Decoration progress management method based on virtual reality

Abstract

The invention provides a decoration progress management method based on virtual reality, and the method comprises the steps: carrying out the immersive roaming of a decoration virtual reality scene through a virtual reality head-mounted display device, obtaining the interaction operation data of a decoration material and a construction scheme of a user, and employing a recommendation algorithm based on user collaborative filtering, recommending an optimized decoration scheme in the decoration virtual reality scene according to the interactive operation data; the optimized decoration scheme is discussed and modified in a virtual space shared by multiple persons through a virtual reality collaboration platform, a distributed data synchronization mechanism is adopted, decoration virtual reality scenes among different users are synchronized, and the synchronization delay of the distributed data synchronization mechanism does not exceed 20 ms; and according to the evaluated decoration scheme, a construction plan in the order system is updated, a decoration task work order is generated, a construction result of each stage is accepted in the decoration virtual reality scene, and an acceptance result is synchronized to the order system.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a decoration progress management method based on virtual reality. Background Art

[0002] The introduction of virtual reality technology in the Internet home improvement service order system faces the problem of data integration and synchronous update between the virtual reality model and the order system. Since the decoration process is dynamic, the layout of the decoration site, construction progress and other information are constantly changing over time. How to capture the changes in the decoration site in real time and update them to the virtual reality model is a major technical challenge. Traditional three-dimensional modeling methods are difficult to cope with the dynamic nature of the decoration scene. Static modeling data cannot reflect the progress of the decoration, and frequent reconstruction of the model will consume a lot of time and computing resources. It is necessary to study incremental dynamic modeling technology, which can update the changes in real time based on the original model, while ensuring modeling efficiency and maximizing synchronization with the actual situation of the decoration site. At the same time, it is also necessary to establish a mapping relationship between the virtual reality model and the order system, automatically associate the decoration tasks with the model components, and realize the visual annotation and presentation of the order progress and decoration problems in the virtual reality scene. Summary of the invention

[0003] The present invention provides a decoration progress management method based on virtual reality, which mainly includes:

[0004] Obtaining point cloud data of the decoration site, the point cloud density of the point cloud data is not less than 1000 points per square meter, the point cloud accuracy of the point cloud data is better than 5mm, the noise level of the point cloud data is less than 1%, and obtaining a video stream of the decoration site, the frame rate of the video stream is not less than 30 frames per second, and the resolution of the video stream is not less than 1080P;

[0005] The incremental dynamic modeling technology of octree space partitioning is adopted, and the changing part of the point cloud data is reconstructed in real time in three dimensions through the Poisson surface reconstruction algorithm to obtain a three-dimensional mesh model of the decoration site. The number of triangles in the three-dimensional mesh model is not less than 1.5 times the total number of points in the point cloud data, and the texture mapping resolution of the three-dimensional mesh model is not less than 4096×4096;

[0006] Performing scene semantic segmentation on the three-dimensional grid model to obtain main structures such as walls, floors, and ceilings, where the semantic segmentation accuracy of the scene semantic segmentation is not less than 90%, and performing decoration object detection on the three-dimensional grid model using a regional convolutional neural network to obtain decoration objects such as doors, windows, pipelines, and furniture, where the average detection accuracy of the decoration object detection is not less than 85%;

[0007] Acquire a pre-built decoration material database, perform material recognition on the decoration object according to the decoration material database, obtain material property information of decoration materials such as paint, coating, tile, floor, wallpaper, etc., wherein the recognition accuracy of the material recognition is not less than 80%, map the material property information to the corresponding part of the three-dimensional grid model, and generate a decoration virtual reality scene;

[0008] Obtain the construction plan in the order system, annotate the decoration objects in the decoration virtual reality scene according to the construction process according to the construction plan, generate a timeline representation of the decoration progress by using a visualization method based on a Gantt chart, and judge whether the actual decoration progress is consistent with the construction plan. If not, highlight the decoration objects with lagging progress in red in the decoration virtual reality scene;

[0009] Immersing the virtual reality scene of decoration through a virtual reality head display device, obtaining interactive operation data of the user's decoration materials and construction plans, and using a recommendation algorithm based on user collaborative filtering to recommend an optimized decoration plan in the virtual reality scene of decoration according to the interactive operation data;

[0010] Through the virtual reality collaborative platform, the optimized decoration plan is discussed and modified in a virtual space shared by multiple people, and a distributed data synchronization mechanism is used to synchronize the decoration virtual reality scenes between different users. The synchronization delay of the distributed data synchronization mechanism does not exceed 20ms;

[0011] According to the evaluated decoration plan, the construction plan in the order system is updated, a decoration task work order is generated, the construction results of each stage are accepted in the decoration virtual reality scene, and the acceptance results are synchronized to the order system.

