions, greatly reducing registration deviation. It also solves the problem of ICP registration failure under high-altitude acquisition by UAVs. Furthermore, on the plane, there is no need to consider the principle of "not placed at the same height". Three targets of different sizes can complete the establishment of the registration benchmark, avoiding the problem of difficult deployment faced by spherical targets. It is especially suitable for narrow feature degradation scenarios such as underground long-distance corridors. On the other hand, adaptive density point cloud denoising is performed on the original indoor point cloud data to make the overall point cloud density of the indoor wall surface tend to be consistent, further improving the subsequent registration accuracy and stability. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 The flowchart illustrates the indoor-outdoor fusion modeling method based on orthogonal trihedral targets provided in this embodiment of the invention.

[0019] Figure 2 The flowchart of feature-region two-level registration based on the target is provided for both outdoor and indoor applications in this embodiment of the invention.

[0020] Figure 3 This is a schematic diagram of the structure of an orthogonal trihedral target provided in an embodiment of the present invention.

[0021] Figure 4 This is a process diagram of UAV oblique photogrammetry and data acquisition provided in an embodiment of the present invention.

[0022] Figure 5 This is a block diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0023] In view of the shortcomings of the prior art, the inventors of this invention, through long-term research and extensive practice, have proposed the technical solution of this invention. The following will further explain and illustrate this technical solution, its implementation process, and its principles.

[0024] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0025] Furthermore, in the description of this invention, it should be understood that the terms "upper," "lower," "inner," "outer," "horizontal," "vertical," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0026] In the description of this specification, the references to terms such as "an embodiment," "a particular embodiment," or "the embodiment" indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0027] Figure 1 This is a flowchart illustrating an indoor / outdoor fusion modeling method based on an orthogonal trihedral target, provided in an embodiment of the present invention. (See attached document.) Figure 1 , combined Figures 2-4 The present invention provides a detailed description of the indoor and outdoor fusion modeling method based on orthogonal trihedral targets provided in the embodiments of the present invention, the method including operation S100-operation S500.

[0028] Operate S100 to place an orthogonal trihedral target at a designated location on the target building. The designated location includes at least the overlapping area between indoor and outdoor areas. The orthogonal trihedral target includes three mutually perpendicular planes that intersect at the same vertex.

[0029] The S200 was operated to acquire outdoor tilted photographs and indoor point cloud data of the target building based on an orthogonal trihedral target.

[0030] The S300 is used to filter regions with densities higher than the uniform density range of indoor point clouds and remove outliers from regions with densities lower than the uniform density range of indoor point clouds. The processed indoor point cloud data is then processed a second time to obtain optimized indoor point cloud data with uniform point cloud density.

[0031] The S400 is operated, and a target-based feature-region two-level registration method is used to fuse and register UAV outdoor oblique photo data and indoor point cloud optimization data in overlapping indoor and outdoor areas to obtain multi-source fused data.

[0032] Operate the S500 to perform reverse modeling on multi-source fusion data and generate an integrated indoor and outdoor model of the target building.

[0033] The indoor-outdoor fusion modeling method based on orthogonal trihedral targets provided in this invention involves collaborative acquisition of multi-source indoor and outdoor data using orthogonal trihedral targets, adaptive density point cloud denoising of the raw indoor point cloud data to achieve a more uniform point cloud density across the indoor walls, and then high-precision fusion of the relative coordinate systems of the multi-source data through registration based on the orthogonal trihedral target region for reverse modeling. This improves the fidelity and efficiency of digital white modeling and offers greater engineering adaptability. The overall process is as follows: Figure 2 As shown.

[0034] In this embodiment of the invention, the orthogonal trihedral targets in the overlapping indoor and outdoor areas do not need to be deliberately laid out according to the principle of non-collinearity and unequal height (that is, the orthogonal trihedral targets in the overlapping indoor and outdoor areas automatically meet the principle of non-collinearity and unequal height due to the different sizes of the devices), and there is no need for dense laying or repeated laying, and the position is flexible.

