Iptv terminal edge computing method and system

Abstract

The present application relates to a kind of IPTV terminal edge computing method and system, it is related to IPTV terminal technical field, including the following steps, IPTV terminal edge computing method first analyzes user interaction behavior, obtains behavior mode.Based on this mode, system carries out dynamic matching and context processing to content request, generates content preference matrix.Then, the task scheduling of local edge node is optimized using the matrix.System subsequently carries out hardware resource mapping, obtains equipment adaptation parameter, and accordingly pipeline adaptation is carried out to video decoder, generates local rendering instruction.Finally, based on these instructions, video frame buffer is rendered asynchronously, and the output of smooth video play stream is solved.The problem that traditional centralization IPTV system often exists network delay, service quality decline when facing large-scale user access.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of IPTV terminal, and particularly relates to an IPTV terminal edge computing method and system. BACKGROUND

[0002] With the rapid development of Internet technology, IPTV services are increasingly popular, and users' requirements for video content quality and smoothness are constantly improving. However, traditional centralized IPTV systems often have problems such as network delay and service quality degradation when facing large-scale user access. At the same time, the diversity of user behavior and preferences also brings challenges to content recommendation and resource scheduling. These factors seriously affect the user's viewing experience and restrict the further development of IPTV services.

[0003] In order to solve these problems, researchers have begun to focus on the application of edge computing technology in the field of IPTV. Edge computing can effectively reduce network delay and improve service response speed by sinking computing and storage resources to the network edge. However, how to effectively utilize the edge computing capabilities of IPTV terminals to realize intelligent and personalized user behavior analysis, content preference prediction, resource optimization scheduling, and video decoding rendering is still a problem to be solved. In addition, how to achieve efficient resource allocation and task scheduling on heterogeneous terminal devices, and how to ensure the smoothness and quality of video playback, are also important challenges in current IPTV edge computing research. SUMMARY

[0004] The purpose of the present application is to at least partially solve one of the problems in the prior art.

[0005] To achieve the above purpose, the present application provides an IPTV terminal edge computing method, comprising the following steps:

[0006] Performing multi-dimensional time series analysis on user interaction behavior data of the IPTV terminal to obtain user behavior patterns;

[0007] Performing dynamic matching and context-aware processing on content request sequences of the IPTV terminal through the user behavior patterns to obtain a content preference distribution matrix;

[0008] Performing task scheduling optimization on resource state information of a preset local edge node based on the content preference distribution matrix to obtain an edge deployment task queue;

[0009] Performing heterogeneous hardware resource mapping analysis on the IPTV terminal based on the edge deployment task queue to obtain device adaptation configuration parameters;

[0010] Adapt the decoding and rendering pipeline of the video decoder module of the IPTV terminal based on the device adaptation configuration parameters to generate local rendering instructions;

[0011] Perform asynchronous rendering processing on the video frame buffer of the IPTV terminal based on the local rendering instructions to obtain a smooth video playback output stream.

[0012] Further, perform multi-dimensional time series analysis on the user interaction behavior data of the IPTV terminal to obtain user behavior patterns, including:

[0013] Perform time slicing processing on the user interaction behavior data of the IPTV terminal based on a sliding window mechanism to obtain multiple time slice data sets, and label each time slice data set with behavior event types to obtain labeled time slice data;

[0014] Assign weights to different behavior event types in the labeled time slice data to obtain weighted time slice data, and extract feature vectors in the time dimension based on the weighted time slice data to obtain time feature vectors;

[0015] Perform clustering analysis on the time feature vectors to obtain initial user behavior patterns, and perform pattern merging and optimization processing on the initial user behavior patterns to obtain user behavior patterns.

[0016] Further, dynamically match the content request sequence of the IPTV terminal with the user behavior patterns and perform context-aware processing to obtain a content preference distribution matrix, including:

[0017] Map each behavior pattern in the user behavior patterns to a semantic label to obtain behavior semantic features, and perform metadata analysis on the content request sequence of the IPTV terminal to extract program content feature vectors;

[0018] Based on the semantic similarity calculation results between the behavior semantic features and the program content feature vectors, construct a behavior-content correlation matrix, and perform context weight modulation on the behavior-content correlation matrix to obtain a context-enhanced preference matching matrix;

[0019] Dynamically assign labels to the content classification system in the pre-set local content library through the context-enhanced preference matching matrix to generate a content preference label sequence, and generate a candidate set of program entries in the local content library based on the content preference label sequence to obtain a content recommendation candidate set;

[0020] The candidate program items are prioritized based on matching degrees between historical playing hot degrees of the program items in the content recommendation candidate set and the user behavior mode, and a content preference distribution matrix is generated.

[0021] Further, the content preference distribution matrix is used to optimize task scheduling of preset local edge node resource state information, to obtain an edge deployment task queue, including:

[0022] The resource state information of the preset local edge node is monitored in real time to obtain an edge node state monitoring table, and resource availability in the edge node state monitoring table is evaluated to obtain a node resource state matrix;

[0023] The content preference distribution matrix and the node resource state matrix are matched to obtain a task resource matching degree table, and available edge nodes in the task resource matching degree table are screened to obtain an available node list;

[0024] The available node list is subjected to multi-dimensional resource constraint analysis to obtain resource scheduling constraint conditions, and the resource scheduling constraint conditions are subjected to task allocation optimization to generate a task allocation strategy graph;

[0025] The available node list is subjected to task deployment planning based on the task allocation strategy graph to obtain the edge deployment task queue.

[0026] Further, the edge deployment task queue is used to analyze heterogeneous hardware resource mapping of the IPTV terminal to obtain device adaptation configuration parameters, including:

[0027] The edge deployment task queue is subjected to task feature fingerprint encoding to obtain a task feature fingerprint vector, and the task feature fingerprint vector is subjected to instruction semantic analysis to obtain an instruction-level computing density matrix;

[0028] The IPTV terminal is subjected to heterogeneous hardware resource marking based on the instruction-level computing density matrix to obtain a heterogeneous resource marking atlas, and an interface feature protocol in the IPTV terminal is extracted based on the heterogeneous resource marking atlas to obtain an interface protocol feature table;

[0029] Whether there is a resource mapping conflict in the interface protocol feature table is detected based on the instruction-level computing density matrix, if there is, a resource mapping conflict result is obtained, and the resource conflict marking sequence is subjected to topology path reconstruction based on the resource mapping conflict result to obtain a conflict-free resource path graph;

[0030] Quantify the task feature fingerprint vector based on the conflict-free resource path map to obtain a device affinity score table, and perform parameter register level configuration mapping on the device affinity score table to obtain device adaptation configuration parameters.

[0031] Further, the video decoder module of the IPTV terminal is subjected to decoding and rendering pipeline adaptation processing based on the device adaptation configuration parameters to generate local rendering instructions, including:

[0032] Perform decoding capability mapping analysis on the video decoder module in the IPTV terminal based on the device adaptation configuration parameters to obtain decoder hardware mapping capability.

[0033] Perform decoding pipeline segment division based on the decoder hardware mapping capability to obtain a multi-level decoding pipeline configuration descriptor.

