GPU acceleration-based meteorological early warning layer super-high-speed rendering display method

Abstract

This invention relates to the field of image processing technology, and more particularly to a GPU-accelerated ultra-high-speed rendering and display method for early weather warning layers. The method analyzes GPU data to determine the aircraft's operational coefficients and analyzes power data to determine external fluctuation coefficients, thereby calculating data anomaly characterization values ​​and determining the acquisition tendency during the data collection process. For normal acquisition tendencies, the method analyzes the transmission quality coefficient of the data during transmission based on the transmitted data, obtains environmental data during data transmission to calculate the coupling influence coefficient, and combines the data anomaly characterization values ​​to calculate rendering anomaly feature values ​​to determine the data transmission status and whether the data can be used for rendering. For data that can be used for rendering, the method calculates the inter-layer similarity to determine the transmitted data and generates a warning rendering layer. This invention improves the accuracy and reliability of warnings by performing multi-source influence analysis on the data to determine the rendering data.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a GPU-accelerated ultra-high-speed rendering and display method for early weather warning layers. Background Technology

[0002] In the field of meteorological early warning and disaster prevention and mitigation, real-time monitoring and early warning based on multi-source data such as radar, satellite, and numerical weather prediction have become key means. With the popularization of high-resolution observation networks and rapidly updating cyclic assimilation systems, the spatiotemporal resolution of meteorological elements such as reflectivity factor, precipitation estimation, thunderstorm identification, and strong wind areas has been improved to the minute or even second level, the hundred-meter level, or even finer. To support the operational needs of short-term and nowcasting, the early warning layer needs to be dynamically overlaid in the GIS framework to present risk areas with multiple thresholds and multiple time effects. Modern graphics processors, with their thousands of parallel computing cores, can perform pixel-level parallel processing on rasterized meteorological data, transferring operations such as color mark mapping, transparency blending, contour generation, and wind plume drawing from the central processor to shaders or general computing platforms, thereby achieving ultra-high-speed rendering of meteorological early warning layers.

[0003] Chinese Patent Publication No. CN110136234A discloses a method and system for rendering massive meteorological data, comprising: drawing a grid layered structure based on geographical distribution; acquiring meteorological data based on time scale and latitude and longitude and placing it into the first layer grid according to latitude and longitude; refining weather data of the upper-layer grids outside the first layer layer layer by layer based on the meteorological data in the previous layer grid; and rendering each layer grid according to time scale based on the meteorological data corresponding to each layer grid to obtain meteorological cloud maps of each layer at each time scale. This method and system, by grouping massive resource distribution data and refining and rendering data layer by layer based on the grouping results, can quickly complete the real-time and dynamic rendering of massive resource distribution data, effectively improving rendering speed and efficiency.

[0004] Chinese Patent Publication No. CN105653350A discloses a weather radar simulation rendering method for flight simulators. The method includes an operation method for a weather radar imaging rendering system based on the GLSL shader language and an implementation method for accelerating image rendering using a GPU. All meteorological data in this invention is sampled in real-time from the visual scene during the simulation, resulting in high real-time imaging performance. It also maintains consistency with the scene data during flight, improving the realism of the imaging. The GLSL-based weather radar simulation rendering method utilizes the parallel computing and rendering acceleration capabilities of the GPU throughout the calculation process. Furthermore, it abandons traditional radar imaging simulation methods such as relying on satellite cloud images and querying meteorological bureau data from the outset, directly collecting meteorological data from the visual simulation process of the simulation system. This significantly simplifies the tedious calculation process of reconstructing satellite cloud image data, improves imaging efficiency, reduces dependence on hardware devices, and lowers equipment costs.

[0005] However, the following problems still exist in the existing technology. Existing technologies focus on efficient multi-layer compositing, but lack prior analysis of the availability of input data. The system assumes that the original data is complete and reliable and directly sends it into the rendering pipeline. However, hidden defects may be introduced during data acquisition, transmission or processing, causing warnings to be generated based on distorted information, resulting in false alarms or missed alarms, and reducing the accuracy and reliability of warnings. Summary of the Invention

[0006] To address this issue, the present invention provides a GPU-accelerated ultra-high-speed rendering and display method for early weather warning layers. This method solves the problem that existing technologies focus on efficient multi-layer synthesis but lack prior analysis of the availability of input data. The system assumes that the original data is complete and reliable and directly sends it into the rendering pipeline. However, data acquisition, transmission, or processing may introduce hidden defects, leading to warnings being generated based on distorted information, resulting in false alarms or missed alarms, and reducing the accuracy and reliability of warnings.