[0035] In a preferred embodiment, all three planes of the orthogonal trihedral target are made of square plates, and the square plates are coated with a matte coating in a checkerboard pattern, such as... Figure 3 As shown.

[0036] The orthogonal trihedral target provided in this embodiment of the invention adopts a matching trihedral structure of different sizes, for example, made of deformation-resistant white polycarbonate material and L-shaped iron sheet, with a matte surface design, and has the following advantages.

[0037] (1) Abandoning the spherical fitting logic, the system adopts a dual feature of "vertex and edge". The vertex is the unique intersection of three orthogonal planes, and the linear edge feature has strong resistance to occlusion and low point cloud density, which greatly improves the stability of feature localization. This structural feature avoids the inherent error of spherical fitting, and the robustness of the edge feature ensures the accuracy of feature localization under complex acquisition conditions, and the registration deviation is greatly reduced.

[0038] (2) The large size of the orthogonal trihedral target can still form a clear point cloud under highly sparse point cloud conditions, providing sufficient matching point pairs, satisfying the convergence condition of the ICP algorithm, and realizing effective registration between the UAV and indoor and outdoor point clouds. This structural feature solves the problem of ICP registration failure under high-altitude acquisition by UAV, and is compatible with multiple types of equipment such as lidar and visual sensors, realizing full-scene fusion of indoor and outdoor, above-ground and underground.

[0039] (3) The spatial transformation parameters can be calculated by combining three targets of different sizes. On a plane, there is no need to consider the principle of "not being placed at the same height". On an inclined plane, it is easy to arrange. There is no need to worry about the risk of the ball rolling down. The difficulty of setting up is less than that of the target ball, which significantly reduces the difficulty of operation. The sharp edge features improve the recognizability of underground scenes and provide stable and unique matching points for SLAM algorithm, reducing the floating of point cloud recognition. It is suitable for modeling long-distance corridor scenes, especially for narrow feature degradation scenes such as underground long-distance corridors.

[0040] In the operation of S200, for example, UAV oblique photogrammetry technology and 3D laser scanning technology are used to acquire outdoor and indoor multi-source 3D data respectively. Through UAV oblique photogrammetry indoor processing and point cloud post-processing, multi-source data such as UAV outdoor oblique photo data and indoor point cloud raw data are generated.

[0041] Specifically, UAV oblique photogrammetry includes, for example, the following steps 1-7, and its overall process is as follows: Figure 4 As shown.

[0042] Step 1: Based on the target area, calculate the UAV's flight altitude, heading overlap, lateral overlap, and oblique photography angle parameters to generate a multi-path flight mission.

[0043] Step 2: Set up image control points within the target area and obtain the precise three-dimensional coordinates of the image control points.

[0044] Step 3: Control the drone equipped with a five-lens oblique photography system to simultaneously collect ground image data from five different perspectives: vertical, forward, backward, left, and right, ensuring that there is a preset overlap between adjacent images.

[0045] Step 4: Based on the acquired multi-view image data and image control point coordinates, obtain the exterior orientation elements and three-dimensional coordinates of the densification points of the image through bundle adjustment.

[0046] Step 5: Use a multi-view image matching algorithm to generate high-density point cloud data and construct a digital surface model.

[0047] Step 6: Map the original image texture information onto the surface of the 3D model to generate a 3D real-world model with realistic texture.

[0048] Step 7: Perform geometric accuracy, texture quality, and integrity checks on the generated 3D model to ensure it meets the mapping accuracy requirements. The data acquired by the UAV oblique photogrammetry is in OSGB format and needs to be converted to LAS format for fusion and registration with the point cloud data.

