Multi-source data processing method and system for dynamic monitoring of land surface

Abstract

A multi-source data processing method and system for dynamic monitoring of land surface. According to the working state of all monitoring ends arranged in the space where the land is located, the task allocation of all monitoring ends is determined; according to the actual running state of the cloud end accessed by all monitoring ends, an elastic resource pool is constructed to determine the resource allocation of the monitoring end; a benchmark database for land surface monitoring is constructed, and data verification is performed on the benchmark database according to the multi-source data generated during the execution of the monitoring task; feature recognition and screening are performed on the benchmark database that has completed data verification to obtain a plurality of credible multi-source data, so as to generate a land surface live state representation map, and the land surface live state representation map is also subjected to element change to obtain a target representation map and upload it to the cloud end. Through elastic resource allocation of the monitoring end executing different tasks, continuous and stable monitoring and acquisition of multi-source data are ensured, and the reliability and effectiveness of land surface monitoring are improved.

Description

Technical Field

[0001] This invention relates to the field of land monitoring, and in particular to a multi-source data processing method and system for dynamic monitoring of land surface. Background Technology

[0002] Land resources are characterized by wide spatial distribution, significant local variations, and susceptibility to external environmental influences. To ensure the rational development and utilization of land resources and to promptly and comprehensively monitor changes in land status during land construction, dynamic monitoring of the land, especially its surface, is necessary. During land surface monitoring, multi-source data on different land surface states are generated. This multi-source data is heterogeneous and large in volume, and different types of multi-source data require different resources for acquisition. To ensure the stability and continuity of multi-source data monitoring, resources are evenly allocated to all monitoring equipment, and real-time monitored multi-source data is directly uploaded, stored, and processed to obtain a real-time characterization of the land surface. However, the above data monitoring and processing methods do not fully consider the differences between different multi-source data. Evenly allocating monitoring resources may lead to missing or erroneous multi-source data, and the lack of data verification reduces the reliability and effectiveness of land surface monitoring, resulting in inaccurate land monitoring results. Summary of the Invention

[0003] Existing land surface dynamic monitoring methods cannot adjust monitoring resource allocation based on multi-source data monitoring real-time conditions, nor can they verify and filter multi-source data, thus reducing the reliability and effectiveness of land surface monitoring and failing to obtain accurate land monitoring results. This invention proposes a multi-source data processing method and system for land surface dynamic monitoring to overcome or at least partially solve the aforementioned problems.

[0004] This disclosure provides several embodiments of a multi-source data processing method for dynamic monitoring of land surface, including:

[0005] Based on the working status of all monitoring terminals deployed within the land area, determine the task allocation for all monitoring terminals; based on the actual operating status of the cloud accessed by all monitoring terminals, construct an elastic resource pool;

[0006] The monitoring task instructions received by the monitoring terminal are parsed to determine the resource allocation of the elastic resource pool to the monitoring terminal.

[0007] A benchmark database for land surface monitoring is constructed, and the benchmark database is validated based on multi-source data generated during the monitoring tasks performed by the monitoring terminal.

[0008] Feature identification and filtering are performed on the benchmark database that has completed the data verification to obtain several reliable multi-source data;

[0009] The aforementioned reliable multi-source data undergoes temporal and spatial correlation processing to obtain a land surface condition representation map; based on the query request, the elements of the land surface condition representation map are modified to obtain a target representation map, which is then uploaded to the cloud.

[0010] Optionally, based on the operational status of all monitoring terminals deployed within the land area, the task allocation for all monitoring terminals is determined, including:

[0011] The historical working status records and actual working status records of all monitoring terminals deployed within the land area are obtained. The monitoring data change trends of the monitoring terminals during the historical monitoring period are extracted from the historical working status records. The monitoring data change trends include the correlation change trends between the monitoring data error rate and monitoring data redundancy of the monitoring terminals during the historical monitoring period and the generated monitoring data types and their data volume.

[0012] Extract the real-time generated monitoring data type and its volume from the work record; compare the real-time generated monitoring data type and its volume with the monitoring data change trend to estimate the time point in the future when the monitoring terminal will be in an unreliable state;

[0013] Based on the temporal relationship between the time points when each monitoring terminal is in an untrusted state, the task allocation time order for all monitoring terminals is determined.