[0012] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0013] The present invention discloses a decoration progress management method based on virtual reality. The method acquires high-precision point cloud data and high-quality video streams, adopts incremental dynamic modeling technology to reconstruct the three-dimensional model of the decoration site in real time, and performs scene semantic segmentation and decoration object detection. Combined with a pre-built material database, the material of the decoration object is identified to generate a decoration virtual reality scene. The decoration objects in the virtual scene are annotated with attributes according to the construction plan to achieve visual management of the decoration progress. An immersive experience is performed through virtual reality equipment, user interaction data is obtained and optimization plans are recommended. Multiple people can discuss modification plans in a virtual collaborative platform and update the construction plan synchronously. The present invention realizes digital management of the entire decoration process and improves decoration efficiency and quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 The present invention is a flowchart of a decoration progress management method based on virtual reality.

[0015] Figure 2 It is a schematic diagram of a decoration progress management method based on virtual reality of the present invention.

[0016] Figure 3 This is another schematic diagram of a renovation progress management method based on virtual reality according to the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0018] like Figure 1 -3. In this embodiment, a decoration progress management method based on virtual reality may specifically include:

[0019] S101. Acquire point cloud data of the decoration site, wherein the point cloud density of the point cloud data is not less than 1000 points per square meter, the point cloud accuracy of the point cloud data is better than 5mm, and the noise level of the point cloud data is less than 1%; acquire a video stream of the decoration site, wherein the frame rate of the video stream is not less than 30 frames per second, and the resolution of the video stream is not less than 1080P.

[0020] According to the layout and area of ​​the decoration site, determine the appropriate point cloud scanning instrument and scanning parameters to obtain raw point cloud data with a density of not less than 1,000 points per square meter and an accuracy better than 5mm. Preprocess the acquired raw point cloud data, remove noise points through the point cloud noise reduction algorithm, control the noise level below 1%, and obtain high-precision point cloud data of the decoration site. According to the lighting conditions and shooting requirements of the decoration site, select appropriate high-definition camera equipment, set appropriate shooting parameters, and obtain high-quality video stream data with a frame rate of not less than 30 frames per second and a resolution of not less than 1080P. The acquired decoration site point cloud data and video stream data are synchronized in time and space, and the coordinate systems of the two are unified through algorithms such as feature point matching to achieve accurate registration of point cloud and video. The synchronized point cloud and video data are fused and processed, and the point cloud is colored by mapping the video frame to the point cloud to generate color point cloud data with real texture and color information. According to the requirements of the decoration design plan, the color point cloud data is semantically segmented to identify and extract different decoration components, such as walls, floors, ceilings, doors and windows, to form a semantic 3D model of the decoration site. The semantic 3D model of the decoration site is compared with the design plan, and the size deviation and process quality problems of the decoration construction are detected through point cloud registration and deviation analysis algorithms, and a construction quality assessment report of the decoration site is generated.

[0021] S102. Adopting the incremental dynamic modeling technology of octree space partitioning, the changing part of the point cloud data is reconstructed in real time in three dimensions through the Poisson surface reconstruction algorithm to obtain a three-dimensional mesh model of the decoration site. The number of triangles in the three-dimensional mesh model is not less than 1.5 times the total number of points in the point cloud data, and the texture mapping resolution of the three-dimensional mesh model is not less than 4096×4096.

