Multi-mode intelligent agent-based OTT large-screen advertisement intelligent monitoring system and multi-mode intelligent agent-based OTT large-screen advertisement intelligent monitoring method

Abstract

The invention relates to the technical field of advertisements, in particular to an OTT large-screen advertisement intelligent monitoring system and method based on a multi-modal agent, and the system comprises a time window analysis module which determines a verification time window based on a page browsing time sequence of an OTT platform in combination with terminal operation data; the operation offset compensation module is used for establishing an operation offset prediction model based on historical operation data and predicting an advertisement verification opportunity; the operation path planning module is used for constructing a menu tree structure of the OTT platform in real time, automatically identifying clickable areas and operation hotspots based on a visual understanding model, calculating an optimal execution path from a current position to a target advertisement position, and calculating verification starting time; the multi-dimensional verification execution module is used for carrying out multi-dimensional real-time advertisement verification on the same advertisement on the different types of equipment based on the verification starting time, comparing verification results of the different types of equipment and obtaining equipment difference data; and the verification strategy adjusting module is used for adjusting the advertisement verification strategy according to the advertisement putting plan and the equipment difference data.

Description

Technical Field

[0001] The present invention relates to the field of advertising technology, and specifically to an intelligent monitoring system and method for OTT large-screen advertising based on a multimodal intelligent agent. Background Art

[0002] With the rapid evolution of the digital marketing ecosystem, OTT (Over-The-Top) content platforms have become a key channel for targeted brand advertising. Faced with increasingly personalized user behavior and increasingly complex interface interaction structures, advertisers and delivery platforms are placing higher demands on intelligent monitoring, proactive verification, and strategic management of ad placement effectiveness.

[0003] Existing advertising monitoring mechanisms generally have the following technical bottlenecks: First, there is a lack of active monitoring capabilities based on platform status, and usually recording can only be performed after the advertisement is passively triggered, resulting in low monitoring efficiency; second, there is a lack of intelligent body collaboration mechanism, and it is impossible to dynamically adjust the monitoring strategy according to the interface hierarchy, content layout or user behavior of different platforms; third, existing solutions mostly rely on single-modal data for identification, which makes it difficult to accurately understand advertising content, brand logos or contextual semantics, resulting in insufficient monitoring accuracy; fourth, there is a lack of adaptability to changes in personalized recommendation logic and interactive interfaces in OTT platforms, making it difficult to support comprehensive supervision and optimization of advertising strategy effectiveness.

[0004] Therefore, how to build a collaborative mechanism between the HID intelligent agent and the multimodal large model to achieve intelligent navigation of the multi-level menu of the OTT platform and active triggering of advertising positions is an urgent problem to be solved.

[0005] To this end, an OTT large-screen advertising intelligent monitoring system and method based on multimodal intelligent agents are proposed. Summary of the Invention

[0006] The purpose of the present invention is to provide an intelligent monitoring system and method for OTT large-screen advertising based on a multimodal intelligent body, so as to realize intelligent navigation of the multi-level menu of the OTT platform and active triggering of advertising positions.

[0007] To achieve the above object, the present invention provides the following technical solutions: The OTT large-screen advertising intelligent monitoring system based on multimodal intelligent agents includes: The time window analysis module analyzes the exposure time interval of advertisements based on the page browsing sequence of the OTT platform and terminal operation data to determine the verification time window; The operation offset compensation module establishes an operation offset prediction model based on historical operation data, and predicts the timing of ad verification based on the current operation status and verification time window; The operation path planning module builds the menu tree structure of the OTT platform in real time, automatically identifies clickable areas and operation hotspots based on the visual understanding model, calculates the optimal execution path from the current location to the target ad position, and optimizes the ad verification timing; and calculates the verification start time based on the optimized ad verification timing; A multi-dimensional verification execution module performs multi-dimensional real-time advertising verification on the same advertisement on different models of devices based on the verification start time to obtain verification results; and compares the verification results of different models of devices to obtain device difference data; The verification strategy adjustment module intelligently adjusts the advertising verification strategy based on the advertising delivery plan and device difference data.

[0008] Preferably, the terminal operation data includes device CPU usage, memory occupancy, network delay time, video buffer status, platform response time and interface loading speed; The process of determining the verification time window includes: parsing the page browsing timing of the OTT platform, extracting the advertising display node and display duration information; calculating the device performance index and the network stability coefficient based on the terminal operation data; calculating the exposure time interval in which the advertisement is expected to appear according to the advertising display node and display duration information, combined with the device performance index; adjusting the start time and duration of the exposure time interval based on the network stability coefficient to obtain the verification time window.

