VR-based building management method and device, and medium

Abstract

The invention discloses a VR-based building management method and device and a medium, and relates to the field of three-dimensional image generation, and the method comprises the steps: obtaining a building information model of a specified building; based on the geometric information, modeling each component in the specified building to obtain a component model; modeling a construction tool required by a specified building to obtain a tool model; obtaining a building safety management task corresponding to the user according to the user information; in the VR environment, assembling the required component models to obtain a target building model; and determining to execute the building safety management task, collecting and analyzing the user behavior, and evaluating the user based on the collected user behavior. And according to different building safety management tasks, the component models are assembled, so that the target building model is obtained, and constructors can conveniently and flexibly cope with different scenes.

Description

Technical Field

[0001] The present application relates to the field of three-dimensional image generation, and specifically to a VR-based building management method, device and medium. Background Art

[0002] With the development of science and technology, virtual reality (VR) technology has gradually come into people's view, and its combination with construction management has received widespread attention in recent years. Through VR technology, architects, engineers, project managers and other relevant personnel can communicate, collaborate and make decisions more efficiently at all stages of project design, construction and operation, thereby improving the efficiency and quality of construction projects.

[0003] In the construction industry, safety management is of vital importance. VR can provide construction workers with a real virtual environment for safety training simulations, helping them experience dangerous scenarios without getting hurt, thereby improving their safety awareness and emergency response capabilities.

[0004] However, in traditional construction management, safety management is usually pre-set with fixed scenarios, and then construction workers perform fixed actions in the fixed scenarios. This makes it difficult to simulate the changing situations in the actual working environment and fails to train construction workers' ability to face various scenarios flexibly. Summary of the invention

[0005] In order to solve the above problems, this application proposes a VR-based building management method, including: Obtaining a building information model of a specified building; wherein the building information model includes geometric information, construction information, and collaboration information; Based on the geometric information, each component in the designated building is modeled to obtain a component model, and the component model is added to a component model library corresponding to the VR environment; and construction tools required for the designated building are modeled to obtain a tool model, and the tool model is added to a tool model library corresponding to the VR environment; Determine that a user logs into the VR environment, identify the user information, and obtain a building safety management task corresponding to the user based on the user information; Based on the building safety management task, according to the construction process information contained in the building information, in the VR environment, the required component models are assembled to obtain a target building model; and the required tool models are added in the VR environment; Determine to execute the building safety management task, and during the task execution, issue instructions to the user through the VR environment, collect and analyze the user behavior, and evaluate the user based on the collected user behavior.

[0006] On the other hand, the present application also proposes a VR-based building management device, comprising: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the VR-based building management method as described in the above example.

[0007] On the other hand, the present application also proposes a non-volatile computer storage medium storing computer executable instructions, wherein the computer executable instructions are configured as the VR-based building management method described in the above example.

[0008] The VR-based building management method proposed in this application can bring the following beneficial effects: Safety management in construction management is crucial. Through VR, a real virtual environment can be provided for construction workers for various safety training simulations, helping them experience dangerous scenarios without getting hurt, thereby improving their safety awareness and emergency response capabilities.

[0009] Construction workers can experience high-risk links such as aerial work and electrical facility operation in person, and improve their ability to respond to emergencies by simulating different safety accident scenarios (such as fire, equipment failure, etc.). In addition, VR training can also reduce the resource and time costs brought by field exercises.

[0010] In order to solve the problem that it is difficult to exercise the flexibility of construction workers in fixed scenes in VR environment, the component models are assembled according to different building safety management tasks to obtain the target building model, which is convenient for construction workers to flexibly respond to different scenes. Moreover, the randomness of the target building model can be further increased by replacing the model, thereby further increasing the flexibility of the scenes faced by construction workers. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings: Figure 1 This is a flow chart of a VR-based building management method in an embodiment of the present application; Figure 2 This is a schematic diagram of the architecture of a VR system in an embodiment of the present application; Figure 3 This is a schematic diagram of the sequence of process nodes in the embodiment of the present application; Figure 4This is a schematic diagram of a VR-based building management device in an embodiment of the present application. DETAILED DESCRIPTION

[0012] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0013] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0014] like Figure 1 As shown, the embodiment of the present application provides a VR-based building management method, including: S101: Acquire a building information model of a designated building; wherein the building information model includes geometric information, construction information, and collaboration information.

