A method for collecting and analyzing big data of urban three-dimensional building space

Abstract

The present invention discloses a method for collecting and analyzing big data of urban three-dimensional building spaces, including: collecting target building data, performing visualization processing on point cloud data, collecting multi-angle images again, fusing BIM model data with image models, obtaining global metadata of urban remote sensing images in a database, preliminarily screening building data, importing a building data set into a fusion model, performing seamless data docking and lossless integration of attributes according to longitude and latitude coordinates to form a complete model, and finally organizing the complete model to generate a data terminal, constituting a full-coverage urban three-dimensional building image data solution, and inputting it into a cloud server to provide an access function. The method for collecting and analyzing big data of urban three-dimensional building spaces of the present invention adopts multi-source data collection and retains the attribute information of each model, so as to achieve seamless data docking and lossless integration of attributes, making the data more complete and rich, and ensuring full coverage of urban three-dimensional buildings.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and particularly to a method for collecting and analyzing big data of urban three-dimensional building spaces. Background Art

[0002] A high-precision spatial big database of cities with three-dimensional buildings as the smallest unit has always been the core foundation of spatial planning and even urban planning. This database has characteristics such as being difficult to obtain, having a long cycle, and small samples. In the research of imaging technology and data mining, discriminating various data elements for different types of data sources is an important part. The establishment of a big database of urban three-dimensional buildings is related to important aspects such as professional information recognition, data extraction, dynamic change prediction, and comprehensive map production. This database is of great significance for urban planning and design, auxiliary decision-making needs, and urban management needs. In the establishment of the database, reliable collection and analysis methods are required.

[0003] The "Comprehensive Collection and Analysis Method of High-Precision Spatial Big Data of Urban Three-Dimensional Buildings" with the patent number "CN201710997120.7" deals with the processing of a large amount of spatial form data, enabling the rapid and efficient acquisition of high-precision spatial big data of urban three-dimensional buildings and the measurement of spatial structure elements, and realizing the comprehensive collection and information synthesis of basic data for urban spatial analysis based on an artificial intelligence system. However, it mainly collects data on terrains, traffic, etc. of buildings, and the collection means generally only utilize infrared remote sensing sub-band technology, lacking multi-source data sources, and also lacking means for data analysis and reconstruction, making it difficult for the collected images to include rich data and difficult to ensure their clarity and accuracy. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for collecting and analyzing big data of urban three-dimensional building spaces. This method retains the attribute information of each model, thereby enabling seamless docking of data, lossless integration of attributes, making the data more complete and rich, and ensuring full coverage of urban three-dimensional buildings.

[0005] To solve the above technical problems, the present invention adopts the following technical solution: A method for collecting and analyzing big data of urban three-dimensional building spaces, the method includes the following content:

[0006] S110. Collect data of the target building, use three-dimensional laser to scan the target building on-site, collect complete original images of the target building, and obtain point cloud data;

[0007] S120. Perform visualization processing on the point cloud data, based on BIM modeling application, import the point cloud data, and obtain BIM model data;

[0008] S130. Perform multi-angle image acquisition, then perform high-resolution processing on the multi-angle images, construct a triangular network to fuse the high-resolution images, and obtain an image model;

[0009] S140. Integrate the BIM model data with the image model, and in the same three-dimensional geographical environment, retain the attribute information of each model to obtain a fused model;

[0010] S150. Obtain the global metadata of urban remote sensing images in the database, and use the characteristics of building features as the screening and classification criteria to screen out the building data in the remote sensing images;

[0011] S160. Conduct a preliminary screening of the building data to ensure that each longitude and latitude contains one scene of images, and then remove duplicates and supplement them to obtain a set of building image data covering the entire region;

[0012] S170. Import the set of building data into the fused model, perform seamless data docking and lossless attribute integration according to the longitude and latitude coordinates to form a complete model;

[0013] S180. Organize the complete model to generate a data terminal, form a data solution for the full-coverage urban three-dimensional building images, and input it to the cloud server to provide an access function.

[0014] For the above-mentioned method for collecting and analyzing urban three-dimensional building space big data, in the S110, it specifically includes using an aerial device to carry a three-dimensional laser scanner to perform on-site scanning and collect complete original images of the target building. The original images include the position, orientation, angle, distance, time, and intensity data of the target building.

