3D modeling method based on digital twin cities

Abstract

The invention belongs to the technical field of three-dimensional city modeling, and discloses a digital twin city-based 3D modeling method, which specifically comprises the following steps of: 1, data acquisition: dividing different regions of a city into five levels, namely a first level, a second level, a third level, a fourth level and a fifth level according to the importance degrees of the different regions of the city, and according to the sequence from the first stage to the fifth stage from top to bottom. According to the method, the urban areas are divided according to the importance degree, three-dimensional image data acquisition is carried out according to the priorities of the urban areas, then the initial three-dimensional urban area model is constructed according to the acquired three-dimensional image data, and the initial three-dimensional urban area model and the existing three-dimensional urban model are superposed. And finally, after three-dimensional model reconstruction is carried out on the changed area, the changed area is integrated with an existing 3D model, so that a brand new 3D city model can be obtained, and the updating efficiency of the city model can be improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of three-dimensional urban modeling, and specifically relates to a 3D modeling method based on a digital twin city. Background Technique

[0002] A digital twin city refers to the digitalization of all elements such as people, objects, events, and water, electricity, and gas in the physical city by using Internet of Things (IoT) technology on the basis of Building Information Modeling (BIM) and urban three-dimensional Geographic Information System (GIS), and then constructing a completely corresponding "virtual city" in the cyber space. This technology constructs a city-level data closed-loop empowerment system through data global identification, accurate state perception, real-time data analysis, scientific model decision-making, and intelligent and accurate execution, realizing city simulation, monitoring, diagnosis, prediction, and control, thereby eliminating the random uncertainty in urban planning, construction, operation, management, and services.

[0003] With the continuous development of three-dimensional modeling technology, the quality of urban modeling is getting higher and higher. However, in the actual use process, due to the continuous development and progress of the city, the buildings in the city are constantly changing. In order to ensure the accuracy of the city model, the model needs to be updated regularly. However, the existing updates generally involve the reconstruction of the city model. Due to the large area of the city, the full reconstruction of the city model is not only time-consuming and laborious, but also costly, which is not conducive to the maintenance of the city model. Therefore, it is necessary to improve this. Summary of the Invention

[0004] The purpose of the present invention is to provide a 3D modeling method based on a digital twin city to solve the problems raised in the above background technique.

[0005] In order to achieve the above purpose, the present invention provides the following technical solution: A 3D modeling method based on a digital twin city, the specific steps are as follows:

[0006] Step 1: Data collection

[0007] According to the importance of different regions of the city, different regions of the city are divided into five levels: the first level, the second level, the third level, the fourth level, and the fifth level. In the order from high to low of the first level to the fifth level, an unmanned aerial vehicle is used for aerial photography to collect stereo image data of the corresponding regions;

[0008] Step 2: Construct an initial three-dimensional urban area model

[0009] Identify the stereoscopic image data based on the features and attributes of the images, eliminate non-building data and invalid data, then perform data cleaning and denoising on the collected stereoscopic image data. Subsequently, perform data format conversion on the processed stereoscopic image data. Finally, use professional aerial 3D modeling software to process the data to generate an initial three-dimensional urban area model;

[0010] Step Three: Model Overlay and Coincidence

[0011] Import the existing 3D city model into 3DS MAX, then import the initial three-dimensional urban area model. Immediately, perform overlay and coincidence on the initial three-dimensional urban model and the existing 3D city model, mark the non-overlapping parts between the two models and record them as the areas to be processed;

[0012] Step Four: Construction of 3D Model for Areas to be Processed

[0013] Locate the position in the actual city corresponding to the areas to be processed, then use three-dimensional laser scanning technology to obtain the spatial three-dimensional data of the real targets at the areas to be processed. Subsequently, using the point cloud data as a reference, use 3DS MAX modeling software to establish a high-precision three-dimensional model, and then combine with BIM technology to perform texture mapping and lighting simulation, thereby constructing a refined 3D model to improve the visualization effect of the model;

