Multi-screen cooperative control method and system for LED display screen

Abstract

The invention relates to the technical field of LED display control, in particular to an LED display screen multi-screen cooperative control method and system. According to the method, multi-dimensional data acquisition is carried out on the display screen, multi-source data support is provided for subsequent dynamic decision making, the time sequence deviation value is obtained by analyzing network delay data so as to realize accurate description of multi-screen time migration, content feature data is converted into synchronous sensitivity scores, and the accuracy of the multi-screen time migration is improved. The control strategy can dynamically adjust the display priority according to the display content, then media feature data and load state data are fused to obtain a comprehensive efficiency evaluation value, and cooperative control data adaptive to the real-time state of the display screen are generated through a multi-dimensional parameter fusion decision of precise time sequence quantization, content perception control and efficiency dynamic evaluation. And virtual synchronous simulation is carried out on the cooperative control data to obtain a cooperative efficiency index, and the cooperative control data is optimized based on the cooperative efficiency index, so that the problem of display asynchronization caused by multi-screen instruction time sequence deviation accumulation is solved.

Description

Technical Field

[0001] The present invention relates to the field of LED display control technology, and in particular to a multi-screen collaborative control method and system for LED display screens. Background Art

[0002] LED displays, as flat-panel display devices using light-emitting diodes as light-emitting elements, have been widely used in many fields due to their significant technological advantages. This display technology features high brightness, high contrast, long service life, and flexibility, and has been deeply integrated into scenarios such as advertising media, live sports events, traffic signal instructions, and stage visual presentations. With the iteration and upgrading of technology, LED displays are evolving towards refinement and efficiency, as evidenced by the continuous reduction of display pitch, significant improvement in dynamic range, continuous reduction in energy consumption, and increasing intelligence. In the future development process, this technology will be deeply integrated with cutting-edge technologies such as 5G communications and artificial intelligence, gradually becoming the core carrier for information visualization in the digital age, providing key display support for the construction of intelligent scenarios.

[0003] Currently, the collaborative control of multiple LED display screens mostly adopts a centralized static scheduling strategy. This strategy is centered on a single master control node and builds a "master-slave" control architecture, where the master control node uniformly generates and distributes control commands for all screens. At the beginning of system startup, a fixed task allocation scheme and timing synchronization parameters are pre-set, and dynamic adjustments cannot be made based on the real-time status of the system during operation. Under this architecture, the master control node adopts a serial processing mode for tasks such as video segmentation and color correction, and must execute them one by one in strict accordance with the order in which the commands are received, resulting in significant timing queuing characteristics in the command processing. This processing mechanism can easily cause timing deviations in multi-screen control commands, especially when dynamic content is switched, as the response delays of multiple screens accumulate, ultimately leading to asynchronous control between the screens, seriously affecting the collaborative control effect. Summary of the Invention

[0004] The main purpose of the present invention is to provide a multi-screen collaborative control method and system for LED display screens, aiming to solve the technical problems in the prior art.

[0005] The present invention proposes a multi-screen collaborative control method for LED display screens, comprising: Obtain network delay data, load status data, and display content data for each LED display; Obtaining a timing deviation value for each LED display screen according to the network delay data; Obtaining content feature data and media feature data of each LED display screen according to the display content data, and obtaining a synchronization sensitivity score of each LED display screen according to the content feature data; Obtaining a comprehensive performance evaluation value of each LED display screen according to the media characteristic data and the load status data; Acquiring collaborative control data according to the comprehensive effectiveness evaluation value, the synchronization sensitivity score, and the timing deviation value; Performing a virtual synchronization simulation according to the collaborative control data to obtain a simulation output parameter set, and obtaining a collaborative efficiency index according to the simulation output parameter set; Determining whether the collaborative efficiency index is greater than a preset threshold; If the collaborative efficiency index is greater than a preset threshold, the LED display screen is controlled according to the collaborative control data; If the collaborative efficiency index is not greater than a preset threshold, the collaborative control data is adjusted according to the collaborative efficiency index to obtain optimized control data, and the LED display screen is controlled according to the optimized control data.

[0006] Preferably, the step of obtaining the timing deviation value of each LED display screen according to the network delay data includes: Acquire multiple master-slave delay sequences according to the network delay data, and acquire a corresponding random error value according to each of the master-slave delay sequences; Obtaining node topological coordinates, node temperature, and physical characteristic parameters of the communication link, and obtaining a topological distance matrix based on the node topological coordinates; Obtaining a medium transmission coefficient and an inherent delay time according to the physical characteristic parameters, and obtaining a plurality of link basic delays according to the medium transmission coefficient, the inherent delay time, a topological distance matrix, and the random error value; Acquire link dynamic delay according to each link basic delay and node temperature; Performing wavelet transform decomposition on each of the master-slave delay sequences to obtain a burst fluctuation component, and obtaining an abnormal fluctuation interval based on the burst fluctuation component, the link dynamic delay, and the master-slave delay sequence; Acquire a dynamic delay offset according to the abnormal fluctuation interval and the link dynamic delay; A plurality of screen refresh cycles are obtained, and a timing deviation value is obtained according to each of the screen refresh cycles and the corresponding dynamic delay offset.

[0007] Preferably, the step of obtaining a synchronization sensitivity score of each LED display screen according to the content feature data includes: Obtaining predicted motion amplitude, rendering area pixel data, and cross-screen annotation data for each display screen based on the content feature data; Acquire sensitivity benchmark parameters, wherein the sensitivity benchmark parameters include a motion sensitivity benchmark value, a color sensitivity benchmark value, and a cross-screen sensitivity benchmark value; Performing Kalman filtering on the predicted motion amplitude of each display screen to obtain a filtered motion value, and obtaining a motion sensitivity score based on the filtered motion value, a motion sensitivity reference value, and the node topology coordinates; Obtaining an inter-frame chromaticity difference rate and a brightness fluctuation value based on pixel data of a rendering area of each display screen, and obtaining a color sensitivity score based on the inter-frame chromaticity difference rate, the color sensitivity reference value, and the brightness fluctuation value; Obtaining cross-screen coverage and motion coherence according to the cross-screen annotation data, and obtaining a cross-screen sensitivity score according to the cross-screen coverage, motion coherence, and the cross-screen sensitivity benchmark value; A synchronization sensitivity score of each display screen is obtained according to the motion sensitivity score, the color sensitivity score, and the cross-screen sensitivity score.

[0008] Preferably, the step of obtaining a comprehensive performance evaluation value of each LED display screen according to the media characteristic data and the load status data includes: Obtaining corresponding media area proportion and calculation complexity according to each of the media feature data; Obtain a media reference weight group, and obtain a mixed weight group for each display screen based on the media area ratio, calculation complexity, and the media reference weight group; Obtaining a corresponding CPU utilization rate, video memory occupancy rate, network I / O queue depth, and cache hit rate based on each load status data, and obtaining a processing performance index, acceleration potential value, and dynamic load balancing degree for each display screen based on the CPU utilization rate, video memory occupancy rate, network I / O queue depth, cache hit rate, and the hybrid weight group; Obtaining a comprehensive performance intensity and load tendency value of each display screen according to the processing performance index, the dynamic load balancing degree, and the acceleration potential value; Obtaining a performance feature weight group, and obtaining a dynamic allocation weight group for each display screen based on the performance feature weight group and the load tendency value; A comprehensive performance evaluation value of each display screen is obtained according to the comprehensive performance intensity and the dynamic allocation weight group.

[0009] Preferably, the step of acquiring collaborative control data according to the comprehensive effectiveness evaluation value, the synchronization sensitivity score and the timing deviation value includes: Obtaining a strategy optimization target and a three-dimensional decision matrix according to the comprehensive effectiveness evaluation value, the synchronization sensitivity score, and the timing deviation value, and obtaining an initial control strategy according to the strategy optimization target and the three-dimensional decision matrix; Acquire a strategy conflict feature set according to the initial control strategy and the topological distance matrix, wherein the strategy conflict feature set includes timing deviation conflict, resource preemption conflict, and resource allocation conflict; Obtaining a timing synchronization instruction according to the timing deviation conflict and the motion sensitivity score; Obtaining a task allocation plan based on the resource preemption conflict and the cross-screen sensitivity score; A resource scheduling rule is obtained according to the resource allocation conflict and the load tendency value.