[0015] Building Information Modeling (BIM) is a full-process management method for building design, construction, operation and management based on digital technology. The BIM model is a data platform that integrates all the information in the project. Through this platform, the whole life cycle management of building planning, design, construction, operation and maintenance can be carried out. The BIM model can achieve informatization, collaboration, visualization and intelligence.

[0016] The designated building refers to the currently selected building, which is selected according to the training needs. You can choose a building that is under construction or has been completed, as long as it can obtain the building information model.

[0017] The geometric information mainly includes the three-dimensional spatial position and size information of each part of the specified building (referred to as the first size information here). For example, it may include the position and size of each floor in the building, as well as the three-dimensional spatial position and first size information of each wall, load-bearing column, door and window on each floor. Based on needs, it may also include the three-dimensional spatial position and first size information of items such as wall tiles, steel bars, electrical equipment, and placed furniture.

[0018] Construction information mainly includes the types of materials that make up the building, the physical characteristics of the building structure, construction process information, etc. Collaboration information mainly includes engineering data, construction drawings, project progress, construction personnel, and construction tools used.

[0019] S102: Based on the geometric information, each component in the designated building is modeled to obtain a component model, and the component model is added to a component model library corresponding to the VR environment; and the construction tools required for the designated building are modeled to obtain a tool model, and the tool model is added to a tool model library corresponding to the VR environment.

[0020] like Figure 2 As shown, the VR system includes multiple modules, namely the VR device worn by the user, the cloud for data processing, the cloud including the component model library, the tool model library and the action library described below, etc. The cloud generates the VR environment and analyzes and processes the data.

[0021] Generally speaking, there are already three-dimensional models corresponding to each component and each construction tool in the building information model (for example, there are usually three-dimensional models corresponding to each room in the building information model. If a room is regarded as a component of a building, it can be directly added to the component model library), and it can be directly added to the corresponding component model library and tool model library.

[0022] However, in order to facilitate the subsequent adjustment of the building model and randomization processing, the components are first defined here, and the components are defined according to the first size information of each object and the composition of each object.

[0023] Specifically, first obtain the unit size information of a single object recorded in the geometric information of the specified building. For example, a single object may include: a single wall brick, a wall, a potted plant, electrical equipment, doors and windows, etc. Any object that has specific size information recorded can be treated as a single object to obtain its unit size information.

[0024] If there is a single area, which is obtained by splicing multiple identical and continuous unit size information, then the single area is regarded as a single component. For example, for a wall, it is composed of multiple identical and continuous wall tiles; for a floor, it is composed of multiple identical and continuous floor tiles; for a lawn, it is composed of multiple identical and continuous lawn blocks or potted plants. Here, the wall, floor, and lawn can all be regarded as a single component.

[0025] Each unit size information in the remaining areas except all the single areas is treated as a single component. In addition to the areas obtained by stitching together multiple objects, each other unit size information (such as electrical equipment, doors and windows, load-bearing columns, lightning rods, etc.) can be treated as a single component.

[0026] At this time, modeling is performed for each component to obtain the corresponding component model. If it already exists in the BIM model, the 3D model can be directly called as the component model.

[0027] Furthermore, in order to increase the randomness of the construction workers in the safety training process, according to the preset component type, a component model that meets the requirements is selected from all component models as the designated component model, and an alternative model of the designated component model is generated.

[0028] The types of safety training may include: high-altitude outdoor construction, fire escape training, electrical equipment use training, etc. For different scenarios of different types of safety training, corresponding types of component models can be set as designated component models, so as to increase the randomness of the scene. For example, for high-altitude outdoor construction, designated component models may include: wall components, glass curtain walls, air conditioner outdoor units, exterior wall lights, etc. For fire escape training, designated component models may include: home appliances inside the building, doors and windows inside the building, lights inside the building, etc. For electrical equipment use training, designated component models may include: the equipment itself, the environment of the room where the equipment is located, etc.

