Method and system for displaying endangered species and products thereof based on VR technology

Abstract

The invention discloses a method and a system for displaying endangered species and products thereof based on a VR (virtual reality) technology. Through obtaining model interaction data, interaction information statistics and feature analysis are carried out, and VR model region clustering division and initial weight distribution are realized. Performing rendering test on each partition, extracting image features by using an auto-encoder, comparing the image features with a system database to calculate an overall influence coefficient, and then adjusting weights to perform model priority ranking and generate a rendering display scheme, thereby realizing efficient and accurate online VR model display.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of VR data display, more particularly, to a VR technology-based endangered species and products thereof display method and system. BACKGROUND

[0002] Endangered species and products thereof are difficult to enter the public eye due to their scarcity, and virtual reality technology (VR) as a new media technology provides a new solution for the display of endangered species and products thereof. However, the existing VR display method still has some technical bottlenecks, such as large data volume, high computing resource demand, low rendering efficiency, and poor online terminal display effect, and the corresponding VR model is not fully dynamically rendered and analyzed, resulting in inconsistent user experience. In order to overcome these difficulties, the present application proposes a VR technology-based endangered species and products thereof display method and system. SUMMARY

[0003] The present application overcomes the defects of the prior art and proposes a VR technology-based endangered species and products thereof display method and system.

[0004] The first aspect of the present application provides a VR technology-based endangered species and products thereof display method, comprising:

[0005] acquiring model interaction data of the endangered species and products thereof through a fixed VR display platform;

[0006] selecting a current VR model, performing interaction information statistics and interaction feature analysis on the current VR model through the model interaction data, performing regional clustering division on the current VR model based on the interaction features, forming a plurality of VR model partitions, and assigning initial weights to the VR model partitions based on the interaction features;

[0007] performing rendering test on each VR model partition, intercepting a multi-vision two-dimensional image of the VR model partition through the VR system at each test, performing feature learning and feature extraction on the multi-vision two-dimensional image based on a self-encoder, and comparing the multi-vision two-dimensional image with overall number display image features of a system database to calculate an overall influence coefficient of each VR model partition on the current VR model;

[0008] adjusting the initial weights through the overall influence coefficient and obtaining partition display weights, performing partition priority sorting on the current VR model through the partition display weights, generating a rendering display scheme, and performing online VR model display through the rendering display scheme.

[0009] In the present application, the model interaction data of the endangered species and products thereof is acquired through a fixed VR display platform, specifically:

[0010] In a historical time period, a model of an endangered species and its product is displayed through a fixed VR display platform, and model interaction data of users is collected;

[0011] The model interaction data includes interaction instructions for the model, model control, display control, and device interaction statistical information.

[0012] In this scheme, the selected current VR model is interacted with the current VR model through model interaction data to perform interaction information statistics and interaction feature analysis, the current VR model is regionally clustered and divided based on the interaction features, a plurality of VR model partitions are formed, and initial weights are assigned to the VR model partitions based on the interaction features. Specifically:

[0013] A current VR model is selected, and the current VR model includes model data corresponding to an endangered species and a product thereof;

[0014] Based on the current VR model, a plurality of unit models are divided, so that the data amount of each unit model is within a preset range;

[0015] The current VR model is interacted with the model interaction data to perform interaction information statistics and interaction feature analysis, and the interaction information statistics include statistics of interaction frequency, operation behavior, and operation time of different unit regions in the current VR model, and interaction feature data of each unit model is generated.

[0016] In this scheme, the selected current VR model is interacted with the current VR model through model interaction data to perform interaction information statistics and interaction feature analysis, the current VR model is regionally clustered and divided based on the interaction features, a plurality of VR model partitions are formed, and initial weights are assigned to the VR model partitions based on the interaction features. In addition, the scheme includes:

[0017] Based on the rendering time and model data amount of the current VR model, the model complexity is evaluated, and the number of partitions K is set according to the model complexity;

[0018] The interaction feature data of each unit model is used as clustering sample data, a clustering model based on Kmeans is imported, K initial clustering centers are set, and the clustering model is initialized;

[0019] In the clustering model, the similarity between the clustering sample data is calculated based on the cosine distance, and the data clustering and grouping are performed in a loop, and the clustering center point is recalculated in each clustering process;

[0020] When the clustering center point no longer moves, the clustering is stopped, and the clustering result is recorded, a plurality of feature groups are generated through the clustering result;

[0021] Based on the plurality of feature groups, the unit models are grouped and mapped, and the unit regions in the same group are combined to form a plurality of VR model partitions;

[0022] For each feature group, the interaction frequency and operation behavior of the corresponding interaction feature are analyzed to determine the user interaction interest degree of each feature group, and an initial weight is set for each VR model partition based on the interaction interest degree.

