A three-dimensional scene loading method, device and equipment

Abstract

The application discloses a three-dimensional scene loading method, device and equipment, relates to the virtual reality technical field, and aims at solving the problem that a three-dimensional model is difficult to realize synchronization with a real scene in the prior art. The three-dimensional scene loading method comprises the following steps: acquiring images of multiple angles of a target place at a fixed time; generating a first three-dimensional model of the target place by using the images of the multiple angles; and loading a three-dimensional scene of the first three-dimensional model corresponding to a second time closest to a first time in the case that an input for the target place is received. The embodiment of the application can load the three-dimensional scene of the newly generated first three-dimensional model, ensure that the loaded three-dimensional scene is closest to a real time, and ensure that the loaded three-dimensional scene is synchronized with a real scene as much as possible.

Description

Technical Field

[0001] This invention relates to the field of virtual reality technology, and in particular to a method, apparatus, and device for loading three-dimensional scenes. Background Technology

[0002] Existing 3D (3D) scenic spot loading technology involves 3D modeling or VR camera shooting, and then loading the data through version or media updates.

[0003] Existing 3D modeling methods often result in immutable 3D models once created. While particle engines simulate various weather conditions, these simulations often fall short of reality due to variations in weather and seasons, lacking realism and failing to provide an immersive experience. In other words, 3D models generated through 3D modeling cannot synchronize with real-world scenes. Similarly, 3D models generated from VR camera footage exhibit noticeable gaps at the stitching points, and updating these models after recording is costly. Therefore, 3D models generated from VR camera footage also struggle to synchronize with real-world scenes. Summary of the Invention

[0004] This invention provides a method, apparatus, and device for loading three-dimensional scenes to solve the problem in the prior art that it is difficult to synchronize three-dimensional models with real-world scenes.

[0005] This invention provides a three-dimensional scene loading method, comprising:

[0006] Timely acquire images of the target location from multiple angles;

[0007] Using the images from the multiple angles, a first three-dimensional model of the target location is generated;

[0008] Upon receiving input for the target location, the three-dimensional scene of the first three-dimensional model corresponding to the second time closest to the first time is loaded;

[0009] Wherein, the first time is the time when the input for the target location is received;

[0010] The second time is the time it takes to generate the first three-dimensional model;

[0011] The second time is prior to the first time.

[0012] Optionally, generating a first three-dimensional model of the target location using the images from the multiple angles includes:

[0013] Based on the preset 3D reconstruction algorithm and the images from the multiple angles, an initial 3D model of the target location is obtained;

[0014] Based on the initial three-dimensional model of the target location and the real-world spatial region data corresponding to the target location, a second three-dimensional model of the target location is generated;

[0015] The target object model in the second three-dimensional model is processed to obtain the first three-dimensional model of the target location;

[0016] Wherein, the target object model is a duplicate motion object model or an abnormal object model. If the target object model is a duplicate motion object model, the target processing is deduplication; if the target object model is an abnormal object model, the target processing is deletion.

[0017] Optionally, obtaining the initial 3D model of the target location based on a preset 3D reconstruction algorithm and the images from the multiple angles includes:

[0018] Using the preset 3D reconstruction algorithm, feature points are extracted from image regions other than the human body region in images from multiple angles to obtain feature points of images from multiple angles.

[0019] By comparing the feature points of the images from the multiple angles, the three-dimensional information of the images from the multiple angles is obtained;

[0020] The three-dimensional information of the images from the multiple angles is processed by meshing and texturing to obtain an initial three-dimensional model of the target location.

[0021] Optionally, when the images of the target location from multiple angles include images of multiple scene regions from different angles, generating a second three-dimensional model of the target location based on the initial three-dimensional model of the target location and the real-world spatial region data corresponding to the target location includes:

[0022] Divide the scene into a first modeling region corresponding to each of the aforementioned scene regions;

[0023] Based on the modeling area corresponding to the initial 3D model of each scene area and the real space data corresponding to each scene area, a first 3D model of the target location is generated.

[0024] The initial 3D model of the scene area is obtained based on the preset 3D reconstruction algorithm and images of the scene area from different angles.

[0025] Optionally, the target object model includes a repeating moving object model; when the images of the target location from multiple angles include images of multiple scene areas from different angles, the target processing of the target object model in the second three-dimensional model to obtain the first three-dimensional model of the target location includes:

[0026] In the second three-dimensional model, candidate detection regions are determined in the second modeling region corresponding to the plurality of scene regions;

[0027] In the second three-dimensional model, the repeating moving object model in the candidate detection region is determined;

[0028] In the second three-dimensional model, other motion object models except the target motion object model are deleted from the duplicate motion object models to obtain the first three-dimensional model;

[0029] The target motion object model is any one of the repeated motion object models.

[0030] Optionally, determining the candidate detection regions among the plurality of scene regions includes:

[0031] Determine the speed of the moving object based on relevant information from the image corresponding to the target scene area;

[0032] Based on the shooting route of the camera equipment, the movement direction of the moving object in the multiple scene areas, and the movement speed of the moving object, the candidate detection area of ​​the target scene area is determined among the multiple scene areas;

[0033] Wherein, the target scene region is any one of the plurality of scene regions;

[0034] The camera device is a device that captures images from the multiple angles.

[0035] Optionally, determining the motion speed of the moving object based on relevant information from the image corresponding to the target scene region includes:

[0036] The speed of the moving object is obtained based on the ground diameter corresponding to the first and second images, the proportion of the moving object's movement in the second image relative to its movement in the first image, and the time interval between the capture of the first and second images.

[0037] The first image is any image from the images corresponding to the target scene area, and the second image was captured at a time adjacent to that of the first image.

[0038] Optionally, determining the candidate detection region of the target scene region among the multiple scene regions based on the shooting route of the camera device, the movement direction of the moving object in the multiple scene regions, and the movement speed of the moving object includes:

[0039] When the shooting route of the camera device is consistent with the movement direction of the moving object in the multiple scene areas, the initial candidate detection area is determined based on the relationship between the minimum speed at which the moving object moves out of the target scene area and the walking speed of the moving object;

[0040] Based on the map corresponding to the target location, the scene area that does not include the walking path indicated by the map is deleted from the initial candidate detection area to obtain the candidate detection area;

[0041] The minimum speed at which the moving object moves out of the target scene area is determined based on the minimum side length of the image corresponding to the target scene area and the shooting time of the camera device capturing the image of the target scene area.

[0042] Optionally, determining the initial candidate detection region based on the relationship between the minimum speed at which the moving object moves out of the target area scene and the walking speed of the moving object includes:

[0043] If the speed of the moving object is less than the minimum speed at which the moving object moves out of the target scene area, the scene area adjacent to the target scene area among the multiple scene areas is taken as the initial candidate detection area.

[0044] When the speed of the moving object is greater than or equal to the minimum speed at which the target object moves out of the target scene area, the scene areas adjacent to the target scene area and the scene areas adjacent to the first scene area among the multiple scene areas are used as the initial candidate detection areas; wherein, the first scene area is a scene area among the multiple scene areas determined based on the minimum speed of the target scene area, the speed of the moving object, and the direction of the moving object's movement in the multiple scene areas.

[0045] Optionally, determining the candidate detection region of the target scene region among the multiple scene regions based on the shooting route of the camera device, the movement direction of the moving object in the multiple scene regions, and the movement speed of the moving object includes:

[0046] If the shooting path of the camera device is inconsistent with the movement direction of the moving object in the multiple scene areas, and if the time for the moving object to move from the target scene area to the second scene area is less than the time for the camera device to move from the target scene area to the second scene area, then the target scene area and the second scene area are taken as the candidate detection areas.

[0047] The relationship between the time it takes for the moving object to travel from the target scene area to the second scene area and the time it takes for the camera device to travel from the target scene area to the second scene area is obtained based on the shooting route of the camera device, the shooting duration of the camera device in each scene area, the movement speed of the moving object, the minimum diameter of each scene area, and the movement direction of the moving object in the multiple scene areas.

[0048] The second scene area is a scene area adjacent to the target scene area, determined based on the movement direction of the moving object in the multiple scene areas.

[0049] Optionally, the method further includes:

[0050] Based on the shooting route of the camera device, determine the first number of scene areas traversed by the camera device from the target scene area to the second scene area;

[0051] Based on the first quantity, the shooting time of the camera device in each scene area, the movement speed of the moving object, the minimum diameter of each scene area, and the movement direction of the moving object in the multiple scene areas, the third scene area among the multiple scene areas is obtained when the camera device moves from the target scene area to the second scene area.

[0052] Based on the positional relationship between the third scene area and the second scene area, the relationship between the time it takes for the moving object to travel from the target scene area to the second scene area and the time it takes for the camera device to travel from the target scene area to the second scene area is obtained.

[0053] Optionally, before performing target processing on the target object model in the second three-dimensional model to obtain the first three-dimensional model of the target location, the method further includes:

[0054] The abnormal object model in the first three-dimensional model is determined according to at least one of the following methods:

[0055] If the number of facial feature points in the object model in the second 3D model exceeds the number of the first preset feature points, and the number of facial feature points is not a multiple of the number of the first preset feature points, then the object model is determined to be the abnormal object model.

[0056] If the number of skeletal feature points of the object model in the second three-dimensional model exceeds the second preset feature points, the object model is determined to be the abnormal object model.

[0057] If continuous pixel color values ​​appear on the object model in the second 3D model, the object model is determined to be the abnormal object model.