[0015] For the above-mentioned method for collecting and analyzing urban three-dimensional building space big data, in the S120, the visualization processing of the point cloud data includes splicing, denoising, classifying, and coloring the point cloud data, and then based on the BIM modeling technology, importing the point cloud data into Revit software for fine modeling, and finally obtaining the BIM model data.

[0016] For the above-mentioned method for collecting and analyzing urban three-dimensional building space big data, in the S130, obtaining multi-angle images includes obtaining the vertical angle of the target building and the image data of the four side view angles of the target building, and then performing high-resolution processing on the multi-angle images. The high-resolution processing includes using the VCA algorithm to extract the noisy endmember spectra in the video image, performing denoising on the extracted endmember spectra using the SSA algorithm, and then using the super-resolution algorithm SRCNN. Based on the convolutional neural network, with the original video image as the input, the resolution is increased to the set value through convolutional operations to obtain the super-resolution output, enhancing the resolution of the video image, and thus extracting the building side texture information.

[0017] For the foregoing method for collecting and analyzing big data of urban three-dimensional building space, in S130, after performing high-resolution processing on the image, through operations of geometric correction and bundle adjustment, dense point cloud data with elevation is obtained. After thinning, a continuous TIN triangular network is constructed. Finally, the image is pasted onto the triangular network to obtain an image model.

[0018] For the foregoing method for collecting and analyzing big data of urban three-dimensional building space, in S140, the fusion of the BIM model data and the image model is to, in the same three-dimensional geographical environment, and by using the method of layer management, retain the attribute information of each model, and also include the overall information volume to obtain a fusion model.

[0019] For the foregoing method for collecting and analyzing big data of urban three-dimensional building space, in S150, it specifically includes obtaining the global metadata of urban remote sensing images in the aerial remote sensing database. The global metadata of the remote sensing images includes data spatio-temporal resolution, reception date, projection type, geometric correction accuracy, radiometric correction parameters, and cloud cover information. Among them, the building data in the remote sensing images is screened out, and the spatio-temporal resolution, geometric correction accuracy, radiometric correction parameters, and cloud cover information are retained.

[0020] For the foregoing method for collecting and analyzing big data of urban three-dimensional building space, in S160, the deduplication and supplementation specifically means that when there are multiple scenes of images with the same longitude and latitude and the same cloud cover, using the screening method of pairwise comparison, the duplicate images are deleted, and it is judged whether the overall image can fully cover the regional buildings. When it can be satisfied, the final building image data set is retained; when it cannot be satisfied, the data in S150 is used for supplementation to achieve full coverage of the regional buildings, and a building image data set with full regional coverage is obtained.

[0021] For the foregoing method for collecting and analyzing big data of urban three-dimensional building space, in S170, it specifically uses the convenient import mechanism provided by the WebGIS application software to import the target building data set into the fusion model, perform seamless data docking and lossless integration of attributes according to the longitude and latitude coordinates, form a complete model, and provide data query statistics and spatial marking functions.

[0022] For the foregoing method for collecting and analyzing big data of urban three-dimensional building space, in S180, the data terminal is a large database with a model as the display unit, which is input to the cloud server, provides an access function, and synchronously provides an upload data channel for subsequent data collection and filling.

[0023] Compared with the prior art, the beneficial effects of the present invention are:

[0024] 1. The present invention uses three-dimensional laser to obtain point cloud data and constructs BIM model data, which facilitates the direct utilization of point cloud data. It uses drones to cooperate with oblique cameras to collect images and construct image models. Through the global metadata of urban remote sensing images, the building data in the remote sensing images is screened out. In summary, multi-source data is collected, and the attribute information of each model is retained, so that the data is seamlessly docked and the attributes are losslessly integrated, making the data more complete and rich, and ensuring full coverage of urban three-dimensional buildings.

[0025] 2. After the present invention collects point cloud data, it performs stitching, denoising, classification, and coloring processing to improve the visualization effect of the point cloud. After obtaining multi-angle images, it uses the VCA algorithm for denoising and unmixing, and uses the super-resolution algorithm SRCNN to enhance the resolution of video images, so as to extract the texture information of the building side. After obtaining the building data in the remote sensing images, it screens and deletes duplicate images to achieve full coverage of regional buildings. In summary, data is analyzed and processed in multiple steps to ensure the clarity and accuracy of the collected data and reduce errors.