[0014] Optionally, based on the actual operational status of all monitoring terminals connected to the cloud, an elastic resource pool is constructed, including:

[0015] Obtain the real-time CPU operating status of all monitoring terminals connected to the cloud, and determine the spare CPU resources of the cloud in maintaining the preset working mode based on the real-time CPU operating status.

[0016] Based on the available CPU resources, construct an elastic resource pool and adjust the actual resource capacity of the elastic resource pool.

[0017] Optionally, parsing the monitoring task instructions received by the monitoring terminal and determining the resource allocation of the elastic resource pool to the monitoring terminal includes:

[0018] Multimodal task parameters are obtained by parsing the monitoring task instructions received from the monitoring terminal; wherein, the multimodal task parameters include spatial range parameters of the monitoring area on the land surface, monitoring accuracy parameters, and monitoring time window parameters;

[0019] Based on the multimodal task parameters, the workload variation range during the monitoring task corresponding to the monitoring task instruction executed by the monitoring terminal is determined; based on the workload variation range, the resource allocation of the elastic resource pool to the monitoring terminal is determined; wherein, the resource allocation includes CPU resource allocation.

[0020] Optionally, a benchmark database of land surface monitoring may be constructed, including:

[0021] A set of historical monitoring data on the land surface is obtained. Based on the label of each historical monitoring data in the set, all historical monitoring data in the set are preprocessed and indexed. The corresponding historical monitoring data are then used to construct a benchmark database. The preprocessing includes standardization preprocessing. The indexing includes adding unique index tags to the historical monitoring data.

[0022] Optionally, the benchmark database is validated based on multi-source data generated during the monitoring task performed by the monitoring terminal, including:

[0023] The multi-source data generated during the monitoring task performed by the monitoring terminal is divided into several single-modal data, and the single-modal data is standardized and preprocessed.

[0024] By comparing the historical monitoring data with the same type of attributes as the single-modal data in the benchmark database with the single-modal data, the deviation between the single-modal data and the historical monitoring data is determined.

[0025] Based on the deviation, the historical monitoring data of the benchmark database are verified and corrected.

[0026] Optionally, feature identification and filtering are performed on the benchmark database that has completed the data verification to obtain several reliable multi-source data, including:

[0027] Feature identification is performed on the benchmark database that has completed the data verification to obtain the physical quantity characteristics, regional characteristics, and temporal characteristics of all monitoring data in the benchmark database that has completed the data verification.

[0028] The physical quantity characteristics, spatial characteristics, and temporal characteristics of the same type of monitoring data with the same physical quantity attributes are analyzed to obtain the spatial and temporal fluctuations of the same type of monitoring data. Based on the spatial and temporal fluctuations, several reliable data are selected from the same type of monitoring data, and the several reliable data selected from the same type of monitoring data with multiple physical quantity attributes are integrated into several reliable multi-source data.

[0029] Optionally, temporal and spatial correlation processing is performed on the aforementioned reliable multi-source data to obtain a land surface condition representation map, including:

[0030] Based on the respective time points and spatial points of the aforementioned trusted multi-source data, determine the degree of temporal correlation and spatial correlation between any two trusted multi-source data among the aforementioned trusted multi-source data.

[0031] Based on the temporal and spatial correlation, the several reliable multi-source data are arranged and mapped to obtain a land surface condition representation map.

[0032] Optionally, based on the query request, the elements of the land surface condition representation map are modified to obtain the target representation map and uploaded to the cloud, including:

[0033] The query target object is obtained by parsing the query request; based on the query target object, the map elements in the map that meet the preset related conditions with the query target object are determined to represent the actual state of the land surface.

[0034] Based on the distribution of the map elements on the land surface in terms of their actual state, the visualization state of the map elements is changed to obtain the target representation map and upload it to the cloud.

[0035] This invention also provides a multi-source data processing system for dynamic monitoring of land surface, comprising:

[0036] The resource pool construction module is used to determine the task allocation for all monitoring terminals based on their working status within the land area; and to construct an elastic resource pool based on the actual operating status of the cloud accessed by all monitoring terminals.

[0037] The resource allocation module is used to parse the monitoring task instructions received by the monitoring terminal and determine the resource allocation of the elastic resource pool to the monitoring terminal.

[0038] The data verification module is used to construct a benchmark database for land surface monitoring and to verify the benchmark database based on multi-source data generated during the monitoring tasks performed by the monitoring terminal.

