Real-scene three-dimensional building model device and method

Abstract

The invention relates to the technical field of live-action three-dimensional models, and provides a live-action three-dimensional model building device and method, and the device comprises a data collection module, a data processing module, a three-dimensional model generation module, a model output module, and a controller. The data acquisition module is used for acquiring live-action images and video data and transmitting the data to the data processing module; the data processing module is used for processing image and video data by utilizing computer vision and deep learning technologies after receiving the data transmitted by the data acquisition module, extracting three-dimensional space information and constructing a three-dimensional model; the three-dimensional model generation module is used for converting three-dimensional space information into a three-dimensional model by utilizing a three-dimensional model construction algorithm after receiving the data processed by the data processing module; a model output module; according to the method, the live-action three-dimensional model can be quickly and accurately established, the problems that the model texture is not clear and the overall color of the model is not coordinated are effectively solved, and the model quality is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of real scene three-dimensional models, and specifically, to a real scene three-dimensional construction model device and method. Background Art

[0002] With the continuous development of technology, three-dimensional modeling technology has become more and more mature and is gradually applied to various fields. Among them, the real scene three-dimensional construction solution is an important application. The real scene three-dimensional construction solution refers to applying three-dimensional modeling technology to the actual scene and simulating the real environment and scene through software. This technology can present scenes such as buildings, landscapes, and roads through three-dimensional modeling, with very high realism and interactivity, which can greatly improve work efficiency and play an important role in promoting the development of all walks of life.

[0003] For example, in the field of architectural design, the real scene three-dimensional construction solution can provide a more intuitive scene display for architects, so as to better understand the actual effect of the design. In the tourism industry, through the real scene three-dimensional construction solution, the real landscape can be presented to provide a real tourism experience for tourists. In urban planning, the real scene three-dimensional construction solution can simulate the real scene of the city and provide a better planning scheme for urban planners.

[0004] After retrieval, the existing patent (publication number: CN 107481321 B) discloses a three-dimensional model generation method and system. The method includes obtaining the two-dimensional position data of the two-dimensional elements corresponding to the target real scene elements and the terrain data of the target real scene elements; generating the three-dimensional position data of the target three-dimensional model required for the target real scene elements according to the two-dimensional position data and the terrain data; triangulating the two-dimensional elements corresponding to the target real scene elements; and finally generating the target three-dimensional model corresponding to the target real scene elements based on the three-dimensional position data and the triangulation result. This technical solution provides a scheme for automatically generating a three-dimensional model according to the actual two-dimensional position data and terrain data, with high data production efficiency and less model data volume. And because the actual two-dimensional position data and terrain data are used as the data source of the three-dimensional model, the generated three-dimensional model has a high position matching degree with the real scene elements, thereby improving the reference value of the three-dimensional electronic map.

[0005] However, it is found that the following problems exist in the implementation of the related technology: Generally, high-quality model results require rich, distinct, and moderately contrasted texture colors; however, due to the influence of many factors such as acquisition equipment, weather conditions, terrain and landforms, and aerial photography status, the original images are prone to problems such as grayness, uneven color, and poor saturation, which in turn affect the quality of the model texture results. Summary of the Invention

[0006] The present invention provides a real - scene three - dimensional construction model device and method, aiming to solve the problem that due to the influence of many factors such as acquisition equipment, weather conditions, topography and aerial photography status in the existing real - scene three - dimensional construction model device, the original images are prone to problems such as grayness, uneven color, poor saturation, etc., which in turn affect the quality of the model texture results.

[0007] The technical solution of the present invention is as follows: A real - scene three - dimensional construction model device includes a data acquisition module, a data processing module, a three - dimensional model generation module, a model output module and a controller;

[0008] Data acquisition module: This module is used to obtain real - scene images and video data and transmit the data to the data processing module;

[0009] Data processing module: After receiving the data transmitted by the data acquisition module, this module uses computer vision and deep - learning technologies to process the image and video data, extract three - dimensional spatial information, and construct a three - dimensional model;

[0010] Three - dimensional model generation module: After receiving the data processed by the data processing module, this module uses a three - dimensional model construction algorithm to convert the three - dimensional spatial information into a three - dimensional model;

[0011] Model output module: This module outputs the generated three - dimensional model in a certain format for subsequent applications;

[0012] Controller: This controller is used to control the working processes of the above - mentioned modules.