[0034] Establish data path based on the multi-level decoding pipeline configuration descriptor on the rendering pipeline hardware interface in the IPTV terminal to obtain a decoding and rendering data path mapping diagram, and perform rendering buffer memory layout planning based on the decoding and rendering data path mapping diagram to obtain a rendering buffer memory allocation strategy.

[0035] Input the rendering buffer memory allocation strategy into the GPU rendering unit of the IPTV terminal to compile a shader program to obtain GPU shader program bytecode, and perform rendering instruction conversion processing based on the GPU shader program bytecode to generate local rendering instructions.

[0036] Further, the video frame buffer of the IPTV terminal is subjected to asynchronous rendering processing based on the local rendering instructions to obtain a smooth video playback output stream, including:

[0037] Perform frame-level rendering task decomposition on the local rendering instructions to obtain a frame rendering task sequence.

[0038] Perform frame rate synchronization control on the video output interface of the IPTV terminal based on the frame rendering task sequence to obtain a video frame rate synchronization control signal, and perform stream data transmission processing based on the video frame rate synchronization control signal to obtain a smooth video playback output stream.

[0039] The application also provides an IPTV terminal edge computing system, including:

[0040] An analysis module is configured to perform multi-dimensional time series analysis on user interaction behavior data of the IPTV terminal to obtain a user behavior pattern.

[0041] A matching module is configured to perform dynamic matching and context-aware processing on a content request sequence of the IPTV terminal based on the user behavior pattern, and obtain a content preference distribution matrix.

[0042] An optimization module is configured to perform task scheduling optimization on resource state information of a preset local edge node based on the content preference distribution matrix, and obtain an edge deployment task queue.

[0043] A mapping module is configured to perform heterogeneous hardware resource mapping analysis on the IPTV terminal based on the edge deployment task queue, and obtain device adaptation configuration parameters.

[0044] A decoding module is configured to perform decoding and rendering pipeline adaptation processing on a video decoder module of the IPTV terminal based on the device adaptation configuration parameters, and generate local rendering instructions.

[0045] A rendering module is configured to perform asynchronous rendering processing on a video frame buffer of the IPTV terminal based on the local rendering instructions, and obtain a smooth video playback output stream.

[0046] The application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method according to any one of the preceding embodiments when executing the computer program.

[0047] The application further provides a computer readable storage medium storing a computer program, wherein the computer program is executed by a processor to implement the steps of the method according to any one of the preceding embodiments.

[0048] The application provides an IPTV terminal edge computing method, which comprises the following steps: performing multi-dimensional time sequence analysis on user interactive behavior data of an IPTV terminal to obtain a user behavior pattern; performing dynamic matching and context-aware processing on a content request sequence of the IPTV terminal based on the user behavior pattern to obtain a content preference distribution matrix; performing task scheduling optimization on resource state information of a preset local edge node based on the content preference distribution matrix to obtain an edge deployment task queue; performing heterogeneous hardware resource mapping analysis on the IPTV terminal based on the edge deployment task queue to obtain device adaptation configuration parameters; performing decoding and rendering pipeline adaptation processing on a video decoder module of the IPTV terminal based on the device adaptation configuration parameters to generate local rendering instructions; and performing asynchronous rendering processing on a video frame buffer of the IPTV terminal based on the local rendering instructions to obtain a smooth video playback output stream. BRIEF DESCRIPTION OF DRAWINGS

[0049] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only constitute some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor based on these drawings.

[0050] Figure 1 is a step schematic diagram of an IPTV terminal edge computing method in an embodiment of the present application;

[0051] Figure 2 is a step schematic diagram of an IPTV terminal edge computing system in an embodiment of the present application;

[0052] Figure 3 is a structural schematic block diagram of a computer device in an embodiment of the present application.

[0053] The object implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0054] The embodiments of the present application will be described in detail below, and examples of the embodiments are shown in the drawings, wherein the same or similar notations represent the same or similar elements or elements with the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and are only used to explain the present application, and cannot be understood as a limitation of the present application. For the step numbers in the following embodiments, they are only set for the convenience of description and explanation, and the order between the steps is not limited in any way, and the execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0055] The IPTV terminal edge computing method according to the embodiments of the present application will be described in detail below with reference to the drawings. First, the IPTV terminal edge computing method according to the embodiments of the present application will be described with reference to the drawings.

[0056] Figure 1 is an IPTV terminal edge computing method in an embodiment of the present application, comprising the following steps:

[0057] Step S1, performing multi-dimensional time series analysis on user interaction behavior data of an IPTV terminal to obtain a user behavior pattern.

[0058] Step S2, performing dynamic matching and context awareness processing on a content request sequence of the IPTV terminal through the user behavior pattern to obtain a content preference distribution matrix.

[0059] Step S3, task scheduling optimization is performed on the preset resource state information of the local edge node based on the content preference distribution matrix, to obtain an edge deployment task queue.

[0060] Step S4, heterogeneous hardware resource mapping analysis is performed on the IPTV terminal based on the edge deployment task queue, to obtain device adaptation configuration parameters.

[0061] Step S5, decoding and rendering pipeline adaptation processing is performed on the video decoder module of the IPTV terminal based on the device adaptation configuration parameters, to generate local rendering instructions.

[0062] Step S6, asynchronous rendering processing is performed on the video frame buffer of the IPTV terminal based on the local rendering instructions, to obtain a smooth video playback output stream.

[0063] Specifically, the IPTV terminal edge computing method proposed by the application realizes efficient and personalized video services through a series of steps. First, in step S1, the system performs multi-dimensional time series analysis on the user interaction behavior data of the IPTV terminal to obtain user behavior patterns. This process may involve collecting user viewing history, click behavior, search records and other data, and using time series analysis techniques such as autoregressive integrated moving average model (ARIMA) or long short-term memory network (LSTM) to identify user viewing habits, preference periods and content interests and other patterns. For example, the system may find that a user tends to watch news and documentaries between 8pm and 10pm on weekdays, and prefers to watch movies and variety shows on weekends.

[0064] Next, in step S2, the system uses the user behavior patterns obtained in step S1 to perform dynamic matching and context-aware processing on the content request sequence of the IPTV terminal, to generate a content preference distribution matrix. This step may involve using machine learning algorithms such as collaborative filtering or content-based recommendation systems, combining real-time behavior and current context (such as time, location, device type, etc.) of the user, to dynamically adjust the recommendation strategy. The content preference distribution matrix may be a multi-dimensional array containing the user's preference for different types of content and the changes in these preferences over time and context. For example, the matrix may show that the user has a higher preference for sports events when using a smart TV on Friday night, and is more inclined to watch short video content when using a mobile device on a weekend morning.

[0065] Step S3 is based on the content preference distribution matrix generated in the previous step, the resource state information of the pre-set local edge node is optimized for task scheduling, and an edge deployment task queue is generated. This process may involve considering the computing power, storage capacity, network bandwidth and other resource states of the edge node, combining the user's content preferences, and through optimization algorithms (such as genetic algorithm or particle swarm optimization) to determine which content should be cached to the local edge node and how to allocate computing resources to process user requests. For example, the system may decide to pre-cache the news channel and popular movies that the user watches most frequently to the local edge node, while reserving sufficient computing resources for the upcoming peak viewing period.