[0039] The identification and filtering module is used to perform feature identification and filtering on the benchmark database that has completed the data verification, and to obtain several reliable multi-source data.

[0040] The data processing module is used to perform time-domain and spatial-domain correlation processing on the several trusted multi-source data to obtain a land surface condition representation map; according to the query request, the elements of the land surface condition representation map are modified to obtain the target representation map and uploaded to the cloud.

[0041] The beneficial effects of the above-mentioned technical solutions provided in the embodiments of the present invention include at least the following:

[0042] This invention provides a multi-source data processing method and system for dynamic land surface monitoring. The method determines task allocation for all monitoring terminals deployed within the land area based on their operational status. It also constructs an elastic resource pool based on the actual cloud operation status of all monitoring terminals, thereby determining resource allocation for the monitoring terminals. A benchmark database for land surface monitoring is built, and the database is validated using multi-source data generated during monitoring tasks. The validated benchmark database undergoes feature identification and filtering to obtain several reliable multi-source data sets, which are used to generate a land surface condition representation map. Furthermore, the map's elements are modified to obtain a target representation map, which is then uploaded to the cloud. By flexibly allocating resources to monitoring terminals performing different tasks, continuous and stable acquisition of multi-source data is ensured. The validation and filtering of multi-source data improves its reliability, enhances the reliability and effectiveness of land surface monitoring, and ensures accurate land monitoring results.

[0043] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0044] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0045] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0046] Figure 1 This is a flowchart illustrating the multi-source data processing method for dynamic monitoring of land surface provided in this embodiment of the invention.

[0047] Figure 2 for Figure 1 A detailed flowchart of a portion of step S100.

[0048] Figure 3 for Figure 1 The detailed flowchart of another part of step S100.

[0049] Figure 4 for Figure 1 The detailed flowchart of step S200.

[0050] Figure 5 for Figure 1 A detailed flowchart of a specific instance of step S200.

[0051] Figure 6 for Figure 5 The detailed flowchart of step S203.

[0052] Figure 7 for Figure 1 The detailed flowchart of step S300.

[0053] Figure 8 for Figure 1 The detailed flowchart of step S400.

[0054] Figure 9 for Figure 1 A detailed flowchart of a portion of steps in step S500.

[0055] Figure 10 for Figure 1 The detailed flowchart of another part of step S500.

[0056] Figure 11 This is a schematic diagram of the structure of a multi-source data processing system for dynamic monitoring of land surface provided in an embodiment of the present invention. Detailed Implementation

[0057] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0058] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," "outer," "far," "near," "front," and "rear," etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the accompanying drawings and are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0059] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0060] Please see Figure 1 As shown, an embodiment of this application provides a multi-source data processing method for dynamic monitoring of land surface. This multi-source data processing method for dynamic monitoring of land surface includes the following steps:

[0061] S100: Determine the task allocation for all monitoring terminals based on their working status within the land area; construct an elastic resource pool based on the actual operating status of the cloud accessed by all monitoring terminals.

[0062] S200: Parse the monitoring task instructions received by the monitoring terminal and determine the resource allocation of the elastic resource pool to the monitoring terminal;

[0063] S300: Construct a benchmark database for land surface monitoring, and verify the benchmark database based on multi-source data generated during the monitoring tasks performed by the monitoring terminals;

[0064] S400: Perform feature identification and filtering on the benchmark database that has completed data verification to obtain several reliable multi-source data;

[0065] S500: Performs temporal and spatial correlation processing on several reliable multi-source data to obtain a land surface condition representation map; according to the query request, modifies the elements of the land surface condition representation map to obtain the target representation map and uploads it to the cloud.

[0066] The beneficial effects of the above embodiments are as follows: This multi-source data processing method for dynamic land surface monitoring determines the task allocation for all monitoring terminals deployed within the land space based on their working status; it constructs an elastic resource pool based on the actual operating status of the cloud accessed by all monitoring terminals, thereby determining the resource allocation for the monitoring terminals; it constructs a benchmark database for land surface monitoring, and verifies the benchmark database based on the multi-source data generated during the execution of monitoring tasks; it performs feature identification and filtering on the benchmark database after data verification to obtain several reliable multi-source data, thereby generating a land surface condition representation map; it also modifies the elements of the land surface condition representation map to obtain a target representation map and uploads it to the cloud. By flexibly allocating resources to monitoring terminals performing different tasks, it ensures continuous and stable acquisition of multi-source data; it also verifies and filters multi-source data, improving the reliability of multi-source data, enhancing the reliability and effectiveness of land surface monitoring, and ensuring accurate land monitoring results.