[0049] Preferably, the process of acquiring the original indoor point cloud data specifically includes: after acquiring the original point cloud using three-dimensional laser scanning technology, statistical outlier removal and radius filtering are performed on the original point cloud to remove isolated noise points caused by occlusion, reflection or equipment error. Then, a voxel grid filter is used to uniformly downsample the point cloud to reduce the amount of data while maintaining geometric features, thereby obtaining the original indoor point cloud data.

[0050] In a preferred embodiment, between operation S200 and operation S300, it is necessary to perform an operation to determine the uniform density range of indoor point clouds. The determination process includes the following steps 11-13.

[0051] Step 11: Calculate the local point cloud density of each point in the original indoor point cloud data.

[0052] Step 12: Calculate the local point cloud density of all points to determine the global average density and density standard deviation.

[0053] Step 13: Determine the range of uniform density of indoor point clouds based on the global average density and density standard deviation.

[0054] Preferably, in step 11, for each point Using kd-trees or fixed-radius spheres ( ) Query its neighborhood point set Calculate local point cloud density This indicates the number of points per unit volume or unit area. ; In a preferred embodiment, the uniform density range of the indoor point cloud is: ; ; in, This represents the lower limit of the uniform density range of indoor point clouds. This represents the upper limit of the uniform density range of indoor point clouds. For global average density, For density standard deviation, The first adjustment coefficient, This is the second adjustment coefficient. and Used to control the strictness of density equilibrium.

[0055] In a preferred embodiment, operation S300 filters regions with densities higher than the uniform density range of indoor point clouds, specifically including the following sub-operations S310-S320.

[0056] In sub-operation S310, for densities higher than the indoor point cloud uniform density range (i.e., greater than...) The filtering radius and filtering threshold are determined based on the local density of the region.

[0057] In sub-operation S320, the region is filtered according to the filtering radius and filtering threshold to retain points with curvature greater than the curvature threshold or normal vector change higher than the change threshold, and redundant points are removed.

[0058] In a preferred embodiment, the filter radius and the filter threshold are respectively: ; ; in, For point clouds The filtering radius of the neighborhood. For point clouds The filtering threshold of the neighborhood For point clouds The radius of the neighborhood, For point clouds Local density of the neighborhood For point clouds Point cloud to point cloud within the neighborhood The mean distance, For point clouds Point cloud to point cloud within the neighborhood The standard deviation of the distance, It is a multiplier. This represents the lower limit of the uniform density range of indoor point clouds. This represents the upper limit of the uniform density range of indoor point clouds. This formula makes the filtering radius smaller and the threshold more stringent in areas with higher density, thereby effectively diluting overly dense point clouds while prioritizing the preservation of geometric feature points (points with large curvature or abrupt changes in normal vectors).

[0059] Preferably, for densities lower than the indoor point cloud uniform density range (i.e., less than) The area is protected by removing only obvious outliers to avoid accidentally deleting valid wall points due to low density. The outlier detection criteria are: ; in, Point To its neighborhood center point distance, and Let be the mean and standard deviation of the global neighborhood distance. To protect the coefficient, this strategy can effectively prevent the accidental deletion of valid point clouds due to local point cloud sparseness.

[0060] In a preferred embodiment, operation S300 performs secondary processing on the processed indoor point cloud raw data (the density difference of the point cloud after secondary processing is controlled within 15%), specifically including: weighted sampling of the processed indoor point cloud raw data to obtain optimized indoor point cloud data; wherein, the resampling weight is: ; in, For point clouds resampling weights, This represents the upper limit of the uniform density range of indoor point clouds. For point clouds Local density of the neighborhood For global average density, For density standard deviation, This is the smoothing coefficient.

[0061] In the operation of S300, other methods can also be used to perform secondary processing on the processed indoor point cloud raw data. For example, density equalization filtering can be performed on the processed indoor point cloud raw data to make the point cloud density after secondary processing more consistent.