[0029] Here, corresponding alternative models are generated for the specified component models of all types of safety training.

[0030] Among them, for the designated component model corresponding to a single area, the means of obtaining the replacement model includes at least one of the following: replacing and / or deleting at least part of the unit size information in the designated component model, or changing the properties of the designated component model. For example, for a wall component, the unit size information of at least part of the wall bricks can be deleted, so that the wall becomes a damaged wall or a wall under construction, or the unit size information of part of the wall bricks can be replaced, so that the wall component becomes a different style (for example, different decoration, different materials, etc.). Similarly, for a lawn component, lawn components of different sizes and different plants can be obtained.

[0031] For the designated component model corresponding to the single unit size information, the means of obtaining the replacement model include at least one of the following: replacing the unit size information, changing the properties of the designated component model. Compared with the designated component model corresponding to the region, since it is composed of only a single object, the unit size information of the internal components cannot be changed. If the size needs to be changed, the overall unit size information of the unit needs to be replaced.

[0032] Changing the properties of a specified component model means that when a component model is generated, the corresponding properties will be assigned to it. For example, for a door and window component, it has the property of whether it can be opened. If it can be opened, the corresponding opening animation can be set in the door and window component, as well as the state transition after opening (including whether it is allowed to pass through, volume transformation, etc.). For a wall component, it has the property of whether it can be edited (including destruction, continued construction, etc.). For electrical equipment components (including lighting components, elevator components, etc.), it has the property of whether it can be started to work. Of course, you can also set common properties for each component, such as the current temperature of the component, the smoothness of the surface material, etc.

[0033] For construction tools, the construction tools required for a specified building can be obtained, and the size information (herein referred to as the second size information) and function information corresponding to the construction tools can be obtained. For example, measuring tools (including level, laser rangefinder, etc.), masonry tools (including shovel, brick cutter, etc.), installation tools (including wrench, cutter, etc.) and other tools, and the corresponding function information is determined according to the purpose and function of each tool.

[0034] At this time, modeling is performed based on the size information to obtain a tool model, and corresponding function labels are generated for the tool model based on the function information. For example, the function labels of a level include: measuring plane, the function labels of a laser rangefinder include: measuring distance, and the function labels of a shovel include: digging materials, throwing materials.

[0035] S103: Determine whether a user logs into the VR environment, identify the user information, and obtain the building safety management task corresponding to the user according to the user information.

[0036] After the user wears the VR device (including VR helmet, controller, etc.), he logs into the VR environment created by the VR platform and logs into his own account. The VR platform identifies the user's account information, obtains the user information, and determines the building safety management tasks that have been issued to the user. Building safety management tasks can include: high-altitude outdoor construction tasks, fire escape tasks, electrical equipment use tasks, etc.

[0037] S104: Based on the building safety management task and according to the construction process information included in the building information, in the VR environment, the required component models are assembled to obtain a target building model; and in the VR environment, the required tool models are added.

[0038] Generally speaking, there may be component models collected from multiple specified buildings in the component model library, so the component models of different specified buildings can be partitioned and classified for storage in the component model library. In the actual assembly process, if there is a demand, a component model of a specified building can be selected, and if there is no demand, it can be randomly selected.

[0039] Based on the building safety management task, determine the required construction progress. For example, for a high-altitude outdoor construction task, if it corresponds to the bricklaying process, its construction progress can be that the main structure of the building has not been completed. If it corresponds to tasks such as installing an air-conditioning outdoor unit, the construction progress can be that the main structure of the building has been completed. If the building safety management task is a fire escape task, the construction progress can be that the main building and the interior decoration are all completed.

[0040] Based on the construction progress, the required component models and the corresponding construction sequence are determined according to the construction process information contained in the build information. For example, if the construction progress is that the main structure of the building has not been completed, the required component model may only include basic components such as some wall components and load-bearing column components. If the construction progress is that the main structure of the building has been completed, it is also necessary to add components such as glass curtain walls and roof structures. If the construction progress is that the main building and interior decoration are all completed, it is also necessary to add components such as internal doors and windows and internal electrical equipment.