[0023] In this scheme, the rendering test is performed on each VR model partition, and a multi-view two-dimensional image of the VR model partition is intercepted by the VR system at each test, feature learning and feature extraction are performed on the multi-view two-dimensional image based on the autoencoder, and the overall number of display image features in the system database is compared to calculate the overall influence coefficient of each VR model partition on the current VR model, specifically:

[0024] The rendering test is performed on each VR model partition, and the rendering test time is recorded;

[0025] A multi-view two-dimensional image of the VR model partition after rendering is intercepted by the VR system at each test, and the multi-view two-dimensional image includes multiple images, each image corresponding to a visual angle;

[0026] The multi-view two-dimensional image is denoised and normalized;

[0027] The multi-view two-dimensional image is feature-extracted by a graph convolution network, and the multi-view features are stored to form a multi-view feature set;

[0028] An autoencoder is constructed, the size of the input layer is set to be consistent with the feature dimension in the multi-view feature set, the multi-view feature set is imported into the autoencoder for feature learning, the imported data is compressed to a low-dimensional representation data based on the encoder, and the low-dimensional representation data is reconstructed by the decoder to generate output data;

[0029] The mean square error is used as a loss function to evaluate the difference between the output data and the multi-view feature set, and the autoencoder is optimized based on the evaluation result, and the generation and training of the autoencoder are repeated until the loss function converges to a predetermined value, and the output data at this time is recorded as a second feature set;

[0030] The overall number of display image features of the current VR model is obtained from the system database, the overall number of display image features is classified based on the set multi-view angle to form a multi-view comparison feature set;

[0031] The second feature set and the multi-view comparison feature set are compared and similarity evaluated, the similarity evaluation is calculated by the Euclidean distance method, and a plurality of feature difference values are obtained based on multi-view analysis;

[0032] The mean value of the plurality of feature difference values is taken as the overall influence coefficient of the VR model partition on the current VR model.

[0033] In the scheme, the initial weight is adjusted by the overall influence coefficient, and the partition display weight is obtained, the current VR model is prioritized by the partition display weight, and the rendering display scheme is generated, and the online VR model display is performed through the rendering display scheme, specifically:

[0034] The initial weight is adjusted by the overall influence coefficient, and the partition display weight is obtained;

[0035] The partition display weight corresponding to each VR model partition is calculated, and the partitions are prioritized according to the partition display weight, and the online rendering priority information of the current VR model is generated;

[0036] According to the partition display weight, the model resource allocation setting is performed on each VR model partition, and the resource allocation scheme is obtained;

[0037] The rendering display scheme for the current VR model is generated through the online rendering priority information and the resource allocation scheme.

[0038] In the scheme, the VR display platform includes a VR system and a display system, the VR system is used for collecting and processing VR model data, and the display system is used for rendering and visualizing VR models.

[0039] In the scheme, the VR display platform establishes a data connection with a user terminal through a preset special network.

[0040] The second aspect of the application also provides a VR technology-based endangered species and product display system, which comprises a memory and a processor, the memory comprises a VR technology-based endangered species and product display program, and the VR technology-based endangered species and product display program is executed by the processor to realize the following steps:

[0041] The model interaction data of the endangered species and products is obtained by fixing the VR display platform;

[0042] The current VR model is selected, the interaction information statistics and interaction feature analysis of the current VR model are performed through the model interaction data, the current VR model is regionally clustered and divided based on the interaction features, a plurality of VR model partitions are formed, and the initial weight of the VR model partition is allocated based on the interaction features;

[0043] Each VR model partition is tested, a multi-vision two-dimensional image of the VR model partition is intercepted by the VR system at each test, the multi-vision two-dimensional image is feature learned and extracted based on the auto-encoder, and is compared with the overall number display image features of the system database, and the overall influence coefficient of each VR model partition on the current VR model is calculated;

[0044] The initial weight is adjusted by the overall influence coefficient, and a partition display weight is obtained, the current VR model is prioritized by the partition display weight, and a rendering display scheme is generated, and the online VR model display is performed through the rendering display scheme.

[0045] The third aspect of the present application also provides a computer readable storage medium, wherein the computer readable storage medium comprises a VR technology-based endangered species and product display program, and the VR technology-based endangered species and product display program, when executed by a processor, implements the steps of the VR technology-based endangered species and product display method according to any one of the above.

[0046] The present application discloses a VR technology-based endangered species and product display method and system. By obtaining model interaction data, interactive information statistics and feature analysis are performed to realize VR model region clustering division and initial weight distribution. Each partition is tested for rendering, image features are extracted using an autoencoder, and overall influence coefficients are calculated by comparing with a system database, and then the weights are adjusted to prioritize the model and generate a rendering display scheme, thereby realizing efficient and accurate online VR model display. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 A flowchart of a VR technology-based endangered species and product display method is shown.

[0048] Figure 2 A block diagram of a VR technology-based endangered species and product display system is shown. DETAILED DESCRIPTION

[0049] In order to more clearly understand the above-mentioned purposes, features and advantages of the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0050] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application, however, the present application can also be implemented in other ways different from those described herein, therefore, the scope of protection of the present application is not limited by the specific embodiments disclosed below.

[0051] Figure 1 A flowchart of a VR technology-based endangered species and product display method is shown.