[0014] Step Five: Model Integration

[0015] Replace the 3D model of the areas to be processed that has been re-modeled into the corresponding areas in the existing 3D city model and integrate them to form a new 3D city model;

[0016] Step Six: Model Verification and Optimization

[0017] Compare and verify the re-constructed 3D city model with the actual city to ensure the accuracy and reliability of the model. Then, according to the verification results, optimize and adjust the model to improve the accuracy and fidelity of the model;

[0018] Step Seven: Integration into the Platform

[0019] Integrate the optimized three-dimensional model into the digital twin city platform to achieve docking with the urban management system.

[0020] As a preferred technical solution of the present invention, the importance levels of different regions of the city in Step One are divided according to the population flow, geographical location, and traffic conditions in the area. The specific calculation formula is:

[0021]

[0022] Among them, V1 represents the magnitude of the pedestrian flow per unit time, Q represents the distance from the city center, V2 represents the magnitude of the vehicle flow per unit time, and the larger the P value, the higher the level of the area.

[0023] As a preferred technical solution of the present invention, when the drone conducts aerial photography in step one, it is necessary to collect images from five different perspectives, namely one vertical perspective and four oblique perspectives, and at the same time select a weather with sufficient light and low wind speed and a time between 10 am and 4 pm. When planning the route, one or several of the surrounding method, the cross-surrounding method, and the cross method are selected.

[0024] As a preferred technical solution of the present invention, when identifying the stereo image data in step two, the SIFT algorithm is used to perform feature recognition processing on the image data. The specific process is to use the Gaussian function to perform blurring processing on the original image at different scales to generate a series of images at different scales, forming a scale space. Then, in the scale space, key points are detected through the Gaussian difference pyramid, the key points are accurately located through the Taylor expansion formula, and the feature direction is determined. Finally, the SIFT feature descriptor is generated.

[0025] As a preferred technical solution of the present invention, when performing data cleaning and denoising in step two, Gaussian filtering is used. The specific process is to generate a discrete Gaussian kernel according to the degree of image noise and the requirement of detail retention, and perform a convolution operation on the generated Gaussian kernel and the image. During the convolution process, each pixel point of the image will perform a weighted summation operation with the Gaussian kernel, and the result obtained is the Gaussian filtering value of the pixel point.

[0026] As a preferred technical solution of the present invention, during the stereo image recognition process in step two, relevant laws, regulations, and privacy policies need to be complied with, and images involving personal privacy and sensitive information are removed.

[0027] As a preferred technical solution of the present invention, when the model superposition and coincidence are performed in step three, it needs to be carried out under a unified coordinate system and scaling ratio, and at the same time ensure that the edge parts between the two models overlap.

[0028] As a preferred technical solution of the present invention, when optimizing the model in step six, deep learning algorithms are used to repair damaged three-dimensional data and optimize the rendering effect of the three-dimensional model.

[0029] As a preferred technical solution of the present invention, the digital twin city platform described in step seven provides an intuitive three-dimensional visualization interface and interactive functions, and can monitor and manage the city in real time.

[0030] The beneficial effects of the present invention are as follows:

[0031] The present invention divides urban areas according to their importance levels, so as to collect stereo image data for the priority levels of urban areas. Then, an initial three-dimensional urban area model is constructed based on the collected stereo image data, and the initial three-dimensional urban area model is superimposed and coincided with the existing 3D urban model, thereby marking the changed areas. Finally, after reconstructing the three-dimensional model for the changed areas and integrating it with the existing 3D model, a new 3D urban model can be obtained, thereby improving the update efficiency of the urban model and reducing the maintenance cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a flowchart of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0033] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0034] As Figure 1 shown, the embodiments of the present invention provide a 3D modeling method based on a digital twin city, and the specific steps are as follows:

[0035] Step 1: Data collection

[0036] According to the importance levels of different areas in the city, different areas of the city are divided into five levels: the first level, the second level, the third level, the fourth level, and the fifth level. In the order from high to low of the first level to the fifth level, an unmanned aerial vehicle is used for aerial photography to collect stereo image data of the corresponding areas;

[0037] Step 2: Construct an initial three-dimensional urban area model

[0038] Identify the stereo image data according to the features and attributes of the images, eliminate non-building data and invalid data, then perform data cleaning and denoising on the collected stereo image data, then perform data format conversion on the processed stereo image data, and finally use professional aerial photography 3D modeling software to process the data, thereby generating an initial three-dimensional urban area model;

[0039] Step 3: Model superposition and coincidence

[0040] Import the existing 3D urban model into 3DS MAX, then import the initial three-dimensional urban area model, and then superimpose and coincide the initial three-dimensional urban modeling and the existing 3D urban model, mark the non-overlapping parts between the two models and record them as areas to be processed;

[0041] Step 4: Construction of the 3D model of the area to be processed

[0042] Locate the position corresponding to the area to be processed in the actual city, then use 3D laser scanning technology to obtain the spatial 3D data of the real-world target at the area to be processed. Subsequently, using the point cloud data as a reference, use 3DS MAX modeling software to build a high-precision 3D model, and then combine with BIM technology for texture mapping and lighting simulation to construct a refined 3D model and improve the visualization effect of the model;

[0043] Step 5: Model integration

[0044] Replace the 3D model of the area to be processed after re-modeling into the corresponding area in the existing 3D city model and integrate it to form a new 3D city model;

[0045] Step 6: Model verification and optimization

[0046] Compare and verify the re-constructed 3D city model with the actual city to ensure the accuracy and reliability of the model. Then, according to the verification results, optimize and adjust the model to improve the accuracy and fidelity of the model;

[0047] Step 7: Integration into the platform

[0048] Integrate the optimized 3D model into the digital twin city platform to achieve docking with the urban management system.

[0049] By collecting the stereo image data of different urban areas, an initial 3D urban area model of the current urban area can be established. Then, overlay and coincide the initial 3D urban area model with the existing 3D city model to identify the changed areas. Immediately afterwards, by re-modeling the changed areas and integrating them with the existing 3D city model, a new 3D city model can be obtained, which can improve the update efficiency of the 3D city model, shorten the time consumed for updating the model, and reduce the model maintenance cost.

[0050] Among them, the importance levels of different areas of the city in Step 1 are divided according to the pedestrian flow, geographical location, and traffic conditions in this area. The specific calculation formula is:

[0051]

[0052] Among them, V1 represents the magnitude of the pedestrian flow per unit time, Q represents the distance from the location to the city center, V2 represents the magnitude of the vehicle flow per unit time, and the larger the P value, the higher the level of this area.

[0053] After grading urban areas according to the P values of the obtained regions, the urban areas can be divided into zones, so that data can be collected according to the size of the priorities. For some unimportant or undeveloped regions, the data collection cycle of these regions can be extended, so that the data collection volume for 3D urban modeling can be effectively reduced, and the cost of updating 3D urban modeling can be lowered.

[0054] Among them, when the drone in step one conducts aerial photography, it is necessary to collect images from five different perspectives, namely one vertical perspective and four oblique perspectives. At the same time, it is selected to conduct the aerial photography on a day with sufficient light and low wind speed and between 10 am and 4 pm. When planning the route, one or several of the circumferential method, the cross-circumferential method and the cross method are selected.

[0055] To ensure the accuracy of recognition by the modeling software, a certain overlap rate needs to be ensured between the aerial photography photos. The forward overlap rate is generally 80%, and the side overlap rate is 75-80%. When planning the flight route, it is necessary to determine the shooting height. In principle, the closer to the object, the higher the modeling fineness, but the flight height also needs to be determined according to the actual situation. At the same time, the flight parameters also need to be adjusted: photo resolution, shooting angle, timed shooting, etc., so as to ensure that a sufficient number and quality of aerial photography images are obtained.

