Abnormality identification method and system based on VR cinema software development environment

Abstract

The invention discloses an exception identification method and system based on a VR cinema software development environment, and relates to the technical field of exception identification. The method comprises the steps of collecting environment operation data in a VR cinema software development environment; constructing a normal behavior model and an anomaly recognition model based on the environment operation data; monitoring and testing the software operation performance in the VR cinema software development process; acquiring real-time operation data, performing anomaly recognition based on the anomaly recognition model, and marking recognized anomaly data; and classifying the abnormal data and feeding back the abnormal data to developers for restoration and optimization. According to the invention, through effective data acquisition, the anomaly recognition model is constructed based on the data, anomaly recognition and real-time response are carried out, the performance of the software development equipment and the VR equipment is monitored, and repair and optimization are carried out according to feedback information, so that the anomaly recognition efficiency and the system stability are improved, and the user experience is improved. And powerful technical support is provided for development, test and optimization of VR cinema software.

Description

Technical Field

[0001] The present invention belongs to the technical field of anomaly recognition, and particularly relates to an anomaly recognition method and system based on a VR cinema software development environment. Background Art

[0002] With the continuous development of virtual reality (VR) technology, VR cinemas, as an innovative way of watching movies, have gradually gained the favor of users. VR cinemas can provide immersive audio-visual experiences, bringing users a completely different movie-watching feeling from traditional cinemas. However, due to the hardware complexity of VR devices and the diversity of software development environments, various abnormal phenomena often occur in VR cinema software during actual applications, such as frame stuttering, image distortion, poor user interaction, insufficient system performance, etc. These abnormal phenomena not only affect the user experience but may also lead to system crashes or failures. Therefore, how to timely detect and solve these problems is an important challenge in the VR cinema software development process.

[0003] Currently, traditional anomaly detection methods mainly rely on manual testing and manual troubleshooting. This method is not only inefficient but also difficult to cover all possible abnormal situations. Therefore, there is an urgent need for an anomaly recognition method based on automated and intelligent means that can monitor and identify abnormal problems in the VR cinema software development environment in real time, discover potential risks in advance, and give corresponding feedback and repair suggestions. Summary of the Invention

[0004] The purpose of the present invention is to provide an anomaly recognition method and system based on a VR cinema software development environment, which can be achieved through the following technical solutions: In a first aspect, an embodiment of the present application provides an anomaly recognition method based on a VR cinema software development environment, including the following steps: Collect environmental operation data in the VR cinema software development environment; Construct a normal behavior model and an anomaly recognition model based on the environmental operation data; Monitor and test the software operation performance during the VR cinema software development process; Obtain real-time operation data and perform anomaly recognition based on the anomaly recognition model, and mark the recognized abnormal data; Classify the abnormal data and feedback the classified data to developers, and the developers repair and optimize the VR cinema software according to the feedback information.

[0005] Preferably, the environmental operation data includes user interaction data, system resource data, image quality data, and video stream data.

[0006] Preferably, constructing the normal behavior model and the anomaly recognition model based on the environmental operation data includes: Dividing the environmental operation data to obtain normal operation data and abnormal operation data; Establishing a normal behavior model based on the normal operation data; Establishing an anomaly recognition model based on the abnormal operation data; Among them, the normal behavior model is used to define the normal scope of various operations and behaviors in the VR cinema software development environment, providing a behavior benchmark and reference for the anomaly recognition model.

[0007] Preferably, establishing the normal behavior model based on the normal operation data includes: Selecting behavior pattern-related features from the normal operation data; Judging the importance among the behavior pattern-related features and selecting important clustering features; Taking the feature data represented by the important clustering features as the input of K-means clustering; Adopting a method combining the elbow method and the silhouette coefficient to select the optimal K value of K-means clustering to divide the feature data into K clusters; Using the selected optimal K value, clustering the feature data using the K-means algorithm; After clustering, analyzing the results of each cluster to understand the normal behavior pattern represented by each cluster; Using the clustering result as the benchmark of the normal behavior pattern and constructing the normal behavior model accordingly; Among them, the important clustering features represent the normal behavior features that contribute to establishing the normal behavior model among the behavior pattern-related features.