[0009] Preferably, the operation offset prediction model includes: a historical operation data collection layer, a time series feature extraction layer, a multiple regression analysis layer and a real-time correction layer; the historical operation data collection layer collects historical operation data of different time periods and different platforms, including historical device CPU usage, historical memory occupancy, historical network delay time, historical video buffer status, historical platform response time and historical interface loading speed; the time series feature extraction layer extracts the periodic change law and trend characteristics of historical operation data; the multiple regression analysis layer establishes an operation offset prediction function based on device status, network quality and platform load to predict the operation offset; the real-time correction layer corrects the prediction result of the operation offset in real time according to the current operation status, and predicts the timing of advertising verification in combination with the verification time window.

[0010] Preferably, the menu tree structure includes: a main interface node of the OTT platform, an application classification node, a submenu node, an advertising position node, and an operation path edge; each node contains location coordinates, clickable status, and hierarchical relationship information; The process of obtaining the verification start time includes: analyzing the current interface status in real time through the visual understanding model to identify the location coordinates of clickable areas and operation hotspots; calculating the operation step sequence and path distance from the current position to the target advertising position based on the menu tree structure to obtain the optimal execution path; combining the terminal operation data to determine the total execution time of the optimal execution path; optimizing the advertising verification timing according to the advertising verification timing and the total execution time of the optimal execution path; based on the optimized advertising verification timing, reversely calculating the precise time point when the verification operation needs to be started as the verification start time.

[0011] Preferably, the visual understanding model includes: an interface element detection layer, which identifies the location and type of UI components; a clickable area analysis layer, which determines the interactive properties and operation priority of interface elements; a platform feature recognition layer, which identifies the interface style and layout features of different OTT platforms; and an operation hotspot positioning layer, which calculates the center coordinates and operation range of the clickable area.

[0012] Preferably, the process of obtaining the device difference data includes: starting advertisement verification based on the verification start time, performing image capture on the display interface of the same advertisement on different models of devices, calculating multi-dimensional verification parameters of the advertisement on different devices, and obtaining verification results; the multi-dimensional verification parameters include image quality parameters, position parameters, content integrity and advertisement duration; image quality parameters include image clarity and color saturation; position parameters include size ratio and position offset parameters; analyzing the difference values ​​of the multi-dimensional verification parameters between different devices, establishing a device performance difference matrix as the device difference data, and quantifying the display effect deviation between devices.

[0013] Preferably, the adjustment process of the advertising verification strategy includes: extracting strategy parameters based on the advertising delivery plan; identifying key difference factors affecting verification accuracy based on the device difference data, and generating a device performance grading table and a compensation coefficient matrix; assigning verification accuracy levels to advertisements of different priorities in combination with the strategy parameters and the device performance grading table; adjusting the verification thresholds of different devices according to the compensation coefficient matrix; and updating the advertising verification strategy based on the adjusted verification thresholds and accuracy levels.

[0014] Preferably, the OTT large-screen advertising intelligent monitoring method based on a multimodal intelligent agent includes: Based on the program timeline of the OTT platform and combined with terminal operation data, the preset exposure time interval of the advertisement is analyzed to determine the verification time window; Build an operation offset prediction model based on historical monitoring data, and predict the timing of ad verification according to the current operation status and verification time window; Build the menu tree structure of the OTT platform in real time, automatically identify clickable areas and operation hotspots based on the visual understanding model, calculate the optimal execution path from the current location to the target ad position, and optimize the ad verification timing; calculate the verification start time based on the optimized ad verification timing; Perform multi-dimensional real-time advertising verification on the same advertisement on different models of devices based on the verification start time to obtain verification results; compare the verification results of different models of devices to obtain device difference data; Intelligently adjust ad verification strategies based on ad delivery plans and device difference data.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. Through the collaborative work of the time window analysis module and the operation offset compensation module, the present invention can accurately predict the timing of advertisements and compensate for system operation offsets, significantly improving the accuracy of monitoring. Based on the timing analysis of page browsing on the OTT platform and combining terminal operation data to establish an operation offset prediction model, it is possible to predict the timing of advertisement verification in advance, avoiding the monitoring lag problem caused by traditional passive waiting. At the same time, through the extraction of time series features of historical operation data and multivariate regression analysis, an accurate prediction function is established, realizing intelligent compensation for factors such as network fluctuations and changes in equipment performance.

[0016] 2. The present invention achieves a technological leap from passive monitoring to active verification through the operation path planning module and the multi-dimensional verification execution module. The visual understanding model can analyze the interface status of the OTT platform in real time, automatically identify clickable areas and operation hotspots, dynamically build a menu tree structure, and calculate the optimal execution path from the current position to the target advertising position. This intelligent path planning not only improves operational efficiency, but also adapts to the interface differences of different OTT platforms and supports automated monitoring of mainstream OTT platforms. The multi-dimensional verification execution module can simultaneously monitor multiple dimensions such as the image quality parameters, location parameters, content integrity and playback duration of the advertisement to form a comprehensive verification system.