[0058] Optionally, after loading the 3D scene of the first 3D model corresponding to the second time closest to the first time when receiving input for the target location, the method further includes:

[0059] Record the user's first position coordinates in the 3D scene;

[0060] The first 3D model is updated periodically, and after obtaining the updated first 3D model, the 3D scene of the updated first 3D model is loaded.

[0061] The first position coordinates are synchronized to the 3D scene of the updated first 3D model.

[0062] This invention also provides a three-dimensional scene loading device, comprising:

[0063] The image acquisition module is used to periodically acquire images of the target location from multiple angles.

[0064] The first processing module is used to generate a first three-dimensional model of the target location using the images from the multiple angles;

[0065] The first scene loading module is used to load the three-dimensional scene of the first three-dimensional model corresponding to the second time closest to the first time when it receives input for the target location;

[0066] Wherein, the first time is the time when the input for the target location is received;

[0067] The second time is the time it takes to generate the first three-dimensional model;

[0068] The second time is prior to the first time.

[0069] Optionally, the first processing module includes:

[0070] The first processing unit is used to obtain an initial three-dimensional model of the target location based on a preset three-dimensional reconstruction algorithm and the images from the multiple angles.

[0071] The second processing unit is used to generate a second three-dimensional model of the target location based on the initial three-dimensional model of the target location and the real spatial area data corresponding to the target location;

[0072] The third processing unit is used to perform target processing on the target object model in the second three-dimensional model to obtain the first three-dimensional model of the target location.

[0073] Wherein, the target object model is a duplicate motion object model or an abnormal object model. When the target object model is the duplicate motion object model, the target processing is deduplication processing. When the target object model is the abnormal object model, the target processing is deletion processing.

[0074] Optionally, the first processing unit is specifically used for:

[0075] Using the preset 3D reconstruction algorithm, feature points are extracted from image regions other than the human body region in images from multiple angles to obtain feature points of images from multiple angles.

[0076] By comparing the feature points of the images from the multiple angles, the three-dimensional information of the images from the multiple angles is obtained;

[0077] The three-dimensional information of the images from the multiple angles is processed by meshing and texturing to obtain an initial three-dimensional model of the target location.

[0078] Optionally, the second processing unit is specifically used for:

[0079] In the case where the images of the target location from multiple angles include images of multiple scene regions from different angles, a first modeling region corresponding to each scene region is defined;

[0080] Based on the modeling area corresponding to the initial 3D model of each scene area and the real space data corresponding to each scene area, a first 3D model of the target location is generated.

[0081] The initial 3D model of the scene area is obtained based on the preset 3D reconstruction algorithm and images of the scene area from different angles.

[0082] Optionally, the target object model includes a repeating motion object model;

[0083] The third processing unit is specifically used for:

[0084] In the case where the images of the target location from multiple angles include images of multiple scene regions from different angles, in the second three-dimensional model, candidate detection regions in the second modeling region corresponding to the multiple scene regions are determined;

[0085] In the second three-dimensional model, the repeating moving object model in the candidate detection region is determined;

[0086] In the second three-dimensional model, other motion object models other than the target motion object model are deleted from the duplicate motion object models to obtain the second three-dimensional model;

[0087] The target motion object model is any one of the repeated motion object models.

[0088] Optionally, the third processing unit is specifically used for:

[0089] Determine the speed of the moving object based on relevant information from the image corresponding to the target scene area;

[0090] Based on the shooting route of the camera equipment, the movement direction of the moving object in the multiple scene areas, and the movement speed of the moving object, the candidate detection area of ​​the target scene area is determined among the multiple scene areas;

[0091] Wherein, the target scene region is any one of the plurality of scene regions;

[0092] The camera device is a device that captures images from the multiple angles.

[0093] Optionally, the third processing unit is specifically used for:

[0094] The speed of the moving object is obtained based on the ground diameter corresponding to the first and second images, the proportion of the moving object's movement in the second image relative to its movement in the first image, and the time interval between the capture of the first and second images.

[0095] The first image is any image from the images corresponding to the target scene area, and the second image was captured at a time adjacent to that of the first image.

[0096] Optionally, the third processing unit is specifically used for:

[0097] When the shooting route of the camera device is consistent with the movement direction of the moving object in the multiple scene areas, the initial candidate detection area is determined based on the relationship between the minimum speed at which the moving object moves out of the target scene area and the walking speed of the moving object;

[0098] Based on the map corresponding to the target location, the scene area that does not include the walking path indicated by the map is deleted from the initial candidate detection area to obtain the candidate detection area;

[0099] The minimum speed at which the moving object moves out of the target scene area is determined based on the minimum side length of the image corresponding to the target scene area and the shooting time of the camera device capturing the image of the target scene area.

[0100] Optionally, the third processing unit is specifically used for:

[0101] If the speed of the moving object is less than the minimum speed at which the moving object moves out of the target scene area, the scene area adjacent to the target scene area among the multiple scene areas is taken as the initial candidate detection area.

[0102] When the speed of the moving object is greater than or equal to the minimum speed at which the target object moves out of the target scene area, the scene areas adjacent to the target scene area and the scene areas adjacent to the first scene area among the multiple scene areas are used as the initial candidate detection areas; wherein, the first scene area is a scene area among the multiple scene areas determined based on the minimum speed of the target scene area, the speed of the moving object, and the direction of the moving object's movement in the multiple scene areas.

[0103] Optionally, the third processing unit is specifically used for:

[0104] If the shooting path of the camera device is inconsistent with the movement direction of the moving object in the multiple scene areas, and if the time for the moving object to move from the target scene area to the second scene area is less than the time for the camera device to move from the target scene area to the second scene area, then the target scene area and the second scene area are taken as the candidate detection areas.

[0105] The relationship between the time it takes for the moving object to travel from the target scene area to the second scene area and the time it takes for the camera device to travel from the target scene area to the second scene area is obtained based on the shooting route of the camera device, the shooting duration of the camera device in each scene area, the movement speed of the moving object, the minimum diameter of each scene area, and the movement direction of the moving object in the multiple scene areas.

[0106] The second scene area is a scene area adjacent to the target scene area, determined based on the movement direction of the moving object in the multiple scene areas.

[0107] Optionally, the third processing unit is further specifically used for:

[0108] Based on the shooting route of the camera device, determine the first number of scene areas traversed by the camera device from the target scene area to the second scene area;

[0109] Based on the first quantity, the shooting time of the camera device in each scene area, the movement speed of the moving object, the minimum diameter of each scene area, and the movement direction of the moving object in the multiple scene areas, the third scene area among the multiple scene areas is obtained when the camera device moves from the target scene area to the second scene area.

[0110] Based on the positional relationship between the third scene area and the second scene area, the relationship between the time it takes for the moving object to travel from the target scene area to the second scene area and the time it takes for the camera device to travel from the target scene area to the second scene area is obtained.

[0111] Optionally, the first processing module further includes:

[0112] The first determining unit is configured to determine the abnormal object model according to at least one of the following methods:

[0113] The second determining unit is used to determine the object model as the abnormal object model when the number of facial feature points of the object model in the first three-dimensional model exceeds the number of the first preset feature points, and the number of the facial feature points is not a multiple of the first preset feature points.

[0114] The third determining unit is used to determine the object model as the abnormal object model when the number of skeletal feature points of the object model in the first three-dimensional model exceeds the number of second preset feature points.

[0115] The fourth determining unit is used to determine that the object model is the abnormal object model when continuous pixel color values ​​appear on the object model in the first three-dimensional model.

[0116] Optionally, the device further includes:

[0117] A recording module is used to record the user's initial position coordinates in the three-dimensional scene.

[0118] The second scene loading module is used to periodically update the first three-dimensional model, and after obtaining the updated first three-dimensional model, load the three-dimensional scene of the updated first three-dimensional model.

[0119] The synchronization module is used to synchronize the first position coordinates to the 3D scene of the updated first 3D model.

[0120] This invention also provides a three-dimensional scene loading device, comprising: a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor; the processor is configured to read the program in the memory to implement the steps of the three-dimensional scene loading method as described above.

[0121] This invention also provides a computer-readable storage medium for storing a computer program, which, when executed by a processor, implements the steps of the three-dimensional scene loading method as described above.

[0122] In this embodiment of the invention, images of a target location from multiple angles are acquired periodically. Using these images, a first three-dimensional model of the target location is generated. Upon receiving input for the target location, the three-dimensional scene of the first three-dimensional model corresponding to the time when the input for the target location was received (first time) and the time when the first three-dimensional model was generated (second time) is loaded. This enables the first three-dimensional model to be generated and updated periodically based on the acquired images of the target location from multiple angles. Furthermore, upon receiving input for the target location, the three-dimensional scene of the most recently generated first three-dimensional model is loaded. In other words, the loaded three-dimensional scene is the one closest to the real time, ensuring that the loaded three-dimensional scene is synchronized with the real scene as much as possible. Attached Figure Description

[0123] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0124] Figure 1 This is one of the flowcharts of the three-dimensional scene loading method provided in the embodiments of the present invention;

[0125] Figure 2 This is one of the schematic diagrams of the scene area in the target location provided in the embodiments of the present invention;

[0126] Figure 3 This is the second schematic diagram of the scene area in the target location provided in the embodiment of the present invention;

[0127] Figure 4 This is the second flowchart of the three-dimensional scene loading method provided in the embodiments of the present invention;

[0128] Figure 5 This is a schematic diagram of the structure of the three-dimensional scene loading device provided in an embodiment of the present invention;

[0129] Figure 6 This is a schematic diagram of the structure of the three-dimensional scene loading device provided in an embodiment of the present invention. Detailed Implementation

[0130] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0131] See Figure 1 , Figure 1 This is a flowchart of the three-dimensional scene loading method provided in the embodiments of the present invention, such as... Figure 1 As shown, it includes the following steps:

[0132] Step 101: Periodically acquire images of the target location from multiple angles;

[0133] Preferably, in this step, a camera device is used to acquire images of the target location from multiple angles within a first preset time interval, according to a first preset time interval.