[0013] Further, the acquisition process of the data acquisition module is as follows: Use a laser scanner, a panoramic camera, and a drone to take multi - angle and high - precision photos of the target area to obtain all - round images and terrain data.

[0014] Further, the laser scanner measures the distance and angle of the object's surface by emitting laser light to obtain the three - dimensional coordinate data of the object.

[0015] Further, the panoramic camera stitches multiple photos into a panoramic image to obtain the two - dimensional image data of the object.

[0016] Further, the drone takes aerial photos by carrying a high - resolution camera to obtain the two - dimensional image data of the object.

[0017] Further, the processing process of the data processing module for the image and video data includes image correction, stitching, denoising, and conversion into three - dimensional model data.

[0018] Among them, the processes of image correction, stitching, and denoising are operated through image - processing software: Photoshop, GIMP.

[0019] Among them, the conversion into 3D model data is operated through geometric modeling software: AutoCAD and SketchUp.

[0020] Furthermore, the process by which the 3D model generation module converts 3D spatial information into a 3D model includes establishing a terrain model, adding textures, and setting lighting.

[0021] Furthermore, the terrain model is established by converting terrain data into a 3D terrain model through GIS technology.

[0022] Furthermore, textures are added by mapping image data onto the model surface through texture mapping technology.

[0023] Furthermore, lighting is set by setting light and material properties to simulate the lighting effects of the real environment.

[0024] Furthermore, the model output module outputs the generated 3D model in the formats of obj and fbx.

[0025] Even further, the present invention also provides a construction method for a real-scene 3D construction model, including the following usage steps:

[0026] S1, Conduct aerial oblique photography to obtain original images and down-looking POS data;

[0027] S2, Analyze and preprocess data, including correcting the distortion difference of the original images and performing equalization of light and color processing, performing aerial triangulation on the down-looking images, and making camera and image description files;

[0028] S3, Process data, perform aerial triangulation on the oblique images to obtain high-precision exterior orientation elements of all images, use the distorted-corrected image data combined with the aerial triangulation encryption results, generate an ultra-high-density point cloud through multi-view image dense matching technology, and then construct an irregular triangular network model based on the point cloud;

[0029] S4, Automatically extract the corresponding texture information on the images after distortion difference correction and equalization of light and color processing, and map it onto the triangular patches of the corresponding model to achieve texture attachment, and finally generate a 3D model result with clear and realistic textures;

[0030] S5, Output the real-scene 3D model result.

[0031] Compared with the prior art, the beneficial effects of the present invention are:

[0032] The present invention proposes a device and method for constructing a real-life three-dimensional model. The method can effectively solve the problems of unclear model texture and inconsistent overall model color, thereby improving the model quality. At the same time, compared with the current mainstream method of achieving model color balancing based on original image preprocessing, the method proposed in the present invention increases the production efficiency of model color balancing by 9.1 times, significantly shortening the production cycle of model color balancing. The device can quickly and accurately establish a real-life three-dimensional model, providing strong technical support for urban planning, smart cities, emergency rescue and other fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0034] Figure 1 This is a diagram of the implementation process of a real-scene 3D construction model proposed by the present invention; Figure 2 This is a comparison diagram of the original image pre-processing model and the model effect after the experiment of a real-scene 3D construction model proposed by the present invention; DETAILED DESCRIPTION

[0035] In order to make the technical means, creative features, objectives and effects achieved by the present invention easier to understand, the present invention is further described below in conjunction with specific implementation methods.

[0036] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of the present invention.

[0037] See also Figure 1 , the present invention provides a technical solution:

[0038] A real-scene 3D construction model device includes a data acquisition module, a data processing module, a 3D model generation module, a model output module and a controller; the data acquisition module: this module is used to obtain real-scene images and video data, and transmit the data to the data processing module; the data processing module: after receiving the data transmitted by the data acquisition module, this module uses computer vision and deep learning technology to process the image and video data, extract three-dimensional spatial information, and construct a three-dimensional model; the three-dimensional model generation module: after receiving the data processed by the data processing module, this module uses a three-dimensional model construction algorithm to convert the three-dimensional spatial information into a three-dimensional model; the model output module: this module outputs the generated three-dimensional model in a certain format for subsequent application; the controller: this controller is used to control the workflow of each of the above modules.

[0039] Further, the acquisition process of the data acquisition module is as follows: the target area is photographed from multiple angles and with high precision by a laser scanner, a panoramic camera, and a drone to obtain omnidirectional images and terrain data.