[0062] In a preferred embodiment, operation S400 involves fusing and registering outdoor tilted photo data and indoor point cloud optimization data of the UAV in the overlapping indoor and outdoor areas, specifically including the following sub-operations S410-S430.

[0063] In sub-operation S410, multiple common point coordinates are acquired in the indoor-outdoor overlapping area where orthogonal trihedral targets are deployed, and coordinate transformation parameters between the UAV outdoor tilted photo data and indoor point cloud optimization data are calculated based on the common point coordinates.

[0064] In sub-operation S420, coarse registration is performed on the outdoor tilted photo data and indoor point cloud optimization data of the UAV based on the coordinate transformation parameters.

[0065] In sub-operation S430, multi-source point cloud registration is performed on the coarsely registered UAV outdoor tilted photo data and indoor point cloud optimized data with the goal of minimizing the sum of the Euclidean distances between the corresponding point cloud pairs of the coarsely registered UAV outdoor tilted photo data and indoor point cloud optimized data.

[0066] Preferably, the embodiments of the present invention adopt a registration method of "coarse registration first, then fine registration": coarse registration is based on manual point selection of orthogonal trihedral targets to roughly align the point cloud positions; ICP fine registration, on the basis of coarse registration, takes the local area point cloud containing the target as the object and achieves high-precision fusion of multi-source data by using the local to drive the overall coordinate transformation.

[0067] Specifically, the implementation process of ICP coarse registration is as follows. It is necessary to acquire the coordinates of multiple common points in the common point cloud region containing the target and convert them into corresponding parameters. These parameters can be rotation, translation, or scaling parameters, but determining the rotation parameters is the most important because they directly affect the accuracy of the coordinates. Specifically, assume the original coordinates of laser point cloud a are... The coordinates of image point cloud b are The coordinates of their corresponding points are used to calculate the fusion of the two data sets, thereby achieving coordinate unification. The rotation matrix R and translation matrix T required for the coordinate transformation of the two data sets can be solved using the following process.

[0068] ; ; ; in, For rotation parameters, These are the translation parameters. Using the least squares method, the rotation and translation parameters required for coordinate transformation in the registration of the two datasets can be calculated, thus achieving the coordinate transformation and preparing for improved accuracy in subsequent data processing.

[0069] Specifically, the implementation process of ICP fine registration is as follows. This algorithm iteratively calculates the rotation and translation matrices to continuously bring the source and target point clouds closer together, minimizing the sum of the Euclidean distances between corresponding point pairs to achieve multi-source point cloud registration. Its basic principle is that for two given point clouds... and Using the optimization idea of ​​least squares, we can find the optimal transformation matrices R and t that minimize the following expression: ; The core idea of ​​this algorithm can be viewed as repeating the following two steps until it converges: Step (1) Construct matching point pairs: For each point in P Find the point in Q that is closest to it. : ; Step (2) Solve for the optimal transformation: Under the constraint of the least squares method, the transformation is updated by minimizing the Euclidean distance between corresponding points: ; in, , , R, t, and T are obtained from the k-th iteration, respectively.

[0070] In the operation of S500, based on multi-source fusion data, a high-fidelity digital white model is built through reverse modeling to restore indoor and outdoor scenes as closely as possible, improve the precision of the digital white model, and make it more conducive to the digital application of fusion modeling. It has better engineering practicality and is applicable to different indoor and outdoor fusion modeling scenarios, with higher dynamic adaptability.

[0071] Taking a building within a school as an example, the implementation process of this method is illustrated, including the following steps a-e.

[0072] Step a: UAV oblique photogrammetry and data format processing. A building within the school was selected as the research object. Image control points were set up around the building using red spray paint, and its 2000 national geodetic coordinate system was acquired using the Global Navigation Satellite System (GNSS). Orthogonal trihedral targets were placed in specific areas, such as the building entrance and exit. A UAV equipped with a five-lens camera was used as the acquisition device, with flight parameters set as follows: flight altitude 90m, gimbal angle 45°, forward overlap 60%, and lateral overlap 70%. The acquired basic data source was processed sequentially through aerial triangulation, dense point cloud generation, texture mapping, and quality checks to generate an OSGB outdoor model. Maintaining the relative coordinates, a LAS model was further generated.