[0022] The three-dimensional point cloud data of the decoration site is obtained, and the octree algorithm is used to perform spatial division on the point cloud data to divide the point cloud data into a number of voxels; the point cloud data in each voxel is locally reconstructed by a Poisson surface reconstruction algorithm to obtain a local three-dimensional mesh model; it is determined whether the point cloud data has changed, and if so, the point cloud data of the changed part is obtained, and step 2 is repeated for the point cloud data of the changed part to obtain a local three-dimensional mesh model of the changed part; the local three-dimensional mesh model of the changed part is merged with the original three-dimensional mesh model to obtain an updated three-dimensional mesh model of the decoration site; for the updated three-dimensional mesh model, the triangular facets are reduced by a mesh simplification algorithm, and the number of simplified triangular facets is controlled to be not less than 5 times the total number of points of the point cloud data; according to the texture information of the three-dimensional mesh model, a texture mapping algorithm is used to generate a texture map of the three-dimensional mesh model, and the resolution of the texture map is controlled to be not less than 4096×4096; the simplified three-dimensional mesh model and the generated texture map are combined to obtain a final three-dimensional model of the decoration site.

[0023] S103, performing scene semantic segmentation on the three-dimensional grid model to obtain main structures such as walls, floors, and ceilings, and the semantic segmentation accuracy of the scene semantic segmentation is not less than 90%; using a regional convolutional neural network to perform decoration object detection on the three-dimensional grid model to obtain decoration objects such as doors, windows, pipelines, and furniture, and the average detection accuracy of the decoration object detection is not less than 85%.

[0024] The three-dimensional mesh model data is obtained and preprocessed, including denoising, smoothing and other operations, to obtain an optimized three-dimensional mesh model; semantic segmentation is performed on the optimized three-dimensional mesh model according to a preset semantic segmentation model to obtain the main structures such as the segmented wall, ground, ceiling, etc.; it is determined whether the semantic segmentation accuracy of the segmented main structure is not less than 90%, if so, the segmentation result is used as the input of the subsequent decoration object detection, if not, return to step 2 to re-semantic segmentation; the pre-trained regional convolutional neural network model is used to detect the decoration objects of the segmented main structure to obtain the location and category information of the decoration objects such as doors, windows, pipelines, furniture, etc.; the average detection accuracy of the decoration object detection is calculated to determine whether it is not less than 85%, if so, the detection result is used as the input of the subsequent decoration design, if not, return to step 4 to re-detect the decoration objects; according to the results of semantic segmentation and decoration object detection, the functional attributes and spatial relationships of each area in the three-dimensional mesh model are determined to provide data support for the subsequent decoration design; the results of semantic segmentation and decoration object detection are visualized to generate intuitive three-dimensional decoration renderings, which are convenient for users to view and modify, and improve the efficiency and quality of decoration design.

[0025] S104, obtaining a pre-built decoration material database, performing material identification on the decoration object according to the decoration material database, obtaining material property information of decoration materials such as paint, coating, tile, floor, wallpaper, etc., wherein the recognition accuracy of the material identification is not less than 80%; mapping the material property information to the corresponding part of the three-dimensional grid model to generate a decoration virtual reality scene.

[0026] A pre-built decoration material database is obtained, which contains the attribute information of various decoration materials, such as the color and glossiness of paint and coating, and the material and texture of tiles and floors. For a given decoration object, an image segmentation algorithm based on deep learning is used to divide the image of the decoration object into different material areas. According to the divided material areas, the texture, color and other features of each area are extracted, and they are matched with the material properties in the decoration material database to identify the material type and properties corresponding to each area. In the process of material identification, the accuracy of material identification is increased to more than 80% by increasing training data and optimizing algorithms. The identified material attribute information is mapped to the three-dimensional mesh model of the decoration object, and for each triangular facet, the corresponding texture, color and other attributes are assigned according to its corresponding material properties. For unrecognized areas, the material texture coordinated with the surrounding area is synthesized by using a texture synthesis algorithm according to the material properties of the surrounding identified areas. Finally, after the decoration material properties are completely mapped to the three-dimensional mesh model, a realistic decoration virtual reality scene is rendered, and users can freely roam in the virtual scene to experience the decoration effect.

[0027] S105, obtaining a construction plan in the order system, and annotating the decoration objects in the decoration virtual reality scene according to the construction process according to the construction plan; generating a timeline representation of the decoration progress by using a visualization method based on a Gantt chart; determining whether the actual decoration progress is consistent with the construction plan, and if not, highlighting the decoration objects with lagging progress in red in the decoration virtual reality scene.

