Method and device for three-dimensional reconstruction of urban buildings

Abstract

The present application relates to the technical field of image data processing, in particular to a kind of urban building three-dimensional reconstruction method and device, method includes: obtaining from the same scene Multi-View Satellite Ortho Image, target area's spaceborne laser height measurement data, digital terrain model and two-dimensional building vector data.Accurate geometric registration is carried out to Multi-View Satellite Ortho Image, and dense matching is executed, and initial parallax point cloud is generated.Combined with the azimuth information of Multi-View Ortho Image, parallax map with accurate plane positioning is constructed.Using spaceborne laser height measurement data and digital terrain model, the elevation value above ground is calculated.Combining elevation value and parallax value, parallax elevation proportion coefficient is estimated, to convert parallax map into height map.Combined with two-dimensional building vector data, the three-dimensional building vector data of target area is generated.Therefore, the problem that the cost is high and data acquisition is limited based on surveying and mapping level remote sensing data in the related art cannot meet the global urban three-dimensional reconstruction demand.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing, and in particular to a method and device for three-dimensional reconstruction of urban buildings. Background Art

[0002] In related technologies, classic photogrammetry 3D reconstruction methods are generally based on surveying-level remote sensing data, such as aerial images with pose information, airborne lidar point clouds, and stereo satellite images with imaging model information.

[0003] However, related technologies based on surveying-level remote sensing data are costly and difficult to collect, and are unable to meet the needs of three-dimensional reconstruction of cities on a global scale, and urgently need to be improved. Summary of the Invention

[0004] The present invention provides a method and device for three-dimensional reconstruction of urban buildings to solve the problem in related technologies that surveying-level remote sensing data is high in cost and difficult to collect, and cannot meet the needs of global urban three-dimensional reconstruction.

[0005] A first aspect of the present invention provides a method for three-dimensional reconstruction of urban buildings, comprising the following steps: obtaining multi-view satellite orthophotos, satellite-borne laser altimetry data of a target area, a digital terrain model (DTM), and two-dimensional building vector data from the same scene; performing precise geometric registration on the multi-view satellite orthophotos so that the positions of ground points in the images are consistent in images of different viewing angles to obtain multi-view orthophotos; generating an initial disparity map representing the displacement characteristics of matching points based on the multi-view orthophotos, and performing forward intersection on the matching points in combination with the azimuth information of the multi-view orthophotos to generate a final disparity map; extracting a sparse three-dimensional point cloud based on the satellite-borne laser altimetry data, and computing the elevation values ​​above the ground surface of corresponding positions through the digital terrain model and the sparse three-dimensional point cloud; obtaining the disparity values ​​of corresponding positions in the final disparity map, and obtaining a disparity elevation scaling coefficient based on the elevation values ​​above the ground surface and the disparity values; generating three-dimensional point cloud data based on the disparity elevation scaling coefficient, and generating three-dimensional building vector data of the target area in combination with the three-dimensional point cloud data and the two-dimensional building vector data.

[0006] Through the above-mentioned technical solution, the embodiments of the present invention can utilize multi-view satellite orthophotos, spaceborne laser altimetry data, digital terrain models, and two-dimensional building vector data from the same scene, avoiding reliance on expensive and limited surveying-grade remote sensing data. This reduces the cost and difficulty of data acquisition and improves applicability and scalability. Precise geometric registration ensures the consistency of ground point positions in the multi-view orthophotos, laying the foundation for subsequent precise parallax calculations and enhancing the accuracy of parallax calculations. By combining the azimuth information of the multi-view orthophotos for forward intersection and generating a final parallax map, the parallax data is further optimized and its reliability is enhanced. Elevation values ​​above the ground surface are calculated using spaceborne laser altimetry data and the digital terrain model, and the parallax-to-elevation scaling factor is derived from this. This allows for more accurate conversion of parallax to elevation, thereby generating high-precision three-dimensional point cloud data. Finally, by combining the three-dimensional point cloud data with the two-dimensional building vector data to generate three-dimensional building vector data, the resulting three-dimensional model is detailed and accurate, providing powerful data support and decision-making basis for a wide range of fields, including urban planning, disaster management, environmental monitoring, and smart city development.

[0007] Optionally, in one embodiment of the present invention, generating an initial disparity map characterizing the displacement characteristics of matching points based on the multi-view orthophoto includes: calculating the disparity amplitude values ​​of the matching points in the x and y directions; and generating the initial disparity map based on the disparity amplitude values.

[0008] Through the above technical solution, the embodiment of the present invention can calculate the disparity amplitude values ​​of the matching points in the x and y directions and generate an initial disparity map based on them, which can quickly and intuitively reflect the displacement characteristics of the matching points in images with different viewing angles.

[0009] Optionally, in one embodiment of the present invention, generating three-dimensional point cloud data according to the disparity-elevation scaling coefficient includes: converting the final disparity map into an elevation map based on the disparity-elevation scaling coefficient; and generating the three-dimensional point cloud data by pixel-by-pixel sampling based on the elevation map.

[0010] Through the above technical solution, the embodiment of the present invention can convert the final disparity map into an elevation map based on the disparity-elevation scale coefficient, and then generate three-dimensional point cloud data by sampling pixel by pixel. The use of the disparity-elevation scale coefficient closely combines the relationship between disparity and elevation, realizes the scientific conversion from disparity map to elevation map, and enables the elevation information to accurately reflect the actual height of the ground objects. The method of generating three-dimensional point cloud data by pixel-by-pixel sampling carefully and comprehensively collects the information in the elevation map, ensures the high precision and high resolution of the three-dimensional point cloud data, fully presents the topographic details of the target area, and effectively improves the efficiency and accuracy of three-dimensional point cloud data generation.

[0011] Optionally, in one embodiment of the present invention, the final disparity map is represented by:

[0012] ,

[0013] wherein, is the compensated matching point position, is the disparity value of the corresponding position.

[0014] Through the technical solution, the embodiment of the application can integrate the compensated matching point position and the corresponding disparity value information, clearly and intuitively define the constituent elements of the final disparity map, and make the expression of the disparity map highly normative and unified. In the data transmission and analysis of different links, errors and misunderstandings caused by unclear expression can be effectively reduced, and the accuracy and efficiency of data processing in the entire urban building three-dimensional reconstruction process are improved.