[0026] 3. The present invention constitutes the collected data into a complete model, organizes and generates a data terminal, forms a large database with the model as the display unit, and inputs it to the cloud server and provides an access function, which is convenient for the direct adoption of data in projects such as urban planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 is the flowchart of the method of the present invention;

[0028] The present invention will be further described below in conjunction with the drawings and specific embodiments. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0029] Embodiment 1 of the present invention: A method for collecting and analyzing big data of urban three-dimensional building space, including the following contents:

[0030] S110. Collect the target building data, use three-dimensional laser to scan the target building on site, collect the complete original images of the target building, and obtain point cloud data;

[0031] S120. Perform visualization processing on the point cloud data. Based on BIM modeling application, import the point cloud data to obtain BIM model data;

[0032] S130. Collect multi-angle images, then perform high-resolution processing on the multi-angle images, construct a triangular mesh to fuse high-resolution images, and obtain an image model;

[0033] S140. Fuse the BIM model data and the image model, count in the same three-dimensional geographical environment, retain the attribute information of each model, and obtain a fusion model;

[0034] S150. Obtain the global metadata of urban remote sensing images in the database, and use the characteristics of building features as the screening and classification criteria to screen out the building data in the remote sensing images;

[0035] S160. Conduct a preliminary screening of the building data to ensure that each longitude and latitude contains one scene of images. Then, remove duplicates and supplement to obtain a set of building image data covering the entire region;

[0036] S170. Import the set of building data into the fusion model, and perform seamless data docking and lossless attribute integration according to the longitude and latitude coordinates to form a complete model;

[0037] S180. Organize the complete model to generate a data terminal, form a data solution for full-coverage urban 3D building images, and input it to the cloud server to provide access functions.

[0038] In this embodiment, first, an aerial device, such as a drone, is used to perform on-site scanning of the target building by means of 3D laser to obtain point cloud data. Then, a BIM model data is constructed to facilitate the direct utilization of the point cloud data. By using a drone, it is possible to obtain images collected by an oblique camera with vertical and four side-looking angle settings, and use these images to construct an image model. Through the global metadata of urban remote sensing images, the building data in the remote sensing images is screened out. By using multi-source data collection, the attribute information of each model is retained, so that data seamless docking and lossless attribute integration are achieved, the data is more complete and rich, and full coverage of urban 3D buildings is ensured.

[0039] According to the aforementioned method for collecting and analyzing urban 3D building spatial big data, as Figure 1 shown, in S110, it specifically includes using an aerial device to carry a 3D laser scanner for on-site scanning to collect complete original images of the target building. The original images include the position, orientation, angle, distance, time, and intensity data of the target building. In this embodiment, the point cloud data obtained from the original images of the target building has accurate spatial information, which is convenient for subsequent accurate analysis of the target building.

[0040] Embodiment 2 of the present invention: A method for collecting and analyzing urban 3D building spatial big data, as Figure 1 shown, in S120, the visualization processing of the point cloud data includes splicing, denoising, classifying, and coloring the point cloud data. Then, based on the BIM modeling technology, the point cloud data is imported into Revit software for fine modeling, and finally, BIM model data is obtained. In this embodiment, modeling is performed based on the processed data information. Ensure that the modeling effect is more in line with the actual situation, making the analysis result more authentic.