[0028] Obtain the construction plan data in the order system, and determine the corresponding decoration objects and attribute annotation rules for each construction process in the construction plan. According to the process time and sequence of the construction plan, a visualization method based on the Gantt chart is used to generate a timeline representation of the decoration progress, and the construction process is mapped and associated with the time nodes on the timeline. Obtain the actual decoration progress data to determine whether the actual decoration progress is consistent with the planned progress in the construction plan. If the actual progress lags behind the planned progress, determine the decoration objects with lagging progress. In the decoration virtual reality scene, obtain the three-dimensional model data of the decoration objects with lagging progress, and display the decoration objects with lagging progress in red by modifying the material properties of the three-dimensional model. Obtain the actual decoration progress data and construction plan data, and determine the degree of deviation between the actual decoration progress and the construction plan through data comparison. If the deviation exceeds the preset threshold, trigger an early warning prompt. According to the logical dependency relationship of the construction plan, the critical path method is used to calculate the earliest start time and the latest completion time of the decoration project, determine the critical path of the project, and optimize the decoration schedule. Obtain the construction resource data of the decoration project, including personnel, materials, equipment, etc., and use the resource balancing algorithm according to the process arrangement of the construction plan to optimize the allocation and scheduling of construction resources and improve the progress and efficiency of decoration.

[0029] S106, immersively roaming the decoration virtual reality scene through a virtual reality head display device to obtain interactive operation data of the user's decoration materials and construction plans; using a recommendation algorithm based on user collaborative filtering to recommend an optimized decoration plan in the decoration virtual reality scene according to the interactive operation data.

[0030] The user's viewpoint position and sight direction data in the decoration virtual reality scene are obtained through the virtual reality head display device. According to the obtained viewpoint position and sight direction data, the three-dimensional decoration scene image under the corresponding viewpoint is rendered in the decoration virtual reality scene to provide an immersive decoration roaming experience. The user's selection operation data on decoration materials and construction plans in the decoration virtual reality scene is obtained. According to the user's selection operation data, the collaborative filtering recommendation algorithm is used to calculate the similarity of different decoration materials and construction plans, and the recommended decoration materials and construction plans with a high degree of matching with the user's preferences are obtained. According to the recommended decoration materials and construction plans, the material mapping and layout parameters of the three-dimensional model are dynamically adjusted in the decoration virtual reality scene, and the optimized decoration virtual reality scene is rendered to provide users with personalized decoration design recommendations. The interactive operation data of different users in the decoration virtual reality scene are obtained, and users with similar preferences are grouped through the clustering algorithm to build user preference portraits, and the corresponding decoration plans are recommended for user groups with different preferences. According to the user's decoration budget and area and other conditions, the recommended decoration plans are cost estimated and optimized, and the decoration plans that meet the user's budget and area requirements and have a high comprehensive score are selected for key recommendation. Obtain user feedback and evaluation data on recommended decoration plans, and use a reinforcement learning algorithm based on user feedback to dynamically adjust the parameter weights of the recommendation strategy to improve the accuracy of decoration plan recommendations and user satisfaction. Based on the user's historical decoration preference data and house type characteristics, a reasoning algorithm based on knowledge graphs is used to automatically generate decoration plans that meet user preferences and house types, providing intelligent decoration design services.

[0031] S107. Discuss and modify the optimized decoration plan in a virtual space shared by multiple people through a virtual reality collaborative platform, and use a distributed data synchronization mechanism to synchronize the decoration virtual reality scenes between different users. The synchronization delay of the distributed data synchronization mechanism does not exceed 20ms.