[0073] Step b, 3D laser scanning and data processing. Indoor data of the building is collected. First, markers are affixed to the ground according to the pre-designed stand location for subsequent accuracy verification. Then, the instrument is set up near the target paper, and data is collected in five separate steps. Orthogonal trihedral targets are placed between the target papers. The collected data is then denoised and downsampled to generate a LAS point cloud.

[0074] Step c: For the original indoor point cloud data, filter the regions with density higher than the uniform density range of the indoor point cloud, remove outliers in the regions with density lower than the uniform density range of the indoor point cloud, and perform secondary processing on the processed original indoor point cloud data to obtain optimized indoor point cloud data with uniform point cloud density.

[0075] Step d: Perform ICP fusion on indoor and outdoor multi-source data based on orthogonal trihedral targets. Select the overlapping area of ​​the outdoor and indoor LAS point clouds with orthogonal trihedral targets, first perform coarse registration based on the orthogonal trihedral targets, and then perform fine ICP registration.

[0076] Step e involves importing the integrated indoor and outdoor data into the modeling software and performing reverse modeling based on the point cloud. By depicting the shape of the building's indoor and outdoor base, and performing operations such as line-to-surface conversion, extrusion, and Boolean operations, an integrated digital model of the building's indoor and outdoor interiors is established.

[0077] This invention experimentally verifies the feasibility of orthogonal trihedral targets in multi-source data fusion. The specific process of the verification experiment is as follows.

[0078] In a specific indoor setting, a FARO 3D laser scanner was used to acquire and stitch data from multiple stations using both the FARO-compatible target sphere and an orthogonal trihedral target. The fusion accuracy was then compared. At the same station interval, with the FARO-compatible target sphere and orthogonal trihedral target, along with sticker markings, placed next to the instrument while keeping other variables constant, data acquisition and subsequent fusion processing were performed separately.

[0079] In this verification experiment, the FARO software SCENE was used for data fusion. The data was compared with data fused from orthogonal trihedral target regions selected within the program. The spacing between the fused sticker markings was used as a comparison factor for fusion accuracy. Taking a sticker with a black and white checkerboard pattern as an example, the actual size was A4 paper, with adhesive on the back. The measuring tape distance was 10.64 m. The SCENE target ball registration and target registration were 10.61 m and 10.62 m respectively. For comparison, the program's "target local registration" result was 10.63 m. Using the measuring tape as the true value, the ICP deviation was -0.01 m, with a relative error of 0.09%, significantly better than the target ball registration's -0.03 m (0.28%) and the target registration's -0.02 m (0.19%).

[0080] Furthermore, targets were comprehensively used in the collaborative acquisition of multi-source data both indoors and outdoors, followed by multi-source data fusion, and the results were compared with those obtained from total station measurements. Specifically, in terms of UAV oblique photogrammetry: the aircraft was planned and took off from the open space west of the building, with a flight altitude of 80m, a forward overlap rate of 70%, and a lateral overlap rate of 80%; in terms of 3D laser scanning: five scanning stations were set up near the points set indoors, and two sets of experimental data were collected using target ball stitching and marker stitching respectively, while other settings remained at default and the same; the coordinates of the outdoor image control points and the two outdoor entry points were collected, and the coordinates of the five indoor point markers were obtained using scattered measurement starting from the coordinates of the two outdoor entry points; the indoor and outdoor points were measured by total station, and then the fused data were obtained by stitching with FARO target balls and targets respectively. Starting from a certain point outdoors, line segments were drawn to the five indoor points respectively, and their planar distances were measured to obtain five sets of comparative experimental data, as shown in Table 1.