[0017] 3. The present invention effectively solves the display difference problem between different types of devices through the device difference data acquisition and verification strategy adjustment module, and establishes an intelligent adaptation mechanism. The system can automatically analyze the multi-dimensional verification parameter differences of the same advertisement on different devices, establish a device performance difference matrix, and quantify the display effect deviation between devices. Based on these difference data, the verification strategy adjustment module can dynamically adjust the verification threshold and judgment criteria for different devices to ensure the consistency and comparability of the monitoring results. At the same time, through the establishment of a compensation coefficient matrix, the system can automatically compensate for the performance differences of different devices, thereby improving the fairness and accuracy of monitoring. This cross-device compatibility mechanism enables the present invention to adapt to the mainstream smart TV brands and models on the market, support unified monitoring management of devices from different manufacturers, provide a technical basis for large-scale advertising monitoring, and effectively reduce system deployment and maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic diagram of the structure of an OTT large-screen advertising intelligent monitoring system based on a multimodal agent provided in an embodiment of the present invention; Figure 2 A schematic diagram of the structure of the operation deviation prediction model provided by an embodiment of the present invention; Figure 3 A schematic diagram of a process for obtaining a verification start time according to an embodiment of the present invention; Figure 4 A flow chart of a method for intelligent monitoring of OTT large-screen advertisements based on a multimodal intelligent agent provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0020] This invention proposes a multimodal agent-based intelligent monitoring system and method for OTT large-screen advertising, enabling intelligent navigation of multi-level menus on OTT platforms and active triggering of advertising spots. To illustrate the effectiveness of the method of the present invention in enabling intelligent navigation of multi-level menus on OTT platforms and active triggering of advertising spots, the following two examples illustrate the effectiveness of the invention.

[0021] Example 1: In the embodiments of the present application, a detailed description is given by taking the monitoring scenario of a sliding screen advertisement placed by an advertising company on a mainstream OTT platform as an example. The company needs to monitor in real time whether the sliding screen advertisements placed in the small window on the platform's program list page are accurately displayed in accordance with the contract requirements.

[0022] Figure 1 This is a specific structural diagram of the OTT large-screen advertising intelligent monitoring system based on multimodal intelligent agents of the present invention, including: time window analysis module, operation offset compensation module, operation path planning module, multi-dimensional verification execution module and verification strategy adjustment module. Figure 1 The following content is described: The time window analysis module analyzes the exposure time interval of advertisements based on the page browsing sequence of the OTT platform and terminal operation data to determine the verification time window; The terminal operation data includes device CPU usage, memory occupancy, network delay time, video buffer status, platform response time and interface loading speed; The process of determining the verification time window includes: parsing the page browsing timing of the OTT platform, extracting the advertising display node and display duration information; calculating the device performance index and the network stability coefficient based on the terminal operation data; calculating the exposure time interval in which the advertisement is expected to appear according to the advertising display node and display duration information, combined with the device performance index; adjusting the start time and duration of the exposure time interval based on the network stability coefficient to obtain the verification time window.

[0023] Specifically, "page browsing timing" is the time sequence of users' operations in the OTT platform interface, including page loading time, sliding operation time, dwell time, etc.; "exposure time interval" is the preset display time period for the advertisement, which is usually determined by the advertising delivery plan; "verification time window" is the time range for the system to perform advertising verification and monitoring operations, which needs to cover the complete display cycle of the advertisement.

[0024] The terminal operation data is collected in real time through the system monitoring API.

[0025] The process of determining the verification time window includes: The system analyzes the page's DOM structure to identify the appearance and duration of ad container elements. For example, on a program guide page on an OTT platform, the system identifies a sliding screen ad container with the ID "ad-banner" and a preset display duration of 5 seconds. The device performance index and network stability coefficient are calculated based on terminal operation data. The device performance index is calculated using a weighted average: device performance index = 0.3 × (100 - CPU usage) + 0.3 × (100 - memory usage) + 0.2 × interface loading speed + 0.2 × platform response time. The network stability coefficient is calculated based on the variance of network latency: stability coefficient = 1 / (1 + network latency variance). Based on the ad display node and display duration information, combined with the device performance index, the system calculates the expected display time range for the ad. For example, a sliding screen ad is scheduled to appear between 3 and 8 seconds after the page loads. Given a current device performance index of 0.8, the system calculates the actual display time range to be 3.5 to 8.5 seconds.

