Visual mapping method and system based on digital restoration technology

Abstract

The present invention discloses a visualization mapping method and system based on digital restoration technology, comprising the following steps: performing three-dimensional visualization digital restoration on design drawings and design parameters to obtain a three-dimensional scene model image, and obtaining a simulated point cloud based on the three-dimensional scene model image; obtaining internal and external real images of the real scene and actual point clouds corresponding to the internal and external real images, and optimizing and verifying the simulated point cloud in the three-dimensional scene model image through the actual point cloud, and outputting an optimized three-dimensional scene model; and feeding back the visualization scene output by the optimized three-dimensional scene model to a terminal. The visualization mapping method reduces costs and improves the efficiency of problem solving.

Description

Technical Field

[0001] The present invention relates to the field of digital restoration technology, and in particular to a visual mapping method and system based on digital restoration technology. Background Art

[0002] Information visualization mainly studies how to assist users in analytical reasoning through interactive visual interfaces, and provides tools and technologies to assist users in analytical decision-making, enabling users to comprehensively analyze information from massive, dynamic, and fuzzy data and gain insight into hidden laws and patterns.

[0003] For effective projection mapping on complex objects such as buildings, extensive data on the complex object must often be collected before designing the imagery to be projected onto it. For example, for buildings, a model of the building can be obtained using high-resolution scanning services (e.g., laser scanners, camera-projector screen learning systems, or traditional / manual site surveys); however, this can be expensive and / or time-consuming to obtain or even use for projection mapping planning and / or pre-visualization, as the associated files can be large, requiring significant processing overhead. Furthermore, once the model is obtained, a geo-site survey must often be performed to further determine projector placement and / or pose, as well as to obtain internal and external structural data, resulting in a time-consuming and expensive process. Summary of the Invention

[0004] Based on the technical problems existing in the background technology, the present invention proposes a visual mapping method and system based on digital restoration technology, which reduces costs and improves the efficiency of solving problems.

[0005] The visual mapping method based on digital restoration technology proposed in the present invention includes the following steps:

[0006] Perform three-dimensional visualization digital restoration of design drawings and design parameters to obtain a three-dimensional scene model image, and obtain a simulated point cloud based on the three-dimensional scene model image;

[0007] Obtain the real internal and external images of the real scene and the actual point clouds corresponding to the real internal and external images, and use the actual point clouds to optimize and verify the simulated point clouds in the 3D scene model image, and output the optimized 3D scene model;

[0008] The visual scene output by the optimized three-dimensional scene model is fed back to the terminal.

[0009] Furthermore, in optimizing and verifying the simulated point cloud in the three-dimensional scene model image using the actual point cloud and outputting the optimized three-dimensional scene model, the following steps are specifically included:

[0010] A voxel grid is obtained by adaptively spacing the minimum bounding box of the three-dimensional point cloud data in the point cloud space, and voxel grids that are not hit by the three-dimensional point cloud data are removed to obtain a first voxel set;

[0011] Removing a single voxel grid without a point hit in the first voxel set to obtain a second voxel set and a three-dimensional point cloud set in the second voxel set;

[0012] Mapping the three-dimensional point cloud set in the second voxel set to the coordinate system of the three-dimensional scene model image, and calculating the error function between the three-dimensional point cloud set in the second voxel set after mapping and the simulated point cloud in the three-dimensional scene model image;

[0013] If the error function value is less than or equal to a preset threshold, the three-dimensional point cloud data of the inner and outer real images is saved, and the three-dimensional point cloud is used as the three-dimensional point cloud data in the third voxel set;

[0014] If the error function value is greater than a preset threshold, the three-dimensional point cloud data set in the second voxel set replaces the corresponding simulated point cloud in the three-dimensional scene model, and the replaced three-dimensional point cloud is used as the three-dimensional point cloud data in the third voxel set;

[0015] A three-dimensional scene model image is reconstructed based on the three-dimensional point cloud data in the third voxel set, and an optimized three-dimensional scene model is output.