[0010] Preferably, the step of adjusting the collaborative control data according to the collaborative efficiency index to obtain optimized control data includes: Obtaining a real-time state vector of each display screen, and obtaining a dynamic control parameter based on the real-time state vector and the collaborative efficiency index; Obtaining a preset rule conversion matrix, and obtaining a task allocation weight matrix based on the preset rule conversion matrix and the dynamic control parameters; Obtaining a dynamic aggregation weight according to the task allocation weight matrix and the simulation output parameter set; Acquire a cluster state difference vector according to the real-time state vector, and acquire a timing compensation vector according to the cluster state difference vector and the dynamic control parameter; Acquire state-exceeding-limit data according to the simulation output parameter set, and acquire degradation strategy data according to the state-exceeding-limit data and the coordinated control data; Optimization control data is acquired according to the timing compensation vector, the dynamic aggregation weight and the degradation strategy data.

[0011] This application also provides an LED display multi-screen collaborative control system, including: Acquisition module, used to obtain network delay data, load status data and display content data of each LED display; A timing analysis module, configured to obtain a timing deviation value of each LED display screen based on the network delay data; a content perception module, configured to obtain content feature data and media feature data of each LED display screen based on the display content data, and obtain a synchronization sensitivity score of each LED display screen based on the content feature data; An efficiency evaluation module, configured to obtain a comprehensive efficiency evaluation value of each LED display screen based on the media characteristic data and the load status data; A collaborative decision-making module, configured to obtain collaborative control data based on the comprehensive effectiveness evaluation value, the synchronization sensitivity score, and the timing deviation value; a virtual simulation module, configured to perform a virtual synchronization simulation based on the collaborative control data, obtain a simulation output parameter set, and acquire a collaborative efficiency index based on the simulation output parameter set; A judgment module, configured to judge whether the collaborative efficiency index is greater than a preset threshold; If the collaborative efficiency index is greater than a preset threshold, the LED display screen is controlled according to the collaborative control data; If the collaborative efficiency index is not greater than a preset threshold, the collaborative control data is adjusted according to the collaborative efficiency index to obtain optimized control data, and the LED display screen is controlled according to the optimized control data.

[0012] Preferably, the timing analysis module includes: a timing analysis unit, configured to obtain a plurality of master-slave delay sequences according to the network delay data, and obtain a corresponding random error value according to each of the master-slave delay sequences; A feature extraction unit, configured to obtain node topological coordinates, node temperature, and physical characteristic parameters of the communication link, and to obtain a topological distance matrix based on the node topological coordinates; A first acquisition unit is configured to acquire a medium transmission coefficient and an intrinsic delay time according to the physical characteristic parameters, and acquire a plurality of link basic delays according to the medium transmission coefficient, the intrinsic delay time, a topological distance matrix, and the random error value; A correction unit, configured to obtain a dynamic link delay according to each of the link basic delays and the node temperature; an analysis unit, configured to perform wavelet transform decomposition on each of the master-slave delay sequences to obtain a burst fluctuation component, and obtain an abnormal fluctuation interval based on the burst fluctuation component, the link dynamic delay, and the master-slave delay sequence; A second acquiring unit, configured to acquire a dynamic delay offset according to the abnormal fluctuation interval and the link dynamic delay; The synthesis unit is used to obtain multiple screen refresh cycles and obtain a timing deviation value according to each screen refresh cycle and the corresponding dynamic delay offset.

[0013] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned LED display multi-screen collaborative control method are implemented.

[0014] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned LED display multi-screen collaborative control method are implemented.

[0015] The beneficial effects of the present invention are as follows: the present invention collects multi-dimensional data of display screens to provide multi-source data support for subsequent dynamic decision-making, thereby solving the problem in the prior art that the centralized strategy cannot perceive the real-time status differences of each display screen due to one-sided data collection, resulting in the disconnection between control instructions and actual needs. Then, through refined analysis of network delay data, the delay is quantified into a timing deviation value, and accurate characterization of multi-screen time offset is achieved to solve the problem in the prior art that static timing parameters cannot adapt to the dynamic changes of network delay, resulting in the accumulation of multi-screen instruction timing deviations and display asynchrony. By converting content feature data into a synchronization sensitivity score, the control strategy can dynamically adjust the display priority according to the display content, and then integrate the media feature data and load status data to generate a quantitative comprehensive performance evaluation value to achieve dynamic characterization of the processing capabilities of each screen, and then through multi-dimensional The fusion decision of the degree parameters is used to generate collaborative control data that adapts to the real-time status of the display screen. Then, through a virtual simulation environment consistent with the physical topology, hardware performance and network characteristics of the actual LED display screen, the collaborative control data including task allocation schemes, timing synchronization instructions and resource scheduling rules are imported into the environment. The multi-screen collaborative process is simulated according to the actual operation logic, and the instruction execution of each screen is tracked and recorded in real time to obtain a simulation output parameter set. The collaborative efficiency index is obtained through the simulation output parameter set, and finally the collaborative efficiency index is compared with the preset threshold. If the collaborative efficiency index is not greater than the preset threshold, the collaborative control data is adjusted according to the collaborative efficiency index to obtain optimized control data. Through this continuously optimized dynamic control method, the continuous accumulation of multi-screen response delays can be further avoided, which ultimately leads to the problem of asynchronous control between screens, thereby improving the multi-screen collaborative control effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 Schematic diagram of a method flow according to an embodiment of the present invention.

[0017] Figure 2 FIG. 1 is a schematic diagram of a system structure according to an embodiment of the present invention.

[0018] Figure 3 This is a schematic diagram of the internal structure of a computer device according to an embodiment of the present application.

[0019] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0020] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0021] like Figure 1 As shown, the present application provides a multi-screen collaborative control method for an LED display screen, comprising: S1. Obtain network delay data, load status data, and display content data for each LED display screen; S2. Obtaining a timing deviation value for each LED display screen based on the network delay data; S3. Obtaining content feature data and media feature data of each LED display screen according to the display content data, and obtaining a synchronization sensitivity score of each LED display screen according to the content feature data; S4. Obtaining a comprehensive performance evaluation value of each LED display screen based on the media characteristic data and the load status data; S5. Acquire collaborative control data according to the comprehensive effectiveness evaluation value, the synchronization sensitivity score, and the timing deviation value; S6. Performing a virtual synchronization simulation based on the collaborative control data to obtain a simulation output parameter set, and obtaining a collaborative efficiency index based on the simulation output parameter set; S7. Determine whether the collaborative efficiency index is greater than a preset threshold; If the collaborative efficiency index is greater than a preset threshold, the LED display screen is controlled according to the collaborative control data; If the collaborative efficiency index is not greater than a preset threshold, the collaborative control data is adjusted according to the collaborative efficiency index to obtain optimized control data, and the LED display screen is controlled according to the optimized control data.