[0008] Preferably, adopting the method combining the elbow method and the silhouette coefficient to select the optimal K value of K-means clustering includes: Determining a range of K values through the elbow method, specifically: Calculating the within-cluster sum of squares corresponding to different K values; Plotting the relationship diagram between the K value and the within-cluster sum of squares; Finding the elbow position where the within-cluster sum of squares significantly decreases and slows down in the relationship diagram, and determining a range of K values based on this elbow position; Calculating the silhouette coefficient of each K value within the range of K values determined by the elbow method, and selecting the K value with the highest silhouette coefficient as the optimal K value of the K-means clustering; If the K values selected by the elbow method and the silhouette coefficient are the same, directly select this K value as the optimal K value; if the K values selected by the two are different, preferentially select the K value with a higher silhouette coefficient as the optimal K value.

[0009] Preferably, establishing the anomaly recognition model based on the abnormal operation data includes: Converting the abnormal operation data according to the data type and dividing it into time series data and numerical data; Processing the time series data using a long short-term memory network to obtain a first processing result; Processing the numerical data using an isolation forest to obtain a second processing result; Fusing the first processing result and the second processing result to obtain the anomaly recognition model, specifically: Taking the first processing result and the second processing result as feature vectors and inputting them into an ensemble model; Training and learning the ensemble model to obtain the anomaly recognition model.

[0010] Preferably, processing the time series data using a long short-term memory network to obtain a first processing result includes: Preprocessing the time series data, including handling missing values, removing outliers, and data standardization; Performing time series segmentation on the preprocessed data, and dividing the obtained data set into a training set, a validation set, and a test set; Designing the network structure, and configuring the input layer, LSTM layer, and output layer of the long short-term memory network; Selecting a loss function and an optimizer, and performing backpropagation training on the training set through the backpropagation algorithm. Adjusting the parameters through multiple iterations until the LSTM model converges; Adopting an early stopping strategy and monitoring the training process through the validation set; After the training and validation processes are completed, using the trained LSTM model to predict the test set to obtain predicted values; Taking the predicted values as the first processing result.

[0011] Preferably, processing the numerical data using an isolation forest to obtain a second processing result includes: Preprocessing the numerical data to obtain a training data set, including data cleaning and data standardization; Constructing an isolation forest model by selecting the hyperparameters of the isolation forest, including selecting the number of trees in the forest, determining the number of samples randomly selected for each tree, and determining the proportion of outliers in the numerical data; Using the training data set to fit the isolation forest model and marking the outliers; Taking the outliers as the second processing result.

[0012] Preferably, the monitoring and testing of the software running performance in the VR cinema software development process include: Determine the performance monitoring objectives, including frame rate, latency, resource consumption, and smoothness; Monitor the performance monitoring objectives in real time and record the performance metric logs; Conduct performance testing based on the performance metric logs to obtain test results, and perform performance optimization based on the test results.

[0013] In a second aspect, an embodiment of the present application provides an anomaly recognition system based on a VR cinema software development environment, applying the anomaly recognition method as described above, including: Data acquisition module: used to acquire the environmental operation data in the VR cinema software development environment; Model construction module: construct a normal behavior model and an anomaly recognition model based on the environmental operation data; Performance monitoring module: used to monitor and test the software running performance in the VR cinema software development process; Anomaly recognition module: used to obtain real-time operation data and perform anomaly recognition based on the anomaly recognition model, and mark the recognized anomaly data; Response optimization module: used to classify the anomaly data and feedback the classified data to the developers, and the developers repair and optimize the VR cinema software according to the feedback information.

[0014] The beneficial effects of the present invention are as follows: The present invention constructs an anomaly recognition model through effective data acquisition, then uses the model for anomaly recognition and real-time response, also monitors the performance of software development devices and VR devices, and repairs the developed software and optimizes the development process according to the feedback information, thereby improving the anomaly recognition efficiency and system stability, and further being able to effectively identify and solve various anomaly problems in the VR cinema, enhancing the software stability and user experience, and providing strong technical support for the development, testing, and optimization of VR cinema software. Description of the Drawings

[0015] For better understanding and implementation, the technical solutions of the present application will be described in detail below with reference to the drawings.