[0041] Embodiment 3 of the present invention: A method for collecting and analyzing urban 3D building spatial big data, asFigure 1 As shown, in S130, obtaining multi-angle images includes obtaining the influence data of the vertical angle of the target building and the four side view angles of the target building, and then performing high-resolution processing on the multi-angle images. The high-resolution processing includes using the VCA algorithm to extract the noisy endmember spectra in the video image, denoising the extracted endmember spectra using the SSA algorithm, and then using the super-resolution algorithm SRCNN. Based on the convolutional neural network, with the original video image as the input, the resolution is increased to the set value through convolutional operations to obtain the super-resolution output, enhancing the resolution of the video image, and thus extracting the building side texture information. In this embodiment, an oblique camera is set on the drone from one vertical and four side view angles to collect images and obtain multi-angle images. Then, high-resolution processing is performed on the images. First, the VCA algorithm is used to extract the noisy endmember spectra in the video image, and the SSA algorithm is used to denoise the extracted endmember spectra. The endmember extraction algorithm is used to extract the endmember spectrum Y, calculate its trajectory matrix X, and then perform singular value decomposition, grouping, and reconstruction to obtain the denoised and unmixed image. Then, the super-resolution algorithm SRCNN is used. Based on the convolutional neural network, with the original video image as the input, the resolution is increased to the set value using the upsampling algorithm, and then convolutional operations with 9*9*128, 3*3*64, and 5*5 are performed in three layers to obtain the super-resolution output, enhancing the resolution of the video image, and thus extracting the building side texture information.

[0042] According to the foregoing method for collecting and analyzing urban three-dimensional building space big data, as Figure 1 shown, in S130, after performing high-resolution processing on the images, through geometric correction and joint adjustment operations, dense point cloud data with elevation is obtained. After thinning, a continuous TIN triangular network is constructed, and finally the image is pasted onto the triangular network to obtain the image model. In this embodiment, after processing the data, through geometric correction and joint adjustment operations, dense point cloud data with elevation is obtained, thinned, and then a continuous TIN triangular network is constructed. Finally, the image is pasted onto the triangular network to finally obtain the image model. Since an oblique camera is set on the drone from one vertical and four side view angles to collect images, construct the image model, use the VCA algorithm for denoising and unmixing, and use the super-resolution algorithm SRCNN to enhance the resolution of the video image, thereby extracting the building side texture information, the data includes side texture information, diversifying the data, and comprehensively reflecting the characteristics of urban three-dimensional buildings.

[0043] Embodiment 3 of the present invention: In a method for collecting and analyzing urban three-dimensional building space big data, as Figure 1As shown, in S140, the integration of BIM model data and image model is to place them in the same three-dimensional geographical environment and use layer management to retain the attribute information of each model and also include the overall information volume to obtain a fused model. In this embodiment, the BIM model data and the image model are integrated, placed in the same three-dimensional geographical environment, and layer management is used. This not only retains the attribute information of each model but also includes the overall information volume to obtain a fused model. The BIM technology can run through the entire process of construction from beginning to end, providing more refined data support for each stage of the process and realizing informatization and intelligentization at the micro level. The method of combining BIM technology and oblique photography technology is to fuse the BIM three-dimensional model and the image model created by the oblique photography technology, place them in the same three-dimensional geographical environment, and use layer management. This not only retains the attribute information of each model but also enriches the information volume of the three-dimensional model spatial database, providing more comprehensive comprehensive basic data for urban management.

[0044] According to the aforementioned method for collecting and analyzing urban three-dimensional building space big data, as Figure 1 shown, in S150, it specifically includes obtaining the global metadata of urban remote sensing images in the aerial remote sensing database. The global metadata of the remote sensing images includes data spatio-temporal resolution, reception date, projection type, geometric correction accuracy, radiation correction parameters, and cloud cover information. Among them, the building data in the remote sensing images is screened out, and the spatio-temporal resolution, geometric correction accuracy, radiation correction parameters, and cloud cover information are retained. In this embodiment, after obtaining the data information, the building data in the remote sensing images is then screened out using the building feature as the screening and classification criterion, and the spatio-temporal resolution, geometric correction accuracy, radiation correction parameters, and cloud cover information are retained to achieve full coverage of regional buildings. In summary, data is analyzed and processed in multiple steps to ensure the clarity and accuracy of the collected data and reduce errors.