[0026] The start and duration of the exposure time interval are adjusted based on the network stability coefficient to obtain the verification time window. When the network stability coefficient is 0.9, the system sets the verification time window to 3.0-9.0 seconds to ensure that it covers the possible impact of network fluctuations.

[0027] By comprehensively considering platform characteristics and device status, the system accurately predicts when ads will appear, avoiding the omissions caused by traditional fixed-time monitoring and improving monitoring accuracy. By incorporating terminal operating data such as CPU, memory, and network latency to calculate device performance and network stability, the verification window is dynamic and context-aware. Instead of blindly checking at preset time points, the system can now perceive the current terminal load and network status, dynamically adjusting the expected ad exposure window. This increases the probability of capturing ads on devices with weak networks or low performance, improving the initial verification success rate.

[0028] Further, if Figure 2 As shown, the operation offset compensation module establishes an operation offset prediction model based on historical operation data, and predicts the timing of advertisement verification according to the current operation status and verification time window; "operation offset" refers to the deviation between the actual operation status of the system and the expected status, including operation delay, processing delay, transmission delay, etc.; "verification timing" refers to the optimal time point for the system to start the verification operation.

[0029] The operation offset prediction model includes: a historical operation data collection layer, a time series feature extraction layer, a multiple regression analysis layer and a real-time correction layer; the historical operation data collection layer collects historical operation data of different time periods and different platforms, including historical device CPU usage, historical memory occupancy, historical network delay time, historical video buffer status, historical platform response time and historical interface loading speed; the time series feature extraction layer extracts the periodic change law and trend characteristics of historical operation data; the multiple regression analysis layer establishes an operation offset prediction function based on device status, network quality and platform load to predict operation offset; the real-time correction layer corrects the prediction result of operation offset in real time according to the current operation status, and predicts the timing of advertisement verification in combination with the verification time window.

[0030] Specifically, the historical operation data collection layer maintains a time series database containing 30 days of historical data, with a data sampling frequency of once per second; The temporal feature extraction layer uses time series analysis to identify differences in usage patterns between weekdays and weekends, and between daytime and nighttime. For example, it found that 7-10 p.m. is the peak network period, with latency increasing by approximately 15%. The multivariate regression analysis layer uses a multivariate linear regression model: prediction offset = α × CPU usage + β × memory usage + γ × network latency + δ × platform load + ε, where α, β, γ, and δ are coefficients obtained through historical data training; ε is the bias term. The real-time correction layer dynamically adjusts the prediction results of the running offset through the Kalman filter algorithm; based on the dynamically adjusted running offset and the verification time window, the timing of advertising verification is predicted.

[0031] By learning from historical data and implementing real-time corrections, we effectively compensate for system operational deviations, reduce verification timing prediction errors, and improve the timeliness of monitoring. By establishing a predictive model that incorporates historical data collection and multivariate regression analysis, we achieve "predictive compensation" for ad verification timing. This not only considers the current operating status but also predicts inherent, periodic system operational delays based on historical data patterns. This upgrades verification timing prediction from "passive adaptation" to "active prediction," improving the accuracy of verification execution at the precise moment of ad exposure.

[0032] Furthermore, the operation path planning module constructs the menu tree structure of the OTT platform in real time, automatically identifies clickable areas and operation hotspots based on the visual understanding model, calculates the optimal execution path from the current position to the target ad position, and optimizes the ad verification timing; and calculates the verification start time based on the optimized ad verification timing; The menu tree structure includes: the main interface node of the OTT platform, application classification nodes, submenu nodes, advertising position nodes and operation path edges; each node contains location coordinates, clickable status and hierarchical relationship information; taking a certain OTT platform as an example, the main interface node contains first-level menus such as "Home", "TV Series", and "Movies", and the "TV Series" node contains second-level submenus such as "Popular" and "New Dramas", and the advertising position node is located in the recommendation area on the right side of the program list page.

[0033] The flowchart of the process of obtaining the verification start time is as follows: Figure 3 Shown, including: The visual understanding model analyzes the current interface state in real time and identifies the location coordinates of clickable areas and operation hotspots. For example, on the homepage of an OTT platform, the "TV Series" button is identified at coordinates (200, 150), with a clickable area of ​​150×50 pixels. The visual understanding model includes: an interface element detection layer, which uses the YOLO object detection algorithm to identify the location and type of UI components, such as buttons, menu items, text boxes, and other interface elements; a clickable area analysis layer, which determines the interactive properties and operation priority of interface elements and determines clickability by analyzing the visual features (color, borders, shadows, etc.) and semantic information of the elements; a platform feature recognition layer, which identifies the interface styles and layout features of different OTT platforms, establishes a platform feature library, and supports automatic adaptation of mainstream OTT platforms; and an operation hotspot positioning layer, which calculates the center coordinates and operation range of the clickable area and uses pixel-level precision positioning to ensure the success rate of operations.