[0134] The camera device can be a fixed surveillance camera or a mobile drone. It can also simultaneously use both a surveillance camera and a drone to capture images of the same target location from multiple angles, obtaining images from multiple perspectives. These images can be photographs or pictures obtained from videos captured by the surveillance camera or drone.

[0135] The camera device captures images of the target location from multiple angles at preset time intervals (i.e., a first preset duration), obtaining images of the target location from multiple angles. Alternatively, the target location is divided into multiple scene areas, and the camera device captures images of each scene area from multiple angles, obtaining images of each scene area within the target location from multiple angles. The preset time interval can be set as needed, for example, based on the size data of the target location, or based on the performance of the camera device; this is not limited in this embodiment. The number of scene areas into which the target location is divided can also be set based on the size data of the target location; this is not limited in this embodiment.

[0136] For example, when the camera device is a drone, the specific process of the camera device taking pictures of the target location from multiple angles at preset time intervals to obtain images of the target location from multiple angles is as follows: the drone takes 4K video of the target location, then the coordinates are recorded by the Global Positioning System (GPS) to obtain GPS coordinates, the video time is associated with the GPS coordinates, and then images of the target location from multiple angles are obtained by using different angles of the same GPS coordinate.

[0137] In this step, acquiring images of the target location from multiple angles using a camera is done at regular time intervals, specifically every first preset time interval. Correspondingly, the camera can also take multiple shots of the target location at these time intervals; for example, if the camera is a drone, it takes one round of multi-angle shots of the target location every first preset time interval. Alternatively, the camera can take multiple shots of a preset number of scene areas within the target location at the first preset time interval. This involves arranging the scene areas in a specific order, and if the camera is a drone, it takes multiple shots of these scene areas sequentially every first preset time interval.

[0138] Images from multiple angles within the first preset time period are used for subsequent modeling processing.

[0139] In this step, the first preset duration can be set independently based on factors such as the size data of the target location, and is not limited in this embodiment. The target location can be a scenic area, a street in a city, or any location other than a scenic area or a street in a city.

[0140] The first time interval can be set based on the size data of the target location or the size data of each scene area, but is not limited in this embodiment.

[0141] The preset quantity can be set based on at least one of the following: the number of scene areas in the target location, the shooting time of the drone for each scene area, the size data of each scene area, and the first preset time.

[0142] Step 102: Using the images from the multiple angles, generate a first three-dimensional model of the target location.

[0143] In this step, images of the target location from multiple angles acquired at regular intervals are used to generate a first three-dimensional model of the target location. Preferably, based on images of the target location from multiple angles within a first preset time period, a 3D scene reconstruction of the target point is performed to obtain the first three-dimensional model of the target location.

[0144] Through the above steps, a first 3D model is generated based on images of the target location from multiple angles acquired at regular intervals, and the first 3D model is updated at regular intervals to achieve synchronization between the generated 3D model and real time as much as possible.

[0145] Step 103: Upon receiving input for the target location, load the 3D scene of the first 3D model corresponding to the second time closest to the first time.

[0146] Wherein, the first time is the time when the input for the target location is received;

[0147] The second time is the time it takes to generate the first three-dimensional model;

[0148] The second time is prior to the first time.

[0149] After generating a first 3D model at regular intervals, this step receives user input for the target location. Based on the first time the user input for the target location is received, the second time that the first 3D model was generated, which is closest to the first time and located before the first time, is determined. The 3D scene in the first 3D model corresponding to the second time is then loaded. That is, when the user logs in, the 3D scene of the newly generated first 3D model is loaded, ensuring that the user experiences the 3D scene that is closest to the real time, thus improving the user experience.

[0150] Optionally, generating a first three-dimensional model of the target location using the images from the multiple angles includes:

[0151] Based on the preset 3D reconstruction algorithm and the images from the multiple angles, an initial 3D model of the target location is obtained;

[0152] A second three-dimensional model of the target location is generated based on the initial three-dimensional model of the target location and the real-world spatial region data corresponding to the target location.

[0153] The target object model in the second three-dimensional model is processed to obtain the first three-dimensional model;

[0154] Wherein, the target object model is a duplicate motion object model or an abnormal object model. When the target object model is the duplicate motion object model, the target processing is deduplication processing. When the target object model is the abnormal object model, the target processing is deletion processing.

[0155] The preset three-dimensional reconstruction algorithm is either the structure from motion (SFM) algorithm or the Neural Radiance Fields (NeRF) algorithm.

[0156] In this embodiment, the SFM algorithm is used as an example to illustrate the preset three-dimensional reconstruction algorithm. Using the SFM algorithm, the target location is modeled using images from multiple angles to obtain an initial three-dimensional model of the target location. After the initial three-dimensional model is generated using the SFM algorithm, the initial three-dimensional model is imported into the 3D engine according to the data of the real-world photographed spatial area (real-world spatial area data) to generate a second three-dimensional model.

[0157] It should be noted that the SFM algorithm has certain requirements for the shooting angle and accuracy of the images. For example, the angle between two adjacent images should be less than or equal to 15 degrees.

[0158] Taking a drone as an example, the specific parameters for image capture are as follows:

[0159] The sensor size is Ls (unit: mm); the distance between the camera on the drone and the subject is D (unit: m); the focal length of the camera on the drone is f (unit: px); the image size is L (unit: px); the ground resolution of the image is R (unit: m / px); the spatial positioning accuracy of the 3D mesh vertices is P;

[0160] Then, the flight altitude H of the drone satisfies the following formula:

[0161] H = D = f * R * L / Ls

[0162] and,

[0163] P = 3 * R

[0164] In this context, * indicates multiplication.

[0165] It should be noted that the target object model can be a repeating motion object model, an abnormal object model, or a preset object model. A repeating motion object model refers to an object model corresponding to the same motion object appearing more than once in the first 3D model. This object can be a person or an animal; for example, multiple person models or multiple animal models appearing in the first 3D model. An abnormal object model refers to an object model corresponding to an object that exhibits an anomaly in the first 3D model. This object can be a person, an animal, or other objects. Anomalies can include repetition of facial features, body parts, animal faces, or animal body parts during modeling, or pixel continuation. A preset target object refers to a preset object model corresponding to a person, animal, or object.

[0166] Specifically, when the target object model is a duplicate motion object model, if a duplicate motion object model is found in the second 3D model, the duplicate motion object model is deduplicated, that is, only one motion object model is kept and the rest are deleted. When the target object model is an abnormal object model, if an abnormal object model is found in the second 3D model, the abnormal object model is deleted. When the target object model is a preset object model, if a preset object model is found in the second 3D model, the preset object model is preset to obtain the first 3D model. Through the above processing, the modeling accuracy of the 3D model for the scene of the target location can be improved, and the scene of the target location can be modeled more accurately through the first 3D model.

[0167] Optionally, obtaining the initial 3D model of the target location based on a preset 3D reconstruction algorithm and the images from the multiple angles includes:

[0168] Using the preset 3D reconstruction algorithm, feature points are extracted from image regions other than the human body region in images from multiple angles to obtain feature points of images from multiple angles.

[0169] By comparing the feature points of the images from the multiple angles, the three-dimensional information of the images from the multiple angles is obtained;

[0170] The three-dimensional information of the images from the multiple angles is processed by meshing and texturing to obtain an initial three-dimensional model of the target location.

[0171] It should be noted that before modeling the target location using the SFM algorithm and images from multiple angles, human detection is performed on each image from multiple angles to detect human body regions and mark or segment them. Then, the SFM algorithm is used to extract feature points from the image regions other than human body regions from multiple angles, obtaining feature points from multiple angles. By comparing the feature points from multiple angles, three-dimensional (3D) information is extracted from the images from multiple angles. The 3D information is then meshed and textured to obtain the initial three-dimensional model of the target location.

[0172] The SFM algorithm is used to extract feature points from image regions other than the human body region in images from multiple angles. This can be done by ignoring the marked human body region when extracting feature points from images from multiple angles, or by directly using the image after segmenting the human body region for feature point extraction.

[0173] By comparing feature points of images from multiple angles, 3D information is extracted from the images from multiple angles. Specifically, depth information is extracted from the images from multiple angles using triangulation.

[0174] For each image from multiple angles, the human detection function is used to detect the human body region in the image and to mark or segment the human body region. This can solve the problem of positioning drift caused by the movement of the human body when the feature points are located on the moving human body.

[0175] Optionally, when the images of the target location from multiple angles include images of multiple scene regions from different angles, generating a second three-dimensional model of the target location based on the initial three-dimensional model of the target location and the real-world spatial region data corresponding to the target location includes:

[0176] Divide the scene into a first modeling region corresponding to each of the aforementioned scene regions;

[0177] Based on the modeling area corresponding to the initial three-dimensional model of each scene area and the real space data corresponding to each scene area, a second three-dimensional model of the target location is generated.

[0178] The initial 3D model of the scene area is obtained based on the preset 3D reconstruction algorithm and images of the scene area from different angles.