[0022] As described in the above steps S1-S7, the present invention provides multi-source data support for subsequent dynamic decision-making by performing multi-dimensional data collection of network delay data, load status data and display content data on the display screen, thereby solving the problem in the prior art that the centralized strategy cannot perceive the real-time status differences of each display screen due to one-sided data collection, resulting in the disconnection between control instructions and actual needs. Among them, network delay data refers to the time delay related data of signal transmission when data communication is carried out between LED display screens, load status data refers to the hardware resource occupancy of each LED display screen control node, and display content data refers to the content information to be displayed on the LED display screen. Then, through the refined analysis of the network delay data, the delay is quantified into a timing deviation value, and the accurate characterization of the time offset of multiple screens is achieved, so as to solve the problem in the prior art that the static timing parameters cannot adapt to the dynamic changes of network delay, resulting in the accumulation of timing deviations of multiple screen instructions and causing display asynchrony. Among them, the timing deviation value refers to the difference between the master node and the slave node, or between different slave nodes. The quantified value of the difference in instruction execution time and frame rendering time between points. Since the static scheduling strategy in the existing technology usually ignores content differences and adopts the same synchronization standard for all display screens, high-priority content (such as objects moving across screens) fails to synchronize due to insufficient resource allocation. Therefore, this solution converts content feature data into a synchronization sensitivity score, so that the control strategy can dynamically adjust the priority according to the content. Among them, content feature data refers to feature information extracted from display content data and closely related to multi-screen synchronization. The synchronization sensitivity score refers to an indicator that quantitatively evaluates the importance of maintaining content synchronization for different display screens during the collaborative control process. Then, the media feature data and load status data are integrated to generate a quantitative comprehensive performance evaluation value to achieve dynamic characterization of the processing capabilities of each screen. Among them, media feature data refers to feature information related to the media type extracted from display content data. The comprehensive performance evaluation value refers to an indicator that quantifies the comprehensive processing capabilities of each LED display node. Then, through the fusion decision of multi-dimensional parameters, collaborative control data adapted to the real-time state is generated. The collaborative control data refers to a set of control strategies generated by the decision algorithm, including task allocation schemes, timing synchronization instructions, resource scheduling rules, etc. Then, through a virtual simulation environment consistent with the physical topology, hardware performance and network characteristics of the actual LED display, the collaborative control data including task allocation schemes, timing synchronization instructions and resource scheduling rules are imported into the environment. The multi-screen collaborative process is simulated according to the actual operation logic, and the execution of instructions on each screen is tracked and recorded in real time to obtain a simulation output parameter set. The simulation output parameter set refers to the parameter set reflecting the multi-screen collaborative effect after the virtual synchronization simulation of the collaborative control data, including simulated timing synchronization error, resource utilization efficiency, cross-screen content consistency, etc. Then, various indicators in the parameter set are normalized to eliminate dimensional differences, and then weighted according to preset weights (such as timing synchronization accounting for 0. 3. Resource efficiency accounts for 0.2, content consistency accounts for 0.4, and conflict resolution accounts for 0.1). The standardized parameters are weighted and summed to obtain a collaborative efficiency index with a value range of [0, 1]. The collaborative efficiency index refers to a comprehensive indicator calculated based on the simulation output parameter set to quantify the multi-screen collaborative effect. Finally, the collaborative efficiency index is compared with the preset threshold. If the collaborative efficiency index is greater than the preset threshold, the LED display is controlled according to the collaborative control data. If the collaborative efficiency index is not greater than the preset threshold, the collaborative control data is adjusted according to the collaborative efficiency index to obtain optimized control data. The optimized control data refers to an improved control strategy obtained by adjusting the collaborative control data, and the LED display is controlled according to the optimized control data. This dynamic control method of continuous optimization can further avoid the continuous accumulation of multi-screen response delays, which ultimately leads to the problem of asynchronous control between screens, thereby improving the multi-screen collaborative control effect.

[0023] In one embodiment, the step S2 of obtaining the timing deviation value of each LED display screen according to the network delay data includes: S21. Acquire multiple master-slave delay sequences according to the network delay data, and acquire a corresponding random error value according to each of the master-slave delay sequences; S22, obtaining node topological coordinates, node temperature, and physical characteristic parameters of the communication link, and obtaining a topological distance matrix according to the node topological coordinates; S23. Obtain a medium transmission coefficient and an inherent delay time according to the physical characteristic parameters, and obtain a plurality of link basic delays according to the medium transmission coefficient, the inherent delay time, the topological distance matrix, and the random error value; S24. Obtaining a dynamic link delay according to each link basic delay and node temperature; S25. Perform wavelet transform decomposition on each of the master-slave delay sequences to obtain a burst fluctuation component, and obtain an abnormal fluctuation interval based on the burst fluctuation component, the link dynamic delay, and the master-slave delay sequence; S26. Obtain a dynamic delay offset according to the abnormal fluctuation interval and the link dynamic delay; S27. Acquire multiple screen refresh cycles, and acquire a timing deviation value according to each screen refresh cycle and the corresponding dynamic delay offset.

[0024] As described in steps S21-S27 above, the present invention obtains multiple master-slave delay sequences in real time based on network delay data through a sliding time window, wherein the master-slave delay sequence refers to a set of network round-trip delay measurement values arranged in chronological order between the master display screen node and the slave display screen node, for example, [12.1, 12.3, 12.0, 18.6, 14.2] ms, and performs log-normal distribution fitting on each master-slave delay sequence, calculates the mean and standard deviation, and obtains a random error value, wherein the random error value refers to the normal fluctuation range in the delay sequence that conforms to statistical laws. This data processing method breaks through the limitations of traditional static threshold judgment and provides a refined error benchmark for subsequent delay decomposition. Then, through G The PS module obtains the node topology coordinates of the communication link, where the node topology coordinates refer to the deployment position of the LED display control node in three-dimensional space, including the master node coordinates and the slave node coordinates, and calculates the Euclidean distance between the master node coordinates and the slave node coordinates based on the node topology coordinates. Then, the topological distance matrix is further constructed based on the Euclidean distance. The topological distance matrix refers to a symmetric square matrix that describes the physical distance between all nodes. When estimating the communication delay between nodes, the traditional method ignores the physical position relationship and only uses the IP hop count to estimate the delay. However, this solution accurately quantifies the inherent delay of signal transmission through spatial perception delay prediction to solve the problem of delay misjudgment caused by the physical transmission path of long-distance nodes. Then, through the formula " ”Calculate the basic link delay, where Indicates the Node to The basic link delay of each node refers to the static theoretical delay without considering the environmental impact. Represents the medium transmission coefficient, which refers to the theoretical delay of signal transmission per unit distance. Inherent delay time refers to the fixed time overhead required for network equipment to process signals. Indicates the Node to The random error value corresponding to the communication link of the node, the formula decomposes the link delay into components directly related to the physical characteristics of the link and a fixed offset, where " The paper demonstrates the linear positive correlation between delay and link length. The medium transmission coefficient, as a transmission coefficient, characterizes the delay contribution per unit length, reflecting the basic physical laws of signal propagation in the medium. The inherent delay time, as a fixed offset, accounts for the impact of other fixed factors besides length (such as initial device processing delay and inherent time consumption of interface conversion). The random error value is used to characterize delay fluctuations caused by uncertain factors such as environmental interference and device fluctuations in actual transmission. This enables interpretable delay modeling and provides a theoretical basis for the accurate quantification of long-distance transmission delay. Then, the temperature of each node is monitored in real time by the temperature sensor, and the temperature of each node is measured by the formula " ”Calculate the dynamic delay of the link, where Indicates the Node to The link dynamic delay of each node refers to the dynamic reference delay after superimposing temperature compensation. It represents the temperature sensitivity coefficient, which can be obtained through the material thermal characteristic parameters of the link transmission medium (such as cable, optical fiber). Indicates the Node to The average temperature of this communication link of nodes, Indicates the Node to The core of this approach is to establish a causal chain of "temperature-medium characteristics-delay": when the temperature is above 25°C, the medium length may elongate or the propagation speed may decrease, resulting in an increase in delay. A correction factor greater than 1 amplifies the basic delay. Conversely, when the temperature drops, the correction factor becomes less than 1, reducing the delay. This correction method considers the direct impact of temperature on the physical quantity of transmission, enabling dynamic delay calculation to quickly respond to environmental changes in engineering practice and ensure accurate delay estimation. Traditional methods typically ignore the impact of temperature, while this solution uses environmentally adaptive delay compensation to avoid synchronization deviations caused by changes in the fiber refractive index due to high temperatures. Then, the db4 wavelet basis is used to perform a three-layer decomposition of the master and slave delay sequences, and the high-frequency burst component is extracted to obtain the burst fluctuation component. The burst fluctuation component refers to the high-frequency abnormal component obtained by wavelet decomposition of the delay sequence, which is used to accurately separate transient anomalies such as sudden network congestion. This multi-resolution time-frequency analysis method can separate burst noise from basic delay, thereby accurately identifying short-term sudden anomalies. Then, when the burst fluctuation component exceeds 10% of the dynamic benchmark delay, it is marked as an abnormal fluctuation interval. For example: if the first Node to The master-slave delay sequence corresponding to the communication link of each node is: [12.1, 12.3, 12.0, 18.6, 14.2] ms. After decomposing this master-slave delay sequence into packets, the basic fluctuation component is: [12.1, 12.2, 12.1, 12.1, 12.2] ms, and the burst fluctuation component is [0, 0.1, -0.1, 6.5, 2]. If the dynamic baseline delay is 12.5 ms, the abnormal fluctuation range can be expressed as [normal, normal, normal, abnormal, abnormal]. In practical applications, the state of the communication system is in dynamic change. Recent abnormalities are often more valuable than historical abnormalities. Therefore, the original delay data and the corresponding data acquisition time are obtained from the master-slave delay sequence in combination with the abnormal fluctuation interval. Then, the time collection interval is obtained based on the difference between the data acquisition time corresponding to the original delay data and the current time, and the time collection interval is obtained through the formula " ” Calculate the time decay weight, where Indicates the The time attenuation weight corresponding to the original delay data, Indicates the The time collection interval corresponding to the original time delay data, Indicates the sequence number of the original time-delayed data. This process reflects the exponential decay characteristic. The closer the data is to the current moment, the higher the weight is and the greater the influence on the final result is. ” Calculate the dynamic delay offset, where, Indicates the dynamic delay offset, which refers to the weighted average deviation of the delay in the abnormal interval from the dynamic benchmark. Indicates the The time attenuation weight corresponding to the original delay data, Indicates the amount of raw delay data, Indicates the sequence number of the original delay data, Indicates the The original delay data, represents the dynamic link delay. This result comprehensively considers the deviation degree and time importance of each data point. Compared with the simple arithmetic average, it can more accurately reflect the abnormal status of the current system. By introducing this method of time-attenuated weighting, it emphasizes recent anomalies, facilitates prioritizing the latest network failures, and reduces historical data interference. Finally, the screen driver chip is queried to obtain the screen refresh cycle, where the screen refresh cycle refers to the time interval for the LED screen to complete a single-frame image refresh. The timing deviation value is obtained based on the screen refresh cycle and the corresponding dynamic delay offset, thereby providing an accurate time compensation basis for the subsequent synchronous control of the master and slave display screen nodes.