[0056] Among them, when identifying the stereo image data in step two, the SIFT algorithm is used to perform feature recognition processing on the image data. The specific process is to use the Gaussian function to perform blurring processing on the original image at different scales to generate a series of images at different scales, forming a scale space. Then in the scale space, key points are detected through the Gaussian difference pyramid, the key points are accurately located through the Taylor expansion formula and the feature direction is determined. Finally, the SIFT feature descriptor is generated.

[0057] The SIFT algorithm aims to detect local key points in the image. These key points are points that will not disappear due to factors such as illumination, scale, and rotation, such as corner points, edge points, bright points in dark areas, and dark points in bright areas. By extracting the features of these key points, the SIFT algorithm can realize functions such as image matching, recognition, and classification.

[0058] Among them, when performing data cleaning and denoising in step two, Gaussian filtering is used. The specific process is to generate a discrete Gaussian kernel according to the degree of image noise and the requirement of detail retention, and perform a convolution operation on the generated Gaussian kernel and the image. During the convolution process, each pixel point of the image will perform a weighted summation operation with the Gaussian kernel, and the result obtained is the Gaussian filtering value of the pixel point.

[0059] Since the Gaussian kernel assigns higher weights to the central point and lower weights to pixels far from the central point, the convolution operation can smooth the image and reduce noise. At the same time, due to the characteristics of the Gaussian function, Gaussian filtering can remove noise while preserving the edge information of the image.

[0060] Among them, in the process of stereo image recognition in step two, relevant laws, regulations and privacy policies need to be complied with, and images involving personal privacy and sensitive information should be removed.

[0061] Aerial photography by drones may involve personal privacy and sensitive information. When collecting data, relevant laws, regulations and privacy policies need to be complied with to ensure that the legitimate rights and interests of others are not violated. At the same time, permission from relevant departments is required to avoid legal penalties for unauthorized flight.

[0062] Among them, when the models in step three are superimposed and coincided, it needs to be carried out under a unified coordinate system and scaling ratio, and at the same time, ensure that the edge parts between the two models overlap.

[0063] By superimposing and coinciding the two models under a unified coordinate system and scaling ratio, the changed parts of urban buildings can be quickly screened out. The overlapping part between the two models indicates that this area has not changed, while the non-overlapping part indicates that this area has changed.

[0064] Among them, when optimizing the model in step six, deep learning algorithms are used to repair damaged 3D data and optimize the rendering effect of the 3D model.

[0065] By using deep learning, the damaged parts in the 3D city model can be effectively removed, thereby further improving the accuracy and realism of the 3D model.

[0066] Among them, the digital twin city platform in step seven provides an intuitive 3D visualization interface and interactive functions, and can monitor and manage the city in real time.

[0067] Through the digital twin city platform, simulations in aspects such as urban planning, construction and management can be carried out, so as to predict urban disasters and simulate the operating status of the city, and thus provide strong data support for urban planning, construction and management.

[0068] It should be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprising", "including" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or elements inherent to such process, method, article or device.