[0179] By dividing the target location into multiple scene areas and using camera equipment to capture images of each scene area from multiple angles, the SFM algorithm is used to model the corresponding scene area according to the pre-assigned number of each scene area, obtaining the initial 3D model of each scene area. Based on the real-world photography space area, the 3D engine divides the first modeling area corresponding to each scene area, and imports the initial 3D model of each scene area into the corresponding first modeling area. Based on the real-world photography space area data (real space data) corresponding to each scene area, the second 3D model of the target location is synthesized.

[0180] The SFM algorithm is used to model each scene region separately using images from multiple angles, according to a pre-assigned number for each scene region. This can be achieved using capture context software, which models each scene region separately according to the pre-assigned number. The capture context software runs based on the SFM algorithm.

[0181] Optionally, the target object model includes a repeating moving object model; when the images of the target location from multiple angles include images of multiple scene areas from different angles, the target processing of the target object model in the second three-dimensional model to obtain the first three-dimensional model of the target location includes:

[0182] In the second three-dimensional model, candidate detection regions are determined in the second modeling region corresponding to the plurality of scene regions;

[0183] In the second three-dimensional model, the repeating moving object model in the candidate detection region is determined;

[0184] In the second three-dimensional model, other motion object models other than the target motion object model are deleted from the duplicate motion object models to obtain the second three-dimensional model;

[0185] The target motion object model is any one of the repeated motion object models.

[0186] When a target location is divided into multiple scene regions and images of each scene region are captured from multiple angles using camera equipment, duplicate moving object models may appear in the second 3D model. To improve processing efficiency, when processing duplicate moving object models in the second 3D model, candidate detection regions are first determined in the second modeling regions corresponding to the multiple scene regions in the second 3D model (each scene region corresponds to a second modeling region in the second 3D model). For the candidate detection regions, corresponding detection algorithms (such as face detection algorithms, pedestrian detection algorithms, or object similarity algorithms) are used to perform deduplication processing on moving objects in each candidate detection region, resulting in duplicate moving object models in the candidate detection regions. When duplicate moving object models are found in the candidate detection regions, only one moving object model (the target moving object model) is retained, and the rest are deleted to obtain the first 3D model.

[0187] The process of determining candidate detection regions among multiple scene regions is explained in detail below:

[0188] Optionally, determining the candidate detection regions among the plurality of scene regions includes:

[0189] Determine the speed of the moving object based on relevant information from the image corresponding to the target scene area;

[0190] Based on the shooting route of the camera equipment, the direction of movement of the moving object in the multiple scene areas, and the speed of the object corresponding to the movement, the candidate detection area of ​​the target scene area is determined among the multiple scene areas;

[0191] Wherein, the target scene region is any one of the plurality of scene regions;

[0192] The camera device is a device that captures images from the multiple angles.

[0193] In multiple scene regions, a target scene region is arbitrarily selected. Based on the relevant information of images from different angles corresponding to the target scene region, the speed of a moving object in the target scene region is determined. The moving object can be a person or an animal. Then, based on the speed of the moving object, the direction of movement (movement path) of the moving object in multiple scene regions, and the shooting path of the camera device (such as a drone) when shooting multiple scene regions, a candidate detection region for the target scene region is determined.

[0194] Furthermore, determining the motion speed of the moving object based on relevant information from the image corresponding to the target scene region includes:

[0195] The speed of the moving object is obtained based on the ground diameter corresponding to the first and second images, the proportion of the moving object's movement in the second image relative to its movement in the first image, and the time interval between the capture of the first and second images.

[0196] The first image is any image from the images corresponding to the target scene area, and the second image was captured at a time adjacent to that of the first image.

[0197] Specifically, when determining the speed of a moving object in the target scene area, the optical flow method is used to calculate the distance between two adjacent images (the first image and the second image). The distance between the moving object and the second image is equal to the ground diameter corresponding to the first image and the second image, multiplied by the proportion of the moving object's movement in the second image relative to its movement in the first image. Then, the speed v of the moving object in the target scene area is obtained by dividing the distance between the moving object and the time interval between the shooting of the first image and the second image.

[0198] The ground diameters corresponding to the first and second images are determined based on the content captured by the first and second images and the corresponding real-world map data. The time interval between capturing the first and second images can be preset, for example, based on the performance of the camera equipment, and is not limited in this embodiment.

[0199] Since the angle between the two images in this embodiment of the invention is less than or equal to 15 degrees, the tracking of the moving object's behavior between the two images will not be lost.

[0200] As a preferred embodiment, determining the candidate detection region of the target scene region among the multiple scene regions based on the shooting route of the camera device, the movement direction of the moving object in the multiple scene regions, and the speed of the moving object, includes:

[0201] When the shooting route of the camera device is consistent with the movement direction of the moving object in the multiple scene areas, the initial candidate detection area is determined based on the relationship between the minimum speed at which the moving object moves out of the target scene area and the walking speed of the moving object;

[0202] Based on the map corresponding to the target location, the scene area that does not include the walking path indicated by the map is deleted from the initial candidate detection area to obtain the candidate detection area;

[0203] The minimum speed at which the moving object moves out of the target scene area is determined based on the minimum side length of the image corresponding to the target scene area and the shooting time of the camera device capturing the image of the target scene area.

[0204] Assuming the two side lengths of the image corresponding to the target scene region are m and n, and the corresponding minimum side length of the image is f(m)min or f(n)min, and the shooting time of the camera in capturing the target scene region is T, then the minimum speed V of the moving object moving out of the target scene region can be calculated according to the following formula:

[0205] V=f(x)min / T

[0206] Where x is m or n.

[0207] When the shooting path of the camera device (such as a drone) is consistent with the movement direction (movement path) of the moving object in multiple scene areas, the initial candidate detection area is determined based on the movement speed v of the moving object in the target scene area and the minimum speed V of the moving object moving out of the target scene area.

[0208] For example, please refer to Figure 2 The target location includes, but is not limited to, scene area 1, scene area 2, scene area 3... scene area 9. If the direction of movement (movement route) of the target object is: from scene area 1 to scene area 2, from scene area 2 to scene area 3... from scene area 8 to scene area 9, and the shooting route of the camera device when shooting multiple scene areas is: from scene area 1 to scene area 2, from scene area 2 to scene area 3... from scene area 8 to scene area 9, then it is determined that the shooting route of the camera device when shooting multiple scene areas is consistent with the direction of movement (movement route) of the moving object in multiple scene areas.

[0209] Based on the map data of the target location, the walking path in multiple scene areas indicated by the map of the target location is obtained. Areas that do not include the walking path in the initial candidate detection area are deleted to obtain the candidate detection area, that is, the candidate detection area is the initial candidate detection area that includes the walking path.

[0210] Please continue reading Figure 2 If the determined initial candidate scene areas include scene area 1, scene area 2, scene area 3, scene area 4, scene area 5, scene area 6, scene area 7, scene area 8, and scene area 9, and the walking paths in the multiple scene areas indicated by the map are as follows: Figure 2As shown, scene regions 6, 5, and 4 include the walking path. Therefore, the final candidate detection regions are scene regions 6, 5, and 4.

[0211] By removing regions that do not include walking paths from the initial candidate detection region, i.e., removing regions where moving objects cannot move, the processing efficiency for processing repetitive moving object models in the first 3D model is further improved.

[0212] Furthermore, determining the initial candidate detection region based on the relationship between the minimum speed at which the moving object moves out of the target area scene and the walking speed of the moving object includes:

[0213] If the speed of the moving object is less than the minimum speed at which the moving object moves out of the target scene area, the scene area adjacent to the target scene area among the multiple scene areas is taken as the initial candidate detection area.

[0214] When the speed of the moving object is greater than or equal to the minimum speed at which the target object moves out of the target scene area, the scene areas adjacent to the target scene area and the scene areas adjacent to the first scene area among the multiple scene areas are used as the initial candidate detection areas; wherein, the first scene area is a scene area among the multiple scene areas determined based on the minimum speed of the target scene area, the speed of the moving object, and the direction of the moving object's movement in the multiple scene areas.

[0215] If the moving object's velocity v is less than the minimum velocity V required for the moving object to move out of the target scene area, the scene areas adjacent to the target scene area are determined as initial candidate detection areas. For example, please refer to [link to relevant documentation]. Figure 2 If the target location includes multiple scene areas including but not limited to scene area 1, scene area 2, scene area 3... scene area 9, scene area 5 is taken as the target scene area. If the speed v of a moving object in scene area 5 is less than the minimum speed V of the moving object moving out of the target scene area, then the scene areas adjacent to scene area 5 (including scene area 1, scene area 2, scene area 3, scene area 4, scene area 6, scene area 7, scene area 8, and scene area 9) are determined as the initial candidate detection areas.

[0216] If the moving object's velocity v is greater than or equal to the minimum velocity V required for the moving object to move out of the target scene area, the first scene area is determined based on the first integer value obtained by dividing the moving object's velocity v by the minimum velocity V (if there is a remainder when the moving object's velocity v is divided by the minimum velocity V, the quotient is rounded up and incremented by one to obtain the first integer value). This first integer value, along with the moving object's direction of motion within the target scene area, determines the first scene area. The scene areas adjacent to the target scene area and the scene areas adjacent to the first scene area are then used as initial candidate detection areas. For example, please refer to [further details]. Figure 2 Taking scene region 5 as the target scene region, if the moving object's speed v is 1.2 m / s and the minimum speed V of the moving object moving out of the target scene region is 1 m / s, then the first integer value is 2, and the moving object's direction of motion in the target scene region is as follows: Figure 2 As indicated by the arrows, the first scene region can be identified as scene region 11. Therefore, the scene regions adjacent to scene region 5 (including scene regions 1, 2, 3, 4, 6, 7, 8, and 9) and the scene regions adjacent to scene region 11 (including scene regions 10, 12, 13, 14, and 15) are all taken as initial candidate detection regions. Among them, the multiple scene regions of the target location include, but are not limited to, scene regions 1, 2, 3, ..., 9, ..., 15.