[0016] Figure 1 It is a step flow chart of an anomaly recognition method based on a VR cinema software development environment provided by an embodiment of the present application; Figure 2 It is a step flow chart of establishing a normal behavior model provided by an embodiment of the present application; Figure 3 It is a step flow chart of establishing an anomaly recognition model provided by an embodiment of the present application; Figure 4 The structural schematic diagram of an anomaly recognition system based on a VR cinema software development environment provided by an embodiment of the present application. Specific embodiments

[0017] To further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the exemplary embodiments will be described in detail herein, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. On the contrary, they are merely examples of methods and systems consistent with some aspects of the present application as detailed in the appended claims.

[0018] The terms used in the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. The singular forms "a", "the", and "said" used in the present application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to any or all possible combinations of one or more of the associated listed items.

[0019] The following, in conjunction with the accompanying drawings and preferred embodiments, details the specific embodiments, features, and their effects according to the present invention.

[0020] Embodiment 1 Please refer to Figure 1 , an embodiment of the present application provides an anomaly recognition method based on a VR cinema software development environment, including the following steps: Collect the environmental operation data in the VR cinema software development environment; Build a normal behavior model and an anomaly recognition model based on the environmental operation data; Monitor and test the software operation performance during the VR cinema software development process; Obtain real-time operation data and perform anomaly recognition based on the anomaly recognition model, and mark the recognized anomaly data; Classify the anomaly data and feedback the classified data to the developers, and the developers repair and optimize the VR cinema software according to the feedback information.

[0021] Specifically, since the existing methods for identifying environmental anomalies in the software development process of VR cinemas mainly rely on manual testing and manual troubleshooting, this method is not only inefficient but also difficult to cover all possible anomaly situations. Therefore, in order to improve the efficiency and accuracy of anomaly identification and the speed of anomaly handling, this application combines means such as automation and artificial intelligence to jointly achieve the anomaly identification of the VR cinema software development environment, specifically including: First, collect the environmental operation data in the VR cinema software development environment, including but not limited to user interaction data, system resource data, image quality data, and video stream data; then build a normal behavior model and an anomaly identification model based on the environmental operation data; then monitor and test the software operation performance during the VR cinema software development process; also obtain real-time operation data and perform anomaly identification based on the above anomaly identification model, and mark the identified anomaly data; finally, classify the anomaly data and feedback the classified data to the developers, and the developers repair and optimize the VR cinema software according to the feedback information.

[0022] This application collects effective data and builds an anomaly identification model based on it, then uses the model for anomaly identification and real-time response, also monitors the performance of software development devices and VR devices, and repairs the developed software and optimizes the development process according to the feedback information, thereby improving the anomaly identification efficiency and system stability, and furthermore, it can effectively identify and solve various anomaly problems in VR cinemas, enhance the software stability and user experience, and provide strong technical support for the development, testing, and optimization of VR cinema software.

[0023] Through automated data collection and intelligent analysis, this application can identify various anomalies in VR cinema software in real time and accurately, significantly improving the efficiency of anomaly detection; through the early warning mechanism for timely feedback, it reduces the manual intervention of developers, reduces the workload and misjudgment rate of manual inspection; through real-time monitoring and anomaly identification, it can timely discover potential performance bottlenecks and functional failures in the system, reducing the risk of system crashes and failures, thereby improving the stability and reliability of VR cinema software; by identifying and repairing various anomalies, it ensures that the immersive experience of users in VR cinemas is not disturbed, enhancing the viewing experience and satisfaction of users.

[0024] In an embodiment provided by this application, the environmental operation data includes user interaction data, system resource data, image quality data, and video stream data.

[0025] Specifically, the above user interaction data includes gestures, voices, motion capture, etc.; the system resource data includes the usage conditions such as memory, CPU, and GPU loads; the image quality data includes data related to images such as texture distortion, image stuttering, and resolution degradation; the video stream data includes playback latency or jitter, etc. The above environmental operation data is used to reflect the operation conditions of the VR cinema software development environment.

[0026] In an embodiment provided by the present application, constructing the normal behavior model and the anomaly recognition model based on the environmental operation data includes: Dividing the environmental operation data to obtain normal operation data and abnormal operation data; Establishing a normal behavior model based on the normal operation data; Establishing an anomaly recognition model based on the abnormal operation data; Among them, the normal behavior model is used to define the normal scope of various operations and behaviors in the VR cinema software development environment, and provides a behavior benchmark and reference for the anomaly recognition model.