[0041] The BIM model stores construction process information, so that the required component models are assembled in sequence according to the construction sequence and the numbers of each component model. During the assembly process, the assembly position is determined by the three-dimensional spatial position information contained in the geometric information.

[0042] Furthermore, in order to allow construction workers to face flexible situations and increase the diversity of scenarios that construction workers can face during safety training, during the assembly process, based on the building safety management task, the specified component model of the corresponding component type is selected, and its corresponding alternative model is used to replace the specified component model to assemble and obtain the target building model.

[0043] For example, in the high-altitude outdoor construction task, the wall component is replaced with its corresponding alternative model (select the alternative model after deletion) so that the construction workers can face walls with different construction progress. For another example, in the fire escape task, the door and window components are replaced with the corresponding alternative model (select the alternative model after changing the properties) so that the construction workers can face doors and windows that can be opened or cannot be opened, so as to test their ability to adapt to changes.

[0044] As for the tool model, in the VR environment, the required construction tools are determined based on the collaboration information. Generally speaking, the collaboration information will include the construction tools used in this construction. At this time, the function label corresponding to the required construction tool is obtained, and other construction tools with at least one identical function label are searched among the remaining construction tools except the required construction tool, so as to add at least one tool model corresponding to the other construction tool in the VR environment.

[0045] For example, the construction tools required this time include tool A (with function label A and function label B) and tool B (with function label C). At this time, among the remaining tools, look for tools with function label A, function label B, or function label C, and then select at least one construction tool (for example, function label C, which has function label B) and add it to the VR environment.

[0046] Since the newly added construction tools also have the same functional labels, the choice of tools is somewhat confusing for construction workers, which can test the construction workers' correct choice of matching tools. Moreover, in some cases, the staff chooses the newly added construction tools and completes the same work. At this time, it can also increase the range of tool selection for construction workers and increase their proficiency in using different construction tools.

[0047] S105: Determine to execute the building safety management task, and during the task execution process, issue instructions to the user through the VR environment, collect and analyze the user behavior, and evaluate the user based on the collected user behavior.

[0048] In the process of building safety management tasks, the user's execution actions are collected in real time through VR equipment, so as to evaluate the user.

[0049] Specifically, each building safety management task has multiple pre-set process nodes, and in each process node, required actions and / or prohibited actions are set. Figure 3 As shown, in the fire escape task, the process nodes include: the first process node (corresponding to the response when the fire is just discovered in the room area), the second process node (the action when judging the timing is right and preparing to go to the escape passage), the third process node (the action performed when escaping in the escape passage), etc. Of course, in the fire escape task, the first process node may also directly correspond to the fourth process node (the action performed when judging the timing is not suitable for escaping and waiting for rescue in the room).

[0050] Taking the second process node as an example, the necessary actions include: covering the mouth and nose with a wet cloth, finding a piece of cloth that can cover the whole body, wetting it and then draping it over the body, etc. The prohibited actions include: opening the window and jumping off the building, blindly opening the fire doors and windows, etc.

[0051] At this point, in each process node, user behavior is collected to determine the user's current execution action.

[0052] If the currently executed action hits a single necessary action, and all the preceding necessary actions of the single necessary action have been executed in the current process node, and the single necessary action is preset with a corresponding model change attribute, then according to the model change attribute, the corresponding model change action is executed in the target building model.

[0053] Still taking the second process node as an example, if the current scenario is suitable for escape, and the user also determines that he can escape, and has performed the two necessary prerequisite actions mentioned above "covering the mouth and nose with a wet cloth, and finding a cloth that can cover the whole body and then wetting it and draping it over the body", then when he continues to perform the next necessary action "open the door and prepare to go to the escape route", it is determined that all the prerequisite necessary actions of the single necessary action have been performed, and the single necessary action is preset with a corresponding model change attribute, which is to change the model corresponding to the door from a closed state to an open state in the target building model. At this time, the model change attribute is executed, and the corresponding model change action is executed in the target building model to open the door.