[0066] In step S4, the system performs heterogeneous hardware resource mapping analysis on the IPTV terminal based on the edge deployment task queue generated in step S3, and obtains device adaptation configuration parameters. This step takes into account that different users may use different types of terminal devices (such as smart TVs, set-top boxes, mobile devices, etc.), each device has its specific hardware configuration and performance characteristics. The system needs to analyze the characteristics of these heterogeneous hardware, such as CPU performance, GPU capability, memory size, etc., and map each task in the task queue to the most suitable hardware resources. For example, for devices equipped with high-performance GPUs, the system may allocate more video decoding and rendering tasks; while for devices with lower performance, it may choose to rely more on edge nodes for preprocessing.

[0067] Step S5 uses the device adaptation configuration parameters obtained in step S4 to perform decoding and rendering pipeline adaptation processing on the video decoder module of the IPTV terminal, and generates local rendering instructions. This process may involve dynamically adjusting video coding parameters (such as bit rate, resolution, frame rate, etc.) according to the hardware capabilities of the device and network conditions, selecting the most suitable decoding algorithm and rendering method. For example, for high-performance devices, the system may choose to use a hardware-accelerated H.265 decoder and an OpenGL ES rendering pipeline; while for low-performance devices, it may choose a more lightweight decoder and software rendering method. The local rendering instructions may include specific decoding parameters, rendering pipeline configurations, buffer settings, etc.

[0068] Finally, in step S6, the system performs asynchronous rendering processing on the video frame buffer of the IPTV terminal based on the local rendering instructions generated in step S5, and finally obtains a smooth video playback output stream. Asynchronous rendering processing allows the system to start processing the next frame while a frame is still being displayed, thereby improving rendering efficiency and reducing stuttering. This step may involve using multi-threading technology to distribute tasks such as decoding, rendering, and display to different threads for parallel processing. At the same time, the system may implement an adaptive buffering strategy to dynamically adjust the buffer size according to network conditions and device performance to balance the relationship between low latency and smooth playback. For example, when network conditions are good, the system may reduce the buffer size to reduce latency; when the network is unstable, the buffer size may be increased to ensure smooth playback.

[0069] Through this series of steps, the IPTV terminal edge computing method of the present application realizes the whole process optimization from user behavior analysis to personalized content recommendation, to resource optimization scheduling and efficient video rendering, effectively improving the quality of IPTV service and user experience.

[0070] In specific embodiments, the multi-dimensional time series analysis of the user interaction behavior data of the IPTV terminal to obtain the user behavior pattern comprises:

[0071] The user interaction behavior data of the IPTV terminal is time-sliced based on a sliding window mechanism to obtain a plurality of time segment data, and each time segment data set is labeled with behavior event types to obtain labeled time segment data;

[0072] The different behavior event types in the labeled time segment data are assigned weights to obtain weighted time segment data, and feature vectors in the time dimension are extracted based on the weighted time segment data to obtain time feature vectors;

[0073] The time feature vectors are subjected to cluster analysis to obtain initial user behavior patterns, and the initial user behavior patterns are subjected to pattern merging and optimization processing to obtain user behavior patterns.

[0074] Specifically, in the IPTV terminal edge computing method, multi-dimensional time series analysis of user interaction behavior data is a key step to obtain user behavior patterns. This process first uses a sliding window mechanism to time-slice the user interaction behavior data of the IPTV terminal. The sliding window mechanism allows the system to create overlapping time slices in continuous time streams, which can capture the continuity and trend of user behavior. For example, a 24-hour window can be set, and the window slides every 1 hour, so that multiple overlapping time slice data can be obtained. For each time slice, the system will mark the behavior event type, which may include watching, pausing, fast forwarding, switching channels, and other types of user behavior. Next, the system assigns weights to different behavior event types in the marked time slice data. The purpose of this step is to reflect the importance of different behavior events and the indication of user preferences. For example, watching a complete program may be assigned a higher weight, while frequently switching channels may be assigned a lower weight. Through this weighting process, the system generates weighted time slice data, which more accurately reflects the user's real interests and behavior patterns. Based on these weighted time slice data, the system extracts feature vectors in the time dimension. This process may involve using statistical methods or machine learning techniques such as principal component analysis (PCA) or autoencoder to extract key features from time series, forming time feature vectors. These feature vectors may contain information about the periodicity, duration, and frequency of user behavior. To further analyze these time feature vectors, the system performs clustering analysis. The purpose of clustering analysis is to classify similar behavior patterns and form initial user behavior patterns. Here, algorithms such as K-means clustering, hierarchical clustering, or density clustering may be used. For example, the system may find that some users have fixed viewing habits in the evening on weekdays, while on weekends they exhibit different patterns. These initial user behavior patterns provide a basic outline of user behavior, but there may still be some subtle differences or repeated patterns. Therefore, the last step is to merge and optimize these initial user behavior patterns. This process may involve pattern similarity analysis, redundant pattern removal, low-frequency pattern filtering, and other operations. Through these processes, the system can obtain more refined and representative user behavior patterns. For example, the originally scattered "workday evening news watching" and "workday evening documentary watching" may be merged into the "workday evening information program watching" pattern. This entire process can be illustrated by a specific example: suppose there is an IPTV user, and the system collects his viewing data for a week. Using a 4-hour sliding window that slides every 1 hour, the system generates multiple time slice data. In each time slice, the system marks the user's behavior, such as "watching news", "pausing", "switching to movie channel", etc. Then, the system weights these behaviors, giving higher weights to "watching a complete program" and lower weights to "quickly switching channels".Next, the system extracts temporal feature vectors from these weighted data, which can include daily viewing peak times, average viewing duration, channel switching frequency, etc. Through clustering analysis, the system can find several initial patterns, such as "weekday evening news viewing", "weekend afternoon movie marathon", and "daily evening variety show". Finally, in the pattern merging and optimization phase, the system can merge "weekday evening news viewing" and "weekday evening documentary viewing" into "weekday evening information program viewing" because they show similar time features and content preferences. At the same time, some patterns with low frequency of occurrence, such as "watching cooking programs at 3am", may be filtered out because they may be accidental behavior rather than stable patterns. Through this series of complex data processing and analysis steps, the system finally obtains a set of refined behavior patterns that accurately reflect the user's habits and preferences. These patterns will provide important decision-making basis for subsequent content recommendation, resource scheduling and personalized services, thereby significantly improving the user experience and efficiency of IPTV services.

[0075] In specific embodiments, the dynamic matching and context-aware processing of the content request sequence of the IPTV terminal by the user behavior pattern obtains a content preference distribution matrix, including:

[0076] Each behavior pattern in the user behavior pattern is mapped to a semantic label to obtain a behavior semantic feature, and the content request sequence of the IPTV terminal is parsed for metadata to extract a program content feature vector;

[0077] Based on the semantic similarity calculation result between the behavior semantic feature and the program content feature vector, a behavior-content correlation matrix is constructed, and the behavior-content correlation matrix is context-weight modulated to obtain a context-enhanced preference matching matrix;

[0078] The content classification system in the pre-set local content library is dynamically labeled by the context-enhanced preference matching matrix to generate a content preference label sequence, and based on the content preference label sequence, a candidate set of program entries in the local content library is generated to obtain a content recommendation candidate set;

[0079] Based on the matching degree of the historical play heat of each program entry in the content recommendation candidate set and the user behavior pattern, the candidate program entries are prioritized to generate a content preference distribution matrix.