[0007] To achieve the above objectives, this invention provides a GPU-accelerated ultra-high-speed rendering and display method for early weather warning layers, comprising: The system acquires GPU data, power data, and transmission data corresponding to the early warning monitoring area. The GPU data includes hotspot temperature and fan speed, the power data includes voltage and electromagnetic values, and the transmission data includes throughput and retry count. The GPU data is analyzed to determine the fuselage operation coefficient corresponding to the early warning monitoring area during the data acquisition process. The power data is analyzed to determine the external fluctuation coefficient during the data acquisition process. Based on the fuselage operation coefficient and the external fluctuation coefficient, the data anomaly characterization value is calculated to determine the acquisition tendency during the data acquisition process. Based on the normal data collection tendency, the transmission quality coefficient of the data during the transmission process is analyzed, environmental data during the data transmission process is obtained to calculate the coupling influence coefficient, and rendering anomaly characteristic value is calculated in combination with the data anomaly characterization value to determine the transmission status of the data and whether the data can be used for rendering. For data that can be used for rendering, calculate the similarity between data layers to determine the transmitted data and generate an early warning rendering layer; The environmental data includes the humidity inside the GPU chassis, the temperature of the power cable sheath, and the vibration speed of the GPU backplate.

[0008] Furthermore, the process of determining the fuselage operating coefficient corresponding to the early warning monitoring area during data acquisition includes, Obtain the rate of change of hotspot temperature and the rate of change of fan speed in a continuous time series; The ratio of the hot spot temperature change rate to the fan speed change rate is determined as the fuselage operating coefficient.

[0009] Furthermore, the process of determining the external fluctuation coefficient during data acquisition includes, Construct voltage time-domain curves of voltage values ​​relative to time, and construct electromagnetic time-domain curves of electromagnetic values ​​relative to time; Determine the maximum rate of change of instantaneous voltage and the maximum rate of change of electromagnetic intensity within a preset time period; The product of the maximum rate of change of instantaneous voltage and the maximum rate of change of electromagnetic intensity is determined as the external fluctuation coefficient.

[0010] Furthermore, the process of calculating the anomaly representation value of the data includes, The ratio of the fuselage operating coefficient to the baseline fuselage operating coefficient is determined as the operating impact factor; The ratio of the external fluctuation coefficient to the benchmark external fluctuation coefficient is determined as the external influence factor; The weighted sum of the operational influencing factors and the external influencing factors is determined to be the data anomaly characterization value.

[0011] Furthermore, the determination of the acquisition bias during the data acquisition process, wherein, If the data anomaly characterization value is greater than the data anomaly characterization value threshold, then the data acquisition tendency during the data acquisition process is determined to be an abnormal acquisition tendency. If the data anomaly characterization value is less than or equal to the data anomaly characterization value threshold, then the data acquisition tendency during the data acquisition process is determined to be a normal acquisition tendency.

[0012] Furthermore, the process of analyzing the transmission quality coefficient of the data during transmission includes, The bandwidth utilization rate is determined as the ratio of the throughput to the theoretical peak bandwidth. The retry rate is defined as the rate at which the retry count increases per unit time. The ratio of the retry rate to the bandwidth utilization rate is determined as the transmission quality coefficient.

[0013] Furthermore, the process of calculating the coupling influence coefficient includes, The ratio of the internal humidity to the reference internal humidity is determined as the first coupling factor; The ratio of the skin temperature to the reference skin temperature is determined as the second coupling factor; The ratio of the vibration velocity to the reference vibration velocity is determined to be the third coupling factor; The average sum of the first coupling factor, the second coupling factor, and the third coupling factor is determined as the coupling influence coefficient.

[0014] Furthermore, the process of calculating the rendering anomaly feature values ​​includes, The ratio of the transmission quality coefficient to the reference transmission quality coefficient is determined as the transmission impact factor; The ratio of the coupling influence coefficient to the benchmark coupling influence coefficient is determined as the coupling influence factor; The ratio of the data anomaly characterization value to the baseline data anomaly characterization value is determined as the anomaly impact factor; The weighted sum of the transmission influence factor, the coupling influence factor, and the anomaly influence factor is determined to be the rendering anomaly feature value.

[0015] Further, the step of determining the transmission status of the data and whether the data can be used for rendering, wherein, If the rendering anomaly feature value is greater than the rendering anomaly feature value threshold, then the data transmission status is determined to be abnormal, and the data is determined to be unusable for rendering. If the rendering anomaly feature value is less than or equal to the rendering anomaly feature value threshold, then the data transmission status is determined to be normal transmission, and the data is determined to be usable for rendering.

[0016] Further, the calculation of inter-layer similarity is used to determine the transmitted data, wherein, Extract pixel blocks at the same spatial location from each warning layer and calculate the structural similarity between adjacent layers; If the structural similarity is greater than the structural similarity threshold, data transmission is performed to complete layer rendering; If the structural similarity is less than or equal to the structural similarity threshold, the data is marked as abnormal data and removed.