[0015] Optionally, in an embodiment of the application, the disparity elevation proportion coefficient is obtained based on the above-ground elevation value and the disparity value, comprising: interpolating the terrain height value of the corresponding position based on the sparse three-dimensional point cloud; calculating the above-ground elevation value of the corresponding position of the sparse three-dimensional point cloud according to the terrain height value; interpolating the disparity value of the corresponding position of the sparse three-dimensional point cloud, and fitting the disparity and the above-ground proportion factor based on the above-ground elevation value and the disparity value to obtain the disparity elevation proportion coefficient.

[0016] Through the technical solution, the embodiment of the application can calculate the above-ground elevation value by interpolating the terrain height value based on the sparse three-dimensional point cloud, interpolate the disparity value and fit the proportion factor to obtain the disparity elevation proportion coefficient. It fully utilizes the sparse three-dimensional point cloud data, fully excavates the effective information in the data through accurate interpolation calculation, comprehensively considers the influence of the terrain factor on the relationship between the disparity and the elevation, and makes the disparity elevation proportion coefficient more accurately reflect the real proportion relationship between the two.

[0017] Optionally, in an embodiment of the application, the representation of the three-dimensional point cloud data is:

[0018] ,

[0019] wherein, is the theoretically correct geographical position after compensation; is the disparity value; is the fitted disparity elevation proportion coefficient.

[0020] Through the technical solution, the embodiment of the application can express the three-dimensional point cloud data, ensure the coherence and accuracy of the data in the three-dimensional reconstruction process, and thus improve the accuracy of the three-dimensional reconstruction.

[0021] The second embodiment of the present invention provides a three-dimensional reconstruction device for urban buildings, including: an acquisition module for acquiring multi-view satellite orthophotos from the same scene, satellite-borne laser altimetry data of the target area, a digital terrain model and two-dimensional building vector data; a matching module for accurately geometrically aligning the multi-view satellite orthophotos so that the positions of ground points in the images are consistent in images of different viewing angles to obtain multi-view orthophotos; a generation module for generating an initial disparity map representing the displacement characteristics of matching points based on the multi-view orthophotos, so as to perform forward registration on the matching points in combination with the azimuth information of the multi-view orthophotos. An intersection is performed to generate a final disparity map; a first calculation module is used to extract a sparse three-dimensional point cloud based on the satellite-borne laser altimetry data, and calculate the elevation value above the surface of the corresponding position through the digital terrain model and the sparse three-dimensional point cloud; a second calculation module is used to obtain the disparity value of the corresponding position in the final disparity map, and obtain a disparity elevation scale coefficient based on the elevation value above the surface and the disparity value; a three-dimensional reconstruction module is used to generate three-dimensional point cloud data according to the disparity elevation scale coefficient, and generate three-dimensional building vector data of the target area by combining the three-dimensional point cloud data and the two-dimensional building vector data.

[0022] Through the above-mentioned technical solution, the embodiments of the present invention can utilize multi-view satellite orthophotos, spaceborne laser altimetry data, digital terrain models, and two-dimensional building vector data from the same scene, avoiding reliance on expensive and limited surveying-grade remote sensing data. This reduces the cost and difficulty of data acquisition and improves applicability and scalability. Precise geometric registration ensures the consistency of ground point positions in the multi-view orthophotos, laying the foundation for subsequent precise parallax calculations and enhancing the accuracy of parallax calculations. By combining the azimuth information of the multi-view orthophotos for forward intersection and generating a final parallax map, the parallax data is further optimized and its reliability is enhanced. Elevation values ​​above the ground surface are calculated using spaceborne laser altimetry data and the digital terrain model, and the parallax-to-elevation scaling factor is derived from this. This allows for more accurate conversion of parallax to elevation, thereby generating high-precision three-dimensional point cloud data. Finally, by combining the three-dimensional point cloud data with the two-dimensional building vector data to generate three-dimensional building vector data, the resulting three-dimensional model is detailed and accurate, providing powerful data support and decision-making basis for a wide range of fields, including urban planning, disaster management, environmental monitoring, and smart city development.

[0023] Optionally, in one embodiment of the present invention, the generating module includes: a first calculating unit, configured to calculate the disparity amplitude values ​​of the matching points in the x and y directions; and a generating unit, configured to generate the initial disparity map according to the disparity amplitude values.

[0024] Through the above technical solution, the embodiment of the present invention can calculate the disparity amplitude values ​​of the matching points in the x and y directions and generate an initial disparity map based on them, which can quickly and intuitively reflect the displacement characteristics of the matching points in images with different viewing angles.

[0025] Optionally, in one embodiment of the present invention, the three-dimensional reconstruction module includes: a conversion unit for converting the final disparity map into an elevation map based on the disparity-elevation scaling coefficient; and a sampling unit for generating the three-dimensional point cloud data by pixel-by-pixel sampling based on the elevation map.

[0026] Through the above technical solution, the embodiment of the present invention can convert the final disparity map into an elevation map based on the disparity-elevation scale coefficient, and then generate three-dimensional point cloud data by sampling pixel by pixel. The use of the disparity-elevation scale coefficient closely combines the relationship between disparity and elevation, realizes the scientific conversion from disparity map to elevation map, and enables the elevation information to accurately reflect the actual height of the ground objects. The method of generating three-dimensional point cloud data by pixel-by-pixel sampling carefully and comprehensively collects the information in the elevation map, ensures the high precision and high resolution of the three-dimensional point cloud data, fully presents the topographic details of the target area, and effectively improves the efficiency and accuracy of three-dimensional point cloud data generation.

[0027] Optionally, in one embodiment of the present invention, the final disparity map is represented by:

[0028] ,

[0029] in, is the position of the matching point after compensation, is the disparity value of the corresponding position.