[0027] Specifically, in this embodiment, the obtained environmental operation data is divided into normal operation data and abnormal operation data, and a normal behavior model is established based on the normal operation data. By defining the normal scope of various operations and behaviors in the VR cinema software development environment, it is used to provide a behavior benchmark and reference for the anomaly recognition model; an anomaly recognition model is also established based on the abnormal operation data. Therefore, the focus of this embodiment is to establish two different models respectively, and regulate the anomaly recognition model through the normal behavior model, providing a basis and a clear benchmark for anomaly detection, so that the anomaly recognition model can accurately and effectively monitor and identify abnormal behaviors, thereby improving the stability, performance, and user experience of the VR cinema software development environment.

[0028] As Figure 2 shown, in an embodiment provided by the present application, establishing the normal behavior model based on the normal operation data includes: Selecting behavior pattern-related features from the normal operation data; the behavior pattern-related features include features related to normal operation and interaction such as operation time, click frequency, number of environmental interactions, frame rate, and loading time; Judging the importance in the behavior pattern-related features and selecting important clustering features; Taking the feature data represented by the important clustering features as the input of K-means clustering; Adopting a method combining the elbow method and the silhouette coefficient to select the optimal K value of K-means clustering to divide the feature data into K clusters; Using the selected optimal K value, using the K-means algorithm to cluster the feature data; After clustering is completed, the results of each cluster are analyzed to understand the normal behavior patterns represented by each cluster; Use the clustering results as a benchmark for normal behavior patterns and build a normal behavior model based on this; Among them, the important clustering features represent normal behavior features in the behavior pattern-related features that help establish a normal behavior model.

[0029] Specifically, in this embodiment, K-means Clustering is used to establish a normal behavior model for defining the normal scope of various operations and behaviors in the VR cinema software development environment. It mainly clusters the data of normal behaviors and uses these clustering results as behavior benchmarks. The key steps in the process of establishing a normal behavior model by K-means Clustering in this embodiment include data collection and preprocessing, feature extraction, selection of an appropriate K value, execution of K-means Clustering, analysis of clustering results, and construction of a normal behavior benchmark. In this way, a benchmark can be defined for normal operations and behaviors in the VR cinema software development environment, and data support can be provided for effectively identifying abnormal behaviors subsequently.

[0030] In an embodiment provided by this application, the method of selecting the optimal K value for K-means Clustering by combining the elbow method and the silhouette coefficient includes: Determine a range of K values through the elbow method, specifically: Calculate the within-cluster sum of squares corresponding to different K values; Draw a relationship graph of K values and the within-cluster sum of squares; Find the elbow position in the relationship graph where the within-cluster sum of squares significantly decreases and slows down, and determine a range of K values based on this elbow position; Calculate the silhouette coefficient of each K value within the range of K values determined by the elbow method, and select the K value with the highest silhouette coefficient as the optimal K value for the K-means Clustering; If the K values selected by the elbow method and the silhouette coefficient are the same, directly select this K value as the optimal K value; if the K values selected by the two are different, preferentially select the K value with a higher silhouette coefficient as the optimal K value.

[0031] Specifically, in this embodiment, the elbow method and the silhouette coefficient are combined to select an appropriate value of K, which can more comprehensively evaluate the quality of the clustering results. Specifically, the elbow method can provide a rough range of K values, while the silhouette coefficient can quantify the quality of clustering for different K values. Specifically, the combined use of the elbow method and the silhouette coefficient can more accurately select the value of K; the elbow method provides a relatively rough range of K values, while the silhouette coefficient can provide more information about the clustering quality. A K value with a high silhouette coefficient is usually the preferred optimal K value, but if the K value given by the elbow method is significantly reasonable and consistent with the silhouette coefficient, then this K value is directly selected. Through the above combination method, this embodiment can more scientifically determine the most suitable number of clusters.

[0032] It can be understood that if the elbow position of the elbow method is different from the K value with the highest silhouette coefficient, then the K value with a relatively high silhouette coefficient is given priority, because it quantifies the clustering quality.

[0033] It should be noted that the Within-Cluster Sum of Squares (WSS) is used in clustering analysis to measure the sum of the squares of the distances from the data points within each cluster to the center of that cluster. The smaller the WSS, the closer the data points within the cluster are, and the better the clustering effect.