[0052] As Figure 1 shown, the first aspect of the present application provides a VR technology-based endangered species and product display method, comprising:

[0053] S102, acquire model interaction data of the endangered species and products thereof through the fixed VR display platform;

[0054] S104, select the current VR model, perform interaction information statistics and interaction feature analysis on the current VR model through the model interaction data, perform regional clustering and division on the current VR model based on the interaction features, form a plurality of VR model partitions, and assign initial weights to the VR model partitions based on the interaction features;

[0055] S106, perform rendering test on each VR model partition, intercept a plurality of visual two-dimensional images of the VR model partition through the VR system at each time of testing, perform feature learning and feature extraction on the plurality of visual two-dimensional images based on the autoencoder, and compare the plurality of visual two-dimensional images with overall number display image features of a system database to calculate an overall influence coefficient of each VR model partition on the current VR model;

[0056] S108, adjust the initial weights through the overall influence coefficient, obtain partition display weights, perform partition priority sorting on the current VR model through the partition display weights, and generate a rendering display scheme to perform online VR model display through the rendering display scheme.

[0057] It should be noted that the fixed VR display platform is a fixed terminal, and a user views and interacts with a VR model through wearing corresponding VR equipment and the terminal. The fixed terminal is used to collect display information and interaction data of the user on the model of the endangered species and products thereof. In addition, based on the exploration of the interaction features, the interaction data analysis and user interest interaction feature analysis can be performed based on other non-VR display platforms. The VR display platform includes a VR system, and the VR system is used to analyze, store VR models, configure a VR visualization process, and configure a visualization scheme.

[0058] According to the embodiment of the present application, the model interaction data of the endangered species and products thereof is acquired through the fixed VR display platform, and specifically,

[0059] In a historical time period, the model of the endangered species and products thereof is displayed through the fixed VR display platform, and model interaction data of a user is collected;

[0060] The model interaction data includes an interaction instruction on the model, model control, display control, and device interaction statistical information.

[0061] It should be noted that the model interaction data can effectively statistically analyze the operation behavior frequency of the user on the VR model, the interaction model region, and the user interest model.

[0062] According to the embodiment of the present application, the selected current VR model is interacted with the model interaction data to perform interaction information statistics and interaction feature analysis, the current VR model is regionally clustered and divided based on the interaction features, a plurality of VR model partitions are formed, and initial weights are assigned to the VR model partitions based on the interaction features, specifically:

[0063] A current VR model is selected, and the current VR model includes model data corresponding to an endangered species and a product thereof;

[0064] Based on the current VR model, a plurality of unit models are divided, so that the data amount of each unit model is within a preset range.

[0065] The current VR model is interacted with the model interaction data to perform interaction information statistics and interaction feature analysis, and the interaction information statistics includes the statistics of interaction frequency, operation behavior, and operation time of different unit regions in the current VR model, and interaction feature data of each unit model is generated.

[0066] It should be noted that the plurality of unit models can be divided based on each component module of the species model or based on a fixed model size, and the data amount stored by each unit model is within a preset range.

[0067] According to the embodiment of the present application, the selected current VR model is interacted with the model interaction data to perform interaction information statistics and interaction feature analysis, the current VR model is regionally clustered and divided based on the interaction features, a plurality of VR model partitions are formed, and initial weights are assigned to the VR model partitions based on the interaction features, and further comprising:

[0068] Based on the rendering time and model data amount of the current VR model, the model complexity is evaluated, and the number K of partitions is set according to the model complexity;

[0069] The interaction feature data of each unit model is used as clustering sample data, a clustering model based on Kmeans is imported, K initial clustering centers are set and the clustering model is initialized;

[0070] In the clustering model, the similarity between the clustering sample data is calculated based on the cosine distance, and the data clustering and grouping are performed in a loop, and the clustering center point is recalculated in each clustering process;

[0071] When the clustering center point is no longer moved, the clustering is stopped, and the clustering result is recorded, and a plurality of feature groups are generated through the clustering result;

[0072] Based on the plurality of feature groups, the unit models are grouped and mapped, and the unit regions in the same group are combined to form a plurality of VR model partitions;

[0073] For each feature group, the interaction frequency and operation behavior of the corresponding interaction feature are analyzed to determine the user interaction interest degree of each feature group, and an initial weight is set for each VR model partition based on the interaction interest degree.

[0074] It should be noted that the size of K is proportional to the model complexity. Each feature group includes interaction feature data of a plurality of unit regions. Each feature group corresponds to a VR model partition. The initial weight is proportional to the interaction interest degree.