[0217] Using both the scene regions adjacent to the target scene region and the scene regions adjacent to the first scene region as initial candidate detection regions can increase fault tolerance.

[0218] As another preferred embodiment, determining the candidate detection region of the target scene region among the multiple scene regions based on the shooting route of the camera device, the movement direction of the moving object in the multiple scene regions, and the movement speed of the moving object includes:

[0219] If the shooting path of the camera device is inconsistent with the movement direction of the moving object in the multiple scene areas, and if the time for the moving object to move from the target scene area to the second scene area is less than the time for the camera device to move from the target scene area to the second scene area, then the target scene area and the second scene area are taken as the candidate detection areas.

[0220] The relationship between the time it takes for the moving object to travel from the target scene area to the second scene area and the time it takes for the camera device to travel from the target scene area to the second scene area is obtained based on the shooting route of the camera device, the shooting duration of the camera device in each scene area, the movement speed of the moving object, the minimum diameter of each scene area, and the movement direction of the moving object in the multiple scene areas.

[0221] The second scene area is a scene area adjacent to the target scene area, determined based on the movement direction of the moving object in the multiple scene areas.

[0222] Specifically, when the shooting route of the camera device (such as a drone) when shooting multiple scene areas is inconsistent with the movement direction (movement route) of the moving object in multiple scene areas, the target scene area and the second scene area are selected as candidate detection areas if the time for the moving object to cross the adjacent scene area (moving from the target scene area to the second scene area) is less than the time for the camera device to move from the target scene area to the second scene area.

[0223] The determination of whether the time taken for the moving object to travel from the target scene area to the second scene area is less than the time taken for the camera device to travel from the target scene area to the second scene area is based on the shooting route of the camera device in multiple scene areas, the shooting duration of the camera device in each scene area, the speed of the moving object in the target scene area, the minimum diameter of each scene area, and the direction of the moving object in multiple scene areas.

[0224] For example, please refer to Figure 3 The target location includes, but is not limited to, scene area 1, scene area 2, scene area 3... scene area 9. If the direction of movement (movement route) of the target object is from scene area 1 to scene area 6, from scene area 6 to scene area 7, and the shooting route of the camera device when shooting multiple scene areas is from scene area 1 to scene area 2, from scene area 2 to scene area 3... from scene area 8 to scene area 9, then it is determined that the shooting route of the camera device when shooting multiple scene areas is inconsistent with the direction of movement (movement route) of the moving object in multiple scene areas.

[0225] It should be noted that the shooting route of the camera equipment is preset; the shooting time of the camera equipment in each scene area is also preset, for example, it is set according to the performance of the camera equipment, and is not limited in this embodiment.

[0226] Furthermore, the method also includes:

[0227] Based on the shooting route of the camera device, determine the first number of scene areas traversed by the camera device from the target scene area to the second scene area;

[0228] Based on the first quantity, the shooting time of the camera device in each scene area, the movement speed of the moving object, the minimum diameter of each scene area, and the movement direction of the moving object in the multiple scene areas, the third scene area among the multiple scene areas is obtained when the camera device moves from the target scene area to the second scene area.

[0229] Based on the positional relationship between the third scene area and the second scene area, the relationship between the time it takes for the moving object to travel from the target scene area to the second scene area and the time it takes for the camera device to travel from the target scene area to the second scene area is obtained.

[0230] Based on the camera's shooting route, determine the first number Num of scene areas between the target scene area and the second scene area during shooting. The duration of the camera's recording in each scene area is T. B Therefore, the time interval between shooting the target scene area and then shooting the second scene area is the first quantity Num multiplied by the shooting time T. B Within this time interval, based on the moving object's speed v, the distance the moving object has traveled can be determined as the first quantity Num multiplied by the shooting time T. B By multiplying the motion speed v and dividing the distance the moving object travels by the minimum diameter of the scene area, the number of scene areas the moving object crosses within the time interval can be determined. Based on the scene area (target scene area) where the moving object is currently located and the direction of the moving object's movement in multiple scene areas, the third scene area that the moving object moves to after the time interval can be determined. If the moving object moves to the third scene area after passing through the second scene area, then the time taken for the moving object to travel from the target scene area to the second scene area is less than the time taken for the camera device to travel from the target scene area to the second scene area.

[0231] For example, please continue reading Figure 3The target location includes multiple scene areas, including but not limited to scene area 1, scene area 2, scene area 3... scene area 9. The movement direction (movement route) of the target object is: from scene area 1 to scene area 6, from scene area 6 to scene area 7. The shooting route of the camera equipment when shooting multiple scene areas is: from scene area 1 to scene area 2, from scene area 2 to scene area 3... from scene area 8 to scene area 9. Scene area 1 is taken as the target scene area, according to... Figure 3 Based on the direction of motion of the moving object shown, the second scene area is defined as scene area 6. The number of scene areas between scene area 1 and scene area 6 for the camera device (drone) is 6. The duration of the camera device taking pictures in each scene area is T. B The time interval between the drone's movement from scene region 1 to scene region 6 is 6 times T. B During this time interval, the moving object travels a distance of 6 times T. B Multiplying the moving object's velocity v by the minimum diameter of each scene region fmin, the number of scene regions the moving object traverses within that time interval is 6 multiplied by T. B Multiply by the moving object's speed v and divide by fmin. Based on this value, determine that the moving object has moved to scene area 7 within the time interval and passed through scene area 6. Then, determine that the time for the moving object to move from the target scene area to the second scene area is less than the time for the camera device to move from the target scene area to the second scene area.

[0232] Optionally, before performing target processing on the target object model in the second three-dimensional model to obtain the first three-dimensional model of the target location, the method further includes:

[0233] The abnormal object model is determined according to at least one of the following methods:

[0234] If the number of facial feature points in the object model in the second 3D model exceeds the number of the first preset feature points, and the number of facial feature points is not a multiple of the number of the first preset feature points, then the object model is determined to be the abnormal object model.

[0235] If the number of skeletal feature points of the object model in the second three-dimensional model exceeds the second preset feature points, the object model is determined to be the abnormal object model.

[0236] If continuous pixel color values ​​appear on the object model in the second 3D model, the object model is determined to be the abnormal object model.

[0237] The following details how, when identifying abnormal object models in the second 3D model, one or more of the following detection methods—facial feature point detection, skeletal feature point detection, and pixel color value continuous detection—are used to detect the same object model. Based on the detection results, it is determined whether the object model is an abnormal object model.

[0238] Specifically, let's take the character model as an example to illustrate:

[0239] If the facial feature point detection method detects that the number of facial feature points of a person exceeds the range of a first preset feature point, and the detected number of facial feature points is not a multiple of the first preset feature point, then the person model is determined to be an abnormal person model. For example, if the first preset feature point is 106, but the detected number of facial feature points is 200, then the person model is an abnormal person model. If the detected number of facial feature points is 212 (possibly due to a face image printed on the person's clothing), then it is not determined whether the person model is an abnormal person model.

[0240] If the skeletal feature points of a person exceed the range of the second preset feature points when the skeletal feature point detection method is used, the person model is determined to be an abnormal person model.

[0241] If the pixel color value continuous detection method detects that the pixel color value of a certain part of a person is a continuous pixel color value, then the person model is determined to be an abnormal person model.

[0242] Optionally, after loading the 3D scene of the first 3D model corresponding to the second time closest to the first time when receiving input for the target location, the method further includes:

[0243] Record the user's first position coordinates in the 3D scene;

[0244] The first 3D model is updated periodically, and after obtaining the updated first 3D model, the 3D scene of the updated first 3D model is loaded.

[0245] The first position coordinates are synchronized to the 3D scene of the updated first 3D model.

[0246] After generating a first 3D model based on images of the target location from multiple angles acquired periodically according to the above steps, the 3D scene corresponding to the first 3D model is set as the latest scene. Furthermore, the first 3D model is generated again periodically based on images of the target location from multiple angles within the corresponding time period, that is, the first 3D model is updated periodically to obtain an updated first 3D model. The 3D scene corresponding to the updated first 3D model is replaced with the latest scene. When the user inputs (login command) for the target location is received, the latest scene is loaded.

[0247] Preferably, when using a camera device to acquire images of the target location from multiple angles within a first preset time interval, a second three-dimensional model is generated based on the images from multiple angles within the first preset time interval every first preset time interval. That is, the second three-dimensional model is updated (reconstructed periodically) according to the first preset time interval, and the three-dimensional scene corresponding to the updated second three-dimensional model is replaced with the latest scene. When the user inputs (login command) for the target location is received, the latest scene is loaded.

[0248] For example, if a first 3D model is generated at 8 o'clock, the 3D scene corresponding to the first 3D model generated at 8 o'clock will be set as the latest scene. The first preset duration is 1 hour. Then, the first 3D model is generated based on images of the target location taken by the camera device from multiple angles between 8 o'clock and 9 o'clock. After that, the 3D scene corresponding to the first 3D model is replaced as the latest scene. If, afterward, before the next first 3D model is generated, a login instruction from the user for the target location is received, the 3D scene corresponding to the first 3D model generated based on images of the target location taken by the camera device from multiple angles between 8 o'clock and 9 o'clock will be loaded.