[0054] Of course, if the currently executed action hits a single prohibited action (for example, blindly opening fire doors and windows), the target building model will be restored to the initial state of the current process node, allowing the user to execute the current process node from the beginning.

[0055] Among them, for each alternative model, since it tests the user's ability to adapt to changes, it is necessary to set corresponding necessary actions and / or prohibited actions, otherwise, the setting of the alternative model is meaningless. For example, still taking the fire escape task as an example, a wall or railing is set as an alternative model, and its corresponding temperature attribute is set to high temperature, and the corresponding prohibited action is set to "touch the wall or railing". Once the user triggers it, the current process node will be restored to the initial state.

[0056] At this time, when evaluating the user, the basic score corresponding to the user is obtained based on the total time the user takes to complete the building safety management task. The shorter the time, the higher the basic score.

[0057] Obtain the prohibited actions that users perform during the execution of building safety management tasks, record the prohibited actions, and deduct points from the basic score based on the prohibited actions. The purpose of recording is to show the user after the user completes the task so that the user can learn from the lesson.

[0058] Of course, it is also possible to obtain other actions that the user has taken in addition to necessary actions and prohibited actions during the execution of the building safety management task. For example, still taking the fire escape task as an example, other actions may include "cutting off the power of household appliances" and "going back and forth in the room area many times". These other actions will not have serious consequences for the escape, but they can indirectly experience whether the current user is calm, so the basic score can be corrected based on other actions taken. For example, "cutting off the power of household appliances" corresponds to a positive correction, and "going back and forth in the room area many times" corresponds to a negative correction. Of course, the other actions here can also include some simpler actions, such as raising your hand, shaking your head, etc., which can be set based on needs.

[0059] Safety management in construction management is crucial. Through VR, a real virtual environment can be provided for construction workers for various safety training simulations, helping them experience dangerous scenarios without getting hurt, thereby improving their safety awareness and emergency response capabilities.

[0060] Construction workers can experience high-risk links such as aerial work and electrical facility operation in person, and improve their ability to respond to emergencies by simulating different safety accident scenarios (such as fire, equipment failure, etc.). In addition, VR training can also reduce the resource and time costs brought by field exercises.

[0061] In order to solve the problem that it is difficult to exercise the flexibility of construction workers in fixed scenes in VR environment, the component models are assembled according to different building safety management tasks to obtain the target building model, which is convenient for construction workers to flexibly respond to different scenes. Moreover, the randomness of the target building model can be further increased by replacing the model, thereby further increasing the flexibility of the scenes faced by construction workers.

[0062] In one embodiment, when collecting and analyzing user behavior, real-time input data generated according to user behavior can be collected through the interface of the VR device worn by the user (such as a VR helmet, a handle, a motion capture device, etc.). The real-time input data is usually in multiple dimensions, such as the user's head data, hand data, leg data, etc., including the user's perspective, head position, hand position, rotation angle, handheld objects, leg movements, etc.

[0063] Define the action library, add target actions defined by user behavior to the action library, and add corresponding action tags to the target actions. In a VR environment, a set of target actions need to be defined. These actions are tasks or interactive behaviors that users need to complete during task execution. For example, "wet the cloth and cover your mouth and nose" mentioned above. Of course, in order to make the action easier to understand, it can be divided into multiple sub-actions to constitute the target action. The sub-actions include: picking up the cloth with your hands, going to the water source, moving the cloth in your hands, touching the cloth with the water source, and moving the cloth to the mouth and nose.

[0064] By analyzing the target action and expressing it as a set of conditions or features, each sub-action is decomposed to obtain the corresponding posture, motion trajectory, time sequence, etc., and then spliced ​​together to achieve the definition of the target action.

[0065] At this time, the real-time input data is preprocessed and feature extracted to obtain multi-dimensional behavior features corresponding to the user behavior. Feature extraction may include: position and speed (e.g., the user's movement trajectory, speed, acceleration in the VR environment), angle and posture (e.g., the rotation angle of the user's arms, legs, and head), and timing data (e.g., the start and end time of each action).