[0080] Specifically, in the IPTV terminal edge computing method, the dynamic matching of content request sequences with user behavior patterns and context-aware processing to obtain the content preference distribution matrix is a complex and delicate process. First, the system performs semantic label mapping on each behavior pattern in the user behavior pattern to obtain behavior semantic features. This step may involve using natural language processing techniques to convert user behavior into understandable semantic labels. For example, "long time watching on weekend evening" may be mapped to "leisure entertainment", "immersive experience", etc. At the same time, the system performs metadata analysis on the content request sequence of the IPTV terminal to extract program content feature vectors. These feature vectors may include program type, theme, actor, director, duration, and other dimensions of information. Next, the system calculates the semantic similarity between the behavior semantic features and the program content feature vectors. This step may use word embedding models (such as Word2Vec or BERT) to calculate semantic similarity, thereby constructing a behavior-content correlation matrix. For example, the behavior semantic feature of "leisure entertainment" may have a high similarity with movie, variety show, etc. Then, the system modulates the context weight of this behavior-content correlation matrix, considering context factors such as time, place, device type, etc., to obtain a context-enhanced preference matching matrix. This modulation can make the recommendation more accurate, for example, when using a mobile device in the morning on weekdays, news content may get a higher weight. Based on the obtained context-enhanced preference matching matrix, the system performs dynamic label assignment on the content classification system in the preset local content library, generating a content preference label sequence. This process can be understood as mapping the user's behavior preference to the classification system of the content library, thereby giving each content category a weight related to the user's preference. For example, if the user behavior pattern shows strong interest in technology programs, then the content of the technology category may get a higher preference label weight. Based on this content preference label sequence, the system further generates a candidate set of program entries in the local content library, obtaining a content recommendation candidate set. This candidate set contains a list of programs that are most likely to meet the user's preferences. Finally, the system needs to further optimize and sort the content recommendation candidate set. This step is based on the historical playback heat of each program entry in the content recommendation candidate set and the matching degree with the user behavior pattern. The historical playback heat may consider the total number of views, recent heat trends, etc., while the matching degree with the user behavior pattern may involve the viewing time, duration, etc. of the program. By integrating these factors, the system prioritizes the candidate program entries and finally generates the content preference distribution matrix. This process can be illustrated by a specific example: suppose there is an IPTV user whose behavior pattern shows that he often watches news and documentaries in the evening on weekdays, and likes to watch movies and sports events on weekends.The system first maps these behavior patterns into semantic tags such as "information acquisition", "knowledge learning", "leisure and entertainment", "sports hobby", etc. At the same time, the system analyzes the user's content request sequence and extracts content feature vectors such as "news", "documentary", "action movie", "soccer match", etc. By calculating the similarity between these semantic tags and content features, the system constructs a preliminary behavior-content correlation matrix. Then, considering that users usually use home smart TVs in the evening on weekdays, while they may use mobile devices more on weekends, the system performs context weight modulation on this matrix to obtain a context-enhanced preference matching matrix. Based on this matrix, the system performs dynamic tag assignment on the programs in the local content library. For example, evening news and international documentaries may have higher "weekday evening" tag weights, while action movies and sports live broadcasts may have higher "weekend" tag weights. The system generates a content preference tag sequence accordingly and filters out programs that meet these tags from the local content library to form a content recommendation candidate set. Finally, the system considers the historical playback heat of these candidate programs. For example, although the user likes to watch documentaries, if a popular news program has recently seen a surge in playback, the system may increase the program's ranking in the recommendation list. At the same time, the system also considers whether the broadcast time of the program matches the user's viewing habits. Through this comprehensive consideration, the system finally generates a content preference distribution matrix, which contains the recommendation priority of various programs in different time periods and on different devices. This detailed process ensures that the IPTV system can provide highly personalized and context-related content recommendations for users, greatly improving user experience and content consumption efficiency.

[0081] In specific embodiments, the task scheduling optimization of the preset local edge node resource state information based on the content preference distribution matrix obtains an edge deployment task queue, including:

[0082] Real-time monitoring of the resource state information of the preset local edge node obtains an edge node state monitoring table, and the resource availability in the edge node state monitoring table is evaluated to obtain a node resource state matrix;

[0083] Matching degree calculation of the content preference distribution matrix and the node resource state matrix obtains a task resource matching degree table, and the available edge nodes in the task resource matching degree table are screened to obtain an available node list;

[0084] Multi-dimensional resource constraint analysis of the available node list obtains resource scheduling constraint conditions, and task allocation optimization of the resource scheduling constraint conditions generates a task allocation strategy graph;

[0085] Task deployment planning of the available node list based on the task allocation strategy graph obtains an edge deployment task queue.

[0086] Specifically, in the IPTV terminal edge computing method, the dynamic matching of content request sequences with user behavior patterns and context-aware processing to obtain the content preference distribution matrix is a complex and delicate process. First, the system performs semantic label mapping on each behavior pattern in the user behavior pattern to obtain behavior semantic features. This step may involve using natural language processing techniques to convert user behavior into understandable semantic labels. For example, "long time watching on weekend evening" may be mapped to "leisure entertainment", "immersive experience", etc. At the same time, the system performs metadata analysis on the content request sequence of the IPTV terminal to extract program content feature vectors. These feature vectors may include program type, theme, actor, director, duration, and other dimensions of information. Next, the system calculates the semantic similarity between the behavior semantic features and the program content feature vectors. This step may use word embedding models (such as Word2Vec or BERT) to calculate semantic similarity, thereby constructing a behavior-content correlation matrix. For example, the behavior semantic feature of "leisure entertainment" may have a high similarity with movie, variety show, etc. Then, the system modulates the context weight of this behavior-content correlation matrix, considering context factors such as time, place, device type, etc., to obtain a context-enhanced preference matching matrix. This modulation can make the recommendation more accurate, for example, when using a mobile device in the morning on weekdays, news content may get a higher weight. Based on the obtained context-enhanced preference matching matrix, the system performs dynamic label assignment on the content classification system in the preset local content library, generating a content preference label sequence. This process can be understood as mapping the user's behavior preference to the classification system of the content library, thereby giving each content category a weight related to the user's preference. For example, if the user behavior pattern shows strong interest in technology programs, then the content of the technology category may get a higher preference label weight. Based on this content preference label sequence, the system further generates a candidate set of program entries in the local content library, obtaining a content recommendation candidate set. This candidate set contains a list of programs that are most likely to meet the user's preferences. Finally, the system needs to further optimize and sort the content recommendation candidate set. This step is based on the historical playback heat of each program entry in the content recommendation candidate set and the matching degree with the user behavior pattern. The historical playback heat may consider the total number of views, recent heat trends, etc., while the matching degree with the user behavior pattern may involve the viewing time, duration, etc. of the program. By integrating these factors, the system prioritizes the candidate program entries and finally generates the content preference distribution matrix. This process can be illustrated by a specific example: suppose there is an IPTV user whose behavior pattern shows that he often watches news and documentaries in the evening on weekdays, and likes to watch movies and sports events on weekends.The system first maps these behavior patterns into semantic tags such as "information acquisition", "knowledge learning", "leisure and entertainment", "sports hobby", etc. At the same time, the system analyzes the user's content request sequence and extracts content feature vectors such as "news", "documentary", "action movie", "soccer match", etc. By calculating the similarity between these semantic tags and content features, the system constructs a preliminary behavior-content correlation matrix. Then, considering that users usually use home smart TVs in the evening on weekdays, while they may use mobile devices more on weekends, the system performs context weight modulation on this matrix to obtain a context-enhanced preference matching matrix. Based on this matrix, the system performs dynamic tag assignment on the programs in the local content library. For example, evening news and international documentaries may have higher "weekday evening" tag weights, while action movies and sports live broadcasts may have higher "weekend" tag weights. The system generates a content preference tag sequence accordingly and filters out programs that meet these tags from the local content library to form a content recommendation candidate set. Finally, the system considers the historical playback heat of these candidate programs. For example, although the user likes to watch documentaries, if a popular news program has recently seen a surge in playback, the system may increase the program's ranking in the recommendation list. At the same time, the system also considers whether the broadcast time of the program matches the user's viewing habits. Through this comprehensive consideration, the system finally generates a content preference distribution matrix, which contains the recommendation priority of various programs in different time periods and on different devices. This detailed process ensures that the IPTV system can provide highly personalized and context-related content recommendations for users, greatly improving user experience and content consumption efficiency.