[0081] Table 1 Comparison of the projection distance of the fused point cloud slices onto the plane.

[0082] Experimental data show that, compared to FARO target ball stitching, target stitching exhibits significant advantages in overall accuracy and stability. Although the absolute length differences between the two are similar, the relative closure error of target stitching is generally smaller, and the error fluctuation is more stable. For example, in segment 13, its relative closure error is -0.0008, significantly better than the target ball stitching's -0.0019, demonstrating higher measurement consistency and error control capabilities. Therefore, local registration of targets, within a high-precision range, balances versatility and cost advantages, providing a reliable geometric benchmark for 3D reverse modeling.

[0083] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the indoor and outdoor fusion modeling method based on orthogonal trihedral targets as described in any of the above embodiments.

[0084] Figure 5 This embodiment illustrates a more specific hardware structure of an electronic device, which may include a processor 510, a memory 520, an input / output interface 530, a communication interface 540, and a bus 550. The processor 510, memory 520, input / output interface 530, and communication interface 540 are interconnected internally via the bus 550.

[0085] The processor 510 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0086] The memory 520 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 520 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 520 and is called and executed by the processor 510.

[0087] Input / output interface 530 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components in the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touch screens, microphones, various sensors, etc., and output devices may include displays, speakers, vibrators, indicator lights, etc.

[0088] The communication interface 540 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0089] Bus 550 includes a pathway for transmitting information between various components of the device, such as processor 510, memory 520, input / output interface 530, and communication interface 540.

[0090] It should be noted that although the above-described device only shows the processor 510, memory 520, input / output interface 530, communication interface 540, and bus 550, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0091] The electronic device described in the above embodiments is used to implement the indoor and outdoor fusion modeling method based on orthogonal trihedral targets as described in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0092] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, the present invention also provides a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the indoor-outdoor fusion modeling method based on orthogonal trihedral targets as described in any of the above embodiments.

[0093] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0094] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to execute the indoor and outdoor fusion modeling method based on orthogonal trihedral targets as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0095] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of the invention (including the claims) is limited to these examples; within the framework of the invention, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of the embodiments of the invention as described above, which are not provided in the details for the sake of brevity.

[0096] Additionally, to simplify the description and discussion, and to avoid obscuring the embodiments of the invention, the well-known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided drawings. Furthermore, the apparatus may be shown in block diagram form to avoid obscuring the embodiments of the invention, and this also takes into account the fact that the details of implementation of these block diagram apparatuses are highly dependent on the platform on which the embodiments of the invention will be implemented (i.e., these details should be fully understood by those skilled in the art). While specific details (e.g., circuits) have been set forth to describe exemplary embodiments of the invention, it will be apparent to those skilled in the art that the embodiments of the invention may be implemented without these specific details or with variations thereof. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0097] Although the invention has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0098] It should be understood that the above embodiments are merely illustrative of the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent changes or modifications made in accordance with the spirit and essence of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for indoor and outdoor fusion modeling based on orthogonal trihedral targets, characterized in that the method... include: An orthogonal trihedral target is set up at a designated location on the target building. The designated location includes at least an overlapping area between indoor and outdoor areas. The orthogonal trihedral target includes three mutually perpendicular planes that intersect at the same vertex. Based on the orthogonal trihedral target, outdoor tilted photographs and raw indoor point cloud data of the UAV targeting the building are obtained respectively. For the original indoor point cloud data, filtering is performed on regions with density higher than the upper limit of the uniform density range of indoor point cloud, and outlier deletion is performed on regions with density lower than the lower limit of the uniform density range of indoor point cloud. The processed original indoor point cloud data is then processed a second time to obtain optimized indoor point cloud data with uniform point cloud density. Using a target-feature-region two-level registration method, outdoor oblique photograph data and indoor point cloud optimization data of UAVs in overlapping indoor and outdoor areas are fused and registered to obtain multi-source fused data; The multi-source fusion data is used for reverse modeling to generate an integrated indoor and outdoor model of the target building; The process involves fusing and registering outdoor oblique drone imagery data and optimized indoor point cloud data in overlapping indoor and outdoor areas. Specifically, this includes: In an indoor-outdoor overlapping area where orthogonal trihedral targets are deployed, multiple common point coordinates are obtained, and coordinate transformation parameters between UAV outdoor tilted photo data and indoor point cloud optimization data are calculated based on the common point coordinates. Based on the coordinate transformation parameters, coarse registration is performed on the outdoor tilted photo data of the UAV and the indoor point cloud optimization data; With the goal of minimizing the sum of the Euclidean distances between corresponding point cloud pairs of the coarsely registered UAV outdoor tilted photo data and the indoor point cloud optimization data, multi-source point cloud registration is performed on the coarsely registered UAV outdoor tilted photo data and the indoor point cloud optimization data.