[0025] In one embodiment, the step S3 of obtaining the synchronization sensitivity score of each LED display screen according to the content feature data includes: S31, obtaining predicted motion amplitude, rendering area pixel data, and cross-screen annotation data for each display screen according to the content feature data; S32. Acquire sensitivity benchmark parameters, wherein the sensitivity benchmark parameters include a motion sensitivity benchmark value, a color sensitivity benchmark value, and a cross-screen sensitivity benchmark value; S33, performing Kalman filtering on the predicted motion amplitude of each display screen to obtain a filtered motion value, and obtaining a motion sensitivity score based on the filtered motion value, the motion sensitivity reference value, and the node topology coordinates; S34. Obtaining an inter-frame chromaticity difference rate and a brightness fluctuation value based on the pixel data of the rendering area of each display screen, and obtaining a color sensitivity score based on the inter-frame chromaticity difference rate, the color sensitivity reference value, and the brightness fluctuation value; S35. Obtaining cross-screen coverage and motion coherence according to the cross-screen annotation data, and obtaining a cross-screen sensitivity score according to the cross-screen coverage, motion coherence, and the cross-screen sensitivity benchmark value; S36. Obtain a synchronization sensitivity score for each display screen according to the motion sensitivity score, the color sensitivity score, and the cross-screen sensitivity score.

[0026] As described in the above steps S31-S36, the present invention first extracts three types of basic data corresponding to each display screen from the content feature data: predicted motion amplitude, rendering area pixel data and cross-screen annotation data, wherein the predicted motion amplitude refers to the maximum displacement of the moving object in the picture to be displayed by the LED display screen calculated by the optical flow algorithm, which is used to quantify the dynamic degree of the picture, the rendering area pixel data refers to the statistical set of chromaticity and brightness of all pixels in the display area assigned to a single LED screen, and the cross-screen annotation data refers to multi-screen collaborative display feature data, including: object cross-screen identifier, cross-screen motion trajectory, color consistency mark, etc., and then obtains the sensitivity benchmark parameters from the system configuration, wherein the sensitivity benchmark parameters refer to a preset benchmark parameter set, including motion sensitivity benchmark value, color sensitivity benchmark value and cross-screen sensitivity benchmark value, and the sensitivity benchmark parameters are basic parameters preset according to the historical data of multi-screen collaborative control and the optimization target. The number is used as the calculation basis for various subsequent sensitivity coefficients. Then, the predicted motion amplitude of each display screen is processed by Kalman filtering. By filtering out noise interference, a filtered motion value reflecting the actual motion state of the object is obtained. According to the relative position of each display screen in the overall display wall, the spatial weight factor corresponding to each display screen is obtained, and the motion perception weight of the display screens at different positions is corrected by the spatial weight factor. Therefore, the filtered motion value, the motion sensitivity reference value and the spatial weight factor are combined to obtain the motion sensitivity score of each display screen. The motion sensitivity score refers to a quantitative value that characterizes the sensitivity of the picture motion to synchronization error. This motion perception quantification method of Kalman filtering noise reduction combined with spatial weight correction solves the problem that the existing technology only evaluates the synchronization requirements by the original motion amplitude, does not consider the influence of noise interference and screen position on perception, resulting in distortion of motion sensitivity evaluation and insufficient multi-screen synchronization accuracy. Then, the pixel matrix of the area to be rendered is obtained according to the pixel data of the rendering area of each display screen, and the pixel matrix of the area to be rendered is converted from the RGB color space to the LAB color space, and the pixel matrix of two consecutive frames of the area to be rendered of the display screen is obtained, and the a and b channel values reflecting the chromaticity information are extracted. Then, the absolute value of the difference between the a and b channel values of the corresponding pixels of the two frames is calculated respectively, and then the absolute value of the difference of all pixels is summed up, and finally divided by the total number of pixels in the area to obtain the inter-frame chromaticity difference rate, where the inter-frame chromaticity difference rate refers to the intensity of the color change between two consecutive frames. The larger the value, the more significant the dynamic color change. Secondly, based on the L channel extracted from the LAB color space, the regional average brightness of all pixels in the area is first calculated, and then the square root of the sum of the squares of the difference between the L value of each pixel and the regional average brightness is divided by the total number of pixels to obtain the brightness standard deviation, and this is used as Luminance fluctuation value, where the luminance fluctuation value refers to the unevenness of pixel luminance within a single frame. This value reflects the degree of brightness dispersion within an area. A larger value indicates a stronger contrast between light and dark. Finally, based on the color sensitivity reference value, the inter-frame chromaticity difference rate calculated above is multiplied by the luminance fluctuation value (normalized to a range of 0-1). The result is then multiplied by the color sensitivity reference value to obtain the color sensitivity score. The color sensitivity score assesses the display's sensitivity to dynamic color changes and differences in light and dark contrast. A higher score indicates that the display's color performance has a more significant impact on the visual experience when rendering content, requiring priority attention in aspects such as color calibration and rendering optimization. This data processing method solves the problem that existing technologies have difficulty accurately quantifying the human eye's sensitivity to dynamic color changes and contrast, resulting in a lack of targeted color synchronization optimization. Then, the cross-screen coverage is obtained based on the cross-screen annotation data, where the cross-screen coverage refers to the proportion of the display area when the object crosses the screen boundary, and the object motion trajectory data in the cross-screen annotation data is parsed, and the coordinate information of the object at the adjacent screen edge is extracted (such as the exit coordinates of the current screen edge and the entry coordinates of the adjacent screen edge), and the Euclidean distance deviation of the two points in three-dimensional space is calculated. At the same time, combined with the velocity vector direction consistency in the object motion trajectory prediction data (that is, the cosine value of the angle between the motion directions of adjacent frames), the distance deviation and direction consistency are weighted according to the preset weights (such as distance deviation accounts for 30% and direction consistency accounts for 70%), and then the result is normalized to the range of [0,1] to obtain motion coherence, where motion coherence refers to a measure of the smoothness of the cross-screen object motion trajectory. The closer the value is to 1, the more coherent the object's cross-screen motion trajectory is. Otherwise, there is a trajectory break or offset. Then, cross-screen sensitivity is used to calculate the smoothness of the cross-screen object motion trajectory. The cross-screen sensitivity score is calculated based on the cross-screen sensitivity benchmark value, combined with cross-screen coverage and motion coherence. Among them, the cross-screen sensitivity score refers to the score that quantifies the consistency of multi-screen collaborative display. This coefficient is used to reflect the sensitivity of the display to cross-screen content. Through this two-dimensional weighted evaluation method of cross-screen coverage and trajectory continuity, the problem that the existing technology does not consider the impact of cross-screen coverage and motion continuity on multi-screen consistency, resulting in misaligned or inconsistent color display of cross-screen content, is solved. Finally, the motion sensitivity score, color sensitivity score and cross-screen sensitivity score are normalized to the same interval, and the preset weight coefficient is read from the system configuration. Then, the linear weighted formula is used to calculate the synchronization sensitivity score, thereby providing a comprehensive and quantitative synchronization priority basis for multi-screen collaborative control, enabling the system to dynamically allocate synchronization resources according to the content characteristics of different screens, and give priority to ensuring the timing and color consistency of high-sensitivity screens.