[0249] For details, please refer to Figure 4Within a first preset time period, images of the target location taken by the camera from multiple angles are used to model and generate a first 3D model 1. The 3D scene 1 corresponding to the first 3D model 1 is then set as the latest scene (new_scene). When a user (Actor1) logs in, 3D scene 1 is loaded, and the user's position coordinates in 3D scene 1 are recorded. Simultaneously, based on images of the target location taken by the camera from multiple angles within the next first preset time period, a first 3D model 2 is modeled and generated. The 3D scene 2 corresponding to the first 3D model 2 is then set as new_scene. That is, 3D scene 1 triggers an asynchronous loading mechanism and begins loading 3D scene 2. At the same time, Actor1's user data (including position coordinates) is synchronized to 3D scene 2 without Actor1 needing to log in again. Also, if a new user (Actor2) logs in at this time, Actor2 will see 3D scene 2. The above steps are repeated for the next first preset time period.

[0250] During the loading process of the aforementioned 3D scene, the update of the 3D scene is imperceptible to the user, but the user can still feel the changes in the environment caused by time through the environmental changes of the 3D scene in real time.

[0251] For a target location with a large size range, in order to save resources when modeling the 3D environment of the target location, the entire 3D scene of the target location will not be loaded at once. Instead, the next part of the 3D scene will be loaded sequentially after a certain period of time. That is, when the first 3D model is reconstructed at a timed interval, the first 3D model will not be reconstructed completely within the first preset time period, but only a part of the first 3D model will be reconstructed.

[0252] It should also be noted that this embodiment of the invention provides a replication mechanism for storing the first 3D model. Each replication stores a 3D scene corresponding to the generated first 3D model. Since the real environment changes over time, the 3D scene data corresponding to the first 3D model stored in each replication is independent and cannot be reused. Based on the above reasons, each replication has the same basic resource requirements for the server. There is no need for too many replications, as too many replications would lead to excessive hardware resources. Preferably, in this embodiment, two replications are set, namely, storing the first 3D model whose generation time is most recent to the current time.

[0253] The three-dimensional scene loading method provided in this embodiment of the invention can further integrate the real environment with the virtual environment to realize virtual reality interaction. When the target location is a scenic spot, users can experience the environmental changes of the scenic spot over real time. When the target location is a city or certain blocks in a city, combined with existing digital twin technology, users can learn about the changes and current status of the city.

[0254] likeFigure 5 As shown, this embodiment of the invention also provides a three-dimensional scene loading device, the device 500 comprising:

[0255] The image acquisition module 501 is used to periodically acquire images of the target location from multiple angles.

[0256] The first processing module 502 is used to generate a first three-dimensional model of the target location using the images from the multiple angles;

[0257] The first scene loading module 503 is used to load the three-dimensional scene of the first three-dimensional model corresponding to the second time closest to the first time when it receives input for the target location;

[0258] Wherein, the first time is the time when the input for the target location is received;

[0259] The second time is the time it takes to generate the first three-dimensional model;

[0260] The second time is prior to the first time.

[0261] Optionally, the first processing module 502 includes:

[0262] The first processing unit is used to obtain an initial three-dimensional model of the target location based on a preset three-dimensional reconstruction algorithm and the images from the multiple angles.

[0263] The second processing unit is used to generate a second three-dimensional model of the target location based on the initial three-dimensional model of the target location and the real spatial area data corresponding to the target location;

[0264] The third processing unit is used to perform target processing on the target object model in the second three-dimensional model to obtain the first three-dimensional model of the target location.

[0265] Wherein, the target object model is a duplicate motion object model or an abnormal object model. When the target object model is the duplicate motion object model, the target processing is deduplication processing. When the target object model is the abnormal object model, the target processing is deletion processing.

[0266] Optionally, the first processing unit is specifically used for:

[0267] Using the preset 3D reconstruction algorithm, feature points are extracted from image regions other than the human body region in images from multiple angles to obtain feature points of images from multiple angles.

[0268] By comparing the feature points of the images from the multiple angles, the three-dimensional information of the images from the multiple angles is obtained;

[0269] The three-dimensional information of the images from the multiple angles is processed by meshing and texturing to obtain an initial three-dimensional model of the target location.

[0270] Optionally, the second processing unit is specifically used for:

[0271] In the case where the images of the target location from multiple angles include images of multiple scene regions from different angles, a first modeling region corresponding to each scene region is defined;

[0272] Based on the modeling area corresponding to the initial 3D model of each scene area and the real space data corresponding to each scene area, a first 3D model of the target location is generated.

[0273] The initial 3D model of the scene area is obtained based on the preset 3D reconstruction algorithm and images of the scene area from different angles.

[0274] Optionally, the target object model includes a repeating motion object model;

[0275] The third processing unit is specifically used for:

[0276] In the case where the images of the target location from multiple angles include images of multiple scene regions from different angles, in the second three-dimensional model, candidate detection regions in the second modeling region corresponding to the multiple scene regions are determined;

[0277] In the second three-dimensional model, the repeating moving object model in the candidate detection region is determined;

[0278] In the second three-dimensional model, other motion object models other than the target motion object model are deleted from the duplicate motion object models to obtain the second three-dimensional model;

[0279] The target motion object model is any one of the repeated motion object models.

[0280] Optionally, the third processing unit is specifically used for:

[0281] Determine the speed of the moving object based on relevant information from the image corresponding to the target scene area;

[0282] Based on the shooting route of the camera equipment, the movement direction of the moving object in the multiple scene areas, and the movement speed of the moving object, the candidate detection area of ​​the target scene area is determined among the multiple scene areas;

[0283] Wherein, the target scene region is any one of the plurality of scene regions;

[0284] The camera device is a device that captures images from the multiple angles.

[0285] Optionally, the third processing unit is specifically used for:

[0286] The speed of the moving object is obtained based on the ground diameter corresponding to the first and second images, the proportion of the moving object's movement in the second image relative to its movement in the first image, and the time interval between the capture of the first and second images.

[0287] The first image is any image from the images corresponding to the target scene area, and the second image was captured at a time adjacent to that of the first image.

[0288] Optionally, the third processing unit is specifically used for:

[0289] When the shooting route of the camera device is consistent with the movement direction of the moving object in the multiple scene areas, the initial candidate detection area is determined based on the relationship between the minimum speed at which the moving object moves out of the target scene area and the walking speed of the moving object;

[0290] Based on the map corresponding to the target location, the scene area that does not include the walking path indicated by the map is deleted from the initial candidate detection area to obtain the candidate detection area;

[0291] The minimum speed at which the moving object moves out of the target scene area is determined based on the minimum side length of the image corresponding to the target scene area and the shooting time of the camera device capturing the image of the target scene area.

[0292] Optionally, the third processing unit is specifically used for:

[0293] If the speed of the moving object is less than the minimum speed at which the moving object moves out of the target scene area, the scene area adjacent to the target scene area among the multiple scene areas is taken as the initial candidate detection area.

[0294] When the speed of the moving object is greater than or equal to the minimum speed at which the target object moves out of the target scene area, the scene areas adjacent to the target scene area and the scene areas adjacent to the first scene area among the multiple scene areas are used as the initial candidate detection areas; wherein, the first scene area is a scene area among the multiple scene areas determined based on the minimum speed of the target scene area, the speed of the moving object, and the direction of the moving object's movement in the multiple scene areas.

[0295] Optionally, the third processing unit is specifically used for:

[0296] If the shooting path of the camera device is inconsistent with the movement direction of the moving object in the multiple scene areas, and if the time for the moving object to move from the target scene area to the second scene area is less than the time for the camera device to move from the target scene area to the second scene area, then the target scene area and the second scene area are taken as the candidate detection areas.

[0297] The relationship between the time it takes for the moving object to travel from the target scene area to the second scene area and the time it takes for the camera device to travel from the target scene area to the second scene area is obtained based on the shooting route of the camera device, the shooting duration of the camera device in each scene area, the movement speed of the moving object, the minimum diameter of each scene area, and the movement direction of the moving object in the multiple scene areas.

[0298] The second scene area is a scene area adjacent to the target scene area, determined based on the movement direction of the moving object in the multiple scene areas.

[0299] Optionally, the third processing unit is further specifically used for:

[0300] Based on the shooting route of the camera device, determine the first number of scene areas traversed by the camera device from the target scene area to the second scene area;

[0301] Based on the first quantity, the shooting time of the camera device in each scene area, the movement speed of the moving object, the minimum diameter of each scene area, and the movement direction of the moving object in the multiple scene areas, the third scene area among the multiple scene areas is obtained when the camera device moves from the target scene area to the second scene area.

[0302] Based on the positional relationship between the third scene area and the second scene area, the relationship between the time it takes for the moving object to travel from the target scene area to the second scene area and the time it takes for the camera device to travel from the target scene area to the second scene area is obtained.

[0303] Optionally, the first processing module 502 further includes:

[0304] The first determining unit is configured to determine the abnormal object model according to at least one of the following methods:

[0305] The second determining unit is used to determine the object model as the abnormal object model when the number of facial feature points of the object model in the first three-dimensional model exceeds the number of the first preset feature points, and the number of the facial feature points is not a multiple of the first preset feature points.