[0075] According to an embodiment of the present application, the rendering test is performed on each VR model partition, a multi-view two-dimensional image of the VR model partition is intercepted by the VR system at each test, feature learning and feature extraction are performed on the multi-view two-dimensional image based on the autoencoder, and the multi-view two-dimensional image is compared with the overall number display image features in the system database to calculate the overall influence coefficient of each VR model partition on the current VR model, specifically:

[0076] The rendering test is performed on each VR model partition, and the rendering test time is recorded;

[0077] A multi-view two-dimensional image of the VR model partition after rendering is intercepted by the VR system at each test, the multi-view two-dimensional image includes a plurality of images, and each image corresponds to a visual angle;

[0078] The multi-view two-dimensional image is denoised and normalized;

[0079] The multi-view two-dimensional image is feature-extracted by a graph convolution network, and the multi-view features are stored to form a multi-view feature set;

[0080] An autoencoder is constructed, the size of the input layer is set to be consistent with the feature dimension in the multi-view feature set, the multi-view feature set is imported into the autoencoder for feature learning, the imported data is compressed to a low-dimensional representation data based on the encoder, the low-dimensional representation data is reconstructed by the decoder, and output data is generated;

[0081] The difference between the output data and the multi-view feature set is evaluated as a loss function, and the autoencoder is optimized based on the evaluation result, the generation and training of the autoencoder are cycled until the loss function converges to a predetermined value, and the output data at this time is recorded as a second feature set;

[0082] The overall number display image features of the current VR model are obtained from the system database, the overall number display image features are classified based on the set multi-view angle to form a multi-view comparison feature set;

[0083] The second feature set is compared with the multi-vision contrast feature set, and similarity evaluation is performed in a Euclidean distance manner, and a plurality of feature difference values are obtained based on multi-vision analysis;

[0084] The mean value of the plurality of feature difference values is used as the overall influence coefficient of the VR model partition on the current VR model.

[0085] It should be noted that the visual angle in the multi-vision two-dimensional image is a plurality of predetermined angles, which can include main view, side view, top view, etc. The multi-vision feature set is the input data of the autoencoder. The second feature set also includes multi-vision features, which are feature data after learning and dimensionality reduction of the multi-vision feature set, and are used for fast view evaluation. The plurality of feature difference values correspond to a plurality of visual angles.

[0086] The multi-vision two-dimensional image is image data close to the real viewing model of the user, and is simulated and evaluated in an informatized form. In the evaluation process, the autoencoder is used to realize feature learning and feature dimensionality reduction of the VR model corresponding to the display image, and then realize fast evaluation of the display effect and accurate regional weight distribution. The higher the overall influence coefficient is, the greater the influence of the corresponding VR model partition on the overall display is, and the higher the rendering priority can be set.

[0087] According to the embodiment of the present application, the initial weight is adjusted by the overall influence coefficient, and the partition display weight is obtained. The current VR model is prioritized by the partition display weight, and a rendering display scheme is generated. The online VR model display is performed by the rendering display scheme, specifically:

[0088] The initial weight is adjusted by the overall influence coefficient, and the partition display weight is obtained.

[0089] The partition display weight corresponding to each VR model partition is calculated, and the partitions are prioritized according to the partition display weight, and the online rendering priority information of the current VR model is generated;

[0090] According to the partition display weight, the model resource allocation setting is performed on each VR model partition, and a resource allocation scheme is obtained.

[0091] The rendering display scheme for the current VR model is generated by the online rendering priority information and the resource allocation scheme.

[0092] It should be noted that the resource allocation scheme includes data transmission priority allocation of different VR model partition data corresponding to the resource data of a VR model in the VR display platform, data compression rate setting, and calculation resource and cache storage matching. The matching of the resource data is adjusted based on the partition display weight.

[0093] The partition display weight is calculated as follows:

[0094] Q E = Q s + K x P E ;

[0095] Wherein, Q E is a partition display weight, K is a correction coefficient, Q S is an initial weight, P E is an overall influence coefficient.

[0096] According to the embodiment of the application, the VR display platform comprises a VR system and a display system, the VR system is used for collecting and processing VR model data, and the display system is used for rendering and visualizing the VR model.

[0097] According to the embodiment of the application, the VR display platform establishes a data connection with the user terminal through a preset special-purpose network.

[0098] It should be noted that the user terminal can be used for online VR model visualization and VR interaction by the user.

[0099] It is worth mentioning here that the traditional VR display platform often lacks an efficient and feasible display scheme, and the model is simply packaged and transmitted, the model priority is not dynamically regulated in the display process, the model display effect is limited by the VR device and the network state, the user experience is poor, and for the model of endangered species and its products, the importance of different regions in the VR model to the overall display is often different, the traditional display platform often does not fully consider the model partition condition, and lacks a fast and accurate display effect evaluation process, hindering the application of the VR platform.

[0100] Based on this, the application analyzes the partition of the VR model corresponding to the endangered species and its products through user interaction information, sets the corresponding weight, analyzes the user display effect and adjusts the weight through the interception of multi-vision two-dimensional images, and further generates a model priority and a display scheme, effectively improving the user experience, especially for users in different interaction situations and different network situations, having better adaptability and improving the comprehensive display experience of the user.

[0101] According to the embodiment of the application, it further comprises:

[0102] Through the rendering display scheme, the online VR model is displayed in the user terminal;

[0103] Through the VR display platform, the VR model partition information displayed by the user is sent to the user terminal;

[0104] In the user terminal, a plurality of user interaction periods are set, and in each user interaction period, a real-time VR model is selected for analysis;

[0105] In the actual rendering display process, the multi-vision two-dimensional image of the corresponding VR model partition is intercepted by the VR system, and the overall influence coefficient is calculated, and a plurality of overall influence coefficients are obtained based on a plurality of user interaction cycles;

[0106] The plurality of overall influence coefficients are fitted based on a linear regression equation, a prediction equation is generated through the fitting process, a prediction time window is set, and a plurality of prediction coefficients are generated based on the prediction equation;

[0107] It is judged whether the plurality of prediction coefficients belong to the pre-influence coefficient range, if not, the real-time VR model information is sent to the VR display platform for secondary rendering test and dynamic adjustment of the VR model partition.