[0069] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A 3D modeling method based on a digital twin city, characterized in that The specific steps are as follows: Step 1: Data collection According to the importance of different regions in the city, different regions of the city are divided into five levels: the first level, the second level, the third level, the fourth level, and the fifth level. In the order from high to low of the first level to the fifth level, use drones for aerial photography to collect three-dimensional image data of the corresponding regions; Step 2: Construct the initial three-dimensional urban area model Identify the three-dimensional image data according to the features and attributes of the images, eliminate non-building data and invalid data, then perform data cleaning and denoising on the collected three-dimensional image data, then perform data format conversion on the processed three-dimensional image data, and finally use professional aerial photography 3D modeling software to process the data, so as to generate the initial three-dimensional urban area model; Step 3: Model overlay and coincidence Import the existing 3D city model into 3DS MAX, then import the initial three-dimensional urban area model, and then overlay and coincide the initial three-dimensional urban modeling and the existing 3D city model, mark the non-overlapping parts between the two models and record them as the areas to be processed; Step 4: Construction of the three-dimensional model of the area to be processed Find the position corresponding to the area to be processed in the actual city, then use three-dimensional laser scanning technology to obtain the spatial three-dimensional data of the real target at the area to be processed. Subsequently, with the point cloud data as a reference, use 3DS MAX modeling software to establish a high-precision three-dimensional model, and then combine BIM technology to perform texture mapping and lighting simulation, so as to construct a refined 3D model and improve the visualization effect of the model; Step 5: Model integration Replace the 3D model of the area to be processed that has been re-modeled into the corresponding area of the existing 3D city model and integrate it to form a new 3D city model; Step 6: Model verification and optimization Compare and verify the re-constructed 3D city model with the actual city to ensure the accuracy and reliability of the model. Then, according to the verification results, optimize and adjust the model to improve the accuracy and fidelity of the model; Step 7: Integration into the platform Integrate the optimized three-dimensional model into the digital twin city platform to achieve docking with the urban management system.

2. The 3D modeling method based on the digital twin city according to claim 1, wherein: The importance of different regions in the city described in Step 1 is divided according to the pedestrian flow, geographical location, and traffic conditions in this area. The specific calculation formula is: Among them, V1 represents the magnitude of the pedestrian flow per unit time, Q represents the distance from the city center, V2 represents the magnitude of the vehicle flow per unit time, and the larger the P value, the higher the level of this area.

3. A 3D modeling method based on a digital twin city according to claim 1, characterized in that: When the drone performs aerial photography in Step 1, it is necessary to collect images from five different perspectives: one vertical perspective and four oblique perspectives. At the same time, it is selected to be carried out on a day with sufficient light and low wind speed and between 10 am and 4 pm. When planning the route, one or several of the circular method, the cross-circular method, and the cross method are selected.

4. A 3D modeling method based on a digital twin city according to claim 1, characterized in that: When identifying the stereoscopic image data in Step 2, the SIFT algorithm is used to perform feature recognition processing on the image data. The specific process is to use the Gaussian function to perform blurring processing on the original image at different scales to generate a series of images at different scales, constituting a scale space. Then, in the scale space, key points are detected through the Gaussian difference pyramid, and the key points are accurately located and the feature directions are determined through the Taylor expansion. Finally, the SIFT feature descriptor is generated.

5. A 3D modeling method based on a digital twin city according to claim 1, characterized in that: When performing data cleaning and denoising in Step 2, Gaussian filtering is used. The specific process is to generate a discrete Gaussian kernel according to the degree of image noise and the requirement of detail retention, and perform a convolution operation on the generated Gaussian kernel and the image. During the convolution process, each pixel point of the image will perform a weighted summation operation with the Gaussian kernel, and the obtained result is the Gaussian filtering value of the pixel point.

6. A 3D modeling method based on a digital twin city according to claim 1, characterized in that: During the stereoscopic image recognition process described in Step 2, relevant laws, regulations and privacy policies need to be complied with, and images involving personal privacy and sensitive information are removed.

7. A 3D modeling method based on a digital twin city according to claim 1, characterized in that: When the model superposition and coincidence are carried out in Step 3, it needs to be carried out under a unified coordinate system and scaling ratio, and at the same time ensure that the edge parts between the two models overlap together.

8. A 3D modeling method based on a digital twin city according to claim 1, characterized in that: When optimizing the model in Step 6, deep learning algorithms are used to repair damaged 3D data and optimize the rendering effect of the 3D model.

9. A 3D modeling method based on a digital twin city according to claim 1, characterized in that: The digital twin city platform described in Step 7 provides an intuitive 3D visualization interface and interactive functions, and can monitor and manage the city in real time.

Images