2. The indoor-outdoor fusion modeling method based on orthogonal trihedral targets according to claim 1, characterized in that, The process of determining the uniform density range of the indoor point cloud includes: Calculate the local point cloud density of each point in the original indoor point cloud data; The local point cloud density of all points is statistically analyzed to calculate the global average density and density standard deviation; The range of uniform density of the indoor point cloud is determined based on the global average density and the density standard deviation.

3. The indoor-outdoor fusion modeling method based on orthogonal trihedral targets according to claim 2, characterized in that, The range of uniform density of the indoor point cloud is: ； ； in, This represents the lower limit of the range of uniform density of the indoor point cloud. This represents the upper limit of the uniform density range of the indoor point cloud. The global average density, The density standard deviation is... The first adjustment coefficient, This is the second adjustment coefficient.

4. The indoor-outdoor fusion modeling method based on orthogonal trihedral targets according to claim 1, characterized in that, Filtering is applied to regions where the density exceeds the upper limit of the uniform density range of indoor point clouds, specifically including: For regions where the density exceeds the upper limit of the uniform density range of indoor point clouds, the filtering radius and filtering threshold are determined based on the local density of the region. The region is filtered according to the filtering radius and the filtering threshold to retain points with curvature greater than the curvature threshold or normal vector change greater than the change threshold, and to remove redundant points.

5. The indoor-outdoor fusion modeling method based on orthogonal trihedral targets according to claim 4, characterized in that, The filtering radius and the filtering threshold are respectively: ； ； in, For point clouds The filtering radius of the neighborhood. For point clouds The filtering threshold of the neighborhood For point clouds The radius of the neighborhood, For point clouds Local density of the neighborhood For point clouds Point cloud to point cloud within the neighborhood The mean distance, For point clouds Point cloud to point cloud within the neighborhood The standard deviation of the distance, It is a multiplier. This represents the lower limit of the range of uniform density of the indoor point cloud. This represents the upper limit of the uniform density range of the indoor point cloud.

6. The indoor-outdoor fusion modeling method based on orthogonal trihedral targets according to claim 1, characterized in that, The processed indoor point cloud raw data undergoes secondary processing, specifically including: weighted sampling of the processed indoor point cloud raw data to obtain the optimized indoor point cloud data; wherein, the resampling weight is: ； in, For point clouds resampling weights, This represents the upper limit of the uniform density range of the indoor point cloud. For point clouds Local density of the neighborhood For global average density, For density standard deviation, This is the smoothing coefficient.

7. The indoor-outdoor fusion modeling method based on orthogonal trihedral targets according to claim 1, characterized in that, The three planes of the orthogonal trihedral target are all made of square plates, and the square plates are coated with a matte coating in a checkerboard pattern.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the indoor and outdoor fusion modeling method based on orthogonal trihedral targets as described in any one of claims 1-7.

9. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to execute the indoor and outdoor fusion modeling method based on orthogonal trihedral targets as described in any one of claims 1-7.

Images