[0017] Compared with existing technologies, this invention analyzes GPU data to determine the chassis operating coefficient and analyzes power data to determine the external fluctuation coefficient, thereby calculating data anomaly characterization values ​​and determining the acquisition tendency during the data acquisition process. For normal acquisition tendencies, it analyzes the transmission quality coefficient of the data during transmission based on the transmitted data, obtains environmental data during the data transmission process to calculate the coupling influence coefficient, and combines the data anomaly characterization values ​​to calculate rendering anomaly feature values ​​to determine the transmission status of the data and whether the data can be used for rendering. For data that can be used for rendering, it calculates the inter-layer similarity of the data to determine the transmitted data and generates an early warning rendering layer. This invention improves the accuracy and reliability of early warnings by performing multi-source influence analysis on the data to determine the rendering data.

[0018] In particular, by analyzing GPU and power data, and considering both internal and external influencing factors, the impact on data acquisition during the process is analyzed. In reality, most data acquisition links are assumed to be ideal and interference-free, directly treating sensor output values ​​as reliable inputs. However, the data acquisition process is affected by both the hardware's own state and external physical fields. When rendering massive amounts of meteorological data at ultra-high speeds, the GPU will be under extreme load for extended periods. At this time, hotspot temperatures often rise irregularly. Fan speeds, limited by mechanical inertia and cooling system design, cannot handle the cooling demands, prematurely triggering GPU throttling or full-speed fan operation. This introduces rendering stutters and power spikes, leading to data acquisition issues. The acquired data contains errors. Furthermore, if voltage data experiences instantaneous fluctuations and other external power facilities are present, the resulting alternating magnetic field will couple into the signal and power lines via electromagnetic induction, further contaminating the current sampling value and electromagnetic reading. This multi-coupling interference of heat, force, and electromagnetic forces leads to serious data errors. If used directly for rendering decisions without analysis, it will cause false alarms, missed alarms, or system oscillations, severely reducing the reliability of the early warning layer. Based on this, this invention considers integrating GPU data and power data to analyze multiple influencing factors in the data acquisition process, in order to determine the acquisition bias in the data acquisition process, providing a data foundation for subsequent analysis of whether the data can be used for rendering, and improving the accuracy and reliability of early warnings.

[0019] In particular, analyzing data transmission quality and environmental factors only for data with a normal acquisition tendency is problematic. In practice, existing technologies often indiscriminately analyze the transmission quality and environmental impact of all acquired data. Even if the data is clearly unreliable due to hardware anomalies or external interference during the acquisition phase, a large amount of computing power is still consumed for subsequent transmission coefficient calculations, coupling impact assessments, and rendering anomaly detection, resulting in wasted computing resources and response delays. At the same time, even for data with a normal acquisition tendency, data loss or delays can still occur during the transmission from the CPU to the GPU memory due to PCIe link signal degradation and a surge in retry counts. Meanwhile, environmental factors can couple with the transmission link, further deteriorating the integrity and timeliness of the data. This can ultimately lead to the inability to identify data that was collected normally but whose transmission was damaged. Sending such damaged data into the rendering pipeline can cause misalignment, screen distortion, or false alarms generated based on erroneous data in the early warning layer. Therefore, this invention considers only further analyzing data that tends to be collected normally. This avoids performing meaningless transmission and environmental layer calculations on invalid data in massive meteorological data scenarios, providing a data foundation for subsequent determination of transmission data, reducing the consumption of redundant computing power, improving rendering response speed and resource utilization efficiency, and improving early warning accuracy and data credibility.

[0020] In particular, for data that can be used for rendering, existing technologies typically send it directly into the fusion pipeline to generate warning layers, assuming that each layer has consistent physical consistency. However, in actual meteorological observations, different layers may have abnormal similarity due to sensor drift, spatiotemporal registration errors, or local interference. For example, adjacent layers may experience abrupt changes in reflectivity or misalignment of warning boundaries at the same spatial location. If these inconsistencies are not verified, these inconsistent layers will be directly superimposed and rendered, resulting in false echoes, broken boundaries, or flickering artifacts in the final warning results. Based on this, the present invention considers calculating layer similarity and retaining only layers with high consistency for fusion rendering, thus achieving data quality maintenance from data acquisition to rendering, improving the accuracy and reliability of warnings. Attached Figure Description