[0032] Step 1: Obtain the user's operation data in the virtual reality scene. Including interactive information such as user perspective position, rotation angle, hand movements, voice commands, etc. Record the user's modification operations on the decoration elements in the virtual scene (such as furniture, walls, floors, etc.), such as moving, rotating, scaling, replacing, etc. Step 2: Encode and compress the obtained user operation data, and use efficient compression algorithms to reduce the amount of data and reduce the network transmission load. Assign a unique identity to each user to distinguish the data of different users. Add timestamp information to the encoded data for subsequent data synchronization and consistency maintenance. Step 3: Broadcast the encoded and compressed user operation data to other users through a distributed data synchronization mechanism. Use a reliable network transmission protocol to ensure the integrity and reliability of data transmission. Dynamically adjust the data transmission rate according to the user's network status to avoid network congestion. Step 4: Receive the operation data broadcast by other users. Decompress and decode the received data to restore the original operation data. Sort according to the timestamp information in the data to ensure the order consistency of the operation. If the timestamp of the received data is earlier than the timestamp of the current scene, the data is discarded. Step 5: Update the local virtual reality scene based on the received operation data. Apply the received operation data to make corresponding modifications to the decoration elements in the virtual scene. Synchronize the scene states between different users through spatial consistency algorithms, such as the state vector synchronization algorithm, to ensure scene consistency. Step 6: Use a prediction algorithm to predict the user's next operation. Predict the user's next possible operation based on the user's historical operation data and current operation trends. Pre-load scene resources related to the predicted operation, such as models, maps, etc., to reduce operation delays. Step 7: Pre-render according to the prediction results. Pre-render the virtual reality scene for the predicted user operation and generate the rendering results in advance. If the user's actual operation is consistent with the predicted result, the pre-rendered result is directly displayed to improve the scene response speed.

[0033] S108. According to the evaluated decoration plan, the construction plan in the order system is updated to generate a decoration task work order; the construction results of each stage are inspected in the decoration virtual reality scene, and the inspection results are synchronized to the order system.

[0034] According to the evaluated decoration plan, the construction plan information is obtained, and the corresponding construction plan data in the order system is updated. For the updated construction plan, the decoration task work order is automatically generated, and the work order information is entered into the order system. The decoration task work order in the order system is obtained, and the corresponding 3D model and material map are loaded in the decoration virtual reality scene according to the work order information. In the decoration virtual reality scene, the construction results of each stage are visually simulated and displayed according to the stage division of the construction plan. For the construction results of each stage, acceptance evaluation is carried out in the virtual reality scene, and the acceptance results are analyzed by artificial intelligence algorithms to determine whether the preset acceptance standards are met. If the acceptance result meets the standard, the acceptance result of this stage is marked as qualified, and the acceptance data is synchronously updated to the task work order corresponding to this stage in the order system. If the acceptance result does not meet the standard, the acceptance result of this stage is marked as unqualified, a rectification notice is generated, and the rectification requirements are synchronized to the order system and task work order, and the time node of the construction plan is updated at the same time.

[0035] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A renovation progress management method based on virtual reality, characterized in that: The method comprises: Obtaining point cloud data of the decoration site, the point cloud density of the point cloud data is not less than 1000 points per square meter, the point cloud accuracy of the point cloud data is better than 5mm, the noise level of the point cloud data is less than 1%, and obtaining a video stream of the decoration site, the frame rate of the video stream is not less than 30 frames per second, and the resolution of the video stream is not less than 1080P; The incremental dynamic modeling technology of octree space partitioning is adopted, and the changing part of the point cloud data is reconstructed in real time in three dimensions through the Poisson surface reconstruction algorithm to obtain a three-dimensional mesh model of the decoration site. The number of triangles in the three-dimensional mesh model is not less than 1.5 times the total number of points in the point cloud data, and the texture mapping resolution of the three-dimensional mesh model is not less than 4096×4096; Performing scene semantic segmentation on the three-dimensional grid model to obtain the main structure of the wall, ground, and ceiling, the semantic segmentation accuracy of the scene semantic segmentation is not less than 90%, and using a regional convolutional neural network to perform decoration object detection on the three-dimensional grid model to obtain doors, windows, pipelines, and furniture decoration objects, the average detection accuracy of the decoration object detection is not less than 85%; Acquire a pre-built decoration material database, perform material recognition on the decoration object according to the decoration material database, obtain material attribute information of paint, coating, tile, floor, wallpaper decoration materials, wherein the recognition accuracy of the material recognition is not less than 80%, map the material attribute information to the corresponding part of the three-dimensional grid model, and generate a decoration virtual reality scene; Obtain the construction plan in the order system, annotate the decoration objects in the decoration virtual reality scene according to the construction process according to the construction plan, generate a timeline representation of the decoration progress by using a visualization method based on a Gantt chart, and judge whether the actual decoration progress is consistent with the construction plan. If not, highlight the decoration objects with lagging progress in red in the decoration virtual reality scene; Immersing the virtual reality scene of decoration through a virtual reality head display device, obtaining interactive operation data of the user's decoration materials and construction plans, and using a recommendation algorithm based on user collaborative filtering to recommend an optimized decoration plan in the virtual reality scene of decoration according to the interactive operation data; Through the virtual reality collaborative platform, the optimized decoration plan is discussed and modified in a virtual space shared by multiple people, and a distributed data synchronization mechanism is used to synchronize the decoration virtual reality scenes between different users. The synchronization delay of the distributed data synchronization mechanism does not exceed 20ms; According to the evaluated decoration plan, the construction plan in the order system is updated, a decoration task work order is generated, the construction results of each stage are accepted in the decoration virtual reality scene, and the acceptance results are synchronized to the order system.