[0040] Further, the laser scanner measures the distance and angle of the object surface by emitting laser light to obtain the three-dimensional coordinate data of the object.

[0041] Further, the panoramic camera stitches multiple photos into a panoramic image to obtain the two-dimensional image data of the object.

[0042] Further, the drone takes aerial photos by carrying a high-resolution camera to obtain the two-dimensional image data of the object.

[0043] Further, the processing process of the data processing module for image and video data includes image correction, stitching, denoising, and conversion into three-dimensional model data.

[0044] Among them, the processes of image correction, stitching, and denoising are operated through image processing software: Photoshop, GIMP.

[0045] Among them, the conversion into three-dimensional model data is operated through geometric modeling software: AutoCAD, SketchUp.

[0046] Further, the process of the three-dimensional model generation module converting three-dimensional space information into a three-dimensional model includes establishing a terrain model, adding textures, and setting lighting.

[0047] Further, the terrain model is established by converting terrain data into a three-dimensional terrain model through GIS technology.

[0048] Further, textures are added by mapping image data to the model surface through texture mapping technology.

[0049] Further, lighting is set by setting light and material properties to simulate the lighting effect of the real environment.

[0050] Further, the model output module outputs the generated three-dimensional model in the formats of obj and fbx.

[0051] Furthermore, the present invention also provides a construction method for a real-scene three-dimensional construction model, including the following usage steps:

[0052] S1, conduct aerial oblique photography to obtain original images and down-looking POS data;

[0053] S2, data analysis and preprocessing, including correction of original image distortion and equalization of light and color, down-looking image aerial triangulation, and production of camera and image description files;

[0054] S3. Data processing: Conduct aerial triangulation on the oblique images to obtain high-precision exterior orientation elements of all images. Using the image data after distortion correction and combining with the results of aerial triangulation encryption, generate an ultra-high-density point cloud through multi-view image dense matching technology, and then construct an irregular triangular network model based on the point cloud;

[0055] S4. Automatically extract the corresponding texture information on the images after distortion difference correction and color equalization processing, and map it to the triangular patches of the corresponding model to achieve texture attachment, and finally generate a three-dimensional model result with clear and realistic texture;

[0056] S5. Output the real-scene three-dimensional model result.

[0057] Embodiment 1

[0058] In this embodiment, an experimental demonstration was carried out on the modeling project of the urban area of Jining City. A technical process for constructing a city real-scene three-dimensional model based on oblique photogrammetry technology was proposed, and the accuracy of the three-dimensional model was evaluated:

[0059] As Figure 2 The comparison chart of the effects of the original image preprocessing model and the model after the experiment; a—the color balance model of the original image preprocessing; b—the model after color balance in Embodiment 1; c—the color balance model of the original image preprocessing; d—the model after color balance in Embodiment 1.

[0061]

[0062] Table 2 Statistical table of the test conditions of each link of the model

[0063] Conclusion: Statistically compare the quality of the model and the efficiency of model color balance after the experiment respectively.

[0064] (1) Statistical comparison of model quality: Compared with the model based on the original image preprocessing under visual conditions, there is no obvious difference in the texture quality of the model after the experiment (see Table 1);

[0065] (2) Statistical comparison of model color balance efficiency: Statistically count the test conditions of each link in the experiment and make a table (see Table 2);

[0066] Comparing the efficiency of the experimental methods from Table 1: The preprocessing cycle of the original image is 1215 working days, and the cycle of Embodiment 1 is 133 working days, and the efficiency is increased by 9.1 times;

[0067] Experimental results show that the construction method of the real-scene three-dimensional construction model can effectively solve the problems of unclear model texture and inconsistent overall model color, thereby improving the model quality. At the same time, compared with the current mainstream method of achieving model color balancing based on original image preprocessing, the method proposed in this invention increases the production efficiency of model color balancing by 9.1 times and greatly shortens the production cycle of model color balancing. This method can quickly and accurately establish real-scene three-dimensional models, providing strong technical support for urban planning, smart cities, emergency rescue and other fields.

[0068] It should be noted that the specific process of this real-scene 3D construction model device is as follows:

[0069] S1, perform aerial oblique photography to obtain original images and downward POS data;

[0070] S2, data analysis and preprocessing, including original image distortion correction and light and color uniformity processing, downward image aerial triangulation solution, and camera and image description file preparation;

[0071] S3, data processing, performs aerial triangulation on oblique images to obtain high-precision exterior orientation elements of all images. Using the distortion-corrected image data combined with aerial triangulation encryption results, an ultra-high-density point cloud is generated through multi-view image dense matching technology. An irregular triangulated network model is then constructed based on the point cloud.