[0087] In specific embodiments, the IPTV terminal is analyzed for heterogeneous hardware resource mapping based on the edge deployment task queue, obtaining device adaptation configuration parameters, including:

[0088] The edge deployment task queue is encoded for task feature fingerprints, obtaining a task feature fingerprint vector, and the task feature fingerprint vector is analyzed for instruction semantics, obtaining an instruction-level computing density matrix;

[0089] The IPTV terminal is marked for heterogeneous hardware resources based on the instruction-level computing density matrix, obtaining a heterogeneous resource marker map, and the interface feature protocol in the IPTV terminal is extracted based on the heterogeneous resource marker map, obtaining an interface protocol feature table;

[0090] Based on the instruction-level computing density matrix, it is detected whether there is a resource mapping conflict in the interface protocol feature table. If there is, a resource mapping conflict result is obtained, and the resource conflict marker sequence is reconfigured for topology path based on the resource mapping conflict result, obtaining a conflict-free resource path graph;

[0091] The task feature fingerprint vector is quantified based on the conflict-free resource path map to obtain a device affinity score table, and the device affinity score table is mapped to a parameter register level configuration to obtain a device adaptation configuration parameter.

[0092] Specifically, in the IPTV terminal edge computing method, mapping heterogeneous hardware resources to the IPTV terminal based on the edge deployment task queue for device adaptation configuration parameter acquisition is a complex and delicate process. First, the system encodes the task feature fingerprints of the edge deployment task queue to obtain the task feature fingerprint vector. This step may involve encoding the features of each task (such as computational complexity, memory demand, I / O intensity, etc.) into a unique digital fingerprint. For example, a video decoding task may be encoded into a vector containing CPU usage, GPU demand, memory occupancy, etc. Then, the system performs instruction semantic analysis on these task feature fingerprint vectors to obtain the instruction-level computational density matrix. This process involves analyzing the computational requirements of each task at the instruction level, which may include the density distribution of different types of instructions such as floating-point operations, integer operations, and memory access. Based on the obtained instruction-level computational density matrix, the system labels the heterogeneous hardware resources of the IPTV terminal to obtain the heterogeneous resource labeling graph. This step involves identifying and labeling various hardware resources in the IPTV terminal, such as CPU cores, GPU units, dedicated video decoders, and memory modules. For example, a high-end IPTV set-top box may be labeled as having a multi-core CPU, a high-performance GPU, and a dedicated video processing unit. Based on this heterogeneous resource labeling graph, the system further extracts the interface feature protocols in the IPTV terminal to obtain the interface protocol feature table. This table may contain information such as communication protocols between hardware modules, bandwidth limitations, and delay characteristics. Next, the system detects whether there are resource mapping conflicts based on the instruction-level computational density matrix and the interface protocol feature table. This step aims to discover potential hardware resource competition problems. For example, the system may find that multiple computationally intensive tasks are competing for limited GPU resources, or that high-bandwidth demand tasks may exceed the transmission capacity of certain interfaces. If a conflict is detected, the system will obtain the resource mapping conflict result and, based on this result, reconstruct the topology path of the resource conflict marking sequence to obtain the conflict-free resource path graph. This process may involve task rescheduling, resource reallocation, or the use of time division multiplexing strategies to solve conflicts. Finally, based on the conflict-free resource path graph, the system quantifies the hardware affinity of the task feature fingerprint vector to obtain the device affinity score table. This step evaluates the matching degree of each task with available hardware resources, which may consider factors such as computational efficiency, energy consumption, and delay. For example, a video decoding task may have high affinity with a dedicated video processing unit, while a complex AI inference task may be more suitable for execution on a GPU. The system then maps the device affinity score table to the parameter register level configuration to finally obtain the device adaptation configuration parameters. These parameters may include specific hardware allocation schemes, clock frequency settings, power management strategies, etc.This process can be illustrated with a specific example: suppose there is a high-end IPTV set-top box whose edge deployment task queue contains tasks such as real-time 4K video decoding, AI-based content recommendation, user interface rendering, etc. The system first encodes these tasks into feature fingerprints to obtain task feature fingerprint vectors. For example, the 4K video decoding task may be encoded as a vector representing high video processing requirements, medium CPU requirements, and low memory requirements. Through instruction semantic analysis, the system generates an instruction-level computational density matrix, showing that the video decoding task mainly involves parallel matrix operations, while the AI recommendation task contains a large number of floating-point operations. Based on this matrix, the system labels the hardware resources of the IPTV set-top box, identifying that it has a quad-core CPU, a medium-performance GPU, and a dedicated 4K video decoder. The system also extracts the interface protocol characteristics between these hardware modules, such as the data transfer bandwidth between the CPU and the GPU, the input / output capabilities of the video decoder, etc. When detecting resource mapping conflicts, the system may find that the AI recommendation task and the user interface rendering task both require GPU resources, while the video decoding task requires exclusive use of the video decoder. To solve this conflict, the system reconfigures the resource allocation strategy, deciding to execute the AI recommendation task during CPU idle periods, while mainly using the GPU for user interface rendering, and ensuring that the video decoder can focus on 4K video processing. Based on this conflict-free resource path graph, the system quantifies hardware affinity. For example, the video decoding task has the highest affinity with the dedicated decoder, the AI recommendation task has higher affinity with the CPU, and the user interface rendering has the highest affinity with the GPU. Finally, the system converts these affinity scores into specific hardware configuration parameters, such as setting the video decoder to work in the highest performance mode, allocating two cores of the CPU to the AI recommendation task, and setting appropriate clock frequencies for the GPU to balance performance and power consumption. Through this meticulous process, the IPTV system can fully utilize the heterogeneous hardware resources of the terminal, find the optimal execution environment for different types of tasks, and significantly improve system performance and user experience. This dynamic resource mapping and optimization strategy enables the IPTV terminal to more efficiently handle complex multimedia tasks, providing users with smooth and high-quality services.