[0067] In another embodiment, such as Figure 2 As shown, step S100, which determines the task allocation for all monitoring terminals based on their operational status within the land area, includes the following steps:

[0068] S111: Obtain the historical working status records and actual working status records of all monitoring terminals deployed within the land space, and extract the monitoring data change trends of the monitoring terminals during the historical monitoring period from the historical working status records; wherein, the monitoring data change trends include the correlation change trends between the monitoring data error rate and monitoring data redundancy of the monitoring terminals during the historical monitoring period and the generated monitoring data types and their data volume.

[0069] S112: Extract the real-time generated monitoring data types and their data volume from the work status records; compare the real-time generated monitoring data types and their data volume with the monitoring data change trend to estimate the time points in the future when the monitoring terminal will be in an unreliable state.

[0070] S113: Determine the time sequence for task allocation to all monitoring terminals based on the temporal relationship between the time points when each monitoring terminal is in an untrusted state.

[0071] The beneficial effects of the above embodiments are that, in actual operation, to meet the monitoring needs of different modal factors on the land surface (such as weather factors, soil factors, etc.), different types of monitoring terminals (such as different types of sensors) are set up at different locations in the land space. Each monitoring terminal operates independently and generates corresponding monitoring data. It is understood that before each monitoring terminal operates, a corresponding monitoring task is assigned to each monitoring terminal, and the monitoring terminal will perform monitoring operations according to the assigned monitoring task content. Whether the monitoring terminal can accurately complete the assigned monitoring task depends on its own working performance status. The aforementioned working performance status may include, but is not limited to, the error rate and redundancy of the monitoring data generated during the execution of the monitoring task. Generally speaking, the worse the working performance status of the monitoring terminal, the higher the error rate and redundancy of the generated monitoring data. To accurately determine the working performance status of the monitoring terminals, the historical working status records and actual working status records of all monitoring terminals deployed in the land space are first obtained. From the historical working status records, the correlation trend between the monitoring data error rate and monitoring data redundancy of the monitoring terminals during the historical monitoring period and the types and amounts of generated monitoring data is extracted. It is understandable that the aforementioned correlation trend refers to the change in the amount of data with errors and redundancy in each type of monitoring data generated by the monitoring terminal during the historical monitoring period over time.

[0072] Next, the real-time generated monitoring data types and their volumes are extracted from the work status records. There is a correlation between the real-time working status and historical working status of the monitoring terminal; that is, the generation of error and redundant data corresponding to the real-time working status is correlated with the generation of error and redundant data corresponding to the historical working status (e.g., temporal correlation). Therefore, by comparing the real-time generated monitoring data types and their volumes with the monitoring data change trends, the estimated time points in the future when the monitoring terminal will be in an unreliable state are determined. These time points refer to the time points when the error rate of the monitoring data generated by the monitoring terminal in the future exceeds a preset error rate threshold and the redundancy exceeds a preset redundancy threshold. Then, based on the temporal relationship between the time points when each monitoring terminal is in an unreliable state, the task allocation time order for all monitoring terminals is determined. That is, monitoring tasks are not assigned to the monitoring terminals when they are in an unreliable state, and monitoring tasks are assigned to the monitoring terminals when they are not in an unreliable state. This ensures that the monitoring terminals only receive monitoring tasks when they are in a reliable state, guaranteeing the reliability of the monitoring task execution.

[0073] In another embodiment, such as Figure 3 As shown, step S100, which involves constructing an elastic resource pool based on the actual cloud operating status of all monitoring terminals, includes the following steps:

[0074] S121: Obtain the real-time CPU operating status of all monitoring terminals connected to the cloud, and determine the spare CPU resources of the cloud in maintaining the preset working mode based on the real-time CPU operating status.

[0075] S122: Based on the available CPU resources, construct an elastic resource pool and adjust the actual resource capacity of the elastic resource pool.