[0027] In one embodiment, the step S4 of obtaining a comprehensive performance evaluation value of each LED display screen according to the media characteristic data and the load status data includes: S41. Obtaining corresponding media area proportion and calculation complexity according to each piece of media feature data; S42: Obtain a media reference weight group, and obtain a mixed weight group for each display screen based on the media area ratio, calculation complexity, and the media reference weight group; S43. Obtaining a corresponding CPU utilization rate, video memory occupancy rate, network I / O queue depth, and cache hit rate based on each load status data, and obtaining a processing performance index, acceleration potential value, and dynamic load balancing degree for each display screen based on the CPU utilization rate, video memory occupancy rate, network I / O queue depth, cache hit rate, and the hybrid weight group; S44. Obtaining a comprehensive performance intensity and load tendency value of each display screen according to the processing performance index, dynamic load balancing degree, and acceleration potential value; S45, obtaining a performance feature weight group, and obtaining a dynamic allocation weight group for each display screen based on the performance feature weight group and the load tendency value; S46 , obtaining a comprehensive performance evaluation value of each display screen according to the comprehensive performance intensity and the dynamic allocation weight group.

[0028] As described in the above steps S41-S46, the present invention obtains the media area ratio and computational complexity corresponding to each display screen through media feature data, wherein the media area ratio refers to the pixel ratio of the dynamic video area and the static image area in the display content of each display screen, including the video stream area ratio and the static image area ratio. The computational complexity refers to the number of instruction cycles required for each display screen to process unit media data, including the video stream computational complexity coefficient and the static image computational complexity coefficient. Then, the media benchmark weight group is obtained, wherein the media benchmark weight group refers to the preset weight configuration of different media types, such as: in the video stream scenario, α=0.7, β=0.3, in the image stream scenario, α=0.3, β=0.7, and then respectively through " "and" The influence weights of video stream resources and static graphic resources are calculated, where: Indicates the impact weight of video stream resources, Indicates the proportion of video stream area, Indicates the proportion of static image and text area. Indicates the computational complexity coefficient of the video stream, Indicates the computational complexity coefficient of static graphics and text, Indicates the influence weight of static graphic resources, and then respectively through " "and" The weight coefficient of the CPU in the mixed media scenario and the weight coefficient of the video memory in the mixed media scenario are calculated, where: Indicates the CPU weight coefficient of the mixed media scene. Indicates the impact weight of video stream resources, Indicates the CPU weight benchmark in the video streaming scenario. Indicates the influence weight of static graphic resources. Indicates the CPU weight benchmark in static graphics scenes. Indicates the weight coefficient of video memory in mixed media scenarios, Indicates the weight benchmark of video memory in the video streaming scenario. This represents the weight benchmark for video memory in static graphics scenes. This dynamic weight adaptation mechanism, based on media area ratio and computational complexity, overcomes the limitations of traditional fixed weight configurations that are incapable of adapting to mixed media scenes, enabling the weight coefficient to be adjusted in real time based on content characteristics. Then through " "Computational processing efficiency index, where Indicates the processing efficiency index, which reflects the theoretical processing capability of the display screen for a specific media type. Indicates the weight coefficient of CPU in mixed media scenarios. Indicates CPU utilization. Indicates the weight coefficient of video memory in mixed media scenarios, Indicates the memory usage. Indicates the network I / O queue depth, Indicates the cache hit rate. The " "and" " represents the "remaining resource rate" of the CPU and GPU, respectively. Larger values indicate more abundant resources. The index is weighted based on the dynamic weight of the media type, reflecting the differentiated dependence of mixed media on the CPU and GPU (e.g., video streaming relies more on GPU decoding, while static graphics rely more on CPU rendering). Because the impact of network queue depth on efficiency has a "diminishing marginal return" characteristic, a logarithmic function is used to reflect that when the queue is short, increasing the queue length has a significant impact on efficiency. When the queue is too long, efficiency has reached saturation, and the impact of further increases is weakened. This integration of three-dimensional hardware characteristics (computing resources, network input, and memory access) into a single processing efficiency index enables comparability across different scenarios. And through ”Calculate the dynamic load balancing degree, where Indicates the dynamic load balancing degree, which refers to the coordination evaluation of the quantified CPU and GPU load differences. Indicates CPU utilization. Indicates the memory usage. Indicates the proportion of static image and text area. Indicates the proportion of video stream area. "Reflects the resource demand ratio between the graphic area and the video area," "Reflects the ratio of the actual CPU load to the actual memory load, which directly corresponds to the actual allocation status of hardware resources. By mapping the "deviation value" to the balance degree: when the actual load ratio is approximately equal to the theoretical demand ratio (the deviation approaches 0), When it approaches 1, it means that the load distribution is highly matched with the regional demand and the balance is good; when the deviation increases (for example, the video area accounts for a high proportion but the video memory occupancy is extremely low), Reduce to visually reflect load imbalance; Since video processing depends on the CPU and network I / O queue depth, the potential is quantified by the remaining CPU resources and network congestion, while graphic processing depends on cache and video memory, so the potential is quantified by cache efficiency and remaining video memory space. In summary, through " Calculate the acceleration potential value, where Represents the acceleration potential value, which is the available performance gain of the quantified hardware acceleration unit. Indicates the weight coefficient of CPU in mixed media scenarios. Indicates CPU utilization. Indicates the network I / O queue depth, Indicates the weight coefficient of video memory in mixed media scenarios, Indicates the cache hit ratio, Represents the video memory utilization rate. The formula uses the logic of "splitting media demand, quantifying resource surplus, and weighted integration by proportion" to output the acceleration potential value, which directly reflects the feasibility and potential space of hardware acceleration of mixed media content under the current hardware load. The higher the acceleration potential value, the closer the match between the remaining hardware resources and media demand, and the more significant the performance improvement after acceleration. Then, the processing efficiency index, dynamic load balancing degree and acceleration potential value are mapped in three-dimensional space to obtain the efficiency feature vector, and the modulus of the efficiency feature vector is calculated to obtain the comprehensive efficiency strength, where the comprehensive efficiency strength refers to the comprehensive processing capability of the quantified display screen in the mixed media scenario. The calculation of the modulus is essentially to geometrically synthesize the contributions of each dimension of the three-dimensional efficiency feature vector through the distance measurement of the Euclidean space. Its physical meaning is to quantify the overall processing capability level of the display screen. The larger the modulus value, the stronger the comprehensive performance of the display screen in the three dimensions of media processing efficiency, resource coordination and hardware acceleration advantage. The angle between the projection of the efficiency feature vector on the processing efficiency index and dynamic load balancing plane and the processing efficiency index is calculated to obtain Load tendency value and angle calculation are calculated by projecting a three-dimensional vector onto a two-dimensional plane of processing efficiency index and dynamic load balancing. Its core purpose is to remove the interference of hardware acceleration potential and analyze the display's resource allocation tendency between "media processing efficiency" and "load balancing." This angle value essentially reflects the display's resource utilization strategy preference: when the angle approaches 0°, it indicates that the display performance is primarily dominated by media processing efficiency; when the angle approaches 90°, dynamic load balancing becomes the core limiting factor. This projection analysis can penetrate the superficial performance improvement brought by hardware acceleration and directly reveal the true state of underlying resource coordination, providing decision-making basis for task scheduling: "Whether to prioritize supplementing CPU resources or optimizing GPU load?" Finally, the performance feature weight group is obtained, wherein the performance feature weight group refers to the predefined scoring weight, and the fuzzy logic controller is used according to the performance feature vector to establish a continuous weight mapping method to correct the performance feature weight group, and obtain the dynamic allocation weight group, wherein the dynamic allocation weight group refers to the actual weight group finally used after optimization, for example: the angle range is divided into three core intervals: calculation dominant area: 0°≤θ<30°, equilibrium transition area: 30°≤θ<60°, graphic dominant area: 60°≤θ≤90°, and the benchmark weight is set at the center point of each interval: when θ=15°: [0.55, 0.3, 0 .15], when θ=45°: [0.25, 0.5, 0.25], when θ=75°: [0.2, 0.3, 0.5]; when θ moves at the boundary of the interval, the weight component changes proportionally, and ensures that the total weight is 1 after each adjustment. Through this weight optimization method, the performance evaluation requirements under different load intensities can be adapted, and the adaptability of weight configuration to complex scenarios is improved. Finally, the comprehensive performance evaluation value of each display screen is obtained according to the comprehensive performance intensity and dynamic allocation weight combination through the weighted summation formula, providing a unified performance quantification basis for task allocation, resource scheduling and load balancing in multi-screen collaborative control.