[0021] Figure 1 A schematic diagram illustrating the steps of a GPU-accelerated ultra-high-speed rendering and display method for early weather warning layers, as described in an embodiment of the invention. Figure 2 This is a logic block diagram illustrating the determination of acquisition bias during the data acquisition process, as described in an embodiment of the invention. Figure 3 A logic block diagram for determining the transmission status of the data and whether the data can be used for rendering, as described in an embodiment of the invention. Figure 4 The logical block diagram for calculating the inter-layer similarity of data in an embodiment of the invention to determine the transmitted data is shown below. Detailed Implementation

[0022] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0023] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0024] Please see Figure 1 The diagram illustrates the steps of a GPU-accelerated ultra-high-speed rendering and display method for early weather warning layers, as described in an embodiment of the invention. The GPU-accelerated ultra-high-speed rendering and display method for early weather warning layers of the present invention includes: Step S1: Obtain GPU data, power data, and transmission data corresponding to the early warning monitoring area. The GPU data includes hotspot temperature and fan speed, the power data includes voltage and electromagnetic values, and the transmission data includes throughput and retry count.

[0025] Specifically, the specific scope of the early warning monitoring area is not limited. It can be flexibly defined as a spatial range of different scales according to the needs of the actual application scenario. In a specific implementation plan, the early warning monitoring area can refer to a municipal administrative area. For example, for a local early warning system deployed by a municipal meteorological bureau, the geographical area covered by all meteorological monitoring stations, radar equipment and the GPU accelerated computing cluster in the central computer room under the jurisdiction of the city constitutes the early warning monitoring area in this scenario. Of course, those skilled in the art can also determine it according to the actual situation, as long as it is reasonable.

[0026] Specifically, there are no restrictions on the method of acquiring GPU data. For example, in practice, the hot spot temperature value reported by the internal temperature sensor of the graphics processor and the real-time speed of the cooling fan can be read in real time through the hardware monitoring interface provided by the operating system. Of course, those skilled in the art can also adopt other methods, as long as the required data can be obtained, which will not be elaborated here.

[0027] Specifically, there are no restrictions on the method of acquiring power data. For example, in implementation, voltage values ​​can be read in real time through the baseboard management controller, and electromagnetic values ​​can be acquired in real time through the deployment of power supply cables and broadband electromagnetic sensor modules. Of course, those skilled in the art can also adopt other methods, as long as the required data can be obtained, which will not be elaborated here.

[0028] Specifically, there are no restrictions on the method of acquiring transmitted data. For example, in implementation, the throughput can be read in real time through the bus performance counter in the operating system kernel, and the retry count can be obtained in real time by reading the error statistics register in the bus configuration space. Of course, those skilled in the art can also adopt other methods, as long as the required data can be obtained, which will not be elaborated here.

[0029] It is understood that, in the embodiments, in order to eliminate the influence of different physical dimensions on subsequent calculations and to enable the coefficients to be compared and weighted on a uniform scale, all calculations involving ratios have been normalized.

[0030] Please continue reading. Figure 1 As shown, in step S2, the GPU data is analyzed to determine the fuselage operation coefficient corresponding to the early warning monitoring area during the data acquisition process, the power data is analyzed to determine the external fluctuation coefficient during the data acquisition process, and the data anomaly characterization value is calculated based on the fuselage operation coefficient and the external fluctuation coefficient to determine the acquisition tendency during the data acquisition process. Specifically, the process of determining the aircraft operating coefficient corresponding to the early warning monitoring area during data acquisition includes, Obtain the rate of change of hotspot temperature and the rate of change of fan speed in a continuous time series; The ratio of the hot spot temperature change rate to the fan speed change rate is determined as the fuselage operating coefficient.

[0031] Specifically, this invention does not consider the case where the fan speed change rate is zero, to avoid the denominator being zero and causing invalid calculations. When the fan speed change rate is greater than zero, the ratio reflects the response efficiency of the heat dissipation system to temperature changes. If the ratio is greater than 1, it indicates that the rate of change of hot spot temperature exceeds the rate of change of fan speed, that is, the temperature rise rate exceeds the adjustment range of heat dissipation capacity. At this time, the heat dissipation effect is insufficient, the system is in a state of heat accumulation or heat dissipation lag, and the reliability of data acquisition needs to be monitored. If the ratio is less than or equal to 1, it indicates that the fan speed change rate is sufficient to follow or lead the temperature change, the heat dissipation capacity is sufficient, and the data acquisition process is less affected by heat.

[0032] It should be noted that the chassis performance coefficient only considers abnormal tendencies where the ratio is greater than 1. For cases where the ratio is less than 1, the system assumes that the heat dissipation is normal and no further thermal risk assessment is performed.

[0033] Specifically, the process of determining the external fluctuation coefficients during data acquisition includes, Construct voltage time-domain curves of voltage values ​​relative to time, and construct electromagnetic time-domain curves of electromagnetic values ​​relative to time; Determine the maximum rate of change of instantaneous voltage and the maximum rate of change of electromagnetic intensity within a preset time period; The product of the maximum rate of change of instantaneous voltage and the maximum rate of change of electromagnetic intensity is determined as the external fluctuation coefficient.