[0034] As Figure 3 shown, in an embodiment provided by the present application, establishing the anomaly recognition model based on the abnormal operation data includes: Converting the abnormal operation data according to the data type and dividing it into time series data and numerical data; Processing the time series data by using a long short-term memory network to obtain a first processing result; Processing the numerical data by using an isolation forest to obtain a second processing result; Fusing the first processing result and the second processing result to obtain the anomaly recognition model, specifically: Taking the first processing result and the second processing result as feature vectors and inputting them into an ensemble model; Training and learning the ensemble model to obtain the anomaly recognition model.

[0035] Specifically, since the abnormal operation data includes different types of data such as user interaction data, system resource data, image quality data, and video stream data, in order to ensure data consistency and the accuracy of subsequent model construction, in this embodiment, the above abnormal operation data is divided into time series data and numerical data. The subdivision of these data categories helps to analyze and optimize system performance, user experience, video quality, etc. from different dimensions. Through in-depth analysis of these data, potential problems can be identified and corresponding optimization adjustments can be made. In addition, in this embodiment, two algorithms, namely Long Short-Term Memory (LSTM) and Isolation Forest, are fused. By combining the anomaly detection capabilities of these two algorithms, a more accurate anomaly detection model can be obtained.

[0036] It should be noted that image quality data and video stream data are usually not directly regarded as traditional numerical data, but they can be converted into numerical data for processing. Image quality data is usually an evaluation value obtained through image processing techniques, and thus is converted into high-dimensional numerical data; video streams also have a time series nature, so their processing involves both numerical data and may combine time series analysis methods.

[0037] In an embodiment provided by the present application, the step of using the long short-term memory network to process the time series data to obtain a first processing result includes: Preprocessing the time series data, including handling missing values, removing outliers, and data standardization; Performing time series segmentation on the preprocessed data, and dividing the segmented data set into a training set, a validation set, and a test set; Designing a network structure, and configuring an input layer, an LSTM layer, and an output layer of the long short-term memory network; Selecting a loss function and an optimizer, and performing backpropagation training on the training set through the backpropagation algorithm. Adjusting the parameters through multiple iterations until the LSTM model converges; Adopting an early stopping strategy and monitoring the training process through the validation set; used to ensure that the LSTM model is not overfitted; After the training and validation processes are completed, using the trained LSTM model to predict the test set to obtain predicted values; Taking the predicted values as the first processing result.

[0038] Specifically, in this embodiment, LSTM is used to process time series data. Through steps such as data preprocessing, LSTM model construction, model training, evaluation, and validation, it helps the LSTM model better understand the patterns and trends in time series data, so as to make accurate predictions and anomaly detections.

[0039] In an embodiment provided by the present application, processing the numerical data using the Isolation Forest to obtain a second processing result includes: Preprocessing the numerical data to obtain a training data set, including data cleaning and data standardization; Constructing an Isolation Forest model by selecting the hyperparameters of the Isolation Forest, including selecting the number of trees in the forest, determining the number of samples randomly selected for each tree, and determining the proportion of outliers in the numerical data; Fitting the Isolation Forest model using the training data set and marking the outliers; Taking the outliers as the second processing result.

[0040] Specifically, in this embodiment, the Isolation Forest is used to process numerical data, mainly for anomaly detection, and its goal is to identify and mark outliers in the data. Through steps such as data preprocessing, constructing and training the Isolation Forest model, and predicting outliers, the Isolation Forest can effectively detect outliers in the data. It should be noted that during the process of constructing the Isolation Forest model, the hyperparameters of the Isolation Forest are selected, including: the number of trees in the forest: the number of trees determines the complexity of the model. More trees can improve the stability and accuracy of the model, but also increase the computational cost; the number of samples randomly selected for each tree: if set to a smaller value, the training speed of the model is faster, but the accuracy may be sacrificed; the proportion of outliers in the data: this parameter helps the model understand the expected proportion of outliers in the data. The common method is to set it according to prior knowledge or through cross-validation, as well as to control randomness to ensure the reproducibility of the experiment.