[0029] In one embodiment, the step S5 of acquiring the collaborative control data according to the comprehensive effectiveness evaluation value, the synchronization sensitivity score, and the timing deviation value includes: S51, obtaining a strategy optimization target and a three-dimensional decision matrix according to the comprehensive effectiveness evaluation value, the synchronization sensitivity score, and the timing deviation value, and obtaining an initial control strategy according to the strategy optimization target and the three-dimensional decision matrix; S52. Acquire a strategy conflict feature set according to the initial control strategy and the topological distance matrix, wherein the strategy conflict feature set includes timing deviation conflict, resource preemption conflict, and resource allocation conflict; S53, obtaining a timing synchronization instruction according to the timing deviation conflict and the motion sensitivity score; S54: Obtain a task allocation plan based on the resource preemption conflict and the cross-screen sensitivity score; S55. Acquire resource scheduling rules according to the resource allocation conflict and the load tendency value.

[0030] As described in the above steps S51-S55, the present invention converts the three types of input parameters, namely the comprehensive performance evaluation value, the synchronization sensitivity score and the timing deviation value, into a strategy optimization target through a multi-objective optimization algorithm, wherein the strategy optimization target refers to a set of collaborative control optimization directions constructed based on multi-dimensional input parameters, specifically including: a synchronization optimization target composed of a synchronization sensitivity score, an efficiency optimization target formed by the comprehensive performance intensity, and a timing optimization target generated by the instruction transmission timing deviation value. Its purpose is to provide a clear goal orientation for subsequent decision-making by clarifying the multi-dimensional optimization direction, ensuring that the collaborative control strategy can meet the content synchronization requirements while taking into account the hardware performance and timing accuracy, so as to solve the problem in the prior art that the multi-screen control strategy only focuses on timing synchronization or performance improvement, resulting in an imbalance in multi-objective optimization; At the same time, the present invention constructs a three-dimensional decision matrix by comprehensively considering the performance evaluation value, synchronization sensitivity score and timing deviation value, which specifically includes the timing deviation dimension, the performance evaluation dimension and the synchronization sensitivity dimension. In the process of constructing the three-dimensional decision matrix, the system inputs the data of the above three dimensions into the parallel gated recurrent unit network for timing feature extraction, and then uses the attention fusion mechanism to dynamically weight the synchronization weight coefficient based on the AHP hierarchical analysis method, the performance balance parameter based on the entropy weight method, and the timing compensation factor based on the Kalman filter prediction in the strategy optimization target to each dimensional feature, and generates a global state representation containing topological dependencies. Subsequently, a distributed strategy network is used to perform strategy exploration, and finally an initial control strategy containing task allocation weights, timing compensation amounts, and resource scheduling priorities is output; Then, a time series anomaly detection algorithm is used to analyze the instruction execution timestamps of each node and the theoretical transmission delay calculated based on the topological distance matrix. When the actual delay exceeds 3 standard deviations of the theoretical value, it is marked as a timing deviation conflict. Based on the resource competition graph model, a graph neural network is used to aggregate GPU memory request, computing unit occupancy and other data. When the total resource request of multiple nodes under the same switch exceeds 90% of the physical capacity, a conflict alarm is triggered. The maximum-minimum fair allocation algorithm is used to calculate the resource overlimit ratio to obtain resource preemption conflicts. A directed resource allocation graph is constructed, and loop dependencies are detected through depth-first search to obtain resource allocation conflicts. Then, timing deviation conflicts, resource preemption conflicts and resource allocation conflicts are standardized to form a four-dimensional feature vector set containing conflict type, associated nodes, severity (0-1 standardized value), and topological location. This converts multi-dimensional policy conflict information into a unified computable representation, providing a quantitative basis for subsequent decision-making. Finally, a weighted fusion algorithm is used to combine the delay difference of the timing deviation conflict with the motion sensitivity score to calculate the compensation amount. The FlowNet2.0 model based on the optical flow method is used to generate 3-5 frames of compensation images. The refresh rate is adjusted with adaptive vertical synchronization to obtain timing synchronization instructions. The cross-screen sensitivity score is used as the feature weight. The task load ratio of each node is determined by solving the cooperative game kernel. Highly sensitive nodes are given priority to obtain computing resources. A quadratic programming with topology constraints is used to ensure that the task migration amount of adjacent nodes does not exceed the threshold specified by the topological distance matrix. The task allocation plan is obtained. The load propensity value and the severity of the resource allocation conflict are input into the dual-depth deterministic policy gradient algorithm. The output is a resource scheduling rule that includes GPU memory quota, computing unit priority, and transmission bandwidth allocation. This method, which integrates content sensitivity characteristics and topological constraints into control policy generation, realizes content-driven refined collaborative control through multi-dimensional dynamic adaptation. It solves the problems of insufficient synchronization accuracy and unbalanced resource allocation caused by the disconnection between control strategy and content characteristics and physical topology in existing technologies.

[0031] In one embodiment, the step S7 of adjusting the collaborative control data according to the collaborative efficiency index to obtain optimized control data includes: S71. Obtain a real-time state vector of each display screen, and obtain a dynamic control parameter based on the real-time state vector and the collaborative efficiency index; S72: Obtain a preset rule conversion matrix, and obtain a task allocation weight matrix based on the preset rule conversion matrix and the dynamic control parameters; S73, obtaining a dynamic aggregation weight according to the task allocation weight matrix and the simulation output parameter set; S74. Obtain a cluster state difference vector according to the real-time state vector, and obtain a timing compensation vector according to the cluster state difference vector and the dynamic control parameter; S75. Acquire state-exceeding-limit data according to the simulation output parameter set, and acquire degradation strategy data according to the state-exceeding-limit data and the coordinated control data; S76. Acquire optimization control data according to the timing compensation vector, the dynamic aggregation weight, and the degradation strategy data.