[0066] After preprocessing the data (e.g., smoothing filtering, normalization), the corresponding behavioral features are extracted.

[0067] At this time, based on the action types set in the process node of the building safety management task (including the prohibited actions, necessary actions, and other actions mentioned above), the behavior characteristics of each dimension are matched with each target action in the action library in order of priority to match and calculate the similarity between the behavior characteristics of a single dimension and the target action. Generally speaking, when the similarity is high enough, it is considered that the target action is hit.

[0068] If the target action is a single action, it can be matched directly. If the target action contains multiple sub-actions, they can be matched in sequence according to the time sequence of the sub-actions. When the multiple sub-actions are matched in sequence, the target action is also matched.

[0069] Generally speaking, due to the large number of action targets and the large number of sub-actions they contain, the amount of calculation is large, which may make it difficult to ensure real-time performance. Therefore, the priority is set. When a prohibited action is detected, the current process node needs to be restarted directly and the user needs to be reminded as quickly as possible, so the highest immediacy is required. The necessary action is used to determine whether to execute the next action, and the immediacy required is also relatively high. Other actions are mainly for scoring, and the lowest immediacy is required. Therefore, the priority order from high to low is: prohibited action, necessary action, other action.

[0070] At this time, based on the similarities corresponding to all dimensions obtained by matching, the target action hit by the real-time input data is determined through multimodal judgment.

[0071] Furthermore, when determining the similarity, for the behavior feature of a single dimension, the real-time input data corresponding to the behavior feature is used as the first time series data, and the target action is used as the second time series data. Then, the first time series data and the second time series data are aligned on the time axis through nonlinear time alignment. Among them, nonlinear time alignment allows matching of different time steps.

[0072] For example, suppose the first time series data , the second first time series data , Refers to the characteristic data at the nth time point, Refers to the feature data at the mth time point.

[0073] By calculating the distance between each time point in the first time series data and the second time series data, a distance matrix corresponding to the first time series data and the second time series data is constructed.

[0074] By calculating each pair and The cosine distance between them is used to construct the distance matrix , each element in the distance matrix D , represents the difference between the feature data at the i-th time point in the first time series data and the feature data at the j-th time point in the second time series data.

[0075] Based on dynamic programming, the minimum cumulative distance is calculated through the distance matrix. The minimum cumulative distance refers to the minimum distance obtained by performing multiple rounds of calculations from the upper left corner to the lower right corner of the matrix, which is used to represent the minimized matching cost of the first time series data and the second time series data.

[0076] The minimum cumulative distance is calculated as: .in, Indicates the minimum cumulative distance between the first i feature data of the first time series data and the first j feature data of the second time series data. Represents the distance between the i-th feature data of the first time series data and the j-th feature data of the second time series data, that is, the parameter in the distance matrix D. It means, in , , Select the minimum value in . And so on, you can accumulate calculations from the upper left corner all the way to the lower right corner, and finally get the minimum cumulative distance of the distance matrix.

[0077] According to the minimum cumulative distance and the preset dynamic coefficient, the similarity between the current single-dimensional behavioral characteristics and the target action is obtained. Generally speaking, the minimum cumulative distance can be directly used as the similarity value. However, considering that when the user first enters the VR environment, some actions may not be performed properly at the beginning due to not adapting to the virtual environment, and as the user gradually adapts to the VR environment, the actions will be performed more effectively. Therefore, a preset dynamic coefficient is set. As the user enters the VR environment for a longer time, the preset dynamic gradually decreases until it is reduced to the preset lower limit value (generally speaking, the lower limit value is 1, that is, the similarity of the user's actions is compensated to a certain extent at the beginning, and finally gradually restored to the most realistic state).

[0078] Furthermore, the similarity can be directly compared with a preset judgment threshold to determine whether the current user behavior hits the target action. For example, the similarity corresponding to all matched dimensions is determined, and based on the weight corresponding to each dimension, the first total similarity corresponding to the real-time input data is obtained by weighted summation.

[0079] Then, the target action hit by the real-time input data is determined based on the first total similarity. For example, if the first total similarity exceeds a preset similarity value, it is considered that the real-time input data hits the target action.