[0034] Specifically, there is no limit to the exact length of the preset time period. In practice, it can be set to a shorter duration that matches the GPU data sampling period, such as 10ms. Of course, those skilled in the art can also determine it according to the actual situation, as long as the required data can be obtained. This will not be elaborated further.

[0035] Specifically, the process of calculating the anomaly representation values ​​of data includes, The ratio of the fuselage operating coefficient to the baseline fuselage operating coefficient is determined as the operating impact factor; The ratio of the external fluctuation coefficient to the benchmark external fluctuation coefficient is determined as the external influence factor; The weighted sum of the operational influencing factors and the external influencing factors is determined to be the data anomaly characterization value.

[0036] Specifically, the baseline fuselage operating coefficient is calculated in advance. Several historical fuselage operating coefficients corresponding to normal rendering states are obtained in advance, and the average of each historical fuselage operating coefficient is determined as the baseline fuselage operating coefficient.

[0037] Specifically, the baseline external fluctuation coefficient is calculated in advance. Several historical external fluctuation coefficients corresponding to normal rendering states are obtained in advance, and the average of each historical external fluctuation coefficient is determined as the baseline external fluctuation coefficient.

[0038] Specifically, the sum of the weight coefficients of the operational impact factor and the external impact factor is 1. When configuring the weights, considering that the state of the data is affected by multiple factors, the weight coefficients of the operational impact factor and the external impact factor are both set to 0.5.

[0039] Specifically, by analyzing GPU and power data, and considering both internal and external influencing factors, the study examines the impact of data acquisition on the data. In practice, data acquisition is often assumed to be ideal and interference-free, with sensor output values ​​considered reliable inputs. However, the data acquisition process is affected by both the hardware's own state and external physical fields. When rendering massive amounts of meteorological data at ultra-high speeds, the GPU operates at its maximum load for extended periods. During this time, hotspot temperatures often rise erratically. Fan speeds, limited by mechanical inertia and the cooling system design, cannot meet the cooling demands, prematurely triggering GPU throttling or full-speed fan operation. This introduces rendering stutters and power spikes, hindering data acquisition. The collected data contains errors. Furthermore, if voltage data experiences instantaneous fluctuations and other external power facilities are present, the resulting alternating magnetic field will couple into the signal and power lines via electromagnetic induction, further contaminating the current sampling values ​​and electromagnetic readings. This multi-coupling interference of heat, force, and electromagnetic forces leads to serious data errors. If used directly for rendering decisions without analysis, it will cause false alarms, missed alarms, or system oscillations, severely reducing the reliability of the early warning layer. Based on this, this invention considers integrating GPU data and power data to analyze multiple influencing factors in the data acquisition process, in order to determine the acquisition bias in the data acquisition process, providing a data foundation for subsequent analysis of whether the data can be used for rendering, and improving the accuracy and reliability of early warnings.

[0040] Please see Figure 2 The diagram shown is a logic block diagram for determining the acquisition tendency during the data acquisition process according to an embodiment of the invention. Specifically, determining the acquisition tendency during the data acquisition process, wherein... If the data anomaly characterization value is greater than the data anomaly characterization value threshold, then the data acquisition tendency during the data acquisition process is determined to be an abnormal acquisition tendency. If the data anomaly characterization value is less than or equal to the data anomaly characterization value threshold, then the data acquisition tendency during the data acquisition process is determined to be a normal acquisition tendency.

[0041] Specifically, the data anomaly representation value threshold represents a boundary where data anomalies exist during the acquisition process. It is calculated in advance by acquiring historical data anomaly representation values ​​corresponding to several normal rendering states in advance, and determining the product of each historical data anomaly representation value and the anomaly precision as the data anomaly representation value threshold. The anomaly precision is determined within the interval [0,1]. In practice, in order to balance sensitivity and stability, the anomaly precision is determined to be 0.9.

[0042] Please continue reading. Figure 1As shown, in step S3, based on the normal acquisition tendency of the data, the transmission quality coefficient of the data during the transmission process is analyzed, environmental data during the data transmission process is obtained to calculate the coupling influence coefficient, and the rendering abnormal feature value is calculated in combination with the data abnormal characterization value to determine the transmission status of the data and whether the data can be used for rendering. Specifically, there are no restrictions on the method of acquiring environmental data. For example, in practice, data can be acquired through capacitive humidity sensors, infrared non-contact temperature sensors, and laser Doppler vibration meters. Of course, those skilled in the art can also use other methods, as long as the required data is obtained. This will not be elaborated further.