[0032] As described in the above steps S71-S76, the present invention collects the GPU utilization, network throughput and frame buffer occupancy of each display screen in real time through embedded sensors, thereby constructing a real-time state vector of each display screen, wherein the real-time state vector refers to a three-dimensional data group that characterizes the current operating state of the LED display screen, and then uses the hyperbolic tangent function to calculate the dynamic control parameters based on the real-time state vector and the collaborative efficiency index, wherein the dynamic control parameters refer to the degree of deviation between the current display screen state and the ideal target. Compared with the traditional linear weighted method, this method of introducing nonlinear state mapping can avoid parameter mutations in extreme states and retain sensitivity when there is a small deviation, and then obtain the preset rule conversion matrix, wherein the preset rule conversion matrix is the default value. The conversion matrix refers to the decision matrix pre-stored in the control system. The column vectors in the preset rule conversion matrix represent the computing resource allocation rules, communication resource allocation rules, and rendering resource allocation rules respectively. The eigenvector is constructed by the linear term, square term, and exponential decay term of the dynamic control parameters. The eigenvector is then transformed by the preset rule conversion matrix and converted into a probability distribution through the softmax function to obtain the task allocation weight matrix. Among them, the task allocation weight matrix refers to the matrix that describes the task load ratio of each display screen. The traditional method can only perform simple polling allocation, while this solution can simultaneously consider the busyness of the computing chip, whether the network transmission is smooth, and the pressure of the picture rendering through the rule matrix, thereby realizing multi-dimensional task allocation; Then, the data of frame delay, resource fluctuation, and task accumulation are parsed through the simulation output parameter set, and the loss function is constructed through these data. The task allocation weight matrix is dynamically recalibrated according to the loss function through the exponential weighted average method to obtain the dynamic aggregation weight. Among them, the dynamic aggregation weight refers to the node importance coefficient that is dynamically adjusted according to the task allocation weight and real-time performance indicators in the federated learning process. This dual-objective aggregation method can not only consider the task allocation weights of multiple display screens, but also integrate the real-time performance indicators of multiple display screens to solve the defects caused by the traditional method of statically allocating tasks according to hardware performance. In the multi-screen collaborative system, the hardware performance (such as GPU speed and network bandwidth) of different LED screens are different, and direct synchronization will lead to fast The screen with the fastest display completes rendering in advance, and waiting for the slow screen in vain results in a waste of resources. The slow screen will drag down the overall display rhythm, causing the picture to freeze or tear. Therefore, this solution obtains the current screen state and the average state of all screens through the real-time state vector, and calculates the difference between the current screen state and the average state of all screens to obtain the cluster state difference vector, where the cluster state difference vector refers to the difference vector between the state vector of a single display screen and the cluster average state. Then, the timing compensation vector is calculated based on the cluster state difference vector and the dynamic control parameters through element-by-element multiplication, where the timing compensation vector refers to the three-dimensional adjustment instruction used to coordinate the timing of multi-screen display. By designing this device-adaptive compensation mechanism, the problem that the fixed compensation strategy cannot adapt to heterogeneous device groups can be solved; Then, the state out-of-limit data is obtained through the simulation output parameter set, wherein the state out-of-limit data refers to the structured alarm information that characterizes that one or more operating parameters of the LED display screen exceed the safety threshold, and a state out-of-limit decision tree is established. The state out-of-limit data and the state out-of-limit decision tree are mapped through predefined strategy rules to obtain degradation strategy data, wherein the degradation strategy data refers to the execution instruction set generated when the system responds to the out-of-limit state. Through this hierarchical degradation strategy, different degradation actions can be triggered according to different simulation output parameter sets to avoid the problem of excessive performance loss caused by global degradation in the existing technology. Finally, the lightweight neural network automatically learns the timing compensation vector, dynamic aggregation weight and degradation strategy data, the complex nonlinear relationship between these parameters, and outputs the picture segmentation scheme, frame synchronization timestamp and resource quota adjustment value to obtain optimized control data.

[0033] This application also provides an LED display multi-screen collaborative control system, including: Acquisition module, used to obtain network delay data, load status data and display content data of each LED display; A timing analysis module, configured to obtain a timing deviation value of each LED display screen based on the network delay data; a content perception module, configured to obtain content feature data and media feature data of each LED display screen based on the display content data, and obtain a synchronization sensitivity score of each LED display screen based on the content feature data; An efficiency evaluation module, configured to obtain a comprehensive efficiency evaluation value of each LED display screen based on the media characteristic data and the load status data; A collaborative decision-making module, configured to obtain collaborative control data based on the comprehensive effectiveness evaluation value, the synchronization sensitivity score, and the timing deviation value; a virtual simulation module, configured to perform a virtual synchronization simulation based on the collaborative control data, obtain a simulation output parameter set, and acquire a collaborative efficiency index based on the simulation output parameter set; A judgment module, configured to judge whether the collaborative efficiency index is greater than a preset threshold; If the collaborative efficiency index is greater than a preset threshold, the LED display screen is controlled according to the collaborative control data; If the collaborative efficiency index is not greater than a preset threshold, the collaborative control data is adjusted according to the collaborative efficiency index to obtain optimized control data, and the LED display screen is controlled according to the optimized control data.

[0034] In one embodiment, the timing analysis module includes: a timing analysis unit, configured to obtain a plurality of master-slave delay sequences according to the network delay data, and obtain a corresponding random error value according to each of the master-slave delay sequences; A feature extraction unit, configured to obtain node topological coordinates, node temperature, and physical characteristic parameters of the communication link, and to obtain a topological distance matrix based on the node topological coordinates; A first acquisition unit is configured to acquire a medium transmission coefficient and an intrinsic delay time according to the physical characteristic parameters, and acquire a plurality of link basic delays according to the medium transmission coefficient, the intrinsic delay time, a topological distance matrix, and the random error value; A correction unit, configured to obtain a dynamic link delay according to each of the link basic delays and the node temperature; an analysis unit, configured to perform wavelet transform decomposition on each of the master-slave delay sequences to obtain a burst fluctuation component, and obtain an abnormal fluctuation interval based on the burst fluctuation component, the link dynamic delay, and the master-slave delay sequence; A second acquiring unit, configured to acquire a dynamic delay offset according to the abnormal fluctuation interval and the link dynamic delay; The synthesis unit is used to obtain multiple screen refresh cycles and obtain a timing deviation value according to each screen refresh cycle and the corresponding dynamic delay offset.

[0035] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned LED display multi-screen collaborative control method are implemented.

[0036] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned LED display multi-screen collaborative control method are implemented.

[0037] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided in this application and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct RAM bus dynamic RAM (DRDRAM), and RAM bus dynamic RAM (RDRAM).

[0038] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, apparatus, article, or method comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, apparatus, article, or method. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, apparatus, article, or method comprising the element.

[0039] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A multi-screen collaborative control method for LED display screens, characterized in that: include: Obtain network delay data, load status data, and display content data for each LED display; Obtaining a timing deviation value for each LED display screen according to the network delay data; Obtaining content feature data and media feature data of each LED display screen according to the display content data, and obtaining a synchronization sensitivity score of each LED display screen according to the content feature data; Obtaining a comprehensive performance evaluation value of each LED display screen according to the media characteristic data and the load status data; Acquiring collaborative control data according to the comprehensive effectiveness evaluation value, the synchronization sensitivity score, and the timing deviation value; Performing a virtual synchronization simulation according to the collaborative control data to obtain a simulation output parameter set, and obtaining a collaborative efficiency index according to the simulation output parameter set; Determining whether the collaborative efficiency index is greater than a preset threshold; If the collaborative efficiency index is greater than a preset threshold, the LED display screen is controlled according to the collaborative control data; If the collaborative efficiency index is not greater than a preset threshold, the collaborative control data is adjusted according to the collaborative efficiency index to obtain optimized control data, and the LED display screen is controlled according to the optimized control data.