[0080] If the target action is other than a necessary action or a prohibited action in the process node of the building safety management task, the determination of the action is not important, so it can be directly determined that the real-time input data hits the target action.

[0081] If the target action is a necessary action or a prohibited action, the judgment of the action is considered to be very important. At this time, another round of judgment is required to splice and fuse the behavioral features of all dimensions to obtain a fused feature. When fusing, they can be spliced ​​in order from front to back.

[0082] At this time, the similarity between the fused feature and the target action is calculated as the second total similarity. If the second total similarity still determines that the real-time input data hits the target action, then it is finally determined that the real-time input data hits the target action. By adding a round of judgment, multiple judgments can be made at two levels to increase the accuracy of the judgment results.

[0083] like Figure 4 As shown, the embodiment of the present application also proposes a VR-based building management device, including: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the VR-based building management method as described in any of the above embodiments.

[0084] The embodiment of the present application also proposes a non-volatile computer storage medium storing computer executable instructions, wherein the computer executable instructions are configured as the VR-based building management method described in any of the above embodiments.

[0085] Each embodiment in this application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device and medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

[0086] The above is only the embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the present application.

Claims

1. A VR-based building management method, characterized in that: include: Obtaining a building information model of a specified building; wherein the building information model includes geometric information, construction information, and collaboration information; Based on the geometric information, each component in the designated building is modeled to obtain a component model, and the component model is added to a component model library corresponding to the VR environment; and construction tools required for the designated building are modeled to obtain a tool model, and the tool model is added to a tool model library corresponding to the VR environment; Determine that a user logs into the VR environment, identify the user information, and obtain a building safety management task corresponding to the user based on the user information; Based on the building safety management task, according to the construction process information contained in the building information, in the VR environment, the required component models are assembled to obtain a target building model; and the required tool models are added in the VR environment; Determine to execute the building safety management task, and during the task execution, issue instructions to the user through the VR environment, collect and analyze the user behavior, and evaluate the user based on the collected user behavior.

2. The VR-based building management method according to claim 1, characterized in that: Based on the geometric information, each component in the designated building is modeled to obtain a component model, which specifically includes: Obtaining unit size information of each object recorded in the geometric information of the designated building; If there is a single region, and the single region is obtained by splicing multiple identical and continuous unit size information, then the single region is taken as a single component; Each unit size information present in the remaining areas except all the individual areas is treated as a single component; Model each component to obtain the corresponding component model; According to the preset component type, a component model that meets the requirements is selected from all component models as a designated component model, and a replacement model of the designated component model is generated; Wherein, for the designated component model corresponding to the single region, the means of obtaining the replacement model includes at least one of the following: replacing and / or deleting at least part of the unit size information in the designated component model, or changing the properties of the designated component model; For a designated component model corresponding to a single unit size information, a means of obtaining a replacement model includes at least one of the following: replacing the unit size information, and changing a property of the designated component model.

3. The VR-based building management method according to claim 2, characterized in that: Based on the building safety management task, according to the construction process information contained in the building information, in the VR environment, the required component models are assembled to obtain the target building model, specifically including: Determining a required construction schedule based on the construction safety management tasks; Based on the construction progress, determining the required component models and the corresponding construction sequence according to the construction process information included in the construction information; Assembling the required component models in sequence according to the construction sequence and the three-dimensional spatial position information contained in the geometric information; During the assembly process, based on the building safety management task, a designated component model of a corresponding component type is selected, and its corresponding replacement model is used to replace the designated component model to assemble and obtain a target building model.

4. The VR-based building management method according to claim 1, characterized in that: Modeling the construction tools required for the specified building to obtain a tool model, specifically including: Obtaining construction tools required for the designated building, and obtaining size information and function information corresponding to the construction tools; Modeling is performed according to the size information to obtain a tool model; and a corresponding function label is generated for the tool model according to the function information; In the VR environment, add the required tool models, including: In the VR environment, determining required construction tools according to the collaboration information; Obtain function labels corresponding to required construction tools, and search for other construction tools with at least one identical function label among the remaining construction tools except the required construction tools, so as to add at least one tool model corresponding to the other construction tools in the VR environment.