[0034] Based on the menu tree structure, the sequence of operation steps and the path distance from the current position to the target ad position are calculated to obtain the optimal execution path; for example, the path from the homepage to the ad position is: homepage → TV series → popular → program list → ad position, which requires a total of 4 steps.

[0035] Combined with terminal operation data, the total execution time of the optimal execution path is determined; the operation time of each step includes click time (50ms), interface response time (dynamically adjusted according to the platform response time), and page loading time (calculated according to network conditions), totaling about 2.5 seconds.

[0036] The ad verification timing is optimized based on the ad verification timing and the total execution time of the optimal execution path; if the ad is predicted to appear in 8 seconds and the path execution takes 2.5 seconds, the optimized verification timing is adjusted to start execution at 5.5 seconds.

[0037] Based on the optimized ad verification timing, the precise time point at which the verification operation needs to be started is reversely calculated as the verification start time; considering the system processing delay of 100ms, the verification start time is finally determined to be 5.4 seconds.

[0038] By constructing a menu tree and calculating the optimal path, we achieve a structured understanding and efficient navigation of complex OTT UIs. This adapts to the interface differences across different platforms, enabling the agent to reliably locate target ad placements in dynamically changing interfaces with varying layouts. Furthermore, by calculating the path duration to infer the launch time, we ensure on-time delivery of operations, improving both the success rate of operations and the automation of broadcast monitoring.

[0039] Furthermore, the multi-dimensional verification execution module performs multi-dimensional real-time advertising verification on the same advertisement on different models of devices based on the verification start time to obtain verification results; and compares the verification results of different models of devices to obtain device difference data; The process of obtaining the device difference data includes: Based on the verification start time, advertisement verification is started, and images of the display interfaces of the same advertisement on different models of devices are captured; the system simultaneously controls smart TVs of different brands and models to capture advertisement display images at the same time point; Calculate multi-dimensional ad verification parameters on different devices to obtain verification results; these multi-dimensional verification parameters include image quality, position, content integrity, and ad duration. Image quality parameters include clarity and color saturation. Clarity is calculated using the SSIM (Structural Similarity) algorithm, with a value ranging from 0 to 1, where values ​​closer to 1 indicate clearer image quality. Color saturation is calculated using the saturation component of the HSV color space. Position parameters include size ratio and position offset. The size ratio is calculated by calculating the ratio of the ad area to the total screen area, while the position offset is calculated by calculating the Euclidean distance between the ad center point and a preset position. Content integrity is tested using optical character recognition (OCR) and image feature matching to determine the integrity of ad text, logos, and image elements. The integrity score ranges from 0 to 1. Ad duration is accurately measured through video frame analysis to accurately measure the actual display duration of the ad.

[0040] Analyze the differences in the multi-dimensional verification parameters between different devices and establish a device performance difference matrix, which serves as the device difference data and quantifies the display effect deviations between devices. For example, the picture clarity of brand A TV is 0.95, brand B TV is 0.88, and brand C TV is 0.92. The difference matrix is ​​used for subsequent compensation calculations.

[0041] By establishing a standardized, multi-dimensional advertising display quality evaluation system to conduct multi-dimensional monitoring and cross-device comparison of advertisements, the final "device performance difference matrix" will transform the vague problem of "poor display effect" into measurable, traceable and comparable structured data, comprehensively evaluate the effectiveness of advertising, provide data support for accurate monitoring, and provide an objective basis for advertising optimization.

[0042] Furthermore, the verification strategy adjustment module intelligently adjusts the advertisement verification strategy based on the advertisement delivery plan and device difference data. The advertisement verification strategy adjustment process includes: Extract strategic parameters based on the advertising delivery plan, including advertising priority (high, medium, and low), delivery frequency (hourly / daily), timeliness requirements (real-time / delayed), budget level, etc.

[0043] Based on the equipment difference data, the key difference factors that affect the verification accuracy are identified, and an equipment performance grading table and a compensation coefficient matrix are generated; the equipment is divided into three levels according to performance: A (high-end), B (mid-end), and C (low-end), and the compensation coefficients of each level of equipment are calculated.

[0044] Based on the policy parameters and the device performance grading table, verification accuracy levels are assigned to advertisements of different priorities; high-priority advertisements are assigned high-precision monitoring (error <1%), medium-priority advertisements are assigned standard accuracy (error <3%), and low-priority advertisements are assigned basic accuracy monitoring (error <5%).