[0030] Through the above technical solution, the present invention integrates the compensated matching point positions and corresponding disparity values, clearly and intuitively defining the components of the final disparity map and ensuring a highly standardized and uniform disparity map representation. This effectively reduces errors and misunderstandings caused by unclear representations during data transmission and analysis across different stages, improving the accuracy and efficiency of data processing throughout the entire 3D reconstruction process for urban buildings.

[0031] Optionally, in one embodiment of the present invention, the second calculation module includes: an interpolation unit for interpolating the terrain height value of the corresponding position based on the sparse three-dimensional point cloud; a second calculation unit for calculating the elevation value above the ground surface of the corresponding position of the sparse three-dimensional point cloud based on the terrain height value; a fitting unit for interpolating the disparity value of the corresponding position based on the sparse three-dimensional point cloud, and fitting the disparity and the above-ground scale factor based on the above-ground height value and the disparity value to obtain the disparity elevation scale coefficient.

[0032] Through the above technical solution, the embodiment of the present invention can calculate the elevation above the ground surface by interpolating terrain height values ​​based on a sparse 3D point cloud. It then interpolates the disparity value and fits the scaling factor to obtain the disparity-to-elevation scaling coefficient. This method fully utilizes sparse 3D point cloud data, uses precise interpolation calculations to fully exploit the effective information in the data, and comprehensively considers the impact of terrain factors on the relationship between disparity and elevation. This enables the disparity-to-elevation scaling coefficient to more accurately reflect the true proportional relationship between the two.

[0033] Optionally, in one embodiment of the present invention, the expression of the three-dimensional point cloud data is:

[0034] ,

[0035] in, The theoretically correct geographical location after compensation; is the disparity value; is the parallax elevation scale coefficient obtained by fitting.

[0036] Through the above technical solution, the embodiment of the present invention can ensure the consistency and accuracy of data during the three-dimensional reconstruction process by representing the three-dimensional point cloud data, thereby improving the accuracy of the three-dimensional reconstruction.

[0037] A third aspect of the present invention provides an electronic device comprising 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 method for three-dimensional reconstruction of urban buildings as described in the above embodiment.

[0038] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned method for three-dimensional reconstruction of urban buildings.

[0039] A fifth aspect of the present invention provides a computer program, which is executed to implement the above method for three-dimensional reconstruction of urban buildings.

[0040] The embodiments of the present invention can utilize a variety of public data, including but not limited to multi-view satellite orthophotos and satellite-borne laser altimetry data, in data acquisition, avoiding the use of expensive and limited surveying-grade remote sensing data, reducing costs and difficulty, and improving applicability and scalability. In terms of data processing, precise geometric alignment lays a solid foundation for disparity calculation, and the final disparity map generated in combination with azimuth information is highly reliable. By calculating the elevation value above the surface and obtaining the disparity elevation scale coefficient, the disparity can be accurately converted to elevation to generate high-precision three-dimensional point cloud data. Generating an initial disparity map can quickly and intuitively reflect the displacement characteristics of the matching points. The conversion of the disparity map to the elevation map and the pixel-by-pixel sampling method to generate three-dimensional point cloud data ensure the accuracy and resolution of the data, fully present the details of the terrain and improve data generation efficiency. The standardized and unified definition of the final disparity map reduces data processing errors. The calculation of the disparity elevation scale coefficient based on sparse three-dimensional point cloud can fully mine data information, consider terrain factors, and make the coefficient more accurate. The reasonable representation of 3D point cloud data ensures data consistency and accuracy. The final constructed 3D model is detailed and precise, which can provide strong data support and decision-making basis for many fields such as urban planning and disaster management.

[0041] Additional aspects and advantages of the present invention will be set forth in part in the description which follows and, in part, will be obvious from the description which follows, or may be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0043] Figure 1 A flowchart of a method for three-dimensional reconstruction of urban buildings according to an embodiment of the present invention;

[0044] Figure 2 is a schematic diagram of input data according to one embodiment of the present invention;

[0045] Figure 3 2. A schematic diagram of parallax compensation for two-view satellite orthophoto images according to a specific embodiment of the present invention;

[0046] Figure 4 A schematic structural diagram of a device for 3D reconstruction of urban buildings according to an embodiment of the present invention;

[0047] Figure 5 FIG. 1 is a diagram illustrating a structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The following describes embodiments of the present invention in detail, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and are not to be construed as limiting the present invention.

[0049] The following describes the three-dimensional reconstruction method and device of urban buildings according to an embodiment of the present invention with reference to the accompanying drawings. In view of the problem that the related technologies mentioned in the above background technology are based on surveying-level remote sensing data with high cost and difficulty in data collection, and cannot meet the needs of global urban three-dimensional reconstruction, the present invention provides a three-dimensional reconstruction method for urban buildings. In this method, the pixel-by-pixel disparity value can be obtained by matching multi-view orthophotos, and the pixel position corresponding to the disparity can be corrected using the plumb line direction compensation technology. The three-dimensional point cloud of the urban building can be restored by estimating the ratio of the disparity value to the height of the ground object in combination with satellite-borne laser altimetry data. Finally, combined with the public two-dimensional building vector data, an accurate and detailed three-dimensional model of the building is constructed, which avoids the use of expensive data, has low cost, and improves the accuracy of the data. Thus, the problem that the related technologies are based on surveying-level remote sensing data with high cost and difficulty in data collection, and cannot meet the needs of global urban three-dimensional reconstruction, is solved.

[0050] Specifically, Figure 1 A schematic flow chart of a method for 3D reconstruction of urban buildings provided by an embodiment of the present invention.

[0051] like Figure 1 As shown, the method for 3D reconstruction of urban buildings includes the following steps:

[0052] In step S101 , multi-view satellite orthophotos from the same scene, spaceborne laser altimetry data of the target area, a digital terrain model, and two-dimensional building vector data are obtained.

[0053] It is understandable that images are obtained through satellite remote sensing technology, with high resolution and accurate geometric information, and can reflect the actual terrain and building distribution of the target area. Figure 2 The following is a schematic diagram of the input data. To ensure the accuracy and consistency of the data, these images should be acquired within the same time period to avoid differences in ground features caused by temporal changes.