[0093] In specific embodiments, the decoding and rendering pipeline adaptation processing of the video decoder module in the IPTV terminal based on the device adaptation configuration parameters generates local rendering instructions, including:

[0094] Based on the device adaptation configuration parameters, the decoding capability mapping analysis of the video decoder module in the IPTV terminal is performed to obtain decoder hardware mapping capabilities;

[0095] Based on the decoder hardware mapping capabilities, decoding pipeline segment division is performed to obtain multi-level decoding pipeline configuration descriptors;

[0096] establishing data path of rendering pipeline hardware interface in IPTV terminal based on the multi-level decoding pipeline configuration descriptor, obtaining decoding rendering data path mapping diagram, and planning rendering buffer memory layout based on the decoding rendering data path mapping diagram, obtaining rendering buffer memory allocation strategy;

[0097] inputting the rendering buffer memory allocation strategy into GPU rendering unit of IPTV terminal to compile shader program, obtaining GPU shader program bytecode, and converting rendering instruction based on the GPU shader program bytecode to generate local rendering instruction.

[0098] Specifically, in the IPTV terminal edge computing method, the decoding and rendering pipeline adaptation processing of the video decoder module of the IPTV terminal based on the device adaptation configuration parameter to generate local rendering instructions is a complex and delicate process. First, the system performs decoding capability mapping analysis on the video decoder module in the IPTV terminal based on the device adaptation configuration parameter to obtain the decoder hardware mapping capability. This step involves evaluating the specific hardware characteristics of the decoder, such as supported video encoding formats (such as H.264, H.265, VP9, etc.), maximum resolution, frame rate capability, parallel decoding capability, etc. For example, for a high-end IPTV set-top box, its decoder may support 4K resolution H.265 video, a maximum frame rate of 60fps, and hardware-level HDR processing capability. Next, based on this decoder hardware mapping capability, the system performs decoding pipeline segment division to obtain a multi-stage decoding pipeline configuration descriptor. This process involves decomposing the entire decoding process into multiple stages, such as video bitstream parsing, macroblock decoding, motion compensation, and deblocking filtering, and determining the parallelism and processing unit allocation of each stage according to the hardware capability. For example, the system may decide to allocate the macroblock decoding stage to multiple parallel processing units to improve decoding efficiency. This multi-stage decoding pipeline configuration descriptor describes the hardware resource allocation, data flow, and processing timing of each decoding stage in detail. Based on the multi-stage decoding pipeline configuration descriptor, the system further establishes data paths for the rendering pipeline hardware interface in the IPTV terminal to obtain a decoding rendering data path mapping diagram. This step involves determining how to efficiently transmit decoded video data to the rendering pipeline, which may include configuring direct memory access (DMA) channels, setting up shared memory areas, etc. For example, the system may establish a direct data path from the video decoder output buffer to the GPU input buffer to minimize data transmission delay. Based on this decoding rendering data path mapping diagram, the system plans the memory layout of the rendering buffer to obtain a rendering buffer memory allocation strategy. This strategy determines how to efficiently allocate and manage various buffers, such as front buffers, back buffers, and depth buffers, in limited memory resources to support smooth video rendering and display. Finally, the system inputs the rendering buffer memory allocation strategy to the GPU rendering unit of the IPTV terminal for shader program compilation to obtain GPU shader program bytecode. This process involves compiling high-level shader programs (such as GLSL or HLSL written programs) into low-level instructions that can be directly executed by the GPU. The shader program may include video post-processing effects such as color correction, sharpening, HDR mapping, etc. For example, for an IPTV terminal that supports HDR, the system may compile a special HDR color mapping shader program. Based on this GPU shader program bytecode, the system performs rendering instruction conversion processing to finally generate local rendering instructions.These local rendering instructions are low-level commands that the GPU can execute directly, including texture loading, vertex processing, pixel shading, and other operations. This process can be illustrated with a specific example: suppose a user is watching a 4K HDR live content. The system first analyzes the decoder capabilities of the IPTV terminal and confirms that it supports 4K H.265 decoding and hardware HDR processing. Based on this capability, the system designs a five-stage decoding pipeline, including bitstream parsing, macroblock decoding, motion compensation, deblocking filtering, and HDR processing. The system allocates four parallel processing units for the macroblock decoding stage to meet the high data throughput requirements of 4K video. Next, the system establishes a zero-copy data path from the decoder output to the GPU input, allowing the GPU to directly access the decoded video frames without additional memory copy operations. In the rendering buffer memory layout, the system allocates sufficient front and back buffers for 4K resolution and reserves additional intermediate buffers for HDR processing. Finally, the system compiles a specialized HDR color mapping shader program to correctly map the HDR content to the display's color space. This shader program is compiled into GPU bytecode and then converted into a series of local rendering instructions. These instructions precisely control the color processing of each pixel, ensuring that the HDR content is presented optimally on the user's display device. Through this meticulous process, the IPTV system can fully utilize the terminal's decoding and rendering capabilities to provide users with high-quality, low-latency video experiences. This dynamic decoding and rendering pipeline adaptation enables the IPTV terminal to flexibly handle video content of different formats and qualities while maximizing hardware resource utilization, significantly improving system performance and user viewing experience.

[0099] In specific embodiments, the IPTV terminal's video frame buffer is asynchronously rendered based on the local rendering instructions to obtain a smooth video playback output stream, including:

[0100] Frame-level rendering task decomposition is performed on the local rendering instructions to obtain a frame rendering task sequence.

[0101] Frame rate synchronization control is performed on the IPTV terminal's video output interface based on the frame rendering task sequence to obtain a video frame rate synchronization control signal, and stream data transmission processing is performed based on the video frame rate synchronization control signal to obtain a smooth video playback output stream.