[0045] The verification thresholds of different devices are adjusted according to the compensation coefficient matrix; for example, the image quality threshold of a Class C device is adjusted from 0.9 to 0.8, and the position offset threshold is adjusted from 5 pixels to 8 pixels.

[0046] Update the ad verification policy based on the adjusted verification threshold and accuracy level.

[0047] By proactively leveraging device variance data, the system can adjust its verification strategies, such as setting looser image quality thresholds for poorly performing devices or assigning stricter verification standards to high-priority ads. This allows the system to continuously evolve during use, enabling intelligent strategy adjustment and personalized configuration, improving the adaptability and accuracy of monitoring and reducing false positives.

[0048] Through the collaborative work of five major modules, automated ad verification with high precision, high robustness, and adaptive optimization capabilities is achieved. First, by combining real-time terminal performance and historical data models, it accurately predicts the precise timing of ad exposure, solving the problem of verification failure caused by device delays and network fluctuations in traditional monitoring, while also improving monitoring efficiency. Second, through visual understanding and path planning capabilities, the system is able to operate complex and changing OTT interfaces, effectively overcoming the problem of script failure caused by UI changes. On this basis, the system can not only confirm the presence of advertisements, but also conduct multi-dimensional quantitative evaluations and discover display differences across devices. Finally, it can use this difference data for self-learning and intelligently adjust subsequent verification strategies, making the entire monitoring system increasingly accurate and efficient as its operating time increases.

[0049] Example 2: In Example 1, the method proposed in the present invention successfully achieved intelligent navigation of the multi-level menu of the OTT platform and active triggering of advertising positions. To further verify the effectiveness of the present invention, an intelligent monitoring method for OTT large-screen advertising based on a multimodal intelligent agent is also proposed in the present embodiment to monitor the advertising placed on the OTT platform by another advertising company. Figure 4 , Figure 4 Detailed flow chart of the method of the present invention.

[0050] Based on the page browsing sequence of the OTT platform, combined with terminal operation data, the advertising exposure time interval is analyzed to determine the verification time window; the terminal operation data includes device CPU usage, memory usage, network latency, video buffer status, platform response time and interface loading speed; The process of determining the verification time window includes: parsing the page browsing timing of the OTT platform, extracting the advertising display node and display duration information; calculating the device performance index and the network stability coefficient based on the terminal operation data; calculating the exposure time interval in which the advertisement is expected to appear according to the advertising display node and display duration information, combined with the device performance index; adjusting the start time and duration of the exposure time interval based on the network stability coefficient to obtain the verification time window.

[0051] Furthermore, an operation offset prediction model is established based on historical operation data to predict the timing of ad verification according to the current operation status and verification time window; The operation offset prediction model includes: a historical operation data collection layer, a time series feature extraction layer, a multiple regression analysis layer and a real-time correction layer; the historical operation data collection layer collects historical operation data of different time periods and different platforms, including historical device CPU usage, historical memory occupancy, historical network delay time, historical video buffer status, historical platform response time and historical interface loading speed; the time series feature extraction layer extracts the periodic change law and trend characteristics of historical operation data; the multiple regression analysis layer establishes an operation offset prediction function based on device status, network quality and platform load to predict operation offset; the real-time correction layer corrects the prediction result of operation offset in real time according to the current operation status, and predicts the timing of advertisement verification in combination with the verification time window.

[0052] The real-time correction layer further includes an abnormal state perception unit, which monitors in real time whether the current terminal operation data has sudden abnormalities that deviate significantly from the historical time series characteristics (such as system pop-ups, instantaneous network congestion); when an abnormality is detected, the operation offset compensation module will temporarily increase the predicted value of the operation offset, and send instructions to the operation path planning module to re-plan or delay execution. This increases the system's ability to cope with emergencies and enhances the system's operational robustness and risk avoidance capabilities in real and complex environments. Through abnormal state perception, the system can handle emergencies that historical models cannot predict (such as system pop-ups, precursors to APP crashes), and immediately take avoidance measures (such as delayed verification, re-planning). This enables the system to work stably in chaotic reality from a prediction model that operates in an ideal environment.

[0053] Furthermore, the menu tree structure of the OTT platform is constructed in real time, clickable areas and operation hotspots are automatically identified based on the visual understanding model, the optimal execution path from the current position to the target ad position is calculated, and the ad verification timing is optimized; the verification start time is calculated based on the optimized ad verification timing; The menu tree structure includes: the main interface node of the OTT platform, application classification node, submenu node, advertising node and operation path edge; each node contains location coordinates, clickable status and hierarchical relationship information; The process of obtaining the verification start time includes: analyzing the current interface status in real time through the visual understanding model to identify the location coordinates of clickable areas and operation hotspots; calculating the operation step sequence and path distance from the current position to the target advertising position based on the menu tree structure to obtain the optimal execution path; combining the terminal operation data to determine the total execution time of the optimal execution path; optimizing the advertising verification timing according to the advertising verification timing and the total execution time of the optimal execution path; based on the optimized advertising verification timing, reversely calculating the precise time point when the verification operation needs to be started as the verification start time.