[0306] The third determining unit is used to determine the object model as the abnormal object model when the number of skeletal feature points of the object model in the first three-dimensional model exceeds the number of second preset feature points.

[0307] The fourth determining unit is used to determine that the object model is the abnormal object model when continuous pixel color values ​​appear on the object model in the first three-dimensional model.

[0308] Optionally, the device further includes:

[0309] A recording module is used to record the user's initial position coordinates in the three-dimensional scene.

[0310] The second scene loading module is used to periodically update the first three-dimensional model, and after obtaining the updated first three-dimensional model, load the three-dimensional scene of the updated first three-dimensional model.

[0311] The synchronization module is used to synchronize the first position coordinates to the 3D scene of the updated first 3D model.

[0312] The apparatus provided in this embodiment of the invention can execute the above-described three-dimensional scene loading method, and its implementation principle and technical effect are similar, so it will not be described again here.

[0313] like Figure 6 As shown, this embodiment of the invention provides a three-dimensional scene loading device, including: a transceiver 610, a memory 620, a bus interface, a processor 600, and a computer program stored in the memory 620 and executable on the processor 600; the processor 600 is used to read the program in the memory 620, and the transceiver 610 is used to receive and send data under the control of the processor 600.

[0314] Processor 600 executes the following procedure:

[0315] Timely acquire images of the target location from multiple angles;

[0316] Using the images from the multiple angles, a first three-dimensional model of the target location is generated;

[0317] Upon receiving input for the target location, the three-dimensional scene of the first three-dimensional model corresponding to the second time closest to the first time is loaded;

[0318] Wherein, the first time is the time when the input for the target location is received;

[0319] The second time is the time it takes to generate the first three-dimensional model;

[0320] The second time is prior to the first time.

[0321] Among them, Figure 6 In this context, the bus architecture may include any number of interconnected buses and bridges, specifically linking various circuits together, represented by one or more processors (processor 600) and memory (memory 620). The bus architecture may also link together various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. A bus interface provides a user interface 630. A transceiver 610 may be multiple elements, including transmitters and transceivers, providing a unit for communicating with various other devices over a transmission medium. Processor 600 is responsible for managing the bus architecture and general processing, and memory 620 may store data used by processor 600 during operation.

[0322] The processor 600 is responsible for managing the bus architecture and general processing, while the memory 620 can store the data used by the processor 600 when performing operations.

[0323] Optionally, the processor 600 is configured to obtain an initial three-dimensional model of the target location based on a preset three-dimensional reconstruction algorithm and images from multiple angles; generate a second three-dimensional model of the target location based on the initial three-dimensional model of the target location and real-world spatial region data corresponding to the target location; and perform target processing on the target object model in the second three-dimensional model to obtain a first three-dimensional model of the target location.

[0324] Wherein, the target object model is a duplicate motion object model or an abnormal object model. When the target object model is the duplicate motion object model, the target processing is deduplication processing. When the target object model is the abnormal object model, the target processing is deletion processing.

[0325] Optionally, the processor 600 is specifically configured to extract feature points from image regions other than the human body region in images from multiple angles using the preset three-dimensional reconstruction algorithm, thereby obtaining feature points of the images from multiple angles; compare the feature points of the images from multiple angles to obtain three-dimensional information of the images from multiple angles; and perform meshing and texturing processing on the three-dimensional information of the images from multiple angles to obtain an initial three-dimensional model of the target location.

[0326] Optionally, the processor 600 is configured to, when the images of the target location from multiple angles include images of multiple scene regions from different angles, divide a first modeling region corresponding to each scene region; and generate a second three-dimensional model of the target location based on the modeling region corresponding to the initial three-dimensional model of each scene region and the real space data corresponding to each scene region.

[0327] The initial 3D model of the scene area is obtained based on the preset 3D reconstruction algorithm and images of the scene area from different angles.

[0328] Optionally, the target object model includes a repeating motion object model;

[0329] The processor 600 is configured to, in the case that the images of the target location from multiple angles include images of multiple scene regions from different angles, determine candidate detection regions in the second modeling region corresponding to the multiple scene regions in the second three-dimensional model; determine duplicate moving object models in the candidate detection regions in the second three-dimensional model; and delete other moving object models other than the target moving object model from the duplicate moving object models in the second three-dimensional model to obtain the first three-dimensional model.

[0330] The target motion object model is any one of the repeated motion object models.

[0331] Optionally, the processor 600 is specifically configured to determine the motion speed of the moving object based on relevant information of the image corresponding to the target scene area; and to determine the candidate detection area of ​​the target scene area among the multiple scene areas based on the shooting route of the camera device, the motion direction of the moving object in the multiple scene areas, and the motion speed of the moving object.

[0332] Wherein, the target scene region is any one of the plurality of scene regions;

[0333] The camera device is a device that captures images from the multiple angles.

[0334] Optionally, the processor 600 is specifically used to obtain the motion speed of the moving object based on the ground diameter corresponding to the first image and the second image, the proportion of the moving object's movement in the second image relative to its movement in the first image, and the time interval between the capture of the first image and the second image.

[0335] The first image is any image from the images corresponding to the target scene area, and the second image was captured at a time adjacent to that of the first image.

[0336] Optionally, the processor 600 is specifically configured to, when the shooting route of the camera device is consistent with the movement direction of the moving object in the multiple scene areas, determine an initial candidate detection area based on the relationship between the minimum speed at which the moving object moves out of the target scene area and the walking speed of the moving object; and delete the scene areas in the initial candidate detection area that do not include the walking path indicated by the map based on the map corresponding to the target location, to obtain the candidate detection area.

[0337] The minimum speed at which the moving object moves out of the target scene area is determined based on the minimum side length of the image corresponding to the target scene area and the shooting time of the camera device capturing the image of the target scene area.

[0338] Optionally, the processor 600 is specifically used for:

[0339] If the speed of the moving object is less than the minimum speed at which the moving object moves out of the target scene area, the scene area adjacent to the target scene area among the multiple scene areas is taken as the initial candidate detection area.

[0340] When the speed of the moving object is greater than or equal to the minimum speed at which the target object moves out of the target scene area, the scene areas adjacent to the target scene area and the scene areas adjacent to the first scene area among the multiple scene areas are used as the initial candidate detection areas; wherein, the first scene area is a scene area among the multiple scene areas determined based on the minimum speed of the target scene area, the speed of the moving object, and the direction of the moving object's movement in the multiple scene areas.

[0341] Optionally, the processor 600 is specifically configured to, when the shooting route of the camera device is inconsistent with the movement direction of the moving object in the multiple scene areas, if the time for the moving object to move from the target scene area to the second scene area is less than the time for the camera device to move from the target scene area to the second scene area, use the target scene area and the second scene area as the candidate detection areas;

[0342] The relationship between the time it takes for the moving object to travel from the target scene area to the second scene area and the time it takes for the camera device to travel from the target scene area to the second scene area is obtained based on the shooting route of the camera device, the shooting duration of the camera device in each scene area, the movement speed of the moving object, the minimum diameter of each scene area, and the movement direction of the moving object in the multiple scene areas.

[0343] The second scene area is a scene area adjacent to the target scene area, determined based on the movement direction of the moving object in the multiple scene areas.

[0344] Optionally, the processor 600 is further configured to: determine a first number of scene areas traversed by the camera device from the target scene area to the second scene area based on the shooting route of the camera device; obtain a third scene area among the multiple scene areas reached by the moving object from the target scene area within the time taken by the camera device from the target scene area to the second scene area based on the first number, the shooting duration of the camera device in each scene area, the movement speed of the moving object, the minimum diameter of each scene area, and the movement direction of the moving object in the multiple scene areas; and obtain a relationship between the time taken by the moving object from the target scene area to the second scene area and the time taken by the camera device from the target scene area to the second scene area based on the positional relationship between the third scene area and the second scene area.

[0345] Optionally, the processor 600 is configured to determine the exception object model according to at least one of the following methods:

[0346] If the number of facial feature points in the object model in the second 3D model exceeds the number of the first preset feature points, and the number of facial feature points is not a multiple of the number of the first preset feature points, then the object model is determined to be the abnormal object model.

[0347] If the number of skeletal feature points of the object model in the second three-dimensional model exceeds the second preset feature points, the object model is determined to be the abnormal object model.

[0348] If continuous pixel color values ​​appear on the object model in the second 3D model, the object model is determined to be the abnormal object model.

[0349] Optionally, the processor 600 is further configured to record the user's first position coordinates in the three-dimensional scene; after the first three-dimensional model is updated periodically to obtain the updated first three-dimensional model, load the three-dimensional scene of the updated first three-dimensional model; and synchronize the first position coordinates to the three-dimensional scene of the updated first three-dimensional model.

[0350] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a computer program instructing the relevant hardware to implement them. The computer program includes instructions to perform some or all of the steps of the above methods; and the computer program can be stored in a readable storage medium, which can be any form of storage medium.

[0351] In addition, a specific embodiment of the present invention also provides a computer-readable storage medium storing a computer program thereon. When the program is executed by a processor, it implements the steps in the above-described three-dimensional scene loading method and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0352] In the several embodiments provided in this application, it should be understood that the disclosed methods and apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0353] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can be physically comprised separately, or two or more units can be integrated into one unit. The integrated unit described above can be implemented in hardware or in the form of hardware plus software functional units.