[0016] Furthermore, in calculating the error function between the three-dimensional point cloud set in the second voxel set after mapping and the simulated point cloud in the three-dimensional scene model image, the calculation process of the error function is as follows:

[0017] Calculate the rotation matrix Mt and translation matrix Mr in the image that scales the 3D point cloud set to approximate the 3D scene model, and denote the number of points in the 3D point cloud set as PC;

[0018] Calculate the mean value Pa1 of the three-dimensional point cloud collection points and the mean value Pa2 of the simulated point cloud;

[0019] According to the scaling factor S of the three-dimensional space, the scaling matrix Ms is obtained;

[0020] The position of P0 is calculated based on the rotation matrix Mt, translation matrix Mr and scaling matrix Ms;

[0021] The vertex coordinate set in the preset three-dimensional scene model image is recorded as P1;

[0022] Traverse the P1 set, find the point in the P0 set that is closest to the P0 position, and calculate the error function The error function value of .

[0023] Furthermore, the mean Pa1 and the mean Pa2 are calculated by the following formula:

[0024] ;

[0025] ;

[0026] in, is the mean of the three-dimensional point cloud collection points and the mean of the simulated point cloud The proportion set of Represents the average value of all points in the 3D point cloud set, represents the set of points in a 3D point cloud set, is the number of points in the 3D point cloud collection.

[0027] The visual mapping method based on digital restoration technology includes the following steps:

[0028] Acquire a real image of the appearance of a real scene, and restore the real image to obtain a restored image;

[0029] By restoring the image and designing the parameters, a simulated point cloud is obtained, and the simulated point cloud is digitally restored to three-dimensional visualization to obtain a three-dimensional scene model image;

[0030] Obtain the real internal and external images of the real scene and the actual point clouds corresponding to the real internal and external images, and use the actual point clouds to optimize and verify the simulated point clouds in the 3D scene model image, and output the optimized 3D scene model;

[0031] Feedback the visualization scene output by the optimized three-dimensional scene model to the terminal;

[0032] Furthermore, in optimizing and verifying the simulated point cloud in the three-dimensional scene model image using the actual point cloud and outputting the optimized three-dimensional scene model, the following steps are specifically included:

[0033] A voxel grid is obtained by adaptively spacing the minimum bounding box of the three-dimensional point cloud data in the point cloud space, and voxel grids that are not hit by the three-dimensional point cloud data are removed to obtain a first voxel set;

[0034] Removing a single voxel grid without a point hit in the first voxel set to obtain a second voxel set and a three-dimensional point cloud set in the second voxel set;

[0035] Mapping the three-dimensional point cloud set in the second voxel set to the coordinate system of the three-dimensional scene model image, and calculating the error function between the three-dimensional point cloud set in the second voxel set after mapping and the simulated point cloud in the three-dimensional scene model image;

[0036] If the error function value is less than or equal to a preset threshold, the three-dimensional point cloud data of the inner and outer real images is saved, and the three-dimensional point cloud is used as the three-dimensional point cloud data in the third voxel set;

[0037] If the error function value is greater than a preset threshold, the three-dimensional point cloud data set in the second voxel set replaces the corresponding simulated point cloud in the three-dimensional scene model, and the replaced three-dimensional point cloud is used as the three-dimensional point cloud data in the third voxel set;

[0038] A three-dimensional scene model image is reconstructed based on the three-dimensional point cloud data in the third voxel set, and an optimized three-dimensional scene model is output.

[0039] A visual mapping system based on digital restoration technology, including a simulation scene construction module, a model optimization module, and a visual interaction module;

[0040] The simulation scene construction module is used to obtain simulation point clouds through design drawings and design parameters, and perform three-dimensional visualization digital restoration on the simulation point clouds to obtain a three-dimensional scene model image;

[0041] The model optimization module is used to obtain the real internal and external images of the real scene and the actual point clouds corresponding to the real internal and external images, and optimize and verify the simulated point clouds in the 3D scene model image through the actual point clouds, and output the optimized 3D scene model;

[0042] The visualization interaction module is used to feed back the visualization scene output by the optimized three-dimensional scene model to the terminal.

[0043] A visual mapping system based on digital restoration technology, including an image acquisition module, a simulation scene construction module, a model optimization module, and a visual interaction module;

[0044] The image acquisition module is used to acquire a real image of the appearance of a real scene, and restore the real image to obtain a restored image;

[0045] The simulation scene construction module is used to obtain a simulation point cloud by restoring the image and design parameters, and to perform three-dimensional visualization digital restoration on the simulation point cloud to obtain a three-dimensional scene model image;

[0046] The model optimization module is used to obtain the real internal and external images of the real scene and the actual point clouds corresponding to the real internal and external images, and optimize and verify the simulated point clouds in the 3D scene model image through the actual point clouds, and output the optimized 3D scene model;

[0047] The visualization interaction module is used to feed back the visualization scene output by the optimized three-dimensional scene model to the terminal.