[0054] The visual understanding model includes: an interface element detection layer, which identifies the location and type of UI components; a clickable area analysis layer, which determines the interactive properties and operation priority of interface elements; a platform feature recognition layer, which identifies the interface styles and layout features of different OTT platforms; and an operation hotspot positioning layer, which calculates the center coordinates and operation range of the clickable area.

[0055] Furthermore, based on the verification start time, multi-dimensional real-time advertising verification is performed on the same advertisement on different models of devices to obtain verification results; the verification results of different models of devices are compared to obtain device difference data; The process of obtaining the device difference data includes: starting advertisement verification based on the verification start time, performing image capture on the display interface of the same advertisement on different models of devices, calculating multi-dimensional verification parameters of the advertisement on different devices, and obtaining verification results; the multi-dimensional verification parameters include image quality parameters, position parameters, content integrity and advertisement duration; image quality parameters include image clarity and color saturation; position parameters include size ratio and position offset parameters; analyzing the difference values ​​of the multi-dimensional verification parameters between different devices, establishing a device performance difference matrix as the device difference data, and quantifying the display effect deviation between devices.

[0056] The device performance difference matrix is ​​a multi-dimensional tensor structure; in addition to the device model dimension, it also includes the OTT application version dimension and the operating system version dimension; the system can analyze the changes in advertising display effects caused by application or system updates on the same device, thereby deepening the adjustment accuracy of the verification strategy from the device level to the application and system version level. The granularity of strategy optimization is deepened from "device" to "advertising material" itself, realizing a refined verification strategy. The system can identify performance issues of specific types of advertising creatives on specific devices and perform targeted dynamic strategy weighting. This makes verification and optimization no longer a one-size-fits-all approach, but can accurately locate and solve specific problems caused by the incompatibility of advertising materials and devices.

[0057] Furthermore, the advertising verification strategy is intelligently adjusted based on the advertising delivery plan and device difference data.

[0058] The adjustment process of the advertising verification strategy includes: extracting strategy parameters based on the advertising delivery plan; identifying key difference factors affecting verification accuracy based on the device difference data, and generating a device performance grading table and a compensation coefficient matrix; combining the strategy parameters and the device performance grading table to assign verification accuracy levels to advertisements of different priorities; adjusting the verification thresholds of different devices according to the compensation coefficient matrix; and updating the advertising verification strategy based on the adjusted verification thresholds and accuracy levels.

[0059] The adjustment process of the ad verification strategy also includes feature analysis of the ad creative itself. The system extracts the material features of the ad to be verified (such as video ads, image ads, and the unique identifier of the ad in the creative library); when different devices have significantly different display effects on specific types of ad creatives (for example, a specific video encoding is prone to screen distortion on a certain model of device), the compensation coefficient matrix will dynamically weight the ad creative features and device model combination to achieve more targeted strategy adjustments for specific problems. It provides in-depth problem attribution capabilities at the "version level". By extending the difference matrix to application versions and system versions, when there is a problem with ad display, the system can not only locate which device it is, but also accurately locate whether the problem occurred after the device updated a certain App or OS version. This provides extremely valuable and highly actionable intelligence for platform developers to fix bugs, shortening the problem investigation and repair cycle.

[0060] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. The OTT large-screen advertising intelligent monitoring system based on multimodal intelligent agents is characterized by: include: The time window analysis module analyzes the exposure time interval of advertisements based on the page browsing sequence of the OTT platform and terminal operation data to determine the verification time window; The operation offset compensation module establishes an operation offset prediction model based on historical operation data, and predicts the timing of ad verification based on the current operation status and verification time window; The operation path planning module builds the menu tree structure of the OTT platform in real time, automatically identifies clickable areas and operation hotspots based on the visual understanding model, calculates the optimal execution path from the current location to the target ad position, and optimizes the timing of ad verification; Calculate verification startup time based on optimized ad verification timing; A multi-dimensional verification execution module performs multi-dimensional real-time advertising verification on the same advertisement on different models of devices based on the verification start time to obtain verification results; and compares the verification results of different models of devices to obtain device difference data; The verification strategy adjustment module intelligently adjusts the advertising verification strategy based on the advertising delivery plan and device difference data.