[0041] It should be noted that when fusing the outputs of the two models in this embodiment, specifically, by training an ensemble model (such as logistic regression, decision tree, etc.), taking the prediction error (predicted value) of the LSTM and the anomaly score (outliers) of the Isolation Forest as feature inputs, and finally obtaining an anomaly recognition model through training and learning to predict whether the data is an anomaly.

[0042] Therefore, generally speaking, the LSTM is good at capturing trends and patterns in time series, while the Isolation Forest is good at identifying outliers. Therefore, by fusing the LSTM and the Isolation Forest in the present application, a more accurate and robust anomaly detection model can be obtained.

[0043] In an embodiment provided by the present application, monitoring and testing the software running performance during the development of the VR cinema software includes: Determining the performance monitoring objectives, including frame rate, latency, resource consumption, and smoothness; Monitor the performance monitoring target in real time and record the performance metric logs; Conduct performance tests based on the performance metric logs to obtain test results, and perform performance optimization based on the test results.

[0044] Specifically, the performance monitoring and testing of the VR cinema software in this embodiment is a continuous process, involving tests in multiple dimensions, including rendering performance, latency, resource consumption, cross-platform compatibility, network bandwidth, user experience, etc. Through the above steps and methods, it can be ensured that the software can provide a high-quality and smooth VR experience in different environments, avoiding problems such as discomfort or degraded experience for users during viewing.

[0045] Embodiment 2 Please refer to Figure 4 , the embodiment of the present application provides an anomaly recognition system based on the VR cinema software development environment, applying an anomaly recognition method based on the VR cinema software development environment as described above, including: Data acquisition module: used to acquire the environmental operation data in the VR cinema software development environment; Model construction module: construct a normal behavior model and an anomaly recognition model based on the environmental operation data; Performance monitoring module: used to monitor and test the software operation performance during the VR cinema software development process; Anomaly recognition module: used to obtain real-time operation data and perform anomaly recognition based on the anomaly recognition model, and mark the recognized anomaly data; Response optimization module: used to classify the anomaly data and feedback the classified data to the developers, and the developers repair and optimize the VR cinema software according to the feedback information.

[0046] In the above embodiments, the descriptions of each embodiment have their own emphases. For parts not detailed or recorded in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0047] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above division of each functional unit and module is used as an example. In actual applications, the above functions can be assigned to different functional units and modules according to needs, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment can be integrated into a processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit. In addition, the specific names of each functional unit and module are only for the convenience of mutual distinction and do not limit the protection scope of this application. The specific working processes of the units and modules in the above system can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated here.

[0048] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or by a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0049] As described above, it is only a preferred embodiment of the present invention and does not impose any form of limitation on the present invention. Although the present invention has been disclosed above with a preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to equivalent embodiments by using the disclosed technical content within the scope of the technical solution of the present invention. However, as long as it does not depart from the content of the technical solution of the present invention, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. An abnormality identification method based on VR cinema software development environment, characterized in that: The steps include: Collect environmental operation data within the VR cinema software development environment; Building a normal behavior model and an abnormality recognition model based on the environmental operation data; Monitor and test the software performance during the VR cinema software development process; Acquire real-time operation data and perform abnormality identification based on the abnormality identification model, and mark the identified abnormal data; The abnormal data is classified and the classified data is fed back to the developer, and the developer repairs and optimizes the VR cinema software according to the feedback information.

2. According to claim 1, an abnormality identification method based on a VR cinema software development environment is characterized in that: The environment operation data includes user interaction data, system resource data, image quality data and video stream data.

3. According to claim 1, the abnormality identification method based on the VR cinema software development environment is characterized in that: The constructing of a normal behavior model and an abnormality recognition model based on the environmental operation data includes: Dividing the environmental operation data to obtain normal operation data and abnormal operation data; Establishing a normal behavior model based on the normal operation data; Establishing an abnormality recognition model based on the abnormal operation data; Among them, the normal behavior model is used to define the normal scope of various operations and behaviors in the VR cinema software development environment, and provide a behavior benchmark and reference for the abnormal identification model.