5. The VR-based building management method according to claim 3, characterized in that: During the task execution, the VR environment is used to issue instructions to the user and collect and analyze the user's behavior, including: Determine a plurality of pre-set process nodes in the building safety management task, and necessary actions and / or prohibited actions set in each process node; At each process node, collect user behaviors and determine the user's current execution action; If the currently executed action hits a single necessary action, and all the preceding necessary actions of the single necessary action have been executed in the current process node, and the single necessary action is preset with a corresponding model change attribute, then according to the model change attribute, the corresponding model change action is executed in the target building model; If the currently executed action hits a single prohibited action, the target building model is restored to the initial state of the current process node; For each alternative model, a corresponding necessary action and / or prohibited action needs to be set.

6. The VR-based building management method according to claim 5, characterized in that: Evaluating the user based on the collected user behavior specifically includes: Based on the total time the user takes to complete the building safety management task, a basic score corresponding to the user is obtained; Obtaining prohibited actions performed by the user during the execution of the building safety management task, recording the prohibited actions, and deducting points from the basic score based on the prohibited actions; Acquire other actions performed by the user in the process of executing the building safety management task, in addition to the necessary actions and the prohibited actions, and modify the basic score based on the other actions performed.

7. The VR-based building management method according to claim 5, characterized in that: Collect and analyze user behavior, including: Collect real-time input data generated by user behavior through the interface of the VR device that the user has worn; An action library is defined, and target actions defined by user behaviors are added to the action library, and corresponding action tags are added to the target actions; Preprocessing the real-time input data and performing feature extraction on the real-time input data to obtain multi-dimensional behavior features corresponding to the user behavior; Based on the action types set in the process node of the building safety management task, the behavior characteristics of each dimension are matched with each target action in the action library in order of priority, so as to match and calculate the similarity between the behavior characteristics of a single dimension and the target action; wherein the action types include: prohibited actions, necessary actions, and other actions; and the priority order from high to low is: prohibited actions, necessary actions, and other actions; Based on the similarities corresponding to all the dimensions obtained by matching, the target action hit by the real-time input data is determined through multimodal judgment.

8. The VR-based building management method according to claim 7, characterized in that: Matching calculates the similarity between the behavior feature of a single dimension and the target action, specifically including: For a behavior feature of a single dimension, the real-time input data corresponding to the behavior feature is used as the first time series data, and the target action is used as the second time series data; Aligning the first time series data with the second time series data on a time axis by nonlinear time alignment; By calculating the distance between each time point in the first time series data and the second time series data, a distance matrix corresponding to the first time series data and the second time series data is constructed; Based on dynamic programming, the minimum cumulative distance is calculated through the distance matrix, and the similarity between the current single-dimensional behavioral characteristics and the target action is obtained according to the minimum cumulative distance and the preset dynamic coefficient; wherein, as the user enters the VR environment for a longer time, the preset dynamic gradually decreases until it decreases to a preset lower limit value.

9. The VR-based building management method according to claim 7, characterized in that: Based on the similarities corresponding to all the dimensions obtained by matching, the target action hit by the real-time input data is determined through multimodal determination, specifically including: Determine the similarities corresponding to all dimensions obtained by matching, and obtain a first total similarity corresponding to the real-time input data by weighted summation based on the weight corresponding to each dimension; Determining a target action hit by the real-time input data according to the first total similarity; If the target action is an action other than a necessary action or a prohibited action in the process node of the building safety management task, determining that the real-time input data hits the target action; If the target action is a necessary action or a prohibited action, the behavioral features of all dimensions are concatenated and fused to obtain the fused features; Calculating the similarity between the fused feature and the target action as a second total similarity; If it is still determined that the real-time input data hits the target action according to the second total similarity, then it is determined that the real-time input data hits the target action.

10. A VR-based building management device, characterized in that: include: at least one processor; as well as, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the VR-based building management method as described in any one of claims 1 to 9.

Images