[0048] A computer-readable storage medium stores a plurality of classification programs, wherein the plurality of classification programs are used to be called by a processor and execute the above-mentioned visual mapping method.

[0049] The advantages of the visualization mapping method and system based on digital restoration technology provided by the present invention are: the visualization mapping method and system based on digital restoration technology provided in the structure of the present invention preliminarily obtain visualization scene information through infrared detection equipment to reduce the error with the real scene, thereby improving the accuracy of analyzing buildings and physical environments using the visualization mapping method of digital restoration technology, and through this technology, a three-dimensional visualization solution can be obtained in solving problems in the physical environment, greatly improving the efficiency of problem solving; in addition, image data collection, data collection, construction structure layout and other data collection are performed on buildings or other related real scenes; the construction of a digital restoration model is conducive to the later engineering inspection and judgment of buildings or related physical environments, and provides visualization analysis and judgment in emergency rescue and other events; thereby improving the accuracy of judgment of urban life and management problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a structural schematic diagram of the present invention. DETAILED DESCRIPTION

[0051] The technical solutions of the present invention are described in detail below through specific embodiments. Many specific details are set forth in the following description to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways than those described herein, and those skilled in the art may make similar modifications without departing from the scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0052] like Figure 1 As shown, the visualization mapping method based on digital restoration technology proposed in the present invention can visualize the three-dimensional image corresponding to the design drawing before the real scene is built, so as to facilitate viewing. Specifically, it includes:

[0053] S01: Perform three-dimensional visualization digital restoration on the design drawings and design parameters to obtain a three-dimensional scene model image, and obtain a simulated point cloud based on the three-dimensional scene model image;

[0054] Before the real scene is built, the scene design parameters are obtained through design drawings and design renderings, and the virtual digital scene model is restored and built using 3D production software to obtain a digital scene that matches the real scene, and the initial point cloud data is obtained from the digitally restored scene. The initial point cloud data is used as the simulation point cloud in this method.

[0055] The design drawings include 3D renderings, CAD construction drawings, structural drawings, etc. The design parameters include location points, distances between location points, structural stiffness, etc. A 3D simulation model is constructed based on the construction data corresponding to the design drawings. The internal structure of the 3D model is filled with relevant data corresponding to the design parameters to obtain a 3D scene model image, which is close to the real scene.

[0056] S02: Acquire real internal and external images of the real scene and actual point clouds corresponding to the real internal and external images, optimize and verify the simulated point cloud in the 3D scene model image using the actual point clouds, and output the optimized 3D scene model;

[0057] The real internal and external images include the appearance image and internal structure image of the scene object. Although the scene object is actually constructed according to the design drawings and design parameters during the actual construction process, the actual objects formed by the construction are still different. Cameras at multiple angles can be set outside and inside the object to obtain the appearance image and internal structure image of the object. The corresponding data in the appearance image and the internal structure image are used as actual point clouds. According to the actual point cloud, the corresponding actual point cloud space can be constructed, and the actual point cloud space is loaded into the three-dimensional scene model image to optimize the three-dimensional scene model image so that the three-dimensional scene model finally output to the terminal is closer to the actual object.

[0058] The optimized three-dimensional scene model makes the scene more intuitive and clear in the subsequent real-scene mapping, and can accurately understand the internal information of the real scene through the visualization of the digital scene.

[0059] S03: Feedback the visualization scene output by the optimized three-dimensional scene model to the terminal.

[0060] The user uses the terminal to scan the QR code image and enter the visualization scene, where he can intuitively see and select the required data and spatial conditions.

[0061] After steps S01 to S03, the constructed three-dimensional scene model image is optimized through the point cloud space to obtain a three-dimensional scene model close to the actual object. The user enters the three-dimensional scene through terminal scanning, realizing digital visualization of buildings and real-world scenes, realizing the digital development of scenes, and also facilitating engineering maintenance, municipal construction and other needs, facilitating urban digital construction, and facilitating urban management.