2. The method according to claim 1, characterized in that The point cloud data of the decoration site is obtained, the point cloud density of the point cloud data is not less than 1000 points per square meter, the point cloud accuracy of the point cloud data is better than 5mm, and the noise level of the point cloud data is less than 1%; Obtain a video stream of the renovation site, the frame rate of which is not less than 30 frames per second, and the resolution of which is not less than 1080P, including: According to the layout and area of ​​the decoration site, determine the appropriate point cloud scanning equipment and scanning parameters to obtain the original point cloud data with a density of not less than 1,000 points per square meter and an accuracy better than 5mm; Pre-process the acquired original point cloud data, remove noise points through point cloud noise reduction algorithm, control the noise level below 1%, and obtain high-precision decoration site point cloud data; According to the lighting conditions and shooting requirements of the decoration site, select appropriate high-definition camera equipment, set appropriate shooting parameters, and obtain high-quality video stream data with a frame rate of no less than 30 frames per second and a resolution of no less than 1080P; The point cloud data and video stream data acquired at the decoration site are synchronized in time and space, and the coordinate systems of the two are unified through the feature point matching algorithm to achieve accurate registration of the point cloud and video; The synchronized point cloud and video data are fused and processed, and the point cloud is colored by mapping the video frame onto the point cloud to generate colored point cloud data with real texture and color information. According to the requirements of the decoration design plan, the color point cloud data is semantically segmented to identify and extract different decoration components, such as walls, floors, ceilings, doors and windows, to form a semantic 3D model of the decoration site; The semantic 3D model of the decoration site is compared with the design plan. Through point cloud registration and deviation analysis algorithms, the dimensional deviation and process quality problems of the decoration construction are detected, and a construction quality assessment report of the decoration site is generated.

3. The method according to claim 2, characterized in that The incremental dynamic modeling technology using octree space partitioning performs real-time 3D reconstruction of the changed part of the point cloud data through the Poisson surface reconstruction algorithm to obtain a 3D mesh model of the decoration site, wherein the number of triangles in the 3D mesh model is not less than 1.5 times the total number of points in the point cloud data, and the texture mapping resolution of the 3D mesh model is not less than 4096×4096, including: Acquire three-dimensional point cloud data of the decoration site, and use an octree algorithm to perform spatial division on the point cloud data to divide the point cloud data into a number of voxels; For the point cloud data in each voxel, local 3D reconstruction is performed using the Poisson surface reconstruction algorithm to obtain a local 3D mesh model. Determine whether the point cloud data has changed. If it has changed, obtain the point cloud data of the changed part, and repeat step 2 for the point cloud data of the changed part to obtain a local three-dimensional mesh model of the changed part; The local three-dimensional mesh model of the changed part is merged with the original three-dimensional mesh model to obtain an updated three-dimensional mesh model of the decoration site; For the updated three-dimensional mesh model, the triangular facets are reduced by a mesh simplification algorithm, and the number of simplified triangular facets is controlled to be no less than 5 times the total number of points in the point cloud data; According to the texture information of the three-dimensional mesh model, a texture mapping algorithm is used to generate a texture map of the three-dimensional mesh model, and a resolution of the texture map is controlled to be not less than 4096×4096; The simplified 3D mesh model and the generated texture map are combined to obtain the final 3D model of the decoration site.