[0043] Specifically, the process of analyzing the transmission quality coefficient of the data during transmission includes, The bandwidth utilization rate is determined as the ratio of the throughput to the theoretical peak bandwidth. The retry rate is defined as the rate at which the retry count increases per unit time. The ratio of the retry rate to the bandwidth utilization rate is determined as the transmission quality coefficient.

[0044] Specifically, theoretical peak bandwidth refers to the maximum data throughput that the current transmission link can support under ideal physical conditions, which is dynamically determined by reading the system interface in real time.

[0045] Specifically, the process of calculating the coupling effect coefficient includes, The ratio of the internal humidity to the reference internal humidity is determined as the first coupling factor; The ratio of the skin temperature to the reference skin temperature is determined as the second coupling factor; The ratio of the vibration velocity to the reference vibration velocity is determined to be the third coupling factor; The average sum of the first coupling factor, the second coupling factor, and the third coupling factor is determined as the coupling influence coefficient.

[0046] Specifically, the baseline internal humidity is calculated in advance. The historical internal humidity corresponding to several normal rendering states is obtained in advance, and the average value of each historical internal humidity is determined as the baseline internal humidity.

[0047] Specifically, the baseline skin temperature is calculated in advance. Several historical skin temperatures corresponding to normal rendering states are obtained in advance, and the average of each historical skin temperature is determined as the baseline skin temperature.

[0048] Specifically, the baseline vibration velocity is calculated in advance. Several historical vibration velocities under normal rendering conditions are obtained in advance, and the average of each historical vibration velocity is determined as the baseline vibration velocity.

[0049] Specifically, the process of calculating rendering anomaly feature values ​​includes, The ratio of the transmission quality coefficient to the reference transmission quality coefficient is determined as the transmission impact factor; The ratio of the coupling influence coefficient to the benchmark coupling influence coefficient is determined as the coupling influence factor; The ratio of the data anomaly characterization value to the baseline data anomaly characterization value is determined as the anomaly impact factor; The weighted sum of the transmission influence factor, the coupling influence factor, and the anomaly influence factor is determined to be the rendering anomaly feature value.

[0050] Specifically, the baseline transmission quality coefficient is calculated in advance. Several historical transmission quality coefficients corresponding to normal rendering states are obtained in advance, and the average of each historical transmission quality coefficient is determined as the baseline transmission quality coefficient.

[0051] Specifically, the baseline coupling influence coefficient is calculated in advance. The historical coupling influence coefficients corresponding to several normal rendering states are obtained in advance, and the average of each historical coupling influence coefficient is determined as the baseline coupling influence coefficient.

[0052] Specifically, the data anomaly representation values ​​corresponding to the baseline fuselage operating coefficient and the baseline external fluctuation coefficient are the baseline data anomaly representation values.

[0053] Specifically, the sum of the weight coefficients of the transmission impact factor, coupling impact factor, and anomaly impact factor is 1. When configuring the weights, considering that transmission and environmental factors will also have some impact on the acquisition process, the weight coefficients of the transmission impact factor and coupling impact factor are both set to 0.3, and the weight coefficient of the anomaly impact factor is set to 0.4.

[0054] Specifically, analyzing data transmission quality and environmental factors only for data with a normal acquisition tendency is problematic. In practice, existing technologies often indiscriminately analyze the transmission quality and environmental impact of all acquired data. Even if data is clearly unreliable due to hardware anomalies or external interference during the acquisition phase, significant computing power is still consumed for subsequent transmission coefficient calculations, coupling impact assessments, and rendering anomaly detection, resulting in wasted computing resources and response delays. Furthermore, even for data with a normal acquisition tendency, data loss or delays can still occur during the transmission from the CPU to the GPU memory due to PCIe link signal degradation and a surge in retry counts. Furthermore, environmental factors can couple with the transmission link, further deteriorating the integrity and timeliness of data. This can ultimately lead to the inability to identify normally collected but poorly transmitted data, sending such damaged data into the rendering pipeline. This results in misaligned warning layers, screen artifacts, or false alarms generated based on erroneous data. Therefore, this invention focuses on further analyzing data that tends to be normally collected. This avoids performing meaningless transmission and environmental layer calculations on invalid data in massive meteorological data scenarios, providing a data foundation for subsequent determination of transmitted data, reducing redundant computing power consumption, improving rendering response speed and resource utilization efficiency, and enhancing warning accuracy and data reliability. Please see Figure 3 The diagram shown illustrates a logic block diagram for determining the transmission status of data and whether the data can be used for rendering, according to an embodiment of the invention. Specifically, determining the transmission status of the data and whether it can be used for rendering involves... If the rendering anomaly feature value is greater than the rendering anomaly feature value threshold, then the data transmission status is determined to be abnormal, and the data is determined to be unusable for rendering. If the rendering anomaly feature value is less than or equal to the rendering anomaly feature value threshold, then the data transmission status is determined to be normal transmission, and the data is determined to be usable for rendering.