[0054] Spaceborne laser altimetry data of the target area is usually obtained through lidar technology, which can provide ground elevation information. This elevation data provides the basis for subsequent height information of surface features.

[0055] A digital terrain model is a digital representation of the Earth's surface topography, including ground elevation information, enabling a better understanding and analysis of the terrain's undulations and variations. This model can be generated using a variety of methods, including but not limited to laser altimetry data and aerial photogrammetry.

[0056] Acquiring 2D building vector data provides the necessary building outlines and positional references for 3D reconstruction. This vector data, typically acquired through geographic information system technology, contains building boundaries, shapes, and other relevant attribute information. By combining this 2D data, the spatial location and structural characteristics of the building can be more accurately defined during the subsequent 3D modeling process.

[0057] Embodiments of the present invention can integrate remote sensing data from multiple sources to provide comprehensive and accurate information for subsequent three-dimensional reconstruction processes.

[0058] In step S102 , precise geometric registration is performed on the multi-view satellite orthophoto images so that the positions of ground points in the images are consistent in images of different viewing angles, thereby obtaining multi-view orthophoto images.

[0059] It can be understood that accurate registration of the ground area of ​​multi-view satellite orthophotos can ensure that the positions of ground points in images from different perspectives are aligned.

[0060] Taking two-view satellite orthophotos as an example, the method of geometric registration is as follows:

[0061] 1) Obtain a set of matching points through the sparse matching algorithm, and record the set of matching points in the first view image as: , the set of matching points in the second view image is recorded as: .in, and ) represents the coordinate pair of matching points in the first and second view images.

[0062] 2) Obtaining the initial matching point set and Finally, these matching points are filtered using the building mask data of the target area. Specifically, the matching points on the buildings in the image are eliminated, and the matching points on the ground are retained to form a ground matching point set:

[0063]

[0064]

[0065] This process ensures that the matching points involved in image registration are all from the ground area, avoiding matching errors caused by differences in building heights, thereby improving the accuracy of registration.

[0066] 3) After filtering out the ground matching point set and After that, the accurate alignment of the two-view images is completed by using a set registration method. Specifically, by using a geometric transformation model, including but not limited to an affine transformation or a perspective transformation, the translation, rotation and scaling parameters of the images are optimized, so that the matching points of the two-view images can be perfectly aligned in the ground area.

[0067] The embodiment of the present application can ensure the consistency of the multi-view satellite orthographic images in the ground area by accurate geometric registration, and provide reliable basic data support for subsequent parallax calculation and height solution.

[0068] In step S103, an initial parallax map representing the displacement characteristics of the matching points is generated according to the multi-view orthographic images, the matching points are intersected in front in combination with the azimuth information of the multi-view orthographic images, and a final parallax map is generated.

[0069] That is to say, the embodiment of the present application can generate a matching point set and by using a dense matching algorithm based on the multi-view orthographic images corrected in step S102, and calculate an initial parallax map according to the displacement characteristics thereof. Subsequently, the matching points are intersected in front in combination with the azimuth information of the images, and a final parallax map with accurate spatial positioning is generated.

[0070] It can be understood that the parallax refers to the position difference of the corresponding points of the same object in different view images, which can reflect the depth information of the object.

[0071] Taking two-view satellite orthographic images as an example, the parallax and of each pair of matching points and is calculated according to the matching point set generated by using a dense matching algorithm. The calculation formula is as follows:

[0072]

[0073] Optionally, in an embodiment of the present application, generating an initial parallax map representing the displacement characteristics of the matching points according to the multi-view orthographic images comprises: calculating the parallax amplitude values of the matching points in the x and y directions; and generating the initial parallax map according to the parallax amplitude values.

[0074] By calculating the parallax of each pair of matching points, a parallax array can be obtained, and according to the parallax data, an initial parallax map can be generated, which can represent the displacement characteristics of the matching points and can intuitively observe the parallax distribution at different positions in the images, thereby providing a reference for subsequent parallax calculation and processing.

[0075] Furthermore, the azimuth angle and each set of matching points are used to perform forward intersection to obtain the correct position corresponding to the matching point, and the position obtained by forward intersection and the matching point position are used to calculate the geometric offset. Then, the intersection point and the coordinates of the matching points on different orthophotos are used to obtain the respective offsets along the plumb line, and the least squares method is used to solve the ratio of the left and right offsets to eliminate gross errors.

[0076] The specific principle is shown as follows Figure 3 As shown. Obtain the azimuth of each image (i.e., the direction of the plumb line in the image) and use the matching points to intersect forward along their respective plumb lines to obtain the correct geographic location of the object where the matching points are located. Correct the positions of all disparity values ​​in this step to obtain a disparity map with higher position accuracy.

[0077] The steps for multi-view orthophoto visual compensation are the same as those for two-view, except that the compensated disparity maps are eventually merged to form a complete disparity map. Taking two-view satellite orthophotos as an example, the method for compensating the disparity position along the plumb line is as follows:

[0078] 1) During the orthophoto matching process, due to the influence of factors such as the tilt of the shooting angle and the undulation of the terrain, the position of the objects in the image may deviate from their true geographic coordinates, resulting in a certain degree of error in the plane position of the matching points. To eliminate this type of deviation, the position of the matching points needs to be compensated accordingly. For each pair of matching points , ), the azimuths of the two images are known and , use the forward intersection formula to calculate the compensation coordinates of the matching point along the plumb line direction Forward intersection is a method of calculating the coordinates of an unknown point using the coordinates and azimuth of a known point. The calculation formula is:

[0079]

[0080] Through the forward intersection calculation, the accurate position of the matching point in the plumb line direction can be obtained, thereby improving the accuracy of the disparity calculation.

[0081] 2) After calculating the compensated coordinates, it is necessary to calculate the offset of each matching point relative to the compensated coordinates and Image matching is used to match the multi-view orthophotos pixel by pixel, extract the precise matching points, and calculate the disparity amplitude of the matching points in the x and y directions, which is used as the disparity value of the corresponding position of the matching points. The calculation formula of the offset is:

[0082]

[0083]

[0084] Get the offset set of all matching points: , These offsets reflect the difference between the original position and the compensated position of the matching points, providing data support for the subsequent optimal scale coefficient solution.