2. The LED display multi-screen collaborative control method according to claim 1, characterized in that: The step of obtaining the timing deviation value of each LED display screen according to the network delay data includes: Acquire multiple master-slave delay sequences according to the network delay data, and acquire a corresponding random error value according to each of the master-slave delay sequences; Obtaining node topological coordinates, node temperature, and physical characteristic parameters of the communication link, and obtaining a topological distance matrix based on the node topological coordinates; Obtaining a medium transmission coefficient and an inherent delay time according to the physical characteristic parameters, and obtaining a plurality of link basic delays according to the medium transmission coefficient, the inherent delay time, a topological distance matrix, and the random error value; Acquire link dynamic delay according to each link basic delay and node temperature; Performing wavelet transform decomposition on each of the master-slave delay sequences to obtain a burst fluctuation component, and obtaining an abnormal fluctuation interval based on the burst fluctuation component, the link dynamic delay, and the master-slave delay sequence; Acquire a dynamic delay offset according to the abnormal fluctuation interval and the link dynamic delay; A plurality of screen refresh cycles are obtained, and a timing deviation value is obtained according to each of the screen refresh cycles and the corresponding dynamic delay offset.

3. The LED display multi-screen collaborative control method according to claim 2, characterized in that: The step of obtaining the synchronization sensitivity score of each LED display screen according to the content feature data includes: Obtaining predicted motion amplitude, rendering area pixel data, and cross-screen annotation data for each display screen based on the content feature data; Acquire sensitivity benchmark parameters, wherein the sensitivity benchmark parameters include a motion sensitivity benchmark value, a color sensitivity benchmark value, and a cross-screen sensitivity benchmark value; Performing Kalman filtering on the predicted motion amplitude of each display screen to obtain a filtered motion value, and obtaining a motion sensitivity score based on the filtered motion value, a motion sensitivity reference value, and the node topology coordinates; Obtaining an inter-frame chromaticity difference rate and a brightness fluctuation value based on pixel data of a rendering area of each display screen, and obtaining a color sensitivity score based on the inter-frame chromaticity difference rate, the color sensitivity reference value, and the brightness fluctuation value; Obtaining cross-screen coverage and motion coherence according to the cross-screen annotation data, and obtaining a cross-screen sensitivity score according to the cross-screen coverage, motion coherence, and the cross-screen sensitivity benchmark value; A synchronization sensitivity score of each display screen is obtained according to the motion sensitivity score, the color sensitivity score, and the cross-screen sensitivity score.

4. The LED display multi-screen collaborative control method according to claim 3, characterized in that: The step of obtaining a comprehensive performance evaluation value of each LED display screen according to the media characteristic data and the load status data includes: Obtaining corresponding media area proportion and calculation complexity according to each of the media feature data; Obtain a media reference weight group, and obtain a mixed weight group for each display screen based on the media area ratio, calculation complexity, and the media reference weight group; Obtaining a corresponding CPU utilization rate, video memory occupancy rate, network I / O queue depth, and cache hit rate based on each load status data, and obtaining a processing performance index, acceleration potential value, and dynamic load balancing degree for each display screen based on the CPU utilization rate, video memory occupancy rate, network I / O queue depth, cache hit rate, and the hybrid weight group; Obtaining a comprehensive performance intensity and load tendency value of each display screen according to the processing performance index, the dynamic load balancing degree, and the acceleration potential value; Obtaining a performance feature weight group, and obtaining a dynamic allocation weight group for each display screen based on the performance feature weight group and the load tendency value; A comprehensive performance evaluation value of each display screen is obtained according to the comprehensive performance intensity and the dynamic allocation weight group.

5. The LED display multi-screen collaborative control method according to claim 4, characterized in that: The step of acquiring collaborative control data according to the comprehensive effectiveness evaluation value, the synchronization sensitivity score, and the timing deviation value includes: Obtaining a strategy optimization target and a three-dimensional decision matrix according to the comprehensive effectiveness evaluation value, the synchronization sensitivity score, and the timing deviation value, and obtaining an initial control strategy according to the strategy optimization target and the three-dimensional decision matrix; Acquire a strategy conflict feature set according to the initial control strategy and the topological distance matrix, wherein the strategy conflict feature set includes timing deviation conflict, resource preemption conflict, and resource allocation conflict; Obtaining a timing synchronization instruction according to the timing deviation conflict and the motion sensitivity score; Obtaining a task allocation plan based on the resource preemption conflict and the cross-screen sensitivity score; A resource scheduling rule is obtained according to the resource allocation conflict and the load tendency value.

6. The LED display multi-screen collaborative control method according to claim 1, characterized in that: The step of adjusting the collaborative control data according to the collaborative efficiency index to obtain optimized control data includes: Obtaining a real-time state vector of each display screen, and obtaining a dynamic control parameter based on the real-time state vector and the collaborative efficiency index; Obtaining a preset rule conversion matrix, and obtaining a task allocation weight matrix based on the preset rule conversion matrix and the dynamic control parameters; Obtaining a dynamic aggregation weight according to the task allocation weight matrix and the simulation output parameter set; Acquire a cluster state difference vector according to the real-time state vector, and acquire a timing compensation vector according to the cluster state difference vector and the dynamic control parameter; Acquire state-exceeding-limit data according to the simulation output parameter set, and acquire degradation strategy data according to the state-exceeding-limit data and the coordinated control data; Optimization control data is acquired according to the timing compensation vector, the dynamic aggregation weight and the degradation strategy data.

7. A multi-screen collaborative control system for LED display screens, characterized in that: include: Acquisition module, used to obtain network delay data, load status data and display content data of each LED display; A timing analysis module, configured to obtain a timing deviation value of each LED display screen based on the network delay data; a content perception module, configured to obtain content feature data and media feature data of each LED display screen based on the display content data, and obtain a synchronization sensitivity score of each LED display screen based on the content feature data; An efficiency evaluation module, configured to obtain a comprehensive efficiency evaluation value of each LED display screen based on the media characteristic data and the load status data; A collaborative decision-making module, configured to obtain collaborative control data based on the comprehensive effectiveness evaluation value, the synchronization sensitivity score, and the timing deviation value; a virtual simulation module, configured to perform a virtual synchronization simulation based on the collaborative control data, obtain a simulation output parameter set, and acquire a collaborative efficiency index based on the simulation output parameter set; A judgment module, configured to judge whether the collaborative efficiency index is greater than a preset threshold; If the collaborative efficiency index is greater than a preset threshold, the LED display screen is controlled according to the collaborative control data; If the collaborative efficiency index is not greater than a preset threshold, the collaborative control data is adjusted according to the collaborative efficiency index to obtain optimized control data, and the LED display screen is controlled according to the optimized control data.

8. The LED display multi-screen collaborative control system according to claim 7, characterized in that: The timing analysis module includes: a timing analysis unit, configured to obtain a plurality of master-slave delay sequences according to the network delay data, and obtain a corresponding random error value according to each of the master-slave delay sequences; A feature extraction unit, configured to obtain node topological coordinates, node temperature, and physical characteristic parameters of the communication link, and to obtain a topological distance matrix based on the node topological coordinates; A first acquisition unit is configured to acquire a medium transmission coefficient and an intrinsic delay time according to the physical characteristic parameters, and acquire a plurality of link basic delays according to the medium transmission coefficient, the intrinsic delay time, a topological distance matrix, and the random error value; A correction unit, configured to obtain a dynamic link delay according to each of the link basic delays and the node temperature; an analysis unit, configured to perform wavelet transform decomposition on each of the master-slave delay sequences to obtain a burst fluctuation component, and obtain an abnormal fluctuation interval based on the burst fluctuation component, the link dynamic delay, and the master-slave delay sequence; A second acquiring unit, configured to acquire a dynamic delay offset according to the abnormal fluctuation interval and the link dynamic delay; The synthesis unit is used to obtain multiple screen refresh cycles and obtain a timing deviation value according to each screen refresh cycle and the corresponding dynamic delay offset.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images