[0055] Specifically, the rendering anomaly feature value threshold represents a boundary where anomalies occur during data transmission. It is calculated in advance by obtaining historical rendering anomaly feature values ​​corresponding to several normal rendering states in advance. The product of the mean of each historical rendering anomaly feature value and the rendering accuracy is determined as the rendering anomaly feature value threshold. The rendering accuracy is determined within the interval [0,1]. In practice, in order to improve the sensitivity of detecting transmission anomalies, the rendering accuracy is determined to be 0.9.

[0056] Please continue reading. Figure 1 As shown, in step S4, for the data that can be used for rendering, the similarity between data layers is calculated to determine the transmitted data and generate an early warning rendering layer. The environmental data includes the humidity inside the GPU chassis, the temperature of the power cable sheath, and the vibration speed of the GPU backplate.

[0057] Please see Figure 4 The diagram shown illustrates a logical block diagram for calculating inter-layer data similarity to determine transmitted data, according to an embodiment of the invention. Specifically, calculating inter-layer data similarity to determine transmitted data involves, wherein... Extract pixel blocks at the same spatial location from each warning layer and calculate the structural similarity between adjacent layers; If the structural similarity is greater than the structural similarity threshold, data transmission is performed to complete layer rendering; If the structural similarity is less than or equal to the structural similarity threshold, the data is marked as abnormal data and removed.

[0058] Specifically, there is no limitation on the method of calculating structural similarity. For example, the SSIM algorithm can be used in implementation to comprehensively calculate the similarity by comparing the brightness, contrast and structure of two image blocks. Of course, those skilled in the art can also use other methods to calculate it, as long as the required data is obtained, which will not be elaborated here.

[0059] Specifically, the structural similarity threshold characterizes whether the data between adjacent layers has a continuous and consistent boundary. It is calculated in advance by obtaining the historical structural similarity of adjacent layers under several normal rendering states in advance, and determining the product of the mean of each historical structural similarity and the structural accuracy as the structural similarity threshold. The structural accuracy is determined in the interval [0, 1]. In practice, in order to balance effective meteorological changes and filter out transmission anomalies, the structural accuracy is set to 0.9.

[0060] Specifically, for data suitable for rendering, existing technologies typically feed it directly into the fusion pipeline to generate warning layers, assuming that each layer has consistent physical uniformity. However, in actual meteorological observations, different layers may exhibit abnormal similarity due to sensor drift, spatiotemporal registration errors, or local interference. For example, adjacent layers may experience abrupt changes in reflectivity or misalignment of warning boundaries at the same spatial location. If these inconsistencies are not verified, these layers will be directly superimposed and rendered, resulting in false echoes, broken boundaries, or flickering artifacts in the final warning results. Based on this, the present invention considers calculating layer similarity and retaining only layers with high consistency for fusion rendering, thus maintaining data quality from data acquisition to rendering and improving the accuracy and reliability of warnings.

[0061] It is understandable that data collection and trend determination, transmission status and rendering feasibility determination, and the generation of early warning layers constitute a tightly coupled and interconnected continuous processing flow, which is inseparable from each other. The absence or interruption of any step will cause the entire early warning monitoring mechanism to fail. Therefore, this method must be executed as a whole in sequence, and cannot be skipped, reversed, or disassembled independently.

[0062] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0063] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A GPU-accelerated ultra-high-speed rendering and display method for early weather warning layers, characterized in that, include: The system acquires GPU data, power data, and transmission data corresponding to the early warning monitoring area. The GPU data includes hotspot temperature and fan speed, the power data includes voltage and electromagnetic values, and the transmission data includes throughput and retry count. The GPU data is analyzed to determine the fuselage operation coefficient corresponding to the early warning monitoring area during the data acquisition process. The power data is analyzed to determine the external fluctuation coefficient during the data acquisition process. Based on the fuselage operation coefficient and the external fluctuation coefficient, the data anomaly characterization value is calculated to determine the acquisition tendency during the data acquisition process. Based on the normal data collection tendency, the transmission quality coefficient of the data during the transmission process is analyzed, environmental data during the data transmission process is obtained to calculate the coupling influence coefficient, and rendering anomaly characteristic value is calculated in combination with the data anomaly characterization value to determine the transmission status of the data and whether the data can be used for rendering. For data that can be used for rendering, calculate the similarity between data layers to determine the transmitted data and generate an early warning rendering layer; The environmental data includes the humidity inside the GPU chassis, the temperature of the power cable sheath, and the vibration speed of the GPU backplate.