[0108] It should be noted that, in the online VR transmission process, due to the influence of network, terminal, interaction and other complex situations, data deviation is prone to occur in high-frequency display model data transmission, which leads to inaccurate partition and unreasonable change of corresponding partition weight distribution, and the influence priority rendering scheme is carried out, therefore, the present application carries out dynamic adjustment on the partition through multi-cycle model overall influence coefficient prediction analysis, evaluates the corresponding prediction value, and adapts to the reasonable setting of model priority analysis and display under complex situations.

[0109] Figure 2 A block diagram of a VR technology-based endangered species and product display system is shown.

[0110] The second aspect of the present application also provides a VR technology-based endangered species and product display system 2, which comprises a memory 21 and a processor 22, wherein the memory 21 comprises a VR technology-based endangered species and product display program, and the VR technology-based endangered species and product display program is executed by the processor 22 to realize the following steps:

[0111] The model interaction data of the endangered species and products are obtained by fixing the VR display platform.

[0112] The current VR model is selected, the current VR model is interacted with the model interaction data, the current VR model is regionally clustered and divided based on the interaction characteristics, a plurality of VR model partitions are formed, and the VR model partitions are assigned with initial weights based on the interaction characteristics.

[0113] Each VR model partition is tested for rendering, the multi-vision two-dimensional image of the VR model partition is intercepted by the VR system at each test, the multi-vision two-dimensional image is feature-learned and feature-extracted based on the autoencoder, and the overall number display image features of the system database are compared to calculate the overall influence coefficient of each VR model partition on the current VR model.

[0114] The initial weight is adjusted by the overall influence coefficient, and a partition display weight is obtained, the current VR model is prioritized by partition display weight, and a rendering display scheme is generated, and the online VR model display is performed through the rendering display scheme.

[0115] It should be noted that the fixed VR display platform is a fixed terminal, and the user views and interacts with the terminal by wearing corresponding VR equipment. The fixed terminal is used to collect user display information and interaction data of the model of endangered species and their products. In addition, based on the exploration of interaction characteristics, the interaction data analysis and user interest interaction characteristic analysis can be performed based on other non-VR display platforms. The VR display platform includes a VR system, and the VR system is used to analyze, store VR models, configure VR visualization processes and configure visualization schemes.

[0116] According to the embodiment of the present application, the model interaction data of endangered species and their products is obtained through the fixed VR display platform, specifically:

[0117] In a historical time period, the model of endangered species and their products is displayed through the fixed VR display platform, and the model interaction data of the user is collected;

[0118] The model interaction data includes interaction instructions, model control, display control, and device interaction statistical information.

[0119] It should be noted that through the model interaction data, the user's operation behavior frequency, interaction model area, and user interest model information of the VR model can be effectively analyzed.

[0120] According to the embodiment of the present application, the selected current VR model is interacted with the current VR model through the model interaction data to perform interaction information statistics and interaction characteristic analysis, the current VR model is regionally clustered and divided based on the interaction characteristics, a plurality of VR model partitions are formed, and the initial weight is assigned to the VR model partition based on the interaction characteristics, specifically:

[0121] Select a current VR model, the current VR model includes model data corresponding to an endangered species and a product thereof;

[0122] Based on the current VR model, a plurality of unit models are divided, so that the data amount of each unit model is within a preset range;

[0123] The interaction information statistics and interaction characteristic analysis of the current VR model are performed through the model interaction data, the interaction information statistics include the statistics of the interaction frequency, operation behavior and operation time of different unit regions in the current VR model, and the interaction characteristic data of each unit model is generated.

[0124] It should be noted that the division of the plurality of unit models can be based on each component module of the species model, or based on a fixed model size, and the amount of data stored by each unit model is within a preset range.

[0125] According to an embodiment of the present application, the selected current VR model is interacted with the current VR model through model interaction data to perform interaction information statistics and interaction feature analysis, the current VR model is regionally clustered and divided based on the interaction features to form a plurality of VR model partitions, and the VR model partitions are assigned initial weights based on the interaction features, and the method further comprises:

[0126] The model complexity is evaluated based on the rendering time and the model data volume of the current VR model, and the number of partitions K is set according to the model complexity;

[0127] The interaction feature data of each unit model is taken as clustering sample data, a Kmeans-based clustering model is imported, K initial clustering centers are set, and the clustering model is initialized;

[0128] In the clustering model, the similarity between the clustering sample data is calculated based on the cosine distance, and the data clustering and grouping are cyclically performed, and the clustering center point is recalculated each time the clustering process is performed;

[0129] When the clustering center point no longer moves, the clustering is stopped, and the clustering result is recorded, and a plurality of feature groups are generated through the clustering result;

[0130] The unit models are grouped and mapped based on the plurality of feature groups, and the unit regions of the same group are combined to form a plurality of VR model partitions;

[0131] For each feature group, the interaction frequency and the operation behavior of the corresponding interaction feature are analyzed, the user interaction interest degree of each feature group is determined, and the initial weight of each VR model partition is set based on the interaction interest degree.