[0062] Step S02: obtaining the real internal and external images of the real scene and the actual point clouds corresponding to the real internal and external images, and optimizing and verifying the simulated point clouds in the 3D scene model image using the actual point clouds, and outputting the optimized 3D scene model. Specifically, optimizing the simulated point clouds in the 3D scene model image in step S01 using the internal structure of the real environment includes:

[0063] S021: obtaining a voxel grid by adaptively spacing the minimum bounding box of the three-dimensional point cloud data in the point cloud space, and removing the voxel grids that are not hit by the three-dimensional point cloud data to obtain a first voxel set;

[0064] The point cloud space is a three-dimensional space constructed by the actual point cloud. This step mainly removes voxel grids that have no three-dimensional point cloud data hits to reduce the number of voxels and ensure that the number of three-dimensional point cloud data in more than 90% of the remaining voxel grids is between 1 and 3, where voxels can be understood as three-dimensional images.

[0065] S022: Remove the single voxel grid without point hits in the first voxel set to obtain a second voxel set and a three-dimensional point cloud set in the second voxel set; to further optimize the voxel set and remove interference points.

[0066] S023: Mapping the three-dimensional point cloud set in the second voxel set to the coordinate system of the three-dimensional scene model image, calculating the error function between the three-dimensional point cloud set in the second voxel set after mapping and the simulated point cloud in the three-dimensional scene model, if the error function value is less than or equal to a preset threshold, proceeding to step S024; if the error function value is greater than the preset threshold, proceeding to step S025;

[0067] The calculation process of the error function is as follows S0231 to S0326:

[0068] S0231: Calculate the rotation matrix Mt and translation matrix Mr in scaling the 3D point cloud set to approximate the 3D scene model image, and denote the number of points in the 3D point cloud set as PC;

[0069] The rotation matrix Mt and translation matrix Mr describe the rotation angle and spatial position of an object in three-dimensional space respectively in three-dimensional graphics.

[0070] S0232: Calculate the mean value Pa1 of the three-dimensional point cloud set points and the mean value Pa2 of the simulated point cloud;

[0071] The mean Pa1 and mean Pa2 are calculated using the following formula:

[0072] ;

[0073] ;

[0074] in, is the mean of the three-dimensional point cloud collection points and the mean of the simulated point cloud The proportion set of Represents the average value of all points in the 3D point cloud set, represents the set of points in a 3D point cloud set, is the number of points in the 3D point cloud collection.

[0075] S0233: Obtain a scaling matrix Ms according to the scaling coefficient S of the three-dimensional space;

[0076] The scaling factor S of the three-dimensional space refers to the ratio of the size of the three-dimensional point cloud to the image size of the three-dimensional scene model.

[0077] S0234: Calculate the position P0 according to the rotation matrix Mt, the translation matrix Mr, and the scaling matrix Ms. The position P0 is the position of the point in the 3D point cloud converted to the 3D scene model image space;

[0078] S0235: The vertex coordinate set in the preset three-dimensional scene model image is recorded as P1;

[0079] S0236: Traverse the P1 set, find the point in the P0 set that is closest to the P0 position, and calculate the error function The error function value of .

[0080] Through steps S0231 to S0236, the three-dimensional scene model image can be reconstructed to be more consistent with the presentation of the actual scene.

[0081] For example: the electrical wires and network cables inside the wall; the wiring position of the urban underground pipeline network, the distance between the lines; the spatial position of the fire-fighting equipment and the distance of the safety exit, etc. can all be measured through steps S0231 to S0236. The difference between the actual point cloud and the simulated point cloud can be used to obtain the construction error value to correct the presentation of the entire real scene.

[0082] S024: Saving the simulated point cloud as three-dimensional point cloud data in the third voxel set;

[0083] S025: replacing the corresponding simulated point cloud in the three-dimensional scene model with the three-dimensional point cloud data set in the second voxel set, and using the replaced three-dimensional point cloud data set as the three-dimensional point cloud data in the third voxel set;

[0084] S026: Reconstruct a three-dimensional scene model image based on the three-dimensional point cloud data in the third voxel set in steps S034 and S025, and output an optimized three-dimensional scene model.

[0085] Through steps S021 to S026, the simulated point cloud in the three-dimensional scene model image is threshold-corrected using the actual point cloud, and the three-dimensional scene model of the terminal is dynamically corrected, so that the final visualized three-dimensional scene model is close to the real scene in the actual construction process.