[0354] The integrated units implemented as software functional units described above can be stored in a computer-readable storage medium. These software functional units, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute some steps of the transmission and reception methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0355] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for loading a three-dimensional scene, characterized in that, include: Timely acquire images of the target location from multiple angles; Using the images from the multiple angles, a first three-dimensional model of the target location is generated; Upon receiving input for the target location, the 3D scene of the first 3D model corresponding to the second time closest to the first time is loaded; wherein, the first time is the time when the input for the target location is received; the second time is the time when the first 3D model is generated; and the second time is prior to the first time. The step of generating a first three-dimensional model of the target location using images from multiple angles includes: Based on the preset 3D reconstruction algorithm and the images from the multiple angles, an initial 3D model of the target location is obtained; Based on the initial three-dimensional model of the target location and the real-world spatial region data corresponding to the target location, a second three-dimensional model of the target location is generated; The target object model in the second three-dimensional model is processed to obtain the first three-dimensional model of the target location; Wherein, the target object model is a duplicate motion object model or an abnormal object model. When the target object model is the duplicate motion object model, the target processing is deduplication processing. When the target object model is the abnormal object model, the target processing is deletion processing.

2. The three-dimensional scene loading method according to claim 1, characterized in that, The step of obtaining an initial 3D model of the target location based on a preset 3D reconstruction algorithm and images from multiple angles includes: Using the preset 3D reconstruction algorithm, feature points are extracted from image regions other than the human body region in images from multiple angles to obtain feature points of images from multiple angles. By comparing the feature points of the images from the multiple angles, the three-dimensional information of the images from the multiple angles is obtained; The three-dimensional information of the images from the multiple angles is processed by meshing and texturing to obtain an initial three-dimensional model of the target location.

3. The three-dimensional scene loading method according to claim 1, characterized in that, In the case where the images of the target location from multiple angles include images of multiple scene areas from different angles, generating a second three-dimensional model of the target location based on the initial three-dimensional model of the target location and the real-world spatial region data corresponding to the target location includes: Divide the scene into a first modeling region corresponding to each of the aforementioned scene regions; Based on the modeling area corresponding to the initial three-dimensional model of each scene area and the real space data corresponding to each scene area, a second three-dimensional model of the target location is generated. The initial 3D model of the scene area is obtained based on the preset 3D reconstruction algorithm and images of the scene area from different angles.

4. The three-dimensional scene loading method according to claim 1, characterized in that, The target object model includes a repeating moving object model; when the images of the target location from multiple angles include images of multiple scene areas from different angles, the target processing of the target object model in the second three-dimensional model to obtain the first three-dimensional model of the target location includes: In the second three-dimensional model, candidate detection regions are determined in the second modeling region corresponding to the plurality of scene regions; In the second three-dimensional model, the repeating moving object model in the candidate detection region is determined; In the second three-dimensional model, other motion object models except the target motion object model are deleted from the duplicate motion object models to obtain the first three-dimensional model; The target motion object model is any one of the repeated motion object models.

5. The three-dimensional scene loading method according to claim 4, characterized in that, Determining the candidate detection regions among the multiple scene regions includes: Determine the speed of the moving object based on relevant information from the image corresponding to the target scene area; Based on the shooting route of the camera equipment, the movement direction of the moving object in the multiple scene areas, and the movement speed of the moving object, the candidate detection area of ​​the target scene area is determined among the multiple scene areas; Wherein, the target scene region is any one of the plurality of scene regions; The camera device is a device that captures images from the multiple angles.

6. The three-dimensional scene loading method according to claim 5, characterized in that, Determining the motion speed of a moving object based on relevant information from the image corresponding to the target scene region includes: The speed of the moving object is obtained based on the ground diameter corresponding to the first and second images, the proportion of the moving object's movement in the second image relative to its movement in the first image, and the time interval between the capture of the first and second images. The first image is any image from the images corresponding to the target scene area, and the second image was captured at a time adjacent to that of the first image.

7. The three-dimensional scene loading method according to claim 5, characterized in that, The step of determining the candidate detection region of the target scene region among the multiple scene regions based on the shooting route of the camera device, the movement direction of the moving object in the multiple scene regions, and the movement speed of the moving object includes: When the shooting route of the camera device is consistent with the movement direction of the moving object in the multiple scene areas, the initial candidate detection area is determined based on the relationship between the minimum speed at which the moving object moves out of the target scene area and the walking speed of the moving object; Based on the map corresponding to the target location, the scene area that does not include the walking path indicated by the map is deleted from the initial candidate detection area to obtain the candidate detection area; The minimum speed at which the moving object moves out of the target scene area is determined based on the minimum side length of the image corresponding to the target scene area and the shooting time of the camera device when shooting the image of the target scene area.

8. The three-dimensional scene loading method according to claim 7, characterized in that, The step of determining the initial candidate detection region based on the relationship between the minimum speed at which the moving object moves out of the target scene region and the walking speed of the moving object includes: If the speed of the moving object is less than the minimum speed at which the moving object moves out of the target scene area, the scene area adjacent to the target scene area among the multiple scene areas is taken as the initial candidate detection area. When the speed of the moving object is greater than or equal to the minimum speed at which the target object moves out of the target scene area, the scene areas adjacent to the target scene area and the scene areas adjacent to the first scene area among the multiple scene areas are used as the initial candidate detection areas; wherein, the first scene area is a scene area among the multiple scene areas determined based on the minimum speed of the target scene area, the speed of the moving object, and the direction of the moving object's movement in the multiple scene areas.

9. The three-dimensional scene loading method according to claim 5, characterized in that, The step of determining the candidate detection region of the target scene region among the multiple scene regions based on the shooting route of the camera device, the movement direction of the moving object in the multiple scene regions, and the movement speed of the moving object includes: If the shooting path of the camera device is inconsistent with the movement direction of the moving object in the multiple scene areas, and if the time for the moving object to move from the target scene area to the second scene area is less than the time for the camera device to move from the target scene area to the second scene area, then the target scene area and the second scene area are taken as the candidate detection areas. The relationship between the time it takes for the moving object to travel from the target scene area to the second scene area and the time it takes for the camera device to travel from the target scene area to the second scene area is obtained based on the shooting route of the camera device, the shooting duration of the camera device in each scene area, the movement speed of the moving object, the minimum diameter of each scene area, and the movement direction of the moving object in the multiple scene areas. The second scene area is a scene area adjacent to the target scene area, determined based on the movement direction of the moving object in the multiple scene areas.

10. The three-dimensional scene loading method according to claim 9, characterized in that, The method further includes: Based on the shooting route of the camera device, determine the first number of scene areas traversed by the camera device from the target scene area to the second scene area; Based on the first quantity, the shooting time of the camera device in each scene area, the movement speed of the moving object, the minimum diameter of each scene area, and the movement direction of the moving object in the multiple scene areas, the third scene area among the multiple scene areas is obtained when the camera device moves from the target scene area to the second scene area. Based on the positional relationship between the third scene area and the second scene area, the relationship between the time it takes for the moving object to travel from the target scene area to the second scene area and the time it takes for the camera device to travel from the target scene area to the second scene area is obtained.

11. The three-dimensional scene loading method according to claim 1, characterized in that, Before performing target processing on the target object model in the second three-dimensional model to obtain the first three-dimensional model of the target location, the method further includes: The abnormal object model is determined according to at least one of the following methods: If the number of facial feature points in the object model in the second 3D model exceeds the number of the first preset feature points, and the number of facial feature points is not a multiple of the number of the first preset feature points, then the object model is determined to be the abnormal object model. If the number of skeletal feature points of the object model in the second three-dimensional model exceeds the second preset feature points, the object model is determined to be the abnormal object model. If continuous pixel color values ​​appear on the object model in the second 3D model, the object model is determined to be the abnormal object model.

12. The three-dimensional scene loading method according to claim 1, characterized in that, Upon receiving input for the target location, after loading the 3D scene of the first 3D model corresponding to the second time closest to the first time, the method further includes: Record the user's first position coordinates in the 3D scene; The first 3D model is updated periodically, and after obtaining the updated first 3D model, the 3D scene of the updated first 3D model is loaded. The first position coordinates are synchronized to the 3D scene of the updated first 3D model.

13. A three-dimensional scene loading device, characterized in that, include: The image acquisition module is used to periodically acquire images of the target location from multiple angles. The first processing module is used to generate a first three-dimensional model of the target location using the images from the multiple angles; The first scene loading module is used to load the 3D scene of the first 3D model corresponding to the second time closest to the first time when it receives input for the target location; wherein, the first time is the time when the input for the target location is received; the second time is the time when the first 3D model is generated; and the second time is located before the first time. The first processing module includes: The first processing unit is used to obtain an initial three-dimensional model of the target location based on a preset three-dimensional reconstruction algorithm and the images from the multiple angles. The second processing unit is used to generate a second three-dimensional model of the target location based on the initial three-dimensional model of the target location and the real spatial area data corresponding to the target location; The third processing unit is used to perform target processing on the target object model in the second three-dimensional model to obtain the first three-dimensional model of the target location. Wherein, the target object model is a duplicate motion object model or an abnormal object model. When the target object model is the duplicate motion object model, the target processing is deduplication processing. When the target object model is the abnormal object model, the target processing is deletion processing.

14. A three-dimensional scene loading device, comprising: A transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor; characterized in that the processor is configured to read the program in the memory to implement the steps of the three-dimensional scene loading method as described in any one of claims 1 to 12.

15. A computer-readable storage medium for storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the three-dimensional scene loading method as described in any one of claims 1 to 12.

Images