[0085] 3) In order to further optimize the position of the matching point, the optimal proportional coefficient of the offset can be solved by the least squares method , so that the proportional relationship of the offset is minimized. The least squares method is a method to find the best function matching method for data by minimizing the sum of squares of errors. Specifically, find a value, so that By solving the optimal scale factor, the offset of the matching point can be adjusted to improve the accuracy of the matching point.

[0086] 4) Calculate the offset difference value Then, these differences were statistically analyzed and their mean values ​​were calculated. and standard deviation , thus according to the statistical principle, we can eliminate These outliers may be caused by image noise, matching errors, etc. Removing them can improve the quality of the data and thus improve the accuracy of disparity calculation.

[0087] After completing the above steps, the compensated matching point positions and corresponding disparity values ​​can be obtained. Based on these data, the final disparity map is generated. The final disparity map is represented by:

[0088] ,

[0089] in, is the position of the matching point after compensation, is the disparity value of the corresponding position.

[0090] The embodiment of the present invention can generate an initial disparity map by calculating the disparity of the matching points in the disparity map generation link, intuitively present the displacement characteristics of the matching points, provide a reference for subsequent processing, and calculate the disparity amplitude value and generate the initial disparity map accordingly. When using the azimuth angle for forward intersection, the correct geographical location of the matching point can be obtained, the disparity value position can be corrected, and the accuracy of the disparity calculation can be improved, effectively overcoming the problem of matching point position deviation caused by factors such as image shooting angle and terrain undulation. The multi-view orthophoto visual compensation step is similar to the two-view and can merge disparity maps, ensuring the versatility and extensibility of the method. In the process of disparity position compensation, operations such as calculating the offset, solving the optimal proportional coefficient, and eliminating abnormal points can optimize the matching point position, improve data quality, and further improve the accuracy of disparity calculation.

[0091] In step S104, a sparse three-dimensional point cloud is extracted based on the satellite-borne laser altimetry data, and the elevation value above the ground surface of the corresponding position is calculated using the digital terrain model and the sparse three-dimensional point cloud.

[0092] It is understandable that satellite-borne laser altimeter data contains rich surface elevation information, but it may contain noise and inaccurate data, so it needs to be processed to extract reliable sparse three-dimensional point clouds. In an embodiment of the present invention, ATL03 (Advanced Topographic Laser Altimeter System Global Geolocated Photon Data) and ATL08 (Land and Vegetation Height) provided by ICESat-2 (Ice, Cloud, and Land Elevation Satellite-2) can be used to obtain reliable elevation points.

[0093] ATL03 data is the most primitive elevation data product from the ICESat-2 satellite. It records the reflection point information of each laser pulse on the ground, including its geographic location (latitude and longitude) and elevation. Because laser pulse propagation is affected by atmospheric scattering, cloud interference, and complex reflections from ground objects, ATL03 data may contain a large number of noise points and inaccurate measurements. Therefore, directly using ATL03 data for analysis and processing may result in significant errors, requiring further processing and screening.

[0094] ATL08 data is a product derived from ATL03 data. Specifically, it is a point cloud data obtained by plane fitting in 100-meter increments along the track direction. This processing method effectively reduces the impact of noise and improves data reliability and accuracy. ATL08 data also includes classification information for different elevations, including -1 (unclassified), 0 (noise), 1 (ground), 2 (canopy), and 3 (top of canopy). This classification information helps screen out reliable elevation points.

[0095] To extract reliable elevation points from the ATL03 data, the ph_segment_id field is used to map ATL08 to ATL03. The ph_segment_id field is a number used to identify the laser pulse segment. This number allows the classification information in the ATL08 data to be associated with the reflection point information in the ATL03 data. The ATL03 data is then classified using the classed_pc_indx field. Point cloud data with classification numbers of -1 (unclassified) and 0 (noise) are filtered out to obtain relatively reliable elevation point data, which forms the basis of the sparse 3D point cloud.

[0096] After obtaining the sparse point cloud, the elevation above the ground at the corresponding location is calculated in conjunction with the digital terrain model. By comparing and calculating the elevation values ​​of the sparse 3D point cloud with the corresponding elevation values ​​in the digital terrain model, the height of each point relative to the ground, i.e., the elevation above the ground, is obtained.

[0097] Embodiments of the present invention utilize satellite-borne laser altimetry data to extract sparse 3D point clouds and mine surface elevation information. ICESat-2's ATL03 and ATL08 products are used. ATL08 is processed to reduce noise and include classification information, facilitating the selection of reliable elevation points. Field association and classification are used to accurately filter out noisy data and construct a high-quality point cloud. In combination with a digital terrain model, elevation values ​​above the surface are calculated, providing accurate and reliable data for 3D modeling and analysis in related fields.

[0098] In step S105 , the disparity value of the corresponding position is obtained in the final disparity map, and a disparity elevation ratio coefficient is obtained based on the elevation value above the ground surface and the disparity value.

[0099] It is understood that obtaining the disparity values ​​corresponding to the satellite-borne laser point cloud from the final disparity map requires the geographic coordinates of the satellite-borne laser point cloud. The satellite-borne laser point cloud data records the precise geographic location information of each point, which is the key basis for locating and obtaining the corresponding disparity values ​​in the final disparity map.

[0100] Specifically, a disparity elevation scale coefficient is obtained based on the elevation value above the ground and the disparity value, including: interpolating the terrain height value of the corresponding position based on the sparse three-dimensional point cloud; calculating the elevation value above the ground for the corresponding position of the sparse three-dimensional point cloud based on the terrain height value; interpolating the disparity value of the corresponding position based on the sparse three-dimensional point cloud, and fitting the disparity and the scale factor above the ground based on the elevation value above the ground and the disparity value, so as to obtain the disparity elevation scale coefficient.