[0086] Another visualization method is the visualization mapping method based on digital restoration technology. After the real scene appearance is built, the entire scene is visualized for easy viewing. Specifically, it includes:

[0087] S10: Acquire a real image of the appearance of a real scene, and restore the real image to obtain a restored image;

[0088] External image data of the top and surrounding appearance of the real scene is collected through infrared detection equipment to obtain a real image of the exterior of the real scene.

[0089] S20: obtaining a simulated point cloud by restoring the image and the design parameters, and performing three-dimensional visualization digital restoration on the simulated point cloud to obtain a three-dimensional scene model image;

[0090] After the appearance of the real scene is constructed, the point cloud data of the constructed scene is directly obtained and a digital restoration scene is built for direct optimization and combination. The obtained point cloud data is used as the simulated point cloud in this method.

[0091] It should be noted that the simulated point cloud obtained before the real scene is built is the data output from the digital scene model constructed through the design drawings; and the simulated point cloud obtained after the real scene appearance is completed is the data corresponding to the real image in the real scene, which can be obtained by setting multiple depth cameras outside the real image.

[0092] The design parameters include location points, distances between location points, structural stiffness, etc. A three-dimensional simulation model of the appearance of the real scene is constructed based on the restored image. The internal structure of the three-dimensional model is filled with relevant data corresponding to the design parameters to obtain a three-dimensional scene model image, which is close to the real scene.

[0093] The specific process of obtaining the three-dimensional scene model image is as follows: load the three simulated point clouds and restored images into three-dimensional software (such as 3dmax, maya), build a three-dimensional scene model, and achieve a rough digital visualization restoration of the 1:1 scene of the building's exterior and internal structure.

[0094] S30: Acquire real internal and external images of the real scene and actual point clouds corresponding to the real internal and external images, optimize and verify the simulated point cloud in the three-dimensional scene model image using the actual point clouds, and output the optimized three-dimensional scene model;

[0095] Although the internal structure of scene objects is actually constructed according to the design parameters during the actual construction process, the actual internal structure formed by the construction is still different. Cameras at multiple angles can be set inside the object to obtain images of the internal structure of the object. The data corresponding to the internal structure image is used as the actual point cloud. Based on the actual point cloud and the restored image, the corresponding actual point cloud space can be constructed, and the actual point cloud space can be loaded into the three-dimensional scene model image to optimize the internal structure of the three-dimensional scene model image, so that the three-dimensional scene model finally output to the terminal is closer to the actual object.

[0096] The simulated point cloud in the three-dimensional scene model image is optimized and checked using the actual point cloud, and the optimized three-dimensional scene model is output, which is consistent with step S02 and will not be repeated here.

[0097] S40: Feeding back the visualized scene output by the optimized three-dimensional scene model to the terminal.

[0098] This step is the same as step S03 and will not be described again here.

[0099] In step 02 and step S30, the simulated point cloud in the three-dimensional scene model image is corrected by the actual point cloud, which can ensure the accuracy of the digital visualization scene and at the same time improve the quality of the operation experience in order to make the subsequent scanning and interaction process smoother.

[0100] At the same time, this embodiment utilizes image data collection, data data collection, construction structure layout and other data collection of buildings or other related real scenes; the construction of digital restoration models is conducive to the later engineering inspection and judgment of buildings or related physical environments, and provides visual analysis and judgment in emergency rescue and other events; thereby improving the accuracy of judgment on problems in urban life and management work.

[0101] Infrared detection equipment can initially capture visual scene information, reducing discrepancies with the real scene. This improves the accuracy of visual mapping methods for analyzing buildings and physical environments using digital restoration technology. Furthermore, this technology enables 3D visualization of solutions to problems within the physical environment, significantly improving problem-solving efficiency. For example, in the event of a fire, digital visualization mapping technology can be used to conduct intuitive, 3D rescue assessments of the affected building, understanding the building structure, emergency exits, and the layout of firefighting and rescue equipment.

[0102] Those skilled in the art will understand that all or part of the steps of implementing the above-mentioned method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above-mentioned method embodiment; and the aforementioned storage medium includes: ROM, RAM, disk or optical disk, etc. Various media that can store program codes.