2. The OTT large-screen advertising intelligent monitoring system based on a multimodal agent according to claim 1 is characterized by: The terminal operation data includes device CPU usage, memory occupancy, network delay time, video buffer status, platform response time and interface loading speed; The process of determining the verification time window includes: parsing the page browsing timing of the OTT platform, extracting the advertising display node and display duration information; calculating the device performance index and the network stability coefficient based on the terminal operation data; calculating the exposure time interval in which the advertisement is expected to appear according to the advertising display node and display duration information, combined with the device performance index; adjusting the start time and duration of the exposure time interval based on the network stability coefficient to obtain the verification time window.

3. The OTT large-screen advertising intelligent monitoring system based on a multimodal agent according to claim 1 is characterized by: The operation offset prediction model includes: a historical operation data collection layer, a time series feature extraction layer, a multiple regression analysis layer and a real-time correction layer; the historical operation data collection layer collects historical operation data of different time periods and different platforms, including historical device CPU usage, historical memory occupancy, historical network delay time, historical video buffer status, historical platform response time and historical interface loading speed; the time series feature extraction layer extracts the periodic change law and trend characteristics of historical operation data; the multiple regression analysis layer establishes an operation offset prediction function based on device status, network quality and platform load to predict operation offset; the real-time correction layer corrects the prediction result of operation offset in real time according to the current operation status, and predicts the timing of advertisement verification in combination with the verification time window.

4. The OTT large-screen advertising intelligent monitoring system based on a multimodal agent according to claim 1 is characterized by: The menu tree structure includes: the main interface node of the OTT platform, application classification node, submenu node, advertising node and operation path edge; each node contains location coordinates, clickable status and hierarchical relationship information; The process of obtaining the verification start time includes: analyzing the current interface status in real time through the visual understanding model to identify the location coordinates of clickable areas and operation hotspots; calculating the operation step sequence and path distance from the current position to the target advertising position based on the menu tree structure to obtain the optimal execution path; combining the terminal operation data to determine the total execution time of the optimal execution path; optimizing the advertising verification timing according to the advertising verification timing and the total execution time of the optimal execution path; based on the optimized advertising verification timing, reversely calculating the precise time point when the verification operation needs to be started as the verification start time.

5. The OTT large-screen advertising intelligent monitoring system based on a multimodal agent according to claim 4 is characterized by: The visual understanding model includes: an interface element detection layer, which identifies the location and type of UI components; a clickable area analysis layer, which determines the interactive properties and operation priority of interface elements; a platform feature recognition layer, which identifies the interface styles and layout features of different OTT platforms; and an operation hotspot positioning layer, which calculates the center coordinates and operation range of the clickable area.

6. The OTT large-screen advertising intelligent monitoring system based on a multimodal agent according to claim 1 is characterized by: The process of obtaining the device difference data includes: starting advertisement verification based on the verification start time, performing image capture on the display interface of the same advertisement on different models of devices, calculating the multi-dimensional verification parameters of the advertisement on different devices, and obtaining the verification results; the multi-dimensional verification parameters include image quality parameters, position parameters, content integrity and advertisement duration; image quality parameters include image clarity and color saturation; position parameters include size ratio and position offset parameters; analyzing the difference values ​​of the multi-dimensional verification parameters between different devices, and establishing a device performance difference matrix as the device difference data.

7. The OTT large-screen advertising intelligent monitoring system based on a multimodal agent according to claim 1 is characterized by: The adjustment process of the advertising verification strategy includes: extracting strategy parameters based on the advertising delivery plan; identifying key difference factors affecting verification accuracy based on the device difference data, and generating a device performance grading table and a compensation coefficient matrix; combining the strategy parameters and the device performance grading table to assign verification accuracy levels to advertisements of different priorities; adjusting the verification thresholds of different devices according to the compensation coefficient matrix; and updating the advertising verification strategy based on the adjusted verification thresholds and accuracy levels.

8. An intelligent monitoring method for OTT large-screen advertising based on a multimodal agent, characterized in that: include: Based on the page browsing sequence of the OTT platform and combined with terminal operation data, the advertising exposure time interval is analyzed to determine the verification time window; Build an operation offset prediction model based on historical operation data, and predict the timing of ad verification according to the current operation status and verification time window; Build the menu tree structure of the OTT platform in real time, automatically identify clickable areas and operation hotspots based on the visual understanding model, calculate the optimal execution path from the current location to the target ad position, and optimize the timing of ad verification; Calculate verification startup time based on optimized ad verification timing; Perform multi-dimensional real-time advertising verification on the same advertisement on different models of devices based on the verification start time to obtain verification results; compare the verification results of different models of devices to obtain device difference data; Intelligently adjust ad verification strategies based on ad delivery plans and device difference data.

Images