[0101] During actual implementation, a search radius is set. Setting the search radius is a parameter that requires careful consideration, as it determines the range within which disparity values ​​are searched in the final disparity map. If the search radius is set too small, the corresponding disparity value may not be found; if it is set too large, excessive irrelevant disparity values ​​may be introduced, affecting the accuracy of the calculation results. In an embodiment of the present invention, after extensive experimentation and verification, the search radius can be set to 5*1e-5 (in degrees). This value ensures that the corresponding disparity value is found while minimizing interference from irrelevant information.

[0102] After determining the search radius, the final disparity map is searched for all disparity values ​​within the search radius, centered on the geographic coordinates of each satellite-borne laser point cloud. Since the disparity values ​​in the final disparity map may not be uniformly distributed, bilinear interpolation can be used to obtain more accurate corresponding disparity values. Bilinear interpolation is a commonly used interpolation method that estimates the disparity value of a point by taking a weighted average of the four adjacent disparity values ​​within the search radius. This method fully utilizes information from surrounding disparity values ​​and improves the accuracy of disparity calculations.

[0103] After obtaining the elevation point cloud above the surface and the corresponding parallax value, it is necessary to find a suitable parallax elevation scale factor. , so that the relationship between the parallax value and the elevation value above the surface reaches the optimal fit.

[0104] The least squares method can be used for optimization. That is, the goal is to find a Value, so that the elevation point cloud above the surface and parallax value go through The sum of squares of the scaled differences is minimized. Specifically, solve the following formula:

[0105]

[0106] The embodiment of the application can utilize the star-borne laser point cloud geographic coordinates to obtain the corresponding parallax value from the final parallax map, rely on the point cloud accurate position information positioning, the method is scientific and reasonable, and provides a reliable basis for subsequent calculation. Secondly, the parallax elevation proportion coefficient is determined by interpolating the terrain height value based on the sparse three-dimensional point cloud, calculating the elevation value above the ground, and interpolating the parallax value, fully utilizing the existing data, deeply mining the data correlation, and making the proportion coefficient more in line with the actual situation. Thirdly, the search radius is set and the parallax value is calculated by using the bilinear interpolation, which can accurately obtain the target parallax value and effectively exclude irrelevant interference, thereby improving the accuracy of the parallax value calculation. Finally, the least square method is used to optimize the solution of the parallax elevation proportion coefficient, the square sum of the difference is minimized, the relationship between the parallax value and the elevation value above the ground is ensured to be optimally fitted, and key and accurate parameters are provided for subsequent parallax-based conversion of the elevation and construction of a high-precision three-dimensional model.

[0107] In step S106, three-dimensional point cloud data is generated according to the parallax elevation proportion coefficient, and three-dimensional building vector data of the target area is generated by combining the three-dimensional point cloud data and the two-dimensional building vector data.

[0108] It can be understood that the parallax elevation proportion coefficient obtained in step S105 is used to convert the parallax value into an elevation value, thereby obtaining three-dimensional point cloud data. The parallax value compensated in step S103 is converted into an elevation value, thereby obtaining three-dimensional point cloud.

[0109] Optionally, in an embodiment of the application, generating three-dimensional point cloud data according to the parallax elevation proportion coefficient comprises: converting the final parallax map into an elevation map based on the parallax elevation proportion coefficient; and generating three-dimensional point cloud data by pixel-by-pixel sampling based on the elevation map.

[0110] Specifically, in step S103, the coordinates after parallax compensation are image coordinates, which are meaningful only in the image plane. In order to accurately reflect the actual geographic space position, it is necessary to convert the image coordinates into geographic coordinates by using the affine transformation parameters of the orthophoto. The affine transformation parameters of the orthophoto contain the translation, rotation and scaling information of the image in the geographic space, and the accurate conversion of the image coordinates to the geographic coordinates can be realized by using these parameters. Specifically, the elevation value above the ground is solved by the following formula:

[0111]

[0112]

[0113] wherein, is the theoretically correct geographic position after compensation; is the parallax value; is the parallax elevation proportion coefficient obtained by fitting; is an affine transformation parameter of the image coordinate of the orthographic image to the geographic coordinate. Through this conversion, not only can the parallax value be converted into the actual elevation value, but also it can be accurately mapped into the geographic space, so as to obtain three-dimensional point cloud data containing geographic position and elevation information. The expression of the three-dimensional point cloud data is:

[0114] .

[0115] Further, by using the three-dimensional point cloud and the building vector data, an optimal plane is solved to obtain three-dimensional building vector data. First, according to the range of the two-dimensional building vector, the three-dimensional point cloud in each building range is processed. The two-dimensional building vector data records the contour information of the building in the plane, and through these information, the range of the three-dimensional point cloud corresponding to each building can be determined.

[0116] In the three-dimensional point cloud range of each building, the overall height of the building is accurately estimated by fitting the optimal plane height. Since the surface of the building can be approximately regarded as composed of multiple planes, by fitting the plane to the three-dimensional point cloud in the building range, the plane equation that best represents the building surface can be found. In the fitting process, considering the complexity of the building structure and the possible noise of the point cloud data, a series of optimization algorithms are used to ensure the accuracy of the plane fitting. For example, a plane fitting algorithm based on least squares method can be used to find the optimal plane parameters by minimizing the sum of squares of distances from the point cloud to the fitted plane.

[0117] After the plane equation of the building surface is determined, the height information of the building at different positions can be calculated according to the parameters of the plane equation, and the corresponding three-dimensional building vector data is generated. The three-dimensional building vector data not only contains the plane contour information of the building, but also adds the height information, which can more truly reflect the three-dimensional form of the building.

[0118] The embodiment of the present application can convert the compensated parallax value into the elevation value by using the parallax elevation scale factor, which can ensure the accuracy of the elevation information in the three-dimensional point cloud data. According to the range of the two-dimensional building vector data, the range of the three-dimensional point cloud is determined, and the optimal plane height is fitted by using the optimization algorithm, which fully considers the complexity of the building structure and the noise of the point cloud, and ensures the accuracy of the estimated overall height of the building. The plane contour and the height information are fused to generate the three-dimensional building vector data, which presents the three-dimensional form of the building.