[0103] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A visual mapping method based on digital restoration technology includes the following steps: Perform three-dimensional visualization digital restoration of design drawings and design parameters to obtain a three-dimensional scene model image, and obtain a simulated point cloud based on the three-dimensional scene model image; Obtain the real internal and external images of the real scene and the actual point clouds corresponding to the real internal and external images, and use the actual point clouds to optimize and verify the simulated point clouds in the 3D scene model image, and output the optimized 3D scene model; Feedback the visualization scene output by the optimized three-dimensional scene model to the terminal; in, The optimization and verification of the simulated point cloud in the 3D scene model image using the actual point cloud, and the output of the optimized 3D scene model, specifically include: A voxel grid is obtained by adaptively spacing the minimum bounding box of the three-dimensional point cloud data in the point cloud space, and voxel grids that are not hit by the three-dimensional point cloud data are removed to obtain a first voxel set; Removing a single voxel grid without a point hit in the first voxel set to obtain a second voxel set and a three-dimensional point cloud set in the second voxel set; Mapping the three-dimensional point cloud set in the second voxel set to the coordinate system of the three-dimensional scene model image, and calculating the error function between the three-dimensional point cloud set in the second voxel set after mapping and the simulated point cloud in the three-dimensional scene model image; If the error function value is less than or equal to a preset threshold, the three-dimensional point cloud data of the inner and outer real images is saved, and the three-dimensional point cloud is used as the three-dimensional point cloud data in the third voxel set; If the error function value is greater than a preset threshold, the three-dimensional point cloud data set in the second voxel set replaces the corresponding simulated point cloud in the three-dimensional scene model, and the replaced three-dimensional point cloud is used as the three-dimensional point cloud data in the third voxel set; A three-dimensional scene model image is reconstructed based on the three-dimensional point cloud data in the third voxel set, and an optimized three-dimensional scene model is output.

2. The visualization mapping method based on digital restoration technology according to claim 1, characterized in that: In calculating the error function between the three-dimensional point cloud set in the second voxel set after mapping and the simulated point cloud in the three-dimensional scene model image, the calculation process of the error function is as follows: Calculate the rotation matrix Mt and translation matrix Mr in the image that scales the 3D point cloud set to approximate the 3D scene model, and denote the number of points in the 3D point cloud set as PC; Calculate the mean value Pa1 of the three-dimensional point cloud collection points and the mean value Pa2 of the simulated point cloud; According to the scaling factor S of the three-dimensional space, the scaling matrix Ms is obtained; The position of P0 is calculated based on the rotation matrix Mt, translation matrix Mr and scaling matrix Ms; The vertex coordinate set in the preset three-dimensional scene model image is recorded as P1; Traverse the P1 set, find the point in the P0 set that is closest to the P0 position, and calculate the error function The error function value of .

3. The visualization mapping method based on digital restoration technology according to claim 2, characterized in that: in, The mean Pa1 and mean Pa2 are calculated using the following formula: ； ； in, is the mean of the three-dimensional point cloud collection points and the mean of the simulated point cloud The proportion set of Represents the average value of all points in the 3D point cloud set, represents the set of points in a 3D point cloud set, is the number of points in the 3D point cloud collection.

4. A visual mapping method based on digital restoration technology includes the following steps: Acquire a real image of the appearance of a real scene, and restore the real image to obtain a restored image; By restoring the image and designing the parameters, a simulated point cloud is obtained, and the simulated point cloud is digitally restored to three-dimensional visualization to obtain a three-dimensional scene model image; Obtain the real internal and external images of the real scene and the actual point clouds corresponding to the real internal and external images, and use the actual point clouds to optimize and verify the simulated point clouds in the 3D scene model image, and output the optimized 3D scene model; Feedback the visualization scene output by the optimized three-dimensional scene model to the terminal; The process of optimizing and verifying the simulated point cloud in the 3D scene model image using the actual point cloud and outputting the optimized 3D scene model specifically includes: A voxel grid is obtained by adaptively spacing the minimum bounding box of the three-dimensional point cloud data in the point cloud space, and voxel grids that are not hit by the three-dimensional point cloud data are removed to obtain a first voxel set; Removing a single voxel grid without a point hit in the first voxel set to obtain a second voxel set and a three-dimensional point cloud set in the second voxel set; Mapping the three-dimensional point cloud set in the second voxel set to the coordinate system of the three-dimensional scene model image, and calculating the error function between the three-dimensional point cloud set in the second voxel set after mapping and the simulated point cloud in the three-dimensional scene model image; If the error function value is less than or equal to a preset threshold, the three-dimensional point cloud data of the inner and outer real images is saved, and the three-dimensional point cloud is used as the three-dimensional point cloud data in the third voxel set; If the error function value is greater than a preset threshold, the three-dimensional point cloud data set in the second voxel set replaces the corresponding simulated point cloud in the three-dimensional scene model, and the replaced three-dimensional point cloud is used as the three-dimensional point cloud data in the third voxel set; A three-dimensional scene model image is reconstructed based on the three-dimensional point cloud data in the third voxel set, and an optimized three-dimensional scene model is output.