4. The method according to claim 3, characterized in that The three-dimensional grid model is subjected to scene semantic segmentation to obtain the main structure of the wall, the ground, and the ceiling, and the semantic segmentation accuracy of the scene semantic segmentation is not less than 90%; the three-dimensional grid model is subjected to decoration object detection by using a regional convolutional neural network to obtain doors, windows, pipelines, and furniture decoration objects, and the average detection accuracy of the decoration object detection is not less than 85%, including: Acquire the three-dimensional mesh model data, perform preprocessing on it, including denoising and smoothing operations, and obtain an optimized three-dimensional mesh model; The optimized three-dimensional mesh model is semantically segmented according to a preset semantic segmentation model to obtain the segmented wall, ground and ceiling main structures; Determine whether the semantic segmentation accuracy of the segmented main structure is not less than 90%. If so, use the segmentation result as the input for subsequent decoration object detection. If not, return to step 2 to perform semantic segmentation again. The pre-trained regional convolutional neural network model is used to detect the decoration objects of the segmented main structure, and the location and category information of doors, windows, pipelines, and furniture decoration objects are obtained; Calculate the average detection accuracy of the decoration object detection and determine whether it is not less than 85%. If so, use the detection result as input for subsequent decoration design. If not, return to step 4 to re-detect the decoration object. According to the results of semantic segmentation and decoration object detection, the functional attributes and spatial relationships of each area in the 3D grid model are determined to provide data support for subsequent decoration design; The results of semantic segmentation and decoration object detection are visualized to generate intuitive three-dimensional decoration renderings, which are convenient for users to view and modify, thereby improving the efficiency and quality of decoration design.

5. The method according to claim 4, characterized in that The method comprises: obtaining a pre-built decoration material database, performing material identification on the decoration object according to the decoration material database, obtaining material attribute information of paint, coating, tile, floor, and wallpaper decoration materials, wherein the recognition accuracy of the material identification is not less than 80%; mapping the material attribute information to the corresponding part of the three-dimensional grid model, and generating a decoration virtual reality scene, including: Get a pre-built decoration material database that contains the property information of various decoration materials, such as the color and glossiness of paints and coatings, and the material and texture of tiles and floors; For a given decoration object, an image segmentation algorithm based on deep learning is used to divide the image of the decoration object into different material areas; According to the divided material areas, the texture and color features of each area are extracted, matched with the material properties in the decoration material database, and the material type and properties corresponding to each area are identified; In the process of material identification, the accuracy of material identification is increased to more than 80% by increasing training data and optimizing algorithms; Map the identified material property information to the three-dimensional mesh model of the decoration object, and assign corresponding texture and color properties to each triangle according to its corresponding material properties; For unrecognized areas, a texture synthesis algorithm is used to synthesize material textures that are coordinated with the surrounding areas based on the material properties of the surrounding recognized areas. Finally, after the decoration material properties are completely mapped to the three-dimensional mesh model, a realistic decoration virtual reality scene is rendered and generated. Users can roam freely in the virtual scene and experience the decoration effect.

6. The method according to claim 5, characterized in that The method includes: obtaining a construction plan in the order system, annotating the decoration objects in the decoration virtual reality scene according to the construction process according to the construction plan; generating a timeline representation of the decoration progress by using a visualization method based on a Gantt chart; judging whether the actual decoration progress is consistent with the construction plan, and if not, highlighting the decoration objects with lagging progress in red in the decoration virtual reality scene, including: Obtain the construction plan data in the order system, and determine the corresponding decoration objects and attribute marking rules for each construction process in the construction plan; According to the process time and sequence of the construction plan, a visualization method based on the Gantt chart is used to generate a timeline representation of the decoration progress, and the construction process is mapped and associated with the time nodes on the timeline; the actual decoration progress data is obtained to determine whether the actual decoration progress is consistent with the planned progress in the construction plan. If the actual progress lags behind the planned progress, the decoration objects with lagging progress are determined; In the decoration virtual reality scene, the three-dimensional model data of the decoration object with delayed progress is obtained, and the decoration object with delayed progress is displayed in red highlight by modifying the material properties of the three-dimensional model; Obtain actual renovation progress data and construction plan data, and determine the degree of deviation between the actual renovation progress and the construction plan through data comparison. If the deviation exceeds the preset threshold, an early warning prompt will be triggered; According to the logical dependency of the construction plan, the critical path method is used to calculate the earliest start time and the latest completion time of the decoration project, determine the critical path of the project, and optimize the decoration schedule; Obtain the construction resource data of the decoration project, including personnel, materials and equipment, and use the resource balancing algorithm according to the process arrangement of the construction plan to optimize the allocation and scheduling of construction resources and improve the progress and efficiency of decoration.