[0132] It should be noted that the size of K is proportional to the model complexity. Each feature group includes interaction feature data of a plurality of unit regions. Each feature group corresponds to a VR model partition. The initial weight is proportional to the interaction interest degree.

[0133] According to an embodiment of the present application, the rendering test is performed on each VR model partition, a multi-vision two-dimensional image of the VR model partition is intercepted through the VR system at each test, the multi-vision two-dimensional image is subjected to feature learning and feature extraction based on the autoencoder, and the overall number of display image features of the system database is compared, and the overall influence coefficient of each VR model partition on the current VR model is calculated, specifically:

[0134] The rendering test is performed on each VR model partition, and the rendering test time is recorded;

[0135] During each test, the VR system captures a multi-visual 2D image of the VR model partition after rendering. The multi-visual 2D image includes multiple images, each corresponding to a visual angle.

[0136] Perform noise reduction and normalization on multi-visual 2D images;

[0137] The multi-visual two-dimensional image is extracted through the graph convolutional network, and the multi-visual features are stored to form a multi-visual feature set;

[0138] Build an autoencoder, set the input layer size to be consistent with the feature dimension of the multi-visual feature set, import the multi-visual feature set into the autoencoder for feature learning, compress the imported data into a low-dimensional representation based on the encoder, reconstruct the low-dimensional representation data through the decoder, and generate output data;

[0139] The difference between the output data and the multi-visual feature set is evaluated using the mean square error as the loss function, and the autoencoder is optimized based on the evaluation results. The generative training of the autoencoder is repeated until the loss function converges to a predetermined value. The output data at this time is recorded as the second feature set.

[0140] Obtain the overall digital display image features of the current VR model from the system database, perform feature classification on the overall digital display image features based on the set multi-visual angles, and form a multi-visual comparison feature set;

[0141] The second feature set is compared with the multi-visual comparison feature set and similarity is evaluated. The similarity evaluation is calculated using the Euclidean distance method, and multiple feature difference values ​​are obtained based on multi-visual analysis.

[0142] The mean of multiple feature difference values ​​is used as the overall influence coefficient of the VR model partition on the current VR model.

[0143] It should be noted that the visual angles in a multi-view 2D image are predefined angles, including front, side, and top views. The multi-view feature set serves as the input data for the autoencoder. The second feature set also includes multi-view features, which are the feature data obtained by learning and reconstructing the reduced dimensionality of the multi-view feature set and are used for rapid view evaluation. Multiple feature difference values ​​correspond to multiple visual angles.

[0144] Multi-view 2D images are image data that closely resembles how users actually view the model. This is simulated and evaluated in an information-based format. During the evaluation process, an autoencoder is used to learn and reduce the features of the VR model's corresponding display image, enabling rapid evaluation of the display effect and precise regional weighting. The higher the overall impact coefficient, the greater the impact of the corresponding VR model partition on the overall display, and the higher the rendering priority that can be set.

[0145] According to the embodiment of the present application, the initial weight is adjusted by the overall influence coefficient, and the partition display weight is obtained, the current VR model is prioritized by the partition display weight, and the rendering display scheme is generated, and the online VR model display is performed by the rendering display scheme, specifically:

[0146] The initial weight is adjusted by the overall influence coefficient, and the partition display weight is obtained;

[0147] The partition display weight corresponding to each VR model partition is calculated, and the partitions are prioritized according to the partition display weight, and the online rendering priority information of the current VR model is generated;

[0148] According to the partition display weight, the model resource allocation setting is performed on each VR model partition, and the resource allocation scheme is obtained;

[0149] The rendering display scheme for the current VR model is generated by the online rendering priority information and the resource allocation scheme.

[0150] It should be noted that the resource allocation scheme includes that in the VR display platform, the data transmission priority allocation is performed on the different VR model partition data corresponding to the resource data of a VR model, the data compression rate is set and calculated, and the computing resources and cache storage are matched, and the matching of the resource data is adjusted based on the partition display weight.

[0151] The partition display weight is calculated as follows:

[0152] Q E = Q S + K x P E ;

[0153] Wherein, Q E is the partition display weight, K is the correction coefficient, Q S is the initial weight, and P E is the overall influence coefficient.

[0154] According to the embodiment of the present application, the VR display platform includes a VR system and a display system, the VR system is used for collecting and processing VR model data, and the display system is used for rendering and visualizing VR model.

[0155] According to the embodiment of the present application, the VR display platform establishes data connection with the user terminal through a preset special network.

[0156] It should be noted that the user terminal can be used for online VR model visualization and VR interaction of the user.