2. The method for ultra-high-speed rendering and display of early weather warning layers based on GPU acceleration according to claim 1, characterized in that, The process of determining the fuselage operation coefficient corresponding to the early warning monitoring area during data acquisition includes: Obtain the rate of change of hotspot temperature and the rate of change of fan speed in a continuous time series; The ratio of the hot spot temperature change rate to the fan speed change rate is determined as the fuselage operating coefficient.

3. The method for ultra-high-speed rendering and display of early weather warning layers based on GPU acceleration according to claim 1, characterized in that, The process of determining the external fluctuation coefficient during data acquisition includes, Construct voltage time-domain curves of voltage values ​​relative to time, and construct electromagnetic time-domain curves of electromagnetic values ​​relative to time; Determine the maximum rate of change of instantaneous voltage and the maximum rate of change of electromagnetic intensity within a preset time period; The product of the maximum rate of change of instantaneous voltage and the maximum rate of change of electromagnetic intensity is determined as the external fluctuation coefficient.

4. The method for ultra-high-speed rendering and display of early weather warning layers based on GPU acceleration according to claim 1, characterized in that, The process of calculating the anomaly representation value of the data includes, The ratio of the fuselage operating coefficient to the baseline fuselage operating coefficient is determined as the operating impact factor; The ratio of the external fluctuation coefficient to the benchmark external fluctuation coefficient is determined as the external influence factor; The weighted sum of the operational influencing factors and the external influencing factors is determined to be the data anomaly characterization value.

5. The method for ultra-high-speed rendering and display of early weather warning layers based on GPU acceleration according to claim 1, characterized in that, The determination of the acquisition tendency during the data acquisition process, wherein... If the data anomaly characterization value is greater than the data anomaly characterization value threshold, then the data acquisition tendency during the data acquisition process is determined to be an abnormal acquisition tendency. If the data anomaly characterization value is less than or equal to the data anomaly characterization value threshold, then the data acquisition tendency during the data acquisition process is determined to be a normal acquisition tendency.

6. The method for ultra-high-speed rendering and display of early weather warning layers based on GPU acceleration according to claim 1, characterized in that, The process of analyzing the transmission quality coefficient of the data during transmission includes, The bandwidth utilization rate is determined as the ratio of the throughput to the theoretical peak bandwidth. The retry rate is defined as the rate at which the retry count increases per unit time. The ratio of the retry rate to the bandwidth utilization rate is determined as the transmission quality coefficient.

7. The method for ultra-high-speed rendering and display of early weather warning layers based on GPU acceleration according to claim 1, characterized in that, The process of calculating the coupling influence coefficient includes: The ratio of the internal humidity to the reference internal humidity is determined as the first coupling factor; The ratio of the skin temperature to the reference skin temperature is determined as the second coupling factor; The ratio of the vibration velocity to the reference vibration velocity is determined to be the third coupling factor; The average sum of the first coupling factor, the second coupling factor, and the third coupling factor is determined as the coupling influence coefficient.

8. The method for ultra-high-speed rendering and display of early weather warning layers based on GPU acceleration according to claim 1, characterized in that, The process of calculating rendering anomaly feature values ​​includes: The ratio of the transmission quality coefficient to the reference transmission quality coefficient is determined as the transmission impact factor; The ratio of the coupling influence coefficient to the benchmark coupling influence coefficient is determined as the coupling influence factor; The ratio of the data anomaly characterization value to the baseline data anomaly characterization value is determined as the anomaly impact factor; The weighted sum of the transmission influence factor, the coupling influence factor, and the anomaly influence factor is determined to be the rendering anomaly feature value.

9. The method for ultra-high-speed rendering and display of early weather warning layers based on GPU acceleration according to claim 1, characterized in that, The process involves determining the transmission status of the data and identifying whether the data can be used for rendering. If the rendering anomaly feature value is greater than the rendering anomaly feature value threshold, then the data transmission status is determined to be abnormal, and the data is determined to be unusable for rendering. If the rendering anomaly feature value is less than or equal to the rendering anomaly feature value threshold, then the data transmission status is determined to be normal transmission, and the data is determined to be usable for rendering.

10. The method for ultra-high-speed rendering and display of early weather warning layers based on GPU acceleration according to claim 1, characterized in that, The calculation of inter-layer similarity is used to determine the transmitted data, wherein, Extract pixel blocks at the same spatial location from each warning layer and calculate the structural similarity between adjacent layers; If the structural similarity is greater than the structural similarity threshold, data transmission is performed to complete layer rendering; If the structural similarity is less than or equal to the structural similarity threshold, the data is marked as abnormal data and removed.

Images