7. The method according to claim 6, characterized in that The method comprises: performing immersive roaming of the decoration virtual reality scene through a virtual reality head display device to obtain interactive operation data of the user's decoration materials and construction plans; adopting a recommendation algorithm based on user collaborative filtering to recommend an optimized decoration plan in the decoration virtual reality scene according to the interactive operation data, including: The user's viewing angle position and sight direction data in the decoration virtual reality scene are obtained through the virtual reality head display device, and a three-dimensional decoration scene image under the corresponding viewing angle is rendered in the decoration virtual reality scene according to the obtained viewing angle position and sight direction data, so as to provide an immersive decoration roaming experience; Obtain the user's selection operation data on decoration materials and construction plans in the decoration virtual reality scene, and use the collaborative filtering recommendation algorithm to calculate the similarity of different decoration materials and construction plans based on the user's selection operation data, and obtain recommended decoration materials and construction plans that have a high degree of match with user preferences; According to the recommended decoration materials and construction plans, the material mapping and layout parameters of the 3D model are dynamically adjusted in the decoration virtual reality scene, and the optimized decoration virtual reality scene is rendered to provide users with personalized decoration design recommendations; Obtain the interactive operation data of different users in the decoration virtual reality scene, group users with similar preferences through clustering algorithms, build user preference portraits, and recommend corresponding decoration plans for user groups with different preferences; According to the user's decoration budget and area conditions, the recommended decoration schemes are cost estimated and optimized, and the decoration schemes that meet the user's budget and area requirements and have high comprehensive scores are selected for key recommendations; Obtain user feedback and evaluation data on recommended decoration plans, and use a reinforcement learning algorithm based on user feedback to dynamically adjust the parameter weights of the recommendation strategy to improve the accuracy of decoration plan recommendations and user satisfaction; According to the user's historical decoration preference data and house type characteristics, a reasoning algorithm based on knowledge graph is used to automatically generate decoration plans that meet the user's preferences and house type, providing intelligent decoration design services.

8. The method according to claim 7, characterized in that The optimized decoration plan is discussed and modified in a virtual space shared by multiple people through a virtual reality collaborative platform, and a distributed data synchronization mechanism is used to synchronize the decoration virtual reality scenes between different users. The synchronization delay of the distributed data synchronization mechanism does not exceed 20ms, including: Step 1: Obtain the user's operation data in the virtual reality scene; Including user viewing angle, rotation angle, hand movements and voice command interaction information; Record users' responses to decoration elements in virtual scenes.

9. The method according to claim 8, characterized in that According to the evaluated renovation plan, the construction plan in the order system is updated to generate a renovation task work order; Accepting the construction results of each stage in the decoration virtual reality scene, and synchronizing the acceptance results to the order system, including: According to the evaluated renovation plan, obtain the construction plan information and update the corresponding construction plan data in the order system; Automatically generate decoration task work orders for the updated construction plan and enter the work order information into the order system; Obtain the decoration task work order in the order system, and load the corresponding 3D model and material map in the decoration virtual reality scene according to the work order information; In the decoration virtual reality scene, according to the phases of the construction plan, the construction results of each phase are visualized and simulated; For each stage of construction results, acceptance evaluation is carried out in a virtual reality scene, and the acceptance results are analyzed through artificial intelligence algorithms to determine whether the preset acceptance standards are met; If the acceptance result meets the standard, the acceptance result of this stage will be marked as qualified, and the acceptance data will be synchronously updated to the task work order corresponding to this stage in the order system; If the acceptance result does not meet the standards, the acceptance result of this stage will be marked as unqualified, a rectification notice will be generated, and the rectification requirements will be synchronized to the order system and task work order, and the time node of the construction plan will be updated at the same time.

Images