[0045] According to the aforementioned method for collecting and analyzing urban three-dimensional building space big data, as Figure 1As shown, in S160, the duplicate removal and supplementation are specifically as follows: when there are multiple images with the same longitude and latitude and the same cloud coverage, a pairwise comparison screening method is used to delete the duplicate images, and it is judged whether the overall images can fully cover the regional buildings. When it can be satisfied, the final building image data set is retained; when it cannot be satisfied, the data in S150 is used for supplementation to achieve full coverage of the regional buildings, and a building image data set with full regional coverage is obtained. In this embodiment, the building data is initially screened to ensure that there is one image for each longitude and latitude, and then duplicate removal and supplementation are performed. Specifically, when there are multiple images with the same longitude and latitude and the same cloud coverage, a pairwise comparison screening method is used to delete the duplicate images. Then, it is judged whether the overall images can fully cover the regional buildings. When it can be satisfied, the final building image data set is retained; when it cannot be satisfied, the data in step five is used for supplementation to achieve full coverage of the regional buildings, and a building image data set with full regional coverage is obtained. Through the global metadata of urban remote sensing images, the building data in the remote sensing images is screened, initially screened to ensure that there is one image for each longitude and latitude, and the duplicate images are deleted to achieve full coverage of the regional buildings, which not only ensures the integrity and accuracy of the data but also further enriches the data of the overall model.

[0046] According to the foregoing method for collecting and analyzing urban three-dimensional building space big data, as Figure 1 shown, in S170, specifically, a portable import mechanism provided by WebGIS application software is used to import the target building data set into the fusion model, and data seamless docking and attribute lossless integration are performed according to longitude and latitude coordinates to form a complete model, providing data query statistics and spatial marking functions. In this embodiment, by importing the data into the fusion model, the reconstruction and direct utilization of the data are realized, resources are saved, and the integrity, security, and consistency of the data are ensured, providing more comprehensive comprehensive basic data for urban management. Ensure the clarity and accuracy of the collected data and reduce errors.

[0047] According to the foregoing method for collecting and analyzing urban three-dimensional building space big data, as Figure 1As shown, in S180, the data terminal is a large database with a model as the display unit, which is input into the cloud server to provide an access function, and simultaneously provides an upload data channel for subsequent data collection and filling. In this embodiment, the complete model is sorted out to generate a data terminal, which is a large database with a model as the display unit, constituting a full-coverage urban three-dimensional building image data solution, input into the cloud server to provide an access function, and simultaneously providing an upload data channel for subsequent data collection and filling. The present invention forms a complete model with the collected data, sorts it out to generate a data terminal, constitutes a large database with a model as the display unit, and inputs it into the cloud server to provide an access function, which is convenient for directly using the data in projects such as urban planning.

[0048] The principle of the method for collecting and analyzing the big data of the urban three-dimensional building space of the present invention is as follows. Refer to Figure 1 , this method uses a drone to perform on-site scanning by means of three-dimensional laser to obtain point cloud data and construct BIM model data for the direct utilization of the point cloud data. An oblique camera is set on the drone to collect images from one vertical and four side views to construct an image model. Through the global metadata of urban remote sensing images, the building data in the remote sensing images is screened out, and multi-source data is collected, retaining the attribute information of each model, so that the data is seamlessly docked and the attributes are losslessly integrated, and the data is more complete and rich, ensuring full coverage of urban three-dimensional buildings. After the present invention collects the point cloud data, it performs stitching, denoising, classification, and coloring processing to improve the visualization effect of the point cloud to highlight the feature information. After obtaining multi-angle images, the VCA algorithm is used for denoising and unmixing, and the super-resolution algorithm SRCNN is used to enhance the resolution of the video image, so as to extract the texture information of the building side. After obtaining the building data in the remote sensing image, preliminary screening is performed to ensure that each longitude and latitude contains a scene image and duplicate images are deleted to achieve full coverage of regional buildings. In summary, the data is analyzed and processed in multiple steps to ensure the clarity and accuracy of the collected data and reduce errors. At the same time, the present invention forms a complete model with the collected data, sorts it out to generate a data terminal, constitutes a large database with a model as the display unit, and inputs it into the cloud server to provide an access function, which is convenient for directly using the data in projects such as urban planning.