[0157] It is worth mentioning here that the traditional VR display platform often lacks an efficient and feasible display scheme, and the model is simply packaged and transmitted, and the model priority is not dynamically regulated during the display process, so that the model display effect is limited by the VR device and the network state, the user experience is poor, and for the model of endangered species and its products, different regions in the VR model are often different in importance to the overall display, and the traditional display platform often does not fully consider the model partitioning condition, and lacks a fast and accurate display effect evaluation process, hindering the application of the VR platform.

[0158] Based on this, the application analyzes the partitioning of the VR model corresponding to the endangered species and its products through user interaction information, sets the corresponding weight, analyzes the user display effect and adjusts the weight through intercepting multi-vision two-dimensional images, and further generates a model priority and a display scheme, effectively improving the user experience, especially for users in different interaction conditions and different network conditions, having better adaptability and improving the comprehensive display experience of users.

[0159] The third aspect of the application also provides a computer readable storage medium, wherein the computer readable storage medium comprises a VR technology-based endangered species and its product display program, and the VR technology-based endangered species and its product display program is executed by a processor to realize the steps of the VR technology-based endangered species and its product display method according to any one of the above.

[0160] The application discloses a VR technology-based endangered species and its product display method and system. By acquiring model interaction data, interaction information statistics and feature analysis are performed to realize VR model region clustering division and initial weight distribution. Each partition is rendered and tested, image features are extracted using a self-encoder, and an overall influence coefficient is calculated by comparing with a system database, and then the weight is adjusted to sort the model priority and generate a rendering display scheme, realizing efficient and accurate online VR model display.

[0161] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only illustrative, for example, the division of the units is only a logical function division, and actual implementation can have another division mode, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed components can be through some interface, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0162] The units described as separate components above can or can not be physically separate, and the components shown as units can or can not be physical units; they can be located in one place or distributed on multiple network units; and part or all of the units can be selected according to actual needs to achieve the purposes of the embodiments.

[0163] In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be realized in the form of hardware or in the form of hardware plus software functional unit.

[0164] Those skilled in the art can understand that all or part of the steps of the above-mentioned method embodiments can be completed by program instruction related hardware, and the aforementioned program can be stored in a computer readable storage medium, and the program executes the steps including the above-mentioned method embodiments when executed; and the aforementioned storage medium includes mobile storage device, read-only memory (ROM), random access memory (RAM), magnetic disc or optical disc, and various storage medium that can store program code.

[0165] Alternatively, the integrated unit of the present application, if implemented in the form of a software function module and sold or used as an independent product, can also be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the embodiments of the present application can be embodied in the form of a software product, which is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in the embodiments of the present application. The aforementioned storage medium includes mobile storage device, ROM, RAM, magnetic disc or optical disc, and various storage medium that can store program code.

[0166] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A method for displaying endangered species and their products based on VR technology, characterized in that: include: Obtain interactive data on models of endangered species and their products through a fixed VR display platform; Select the current VR model, perform interaction information statistics and interaction feature analysis on the current VR model through model interaction data, perform regional clustering on the current VR model based on the interaction features to form multiple VR model partitions, and assign initial weights to the VR model partitions based on the interaction features; Perform a rendering test on each VR model partition. During each test, the VR system captures a multi-visual 2D image of the VR model partition. Feature learning and feature extraction are performed on the multi-visual 2D image based on the autoencoder. The image features are compared with the overall digital display image features in the system database to calculate the overall influence coefficient of each VR model partition on the current VR model. The initial weight is adjusted by the overall influence coefficient, and the partition display weight is obtained. The partition display weight is used to sort the partition priorities of the current VR model, and a rendering display plan is generated. The online VR model is displayed through the rendering display plan.

2. The method for displaying endangered species and their products based on VR technology according to claim 1, characterized in that: The method of obtaining interactive model data of endangered species and their products through a fixed VR display platform is as follows: During a historical period, models of endangered species and their products are displayed on a fixed VR display platform, and user interaction data on the models is collected; Model interaction data includes interaction instructions for the model, model control, display control, and device interaction statistics.

3. The method for displaying endangered species and their products based on VR technology according to claim 1, characterized in that: The current VR model is selected, and interaction information statistics and interaction feature analysis are performed on the current VR model through model interaction data. Based on the interaction features, the current VR model is clustered to form multiple VR model partitions, and initial weights are assigned to the VR model partitions based on the interaction features, specifically: Selecting a current VR model, wherein the current VR model includes model data corresponding to an endangered species and a product thereof; Based on the current VR model, multiple unit models are divided so that the data volume of each unit model is within the preset range; The model interaction data is used to perform interaction information statistics and interaction feature analysis on the current VR model. The interaction information statistics include statistics on the interaction frequency, operation behavior, and operation time of different unit areas in the current VR model, and generate interaction feature data for each unit model.