[0102] Specifically, in the IPTV terminal edge computing method, the asynchronous rendering of video frame buffers based on local rendering instructions to obtain a smooth video playback output stream is a key step. This process involves complex frame-level processing and synchronization control. First, the system decomposes the local rendering instructions into frame-level rendering tasks to obtain a sequence of frame rendering tasks. This step involves subdividing the overall rendering instructions into specific operations for each frame, which may include texture loading, geometric transformation, pixel shading, and a series of fine rendering tasks. For example, for a 4K HDR video stream, the rendering task for each frame may include loading the original video data, applying HDR color mapping, performing image post-processing (such as noise reduction or sharpening), and finally outputting to the display buffer. This frame rendering task sequence accurately describes the complete processing flow of each frame from raw data to final display. Next, based on this frame rendering task sequence, the system performs frame rate synchronization control on the video output interface of the IPTV terminal to obtain a video frame rate synchronization control signal. This step is crucial because it ensures that the rendering process is synchronized with the refresh rate of the display device, thereby avoiding picture tearing or stuttering. Frame rate synchronization control involves multiple aspects, including vertical synchronization (V-Sync), adaptive synchronization technology (such as FreeSync or G-Sync), and possibly frame rate multiplication or interpolation techniques. For example, if the video source is a 24fps movie content and the display is a 60Hz refresh rate, the system may use a 3:2 pull-down technique to match the two different frame rates, ensuring smooth and non-obvious jitter. Based on the generated video frame rate synchronization control signal, the system performs streaming data transmission processing to ultimately obtain a smooth video playback output stream. This process involves accurately transmitting the rendered frame data to the display device according to the rhythm of the synchronization signal. Streaming data transmission processing may involve multiple buffering techniques (such as triple buffering) to smooth the time difference between rendering and display, and may also include frame interpolation or motion compensation techniques to further improve the smoothness of the video. This process can be illustrated by a specific example: suppose a user is watching a 4K HDR movie on an IPTV terminal, and the original frame rate of the movie is 24fps, while the user's display supports a 60Hz refresh rate. The system first decomposes the local rendering instructions into specific tasks for each frame. For each frame, the task sequence may include: loading the original 4K HDR data from the decoding buffer, applying HDR to SDR color mapping (if the display does not support HDR), performing image enhancement processing (such as noise reduction and sharpening), and finally writing the processed frame data to the display buffer. In terms of frame rate synchronization control, the system detects the mismatch between the 24fps video source and the 60Hz display. To solve this problem, the system implements a 3:2 pull-down technique.This means that in every 5 display cycles, the system will show 3 frames twice and the other 2 frames once (2-3-2-3 pattern), thus smoothly mapping the 24fps content onto a 60Hz display. The frame rate synchronization control signals generated by the system precisely control this process, ensuring that each frame is displayed at the correct time point. During the streaming data transmission processing phase, the system uses a triple buffering technique to manage the rendering and display process. While the GPU is rendering a new frame, the display controller is reading another completed frame for display, and the third buffer serves as an intermediate buffer to store the next frame to be displayed. This method effectively reduces rendering latency and ensures the continuity of the display process. In addition, to further improve picture quality, the system may apply motion compensation techniques during the conversion from 24fps to 60Hz. This involves analyzing the motion information between adjacent frames and inserting computationally generated intermediate frames between the original frames, making the motion appear smoother and more natural. This advanced processing can significantly enhance the viewing experience of fast motion scenes, reducing motion blur and the sense of jitter. Through this series of meticulous processing steps, the IPTV system can convert the original 24fps movie content into a high-quality video stream that plays smoothly on a 60Hz display. Users can enjoy delicate, smooth 4K pictures with rich colors and natural motion, without noticeable stuttering or screen tearing. This highly optimized rendering and display process fully utilizes the hardware capabilities of the IPTV terminal, providing users with a near-cinema-level viewing experience, while also demonstrating the powerful potential of edge computing in improving multimedia processing and display quality.

[0103] The above describes an IPTV terminal edge computing method in an embodiment of the application. The following describes an IPTV terminal edge computing system in an embodiment of the application. Please refer to Figure 2 An embodiment of an IPTV terminal edge computing system in an embodiment of the application includes:

[0104] An analysis module 21 is configured to perform multi-dimensional time series analysis on user interaction behavior data of the IPTV terminal to obtain a user behavior pattern.

[0105] A matching module 22 is configured to perform dynamic matching and context awareness processing on a content request sequence of the IPTV terminal by using the user behavior pattern to obtain a content preference distribution matrix.

[0106] An optimization module 23 is configured to perform task scheduling optimization on preset resource state information of a local edge node based on the content preference distribution matrix to obtain an edge deployment task queue.

[0107] A mapping module 24 is configured to perform heterogeneous hardware resource mapping analysis on the IPTV terminal based on the edge deployment task queue to obtain device adaptation configuration parameters.

[0108] A decoding module 25 is configured to perform decoding and rendering pipeline adaptation processing on a video decoder module of the IPTV terminal based on the device adaptation configuration parameter, and generate local rendering instructions;

[0109] A rendering module 26 is configured to perform asynchronous rendering processing on a video frame buffer of the IPTV terminal based on the local rendering instructions, and obtain a smooth video playing output stream.

[0110] In the embodiment, the specific implementation of each module in the system embodiment is described above in the method embodiment, and will not be described here.

[0111] Referring to Figure 3 , the embodiment of the present application also provides a computer device, and the internal structure of the computer device can be as shown in Figure 3 . The computer device includes a processor, a memory, a display screen, an input device, a network interface and a database connected through a system bus. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium. The database of the computer device is configured to store corresponding data in the embodiment. The network interface of the computer device is configured to communicate with an external terminal through a network connection. The computer program is executed by the processor to implement the above method.

[0112] Those skilled in the art can understand Figure 3 that the structure shown in the embodiment is only a block diagram of part of the structure related to the present application scheme, and does not constitute a limitation on the computer device to which the present application scheme is applied.

[0113] The embodiment of the present application also provides a computer readable storage medium having a computer program stored thereon, and the computer program is executed by the processor to implement the above method. It can be understood that the computer readable storage medium in the embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0114] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiment methods. Any reference to memory, storage, database or other medium provided by the present application and used in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM, etc.

[0115] It should be noted that in this document, the terms "comprise", "comprising", or any other variant thereof are intended to cover non-exclusive inclusions, so that processes, devices, articles or methods that include a series of elements not only include those elements, but also include other elements not explicitly listed, or inherent to such processes, devices, articles or methods. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of other identical elements in the process, device, article or method that includes the element.

[0116] The above description is only the preferred embodiment of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation, or direct or indirect application in other related technical fields, based on the content of the present application specification and drawings, are also included in the patent protection scope of the present application.

Claims

1. An IPTV terminal edge computing method, characterized in that, The method comprises the following steps: Multi-dimensional time series analysis is performed on user interaction behavior data of the IPTV terminal to obtain a user behavior pattern; Dynamic matching and context-aware processing are performed on a content request sequence of the IPTV terminal based on the user behavior pattern to obtain a content preference distribution matrix; Task scheduling optimization is performed on preset resource state information of a local edge node based on the content preference distribution matrix to obtain an edge deployment task queue; Heterogeneous hardware resource mapping analysis is performed on the IPTV terminal based on the edge deployment task queue to obtain device adaptation configuration parameters; Decoding and rendering pipeline adaptation processing is performed on a video decoder module of the IPTV terminal based on the device adaptation configuration parameters to generate local rendering instructions; Asynchronous rendering processing is performed on a video frame buffer of the IPTV terminal based on the local rendering instructions to obtain a smooth video playback output stream; The heterogeneous hardware resource mapping analysis based on the edge deployment task queue and the device adaptation configuration parameters comprises: Task feature fingerprint encoding is performed on the edge deployment task queue to obtain a task feature fingerprint vector, and instruction semantic analysis is performed on the task feature fingerprint vector to obtain an instruction-level calculation density matrix; Heterogeneous hardware resource labeling is performed on the IPTV terminal based on the instruction-level calculation density matrix to obtain a heterogeneous resource labeling graph, and interface feature protocols in the IPTV terminal are extracted based on the heterogeneous resource labeling graph to obtain an interface protocol feature table; It is detected whether there is a resource mapping conflict in the interface protocol feature table based on the instruction-level calculation density matrix, and if there is, a resource mapping conflict result is obtained, and a topology path reconstruction is performed on the resource conflict labeling sequence based on the resource mapping conflict result to obtain a conflict-free resource path graph; Hardware affinity quantization is performed on the task feature fingerprint vector based on the conflict-free resource path graph to obtain a device affinity score table, and parameter register-level configuration mapping is performed on the device affinity score table to obtain the device adaptation configuration parameters.