[0119] The three-dimensional reconstruction method for urban buildings proposed in an embodiment of the present invention can utilize multiple data types, such as multi-view satellite orthophotos and satellite-borne laser altimetry data, to avoid reliance on expensive and limited surveying-grade remote sensing data, reduce the cost and difficulty of data acquisition, and improve applicability and scalability. Precise geometric registration improves the accuracy of parallax calculations, and the combination of azimuth information optimizes parallax data and enhances the reliability of the parallax map. With the help of satellite-borne laser altimetry data and digital terrain models, parallax is accurately converted to elevation to generate high-precision three-dimensional point cloud data. Finally, the three-dimensional point cloud is combined with two-dimensional building vector data to generate detailed and accurate three-dimensional building vector data, providing powerful data support and decision-making basis for multiple fields such as urban planning.

[0120] Next, a device for 3D reconstruction of urban buildings according to an embodiment of the present invention will be described with reference to the accompanying drawings.

[0121] Figure 4 4 is a block diagram of a device for 3D reconstruction of urban buildings according to an embodiment of the present invention.

[0122] like Figure 4 As shown, the urban building 3D reconstruction device 10 includes: an acquisition module 100 , a matching module 200 , a generation module 300 , a first calculation module 400 , a second calculation module 500 and a 3D reconstruction module 600 .

[0123] Specifically, the acquisition module 100 is used to acquire multi-view satellite orthophotos from the same scene, spaceborne laser altimetry data of the target area, a digital terrain model and two-dimensional building vector data.

[0124] The matching module 200 is used to perform precise geometric registration on the multi-view satellite orthophotos so that the positions of ground points in the images are consistent in images of different viewing angles, thereby obtaining multi-view orthophotos.

[0125] The generating module 300 is used to generate an initial disparity map representing the displacement characteristics of the matching points based on the multi-view orthophotos, and to perform forward intersection on the matching points in combination with the azimuth information of the multi-view orthophotos to generate a final disparity map.

[0126] The first calculation module 400 is used to extract a sparse three-dimensional point cloud based on the satellite-borne laser altimetry data, and calculate the elevation value above the ground surface of the corresponding position through the digital terrain model and the sparse three-dimensional point cloud.

[0127] The second calculation module 500 is configured to obtain the disparity value of the corresponding position in the final disparity map, and obtain a disparity-elevation ratio coefficient based on the elevation value above the ground surface and the disparity value.

[0128] The 3D reconstruction module 600 is used to generate 3D point cloud data according to the parallax elevation scale coefficient, and to generate 3D building vector data of the target area by combining the 3D point cloud data and the 2D building vector data.

[0129] Optionally, in one embodiment of the present invention, the generation module 300 includes: a first calculation unit and a generation unit.

[0130] The first calculation unit is used to calculate the disparity amplitude values ​​of the matching points in the x and y directions.

[0131] The generating unit is configured to generate an initial disparity map according to the disparity amplitude value.

[0132] Optionally, in one embodiment of the present invention, the 3D reconstruction module 600 includes: a conversion unit and a sampling unit.

[0133] The conversion unit is used to convert the final disparity map into an elevation map based on the disparity-elevation scale coefficient.

[0134] The sampling unit is used to generate three-dimensional point cloud data through pixel-by-pixel sampling based on the elevation map.

[0135] Optionally, in one embodiment of the present invention, the final disparity map is represented by:

[0136] ,

[0137] in, is the position of the matching point after compensation, is the disparity value of the corresponding position.

[0138] Optionally, in one embodiment of the present invention, the second calculation module 500 includes: an interpolation unit, a second calculation unit and a fitting unit.

[0139] The interpolation unit is used to interpolate the terrain height value of the corresponding position based on the sparse three-dimensional point cloud.

[0140] The second calculation unit is used to calculate the elevation value above the ground surface of the corresponding position of the sparse three-dimensional point cloud according to the terrain height value.

[0141] The fitting unit is used to interpolate the disparity value of the corresponding position based on the sparse three-dimensional point cloud, and fit the disparity and the scale factor above the ground based on the height value above the ground and the disparity value to obtain the disparity height scale coefficient.

[0142] Optionally, in one embodiment of the present invention, the expression of the three-dimensional point cloud data is:

[0143] ,

[0144] in, The theoretically correct geographical location after compensation; is the disparity value; is the parallax elevation scale coefficient obtained by fitting.

[0145] It should be noted that the aforementioned explanation of the embodiment of the method for three-dimensional reconstruction of urban buildings is also applicable to the device for three-dimensional reconstruction of urban buildings of the embodiment, which will not be repeated here.

[0146] The device for three-dimensional reconstruction of urban buildings according to the embodiment of the present application can utilize various data such as multi-view satellite orthographic images and spaceborne laser altimetry data, avoid relying on expensive and limited surveying and mapping level remote sensing data, reduce the cost and difficulty of data acquisition, and improve the applicability and generalizability. Precise geometric registration improves the accuracy of parallax calculation, and the parallax data are optimized in combination with azimuth information to enhance the reliability of the parallax map. With the aid of spaceborne laser altimetry data and digital terrain model, the parallax is accurately converted into elevation to generate high-precision three-dimensional point cloud data. Finally, the three-dimensional point cloud and two-dimensional building vector data are combined to generate detailed and accurate three-dimensional building vector data, which provides strong data support and decision basis for many fields such as urban planning.

[0147] Figure 5 The electronic device provided in the embodiment of the present application has the structure shown in the structural schematic diagram of the electronic device. The electronic device can include:

[0148] The memory 501, the processor 502, and the computer program stored in the memory 501 and executable on the processor 502.

[0149] The processor 502 implements the method for three-dimensional reconstruction of urban buildings provided in the above embodiments when executing the program.

[0150] Further, the electronic device further includes:

[0151] The communication interface 503 is used for communication between the memory 501 and the processor 502.

[0152] The memory 501 is used to store the computer program executable on the processor 502.

[0153] The memory 501 can include a high-speed RAM memory, and can also include a non-volatile memory such as at least one disk memory.