5. Visual mapping system based on digital restoration technology, characterized by: Includes simulation scenario construction module, model optimization module and visualization interaction module; The simulation scene construction module is used to obtain simulation point clouds through design drawings and design parameters, and perform three-dimensional visualization digital restoration on the simulation point clouds to obtain a three-dimensional scene model image; The model optimization module is used to obtain the real internal and external images of the real scene and the actual point clouds corresponding to the real internal and external images, and optimize and verify the simulated point clouds in the 3D scene model image through the actual point clouds, and output the optimized 3D scene model; The visualization interaction module is used to feed back the visualization scene output by the optimized three-dimensional scene model to the terminal; The model optimization module is specifically used to: A voxel grid is obtained by adaptively spacing the minimum bounding box of the three-dimensional point cloud data in the point cloud space, and voxel grids that are not hit by the three-dimensional point cloud data are removed to obtain a first voxel set; Removing a single voxel grid without a point hit in the first voxel set to obtain a second voxel set and a three-dimensional point cloud set in the second voxel set; Mapping the three-dimensional point cloud set in the second voxel set to the coordinate system of the three-dimensional scene model image, and calculating the error function between the three-dimensional point cloud set in the second voxel set after mapping and the simulated point cloud in the three-dimensional scene model image; If the error function value is less than or equal to a preset threshold, the three-dimensional point cloud data of the inner and outer real images is saved, and the three-dimensional point cloud is used as the three-dimensional point cloud data in the third voxel set; If the error function value is greater than a preset threshold, the three-dimensional point cloud data set in the second voxel set replaces the corresponding simulated point cloud in the three-dimensional scene model, and the replaced three-dimensional point cloud is used as the three-dimensional point cloud data in the third voxel set; A three-dimensional scene model image is reconstructed based on the three-dimensional point cloud data in the third voxel set, and an optimized three-dimensional scene model is output.

6. Visual mapping system based on digital restoration technology, characterized by: It includes image acquisition module, simulation scene construction module, model optimization module and visualization interaction module; The image acquisition module is used to acquire a real image of the appearance of a real scene, and restore the real image to obtain a restored image; The simulation scene construction module is used to obtain a simulation point cloud by restoring the image and design parameters, and to perform three-dimensional visualization digital restoration on the simulation point cloud to obtain a three-dimensional scene model image; The model optimization module is used to obtain the real internal and external images of the real scene and the actual point clouds corresponding to the real internal and external images, and optimize and verify the simulated point clouds in the 3D scene model image through the actual point clouds, and output the optimized 3D scene model; The visualization interaction module is used to feed back the visualization scene output by the optimized three-dimensional scene model to the terminal; The model optimization module is specifically used to: A voxel grid is obtained by adaptively spacing the minimum bounding box of the three-dimensional point cloud data in the point cloud space, and voxel grids that are not hit by the three-dimensional point cloud data are removed to obtain a first voxel set; Removing a single voxel grid without a point hit in the first voxel set to obtain a second voxel set and a three-dimensional point cloud set in the second voxel set; Mapping the three-dimensional point cloud set in the second voxel set to the coordinate system of the three-dimensional scene model image, and calculating the error function between the three-dimensional point cloud set in the second voxel set after mapping and the simulated point cloud in the three-dimensional scene model image; If the error function value is less than or equal to a preset threshold, the three-dimensional point cloud data of the inner and outer real images is saved, and the three-dimensional point cloud is used as the three-dimensional point cloud data in the third voxel set; If the error function value is greater than a preset threshold, the three-dimensional point cloud data set in the second voxel set replaces the corresponding simulated point cloud in the three-dimensional scene model, and the replaced three-dimensional point cloud is used as the three-dimensional point cloud data in the third voxel set; A three-dimensional scene model image is reconstructed based on the three-dimensional point cloud data in the third voxel set, and an optimized three-dimensional scene model is output.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a plurality of classification programs, which are used to be called by a processor and execute the visual mapping method according to claim 1 or 4.

Images