Claims

1. A method for collecting and analyzing big data of urban three-dimensional building space, characterized in that, The method includes the following steps: S110. Collect target building data. Use 3D laser to conduct on-site scanning of the target building, collect complete original images of the target building, and obtain point cloud data; S120. Conduct visualization processing on the point cloud data. Based on BIM modeling application, import the point cloud data to obtain BIM model data; S130. Collect multi-angle images, then perform high-resolution processing on the multi-angle images, construct a triangular network to fuse high-resolution images, and obtain an image model; S140. Fuse the BIM model data and the image model, count in the same 3D geographical environment, retain the attribute information of each model, and obtain a fused model; S150. Obtain the global metadata of urban remote sensing images in the database, and use the building feature as the screening and classification criterion to screen out the building data in the remote sensing images; S160. Conduct preliminary screening on the building data to ensure that each longitude and latitude contains one image, and then remove duplicates and supplement to obtain a collection of building image data covering the entire region; S170. Import the building data set into the fused model, perform seamless data docking and lossless attribute integration according to longitude and latitude coordinates to form a complete model; S180. Organize the complete model to generate a data terminal, form a data solution for the full-coverage urban 3D building images, input it into the cloud server, and provide an access function; In the step S130, obtaining multi-angle images includes obtaining the impact data of the vertical angle of the target building and the four side view angles of the target building, and then performing high-resolution processing on the multi-angle images. The high-resolution processing includes using the VCA algorithm to extract the noisy endmember spectrum in the video image, denoising the extracted endmember spectrum using the SSA algorithm, and then using the super-resolution algorithm SRCNN. Based on the convolutional neural network, with the original video image as the input, the resolution is increased to the set value through convolutional operations to obtain a super-resolution output, enhancing the resolution of the video image, so as to extract the texture information of the building side; In the step S130, after performing high-resolution processing on the image, through geometric correction and joint adjustment operations, dense point cloud data with elevation is obtained. After thinning, a continuous TIN triangular network is constructed, and finally the image is pasted onto the triangular network to obtain an image model.

2. The method for collecting and analyzing big data of urban three-dimensional building space according to claim 1, characterized in that, In the step S110, it specifically includes using an aerial device to carry a 3D laser scanner to conduct on-site scanning, collect complete original images of the target building, and the original images include the position, orientation, angle, distance, time, and intensity data of the target building.

3. The method for collecting and analyzing big data of urban three-dimensional building space according to claim 1, characterized in that, In the step S120, the visualization processing of the point cloud data includes splicing, denoising, classifying, and coloring the point cloud data, and then based on BIM modeling technology, importing the point cloud data into Revit software for fine modeling, and finally obtaining BIM model data.

4. The method for collecting and analyzing big data of urban three-dimensional building space according to claim 1, characterized in that, In the step S140, the fusion of the BIM model data and the image model is to, in the same 3D geographical environment, and using the layer management method, retain the attribute information of each model, and also include the overall information volume to obtain a fused model.

5. The method for collecting and analyzing big data of urban three-dimensional building space according to claim 1, characterized in that, In the S150, it specifically includes obtaining the global metadata of urban remote sensing images in the aerial remote sensing database. The global metadata of the remote sensing images includes data spatio-temporal resolution, reception date, projection type, geometric correction accuracy, radiometric correction parameters, and cloud cover information. Among them, the building data in the remote sensing images is screened out, and the spatio-temporal resolution, geometric correction accuracy, radiometric correction parameters, and cloud cover information are retained.

6. The method for collecting and analyzing big data of urban three-dimensional building space according to claim 1, characterized in that, In the S160, the duplicate removal and supplementation specifically means that when there are multiple scenes of images with the same longitude and latitude and the same cloud cover, a pairwise comparison screening method is used to delete the duplicate images. It is judged whether the overall image can fully cover the regional buildings. When it can be satisfied, the final building image data set is retained; when it cannot be satisfied, the data in S150 is used for supplementation to achieve full coverage of the regional buildings, and a building image data set with full regional coverage is obtained.

7. The method for collecting and analyzing big data of urban three-dimensional building space according to claim 1, characterized in that, In the S170, it specifically means using the convenient import mechanism provided by the WebGIS application software to import the target building data set into the fusion model, and performing seamless data docking and lossless attribute integration according to the longitude and latitude coordinates to form a complete model, and providing data query and statistics, and spatial marking functions.

8. The method for collecting and analyzing big data of urban three-dimensional building space according to claim 1, characterized in that, In the S180, the data terminal is a large database with the model as the display unit, which is input to the cloud server, provides an access function, and simultaneously provides an upload data channel for subsequent data collection and filling.

Images