[0072] S4 automatically extracts the texture information corresponding to the position on the image after distortion correction and uniform light and color processing, and maps it to the corresponding triangular facets of the model to achieve texture attachment, ultimately generating a 3D model with clear and realistic textures;

[0073] S5, output of real-scene 3D model results.

[0074] The model produced is complete, with realistic textures, especially good details on the facades. It can construct the hanging and hollow parts of the building, fully and realistically reflecting the characteristics of the landforms. At the same time, the process of constructing the real-scene 3D model is highly automated and does not require manual labor. The construction method of this real-scene 3D construction model can effectively solve the problems of unclear model textures and inconsistent overall model colors, improving the model quality. At the same time, compared with the current mainstream method of achieving model color balancing based on raw image preprocessing, the method proposed by the present invention increases the production efficiency of model color balancing by 9.1 times, significantly shortening the production cycle of model color balancing. This method can quickly and accurately establish real-scene 3D models, providing strong technical support for urban planning, smart cities, emergency rescue and other fields.

[0075] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A real - scene three - dimensional construction model device, comprising a data acquisition module, a data processing module, a three - dimensional model generation module, a model output module and a controller, characterized in that: Data acquisition module: This module is used to acquire real - scene images and video data and transmit the data to the data processing module; Data processing module: After receiving the data transmitted by the data acquisition module, this module uses computer vision and deep - learning technologies to process the image and video data, extract three - dimensional spatial information, and construct a three - dimensional model; Three - dimensional model generation module: After receiving the data processed by the data processing module, this module uses a three - dimensional model construction algorithm to convert the three - dimensional spatial information into a three - dimensional model; Model output module: This module outputs the generated three - dimensional model in a certain format for subsequent applications; Controller: This controller is used to control the working processes of the above - mentioned modules.

2. The real-scene three-dimensional construction model device according to claim 1, characterized in that: The acquisition process of the data acquisition module is as follows: Use a laser scanner, a panoramic camera, and a drone to take multi - angle and high - precision photos of the target area to obtain all - round images and terrain data; The laser scanner measures the distance and angle of the object's surface by emitting laser light to obtain the three - dimensional coordinate data of the object; The panoramic camera stitches multiple photos into a panoramic image to obtain the two - dimensional image data of the object; The drone takes aerial photos by carrying a high - resolution camera to obtain the two - dimensional image data of the object.

3. The real-scene three-dimensional construction model device according to claim 1, characterized in that: The processing process of the data processing module for image and video data includes image correction, stitching, denoising, and conversion into three - dimensional model data; Among them, the processes of image correction, stitching, and denoising are operated through image - processing software: Photoshop, GIMP; Among them, the conversion into three - dimensional model data is operated through geometric modeling software: AutoCAD, SketchUp.

4. A real-scene three-dimensional construction model device according to claim 1, characterized in that: The process by which the three - dimensional model generation module converts three - dimensional spatial information into a three - dimensional model includes establishing a terrain model, adding textures, and setting lighting; The establishment of the terrain model converts terrain data into a three - dimensional terrain model through GIS technology; The addition of textures maps the image data to the model surface through texture mapping technology; The setting of lighting simulates the lighting effect of the real environment by setting light and material properties.

5. A real-scene three-dimensional construction model device according to claim 1, characterized in that: The model output module outputs the generated three - dimensional model in the formats of obj and fbx.

6. A construction method for a real - scene three - dimensional construction model, characterized in that: It includes the following usage steps: S1, Conduct aerial oblique photography to obtain original images and down - looking POS data; S2, Data analysis and pre - processing, including original image distortion correction and equalization of light and color processing, down - looking image aerial triangulation, and production of camera and image description files; S3, Data processing, conduct aerial triangulation on the oblique images to obtain high - precision exterior orientation elements of all images, use the distorted - corrected image data combined with the aerial triangulation encryption results, generate an ultra - high - density point cloud through multi - view image dense matching technology, and then construct an irregular triangular network model based on the point cloud; S4, Automatically extract the texture information corresponding to the position on the images after distortion correction and equalization of light and color processing, and map it to the triangular patches of the corresponding model to achieve texture attachment, and finally generate a three - dimensional model result with clear and vivid textures; S5, Output of real scene 3D model results.

Images