[0154] If the memory 501, processor 502, and communication interface 503 are implemented independently, the communication interface 503, memory 501, and processor 502 can be interconnected via a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Buses can be divided into address buses, data buses, control buses, etc. For ease of representation, Figure 5 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0155] Optionally, in a specific implementation, if the memory 501, the processor 502 and the communication interface 503 are integrated on a chip, the memory 501, the processor 502 and the communication interface 503 can communicate with each other through an internal interface.

[0156] The processor 502 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention.

[0157] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which implements the above-mentioned method for three-dimensional reconstruction of urban buildings when executed by a processor.

[0158] An embodiment of the present invention further provides a computer program, which is executed to implement the above method for three-dimensional reconstruction of urban buildings.

[0159] In the description of this specification, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0160] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of technical features indicated. Thus, a feature specified as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "N" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0161] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or N executable instructions for implementing a custom logical function or step of a process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may be performed out of the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention pertain.

[0162] The logic and / or steps represented in a flowchart or otherwise described herein, for example, can be considered a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" is any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (not exhaustive) of computer-readable media include: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic devices, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program can be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing it in other suitable ways as necessary, and then storing it in a computer memory.

[0163] It should be understood that various components of the present invention may be implemented using hardware, software, firmware, or a combination thereof. In the aforementioned embodiment, the N steps or methods may be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented using hardware, as in another embodiment, it may be implemented using any one or a combination of the following technologies known in the art: discrete logic circuits having logic gates for implementing logic functions on data signals, application-specific integrated circuits having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0164] Those skilled in the art will understand that all or part of the steps in the method of the above embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0165] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing module, or each unit may exist physically separately, or two or more units may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or in the form of software functional modules. If the integrated modules are implemented in the form of software functional modules and sold or used as independent products, they may also be stored in a computer-readable storage medium.

[0166] The storage medium mentioned above may be a read-only memory, a magnetic disk, or an optical disk, etc. Although the embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and are not to be construed as limiting the present invention. Persons skilled in the art may make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for 3D reconstruction of urban buildings, characterized in that: The following steps are involved: Acquire multi-view satellite orthophotos from the same scene, spaceborne laser altimetry data of the target area, digital terrain models, and 2D building vector data; Performing precise geometric registration on the multi-view satellite orthophotos so that the positions of ground points in the images are consistent in images of different viewing angles, thereby obtaining multi-view orthophotos; generating an initial disparity map representing the displacement characteristics of matching points based on the multi-view orthophotos, and performing a forward intersection on the matching points in combination with the azimuth information of the multi-view orthophotos to generate a final disparity map; Extracting a sparse three-dimensional point cloud based on the satellite-borne laser altimetry data, and calculating the elevation above the ground surface of corresponding locations using the digital terrain model and the sparse three-dimensional point cloud; Obtaining a disparity value of a corresponding position in the final disparity map, and obtaining a disparity elevation ratio coefficient based on the elevation value above the ground surface and the disparity value; Three-dimensional point cloud data is generated according to the parallax elevation scale coefficient, and three-dimensional building vector data of the target area is generated by combining the three-dimensional point cloud data and the two-dimensional building vector data.

2. The method for 3D reconstruction of urban buildings according to claim 1, characterized in that: Generating an initial disparity map representing the displacement characteristics of matching points according to the multi-view orthophotos includes: Calculating the disparity amplitude values ​​of the matching points in the x and y directions; The initial disparity map is generated according to the disparity amplitude value.

3. The method for 3D reconstruction of urban buildings according to claim 1, wherein: Generating three-dimensional point cloud data according to the parallax elevation scale coefficient includes: Converting the final disparity map into an elevation map based on the disparity-elevation scaling factor; The three-dimensional point cloud data is generated by pixel-by-pixel sampling based on the elevation map.

4. The method for 3D reconstruction of urban buildings according to claim 1, wherein: The expression formula of the final disparity map after compensating the positions of the matching points is: ， in, is the position of the matching point after compensation, is the disparity value of the corresponding position.

5. The method for 3D reconstruction of urban buildings according to claim 1, characterized in that: The parallax elevation ratio coefficient obtained based on the elevation value above the ground surface and the parallax value includes: Interpolating terrain height values ​​at corresponding locations based on the sparse three-dimensional point cloud; Calculating the elevation above the ground surface of the corresponding position of the sparse three-dimensional point cloud according to the terrain height value; The disparity value of the corresponding position is interpolated based on the sparse three-dimensional point cloud, and the disparity and the scale factor above the ground are fitted based on the height above the ground and the disparity value to obtain the disparity elevation scale coefficient.

6. The method for 3D reconstruction of urban buildings according to claim 1, characterized in that: The expression of the three-dimensional point cloud data is: ， in, The theoretically correct geographical location after compensation; is the disparity value of the corresponding position; is the parallax elevation scale coefficient obtained by fitting.

7. A three-dimensional reconstruction device for urban buildings, characterized in that: include: The acquisition module is used to obtain multi-view satellite orthophotos from the same scene, spaceborne laser altimetry data of the target area, digital terrain models and two-dimensional building vector data; A matching module is used to perform precise geometric registration on the multi-view satellite orthophotos so that the positions of ground points in the images are consistent in images of different viewing angles, thereby obtaining multi-view orthophotos; a generating module, configured to generate an initial disparity map representing the displacement characteristics of matching points based on the multi-view orthophotos, and to perform forward intersection on the matching points in combination with the azimuth information of the multi-view orthophotos to generate a final disparity map; A first computing module is configured to extract a sparse three-dimensional point cloud based on the satellite-borne laser altimetry data, and calculate the elevation above the ground surface of a corresponding position using the digital terrain model and the sparse three-dimensional point cloud; A second calculation module is configured to obtain a disparity value of a corresponding position in the final disparity map, and obtain a disparity-elevation ratio coefficient based on the elevation value above the ground surface and the disparity value; A three-dimensional reconstruction module is used to generate three-dimensional point cloud data according to the parallax elevation scale coefficient, and to generate three-dimensional building vector data of the target area by combining the three-dimensional point cloud data and the two-dimensional building vector data.

8. An electronic device, characterized in that: include: 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 method for three-dimensional reconstruction of urban buildings according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the three-dimensional reconstruction method of urban buildings according to any one of claims 1 to 6.

Images