4. According to claim 3, the abnormality identification method based on the VR cinema software development environment is characterized in that: The establishing of a normal behavior model based on the normal operating data includes: selecting behavior pattern related features from the normal operation data; Making importance judgments among the behavior pattern related features and selecting important clustering features; Using the feature data represented by the important clustering features as input for K-means clustering; Selecting an optimal K value of K-means clustering by combining the elbow rule and the silhouette coefficient to divide the feature data into K clusters; Using the selected optimal K value, clustering the feature data using a K-means algorithm; After clustering is completed, the results of each cluster are analyzed to understand the normal behavior pattern represented by each cluster; Use the clustering results as a baseline for normal behavior patterns and use them to build a normal behavior model; The important clustering features represent normal behavior features that are helpful in establishing a normal behavior model among the behavior pattern related features.

5. According to claim 4, an abnormality identification method based on a VR cinema software development environment is characterized in that: The method of selecting the optimal K value of K-means clustering by combining the elbow rule and the silhouette coefficient includes: A K value range is determined by the elbow rule, specifically: Calculate the intra-cluster sum of squares corresponding to different K values; Plot the relationship between K value and the sum of squares within the cluster; Find the elbow position where the sum of squares within the cluster decreases significantly in the relationship graph, and determine a K value range based on this elbow position; Calculating the silhouette coefficient of each K value within the K value range determined by the elbow rule, and selecting the K value with the highest silhouette coefficient as the optimal K value for the K-means clustering; If the K value selected by the elbow rule and the silhouette coefficient is consistent, then this K value is directly selected as the optimal K value; if the K values ​​selected by the two are inconsistent, then the K value with a higher silhouette coefficient is preferentially selected as the optimal K value.

6. According to claim 3, the abnormality identification method based on the VR cinema software development environment is characterized in that: The establishing of an abnormality identification model based on the abnormal operation data comprises: Convert the abnormal operation data into time series data and numerical data according to the data type; Processing the time series data using a long short-term memory network to obtain a first processing result; Using isolation forest to process the numerical data to obtain a second processing result; The first processing result and the second processing result are integrated to obtain the abnormality recognition model, which is specifically: Inputting the first processing result and the second processing result into an integrated model as feature vectors; The integrated model is trained and learned to obtain the abnormality recognition model.

7. The abnormality identification method based on the VR cinema software development environment according to claim 6 is characterized in that: The adopting the long short-term memory network to process the time series data to obtain a first processing result includes: Preprocessing the time series data, including processing missing values, removing outliers, and standardizing data; Perform time series segmentation on the preprocessed data, and divide the segmented data set into training set, validation set and test set; Design a network structure and configure the input layer, LSTM layer and output layer of the long short-term memory network; Select the loss function and optimizer, perform back-propagation training on the training set through the back-propagation algorithm, and adjust the parameters after multiple iterations until the LSTM model converges; Use early stopping strategy and monitor the training process through validation set; After the training and validation process is completed, the trained LSTM model is used to predict the test set to obtain the predicted value; The predicted value is used as the first processing result.

8. The abnormality identification method based on the VR cinema software development environment according to claim 6 is characterized in that: The using isolation forest to process the numerical data to obtain a second processing result includes: Preprocessing the numerical data to obtain a training data set, including data cleaning and data standardization; constructing an isolation forest model by selecting hyperparameters of the isolation forest, including selecting the number of trees in the forest, determining the number of randomly selected samples for each tree, and determining the proportion of outliers in the numerical data; Fitting the isolation forest model using the training data set and marking outliers; The abnormal value is used as the second processing result.

9. The abnormality identification method based on the VR cinema software development environment according to claim 1 is characterized in that: The monitoring and testing of software performance during the VR cinema software development process includes: Determine performance monitoring goals, including frame rate, latency, resource consumption, and smoothness; Monitor the performance monitoring target in real time and record performance indicator logs; A performance test is performed according to the performance indicator log to obtain a test result, and performance optimization is performed according to the test result.

10. An abnormality identification system based on a VR cinema software development environment, applying the abnormality identification method according to any one of claims 1 to 9, characterized in that: include: Data collection module: used to collect environmental operation data within the VR cinema software development environment; Model building module: building a normal behavior model and an abnormality recognition model based on the environment operation data; Performance monitoring module: used to monitor and test the software running performance during the VR cinema software development process; Anomaly recognition module: used to obtain real-time operation data and perform anomaly recognition based on the anomaly recognition model, and mark the recognized abnormal data; Response optimization module: used to classify the abnormal data and feed back the classified data to the developer, and the developer repairs and optimizes the VR cinema software according to the feedback information.

Images