2. The IPTV terminal edge computing method according to claim 1, characterized in that, The multi-dimensional time series analysis on the user interaction behavior data of the IPTV terminal to obtain the user behavior pattern comprises: Time slicing processing is performed on the user interaction behavior data of the IPTV terminal based on a sliding window mechanism to obtain a plurality of time slice data, and behavior event type labeling is performed on each time slice data set to obtain labeled time slice data; Weight assignment is performed on different behavior event types in the labeled time slice data to obtain weighted time slice data, and feature vector extraction is performed on the weighted time slice data in the time dimension to obtain a time feature vector; Cluster analysis is performed on the time feature vector to obtain an initial user behavior pattern, and mode merging and optimization processing are performed on the initial user behavior pattern to obtain the user behavior pattern.

3. The IPTV terminal edge computing method according to claim 1, characterized in that, The dynamic matching and context-aware processing on the content request sequence of the IPTV terminal based on the user behavior pattern to obtain the content preference distribution matrix comprises: The behavior semantic features are obtained by performing semantic label mapping on each of the user behavior patterns, and the IPTV terminal content request sequence is subjected to metadata analysis to extract a program content feature vector; A behavior-content correlation degree matrix is constructed based on a semantic similarity calculation result between the behavior semantic features and the program content feature vector, and context-enhanced preference matching matrix is obtained by context weight modulation on the behavior-content correlation degree matrix; A content preference label sequence is generated by dynamically assigning labels to a content classification system in a preset local content library through the context-enhanced preference matching matrix, and a program item candidate set in the local content library is generated based on the content preference label sequence to obtain a content recommendation candidate set; The program items in the candidate set are prioritized based on the matching degree between the historical playback heat of each program item in the content recommendation candidate set and the user behavior patterns to generate a content preference distribution matrix.

4. The IPTV terminal edge computing method of claim 1, wherein, The resource state information of the preset local edge node is subjected to task scheduling optimization based on the content preference distribution matrix to obtain an edge deployment task queue, including: The resource state information of the preset local edge node is monitored in real time to obtain an edge node state monitoring table, and the resource availability in the edge node state monitoring table is evaluated to obtain a node resource state matrix; The content preference distribution matrix and the node resource state matrix are subjected to matching degree calculation to obtain a task resource matching degree table, and the available edge nodes in the task resource matching degree table are screened to obtain an available node list; The available node list is subjected to multi-dimensional resource constraint analysis to obtain resource scheduling constraint conditions, and the task allocation strategy graph is generated by task allocation optimization on the resource scheduling constraint conditions; The available node list is subjected to task deployment planning based on the task allocation strategy graph to obtain an edge deployment task queue.

5. The IPTV terminal edge computing method according to claim 1, characterized in that, The video decoder module of the IPTV terminal is subjected to decoding and rendering pipeline adaptation processing based on the device adaptation configuration parameters to generate local rendering instructions, including: The decoding capability mapping analysis of the video decoder module in the IPTV terminal is performed based on the device adaptation configuration parameters to obtain a decoder hardware mapping capability; The decoding pipeline paragraph division is performed based on the decoder hardware mapping capability to obtain a multi-level decoding pipeline configuration descriptor; The data path establishment of the rendering pipeline hardware interface in the IPTV terminal is performed based on the multi-level decoding pipeline configuration descriptor to obtain a decoding rendering data path mapping graph, and the rendering buffer memory layout planning is performed based on the decoding rendering data path mapping graph to obtain a rendering buffer memory allocation strategy; The rendering buffer memory allocation strategy is input into the GPU rendering unit of the IPTV terminal for shader program compilation to obtain a GPU shader program bytecode, and the rendering instruction conversion processing is performed based on the GPU shader program bytecode to generate local rendering instructions.

6. The IPTV terminal edge computing method according to claim 1, characterized in that, The asynchronous rendering processing of the video frame buffer of the IPTV terminal based on the local rendering instruction obtains a smooth video playing output stream, and comprises: frame-level rendering task decomposition is performed on the local rendering instruction to obtain a frame rendering task sequence; frame rate synchronization control is performed on a video output interface of the IPTV terminal based on the frame rendering task sequence to obtain a video frame rate synchronization control signal, and streaming data transmission processing is performed based on the video frame rate synchronization control signal to obtain a smooth video playing output stream.

7. An IPTV terminal edge computing system, characterized by, Comprise: The analysis module is configured to perform multi-dimensional time sequence analysis on user interaction behavior data of the IPTV terminal to obtain a user behavior pattern. The matching module is configured to perform dynamic matching and context awareness processing on a content request sequence of the IPTV terminal based on the user behavior pattern to obtain a content preference distribution matrix. The optimization module is configured to perform task scheduling optimization on preset resource state information of a local edge node based on the content preference distribution matrix to obtain an edge deployment task queue. The mapping module is configured to perform heterogeneous hardware resource mapping analysis on the IPTV terminal based on the edge deployment task queue to obtain device adaptation configuration parameters. The decoding module is configured to perform decoding and rendering pipeline adaptation processing on a video decoder module of the IPTV terminal based on the device adaptation configuration parameters to generate a local rendering instruction. The rendering module is configured to perform asynchronous rendering processing on a video frame buffer of the IPTV terminal based on the local rendering instruction to obtain a smooth video playing output stream. The heterogeneous hardware resource mapping analysis on the IPTV terminal based on the edge deployment task queue to obtain device adaptation configuration parameters comprises: Task feature fingerprint encoding is performed on the edge deployment task queue to obtain a task feature fingerprint vector, and instruction semantic analysis is performed on the task feature fingerprint vector to obtain an instruction-level calculation density matrix. Heterogeneous hardware resource labeling is performed on the IPTV terminal based on the instruction-level calculation density matrix to obtain a heterogeneous resource labeling graph, and an interface feature protocol in the IPTV terminal is extracted based on the heterogeneous resource labeling graph to obtain an interface protocol feature table. It is detected whether there is a resource mapping conflict in the interface protocol feature table based on the instruction-level calculation density matrix, and if there is, a resource mapping conflict result is obtained, and a topology path reconstruction is performed on the resource conflict labeling sequence based on the resource mapping conflict result to obtain a conflict-free resource path graph. Hardware affinity quantization is performed on the task feature fingerprint vector based on the conflict-free resource path graph to obtain a device affinity score table, and parameter register-level configuration mapping is performed on the device affinity score table to obtain device adaptation configuration parameters.

8. A computer device comprising a memory and a processor, the memory having stored therein a computer program, characterized in that, The processor executes the computer program to implement the steps of the method of any one of claims 1 to 6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method of any one of claims 1 to 6.

Images