4. The method for displaying endangered species and their products based on VR technology according to claim 3, characterized in that: The method further includes: selecting a current VR model, performing interaction information statistics and interaction feature analysis on the current VR model through model interaction data, performing regional clustering on the current VR model based on the interaction features to form a plurality of VR model partitions, and assigning initial weights to the VR model partitions based on the interaction features. Evaluate the model complexity based on the rendering time and model data volume of the current VR model, and set the number of partitions K based on the model complexity; The interaction feature data of each unit model is used as clustering sample data, the Kmeans-based clustering model is imported, K initial cluster centers are set and the clustering model is initialized; In the clustering model, the similarity between clustered sample data is calculated based on the cosine distance, and the data clustering and grouping are performed cyclically, and the cluster center point is recalculated each time the clustering process; When the cluster center no longer moves, clustering is stopped and the clustering results are recorded. Multiple feature groups are generated based on the clustering results. Based on the multiple feature groups, the unit models are grouped and mapped, and the unit areas of the same group are combined to form multiple VR model partitions; For each feature group, the interaction frequency and operation behavior of the corresponding interaction features are analyzed to determine the user interaction interest of each feature group, and the initial weight is set for each VR model partition based on the interaction interest.

5. The method for displaying endangered species and their products based on VR technology according to claim 1, characterized in that: The rendering test is performed on each VR model partition. During each test, a multi-visual two-dimensional image of the VR model partition is captured by the VR system. Feature learning and feature extraction are performed on the multi-visual two-dimensional image based on the autoencoder, and compared with the overall digital display image features of the system database. The overall influence coefficient of each VR model partition on the current VR model is calculated, specifically: Perform rendering tests on each VR model partition and record the rendering test time; During each test, the VR system captures a multi-visual 2D image of the VR model partition after rendering. The multi-visual 2D image includes multiple images, each corresponding to a visual angle. Perform noise reduction and normalization on multi-visual 2D images; The multi-visual two-dimensional image is extracted through the graph convolutional network, and the multi-visual features are stored to form a multi-visual feature set; Build an autoencoder, set the input layer size to be consistent with the feature dimension of the multi-visual feature set, import the multi-visual feature set into the autoencoder for feature learning, compress the imported data into a low-dimensional representation based on the encoder, reconstruct the low-dimensional representation data through the decoder, and generate output data; The difference between the output data and the multi-visual feature set is evaluated using the mean square error as the loss function, and the autoencoder is optimized based on the evaluation results. The generative training of the autoencoder is repeated until the loss function converges to a predetermined value. The output data at this time is recorded as the second feature set. Obtain the overall digital display image features of the current VR model from the system database, perform feature classification on the overall digital display image features based on the set multi-visual angles, and form a multi-visual comparison feature set; The second feature set is compared with the multi-visual comparison feature set and similarity is evaluated. The similarity evaluation is calculated using the Euclidean distance method, and multiple feature difference values ​​are obtained based on multi-visual analysis. The mean of multiple feature difference values ​​is used as the overall influence coefficient of the VR model partition on the current VR model.

6. The method for displaying endangered species and their products based on VR technology according to claim 5, characterized in that: The initial weight is adjusted by the overall influence coefficient, and the partition display weight is obtained. The current VR model is prioritized by the partition display weight, and a rendering display scheme is generated. The online VR model is displayed by the rendering display scheme, specifically: Adjust the initial weights by the overall influence coefficient and obtain the partition display weights; Calculate the partition display weight corresponding to each VR model partition, prioritize the partitions based on the partition display weight, and generate online rendering priority information for the current VR model; According to the partition display weight, model resource allocation is set for each VR model partition to obtain a resource allocation plan; Generate a rendering display plan for the current VR model through online rendering priority information and resource allocation plan.

7. The method for displaying endangered species and their products based on VR technology according to claim 1, characterized in that: The VR display platform includes a VR system and a display system. The VR system is used to collect and process VR model data, and the display system is used to render and visualize VR models.

8. The method for displaying endangered species and their products based on VR technology according to claim 1, characterized in that: The VR display platform establishes a data connection with the user terminal through a preset dedicated network.

9. A VR-based display system for endangered species and their products, characterized by: The system includes: a memory and a processor. The memory includes a VR technology-based display program for endangered species and their products. When the VR technology-based display program for endangered species and their products is executed by the processor, the following steps are implemented: Obtain interactive data on models of endangered species and their products through a fixed VR display platform; Select the current VR model, perform interaction information statistics and interaction feature analysis on the current VR model through model interaction data, perform regional clustering on the current VR model based on the interaction features to form multiple VR model partitions, and assign initial weights to the VR model partitions based on the interaction features; Perform a rendering test on each VR model partition. During each test, the VR system captures a multi-visual 2D image of the VR model partition. Feature learning and feature extraction are performed on the multi-visual 2D image based on the autoencoder. The image features are compared with the overall digital display image features in the system database to calculate the overall influence coefficient of each VR model partition on the current VR model. The initial weight is adjusted by the overall influence coefficient, and the partition display weight is obtained. The partition display weight is used to sort the partition priorities of the current VR model, and a rendering display plan is generated. The online VR model is displayed through the rendering display plan.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a display program for endangered species and their products based on VR technology. When the display program for endangered species and their products based on VR technology is executed by a processor, the steps of the method for displaying endangered species and their products based on VR technology as described in any one of claims 1 to 8 are implemented.

Images