[0076] The beneficial effects of the above embodiments are that the monitoring terminal requires CPU resources to perform monitoring tasks. When the monitoring terminal receives a monitoring task, the cloud allocates CPU resources to it, and the size of the allocated CPU resources directly affects the performance of the monitoring task. Generally speaking, the larger the allocated CPU resources, the more efficiently the monitoring terminal can process the monitoring task. Considering that the total amount of CPU resources in the cloud is limited, in order to ensure that each monitoring terminal is allocated relatively sufficient CPU resources, it is necessary to first build an elastic resource pool to manage and allocate the CPU resources in the cloud in a unified manner. Specifically, the real-time CPU running status of all monitoring terminals connected to the cloud is first obtained. Based on the real-time CPU running status, the spare CPU resources of the cloud in maintaining a preset working mode are determined. The preset working mode refers to the working mode corresponding to the cloud maintaining its minimum operating requirements (such as a preset minimum amount of computation). The spare CPU resources are obtained by subtracting the amount of CPU resources required to maintain the preset working mode from the total amount of CPU resources of the cloud itself. The spare CPU resources are then placed in the elastic resource pool, thereby using the elastic resource pool to flexibly allocate CPU resources to the monitoring terminals.

[0077] In another embodiment, such as Figure 4 As shown, step S200: Parsing the monitoring task instruction received by the monitoring terminal and determining the resource allocation of the elastic resource pool to the monitoring terminal includes the following steps:

[0078] S210: The monitoring task instructions received from the monitoring terminal are parsed to obtain multimodal task parameters; among which, the multimodal task parameters include the spatial range parameters of the monitoring area on the land surface, the monitoring accuracy parameters, and the monitoring time window parameters;

[0079] S220: Based on the multimodal task parameters, determine the workload variation range during the monitoring task period corresponding to the monitoring terminal executing the monitoring task instruction; based on the workload variation range, determine the resource allocation of the elastic resource pool to the monitoring terminal; wherein, the resource allocation includes CPU resource allocation.

[0080] The beneficial effects of the above embodiments are as follows: Different monitoring tasks assigned to the monitoring terminal result in different CPU resources required during the execution of those tasks. To ensure the monitoring terminal has sufficient CPU resources to execute the monitoring tasks without experiencing CPU resource surplus, it is necessary to predict the workload during the execution of the assigned monitoring tasks. This prediction serves as the basis for CPU resource allocation. Therefore, the monitoring task instructions received from the monitoring terminal are parsed to obtain parameters such as the spatial range of the monitoring area (e.g., the area of ​​the monitoring area), monitoring accuracy parameters (e.g., the minimum monitoring resolution for weather factor monitoring, soil condition monitoring, and topographic monitoring), and monitoring time window parameters (e.g., the duration of the monitoring task execution). Based on these multimodal task parameters, the workload variation range during the monitoring task corresponding to the monitoring terminal's execution of the monitoring task instructions is determined. This determines the CPU resource allocation of the elastic resource pool to the monitoring terminal, ensuring that the monitoring terminal can obtain accurate CPU resources from the elastic resource pool to maintain the normal execution of its monitoring tasks.

[0081] The following example explains steps S200, namely S210 and S220, as follows: Figure 5 As shown, the above technical solution, which "parses the monitoring task instructions received from the monitoring terminal to obtain multimodal task parameters; wherein, the multimodal task parameters include spatial range parameters of the monitoring area on the land surface, monitoring accuracy parameters, and monitoring time window parameters; based on the multimodal task parameters, determines the workload variation range during the monitoring task corresponding to the monitoring terminal executing the monitoring task instructions; based on the workload variation range, determines the resource allocation of the elastic resource pool to the monitoring terminal; wherein, the resource allocation includes CPU resource allocation," can be specifically implemented as follows:

[0082] S201: Extract the following three types of parameters from the monitoring task instruction received from the monitoring terminal:

[0083] The monitoring area spatial range parameter A, whose unit is, for example, m², represents the land surface area to be monitored; the monitoring accuracy parameter P, whose unit is, for example, m / pixel, represents the spatial resolution of image or data acquisition; and the monitoring time window parameter T, whose unit is, for example, hours, represents the duration of task execution.

[0084] These three parameters together determine the complexity, data volume, and execution intensity of the monitoring task, and form the basis for subsequent workload calculations.

[0085] S 202: Based on the monitoring area spatial range parameter A, monitoring accuracy parameter P, and monitoring time window parameter T, determine the normalized task complexity index C_task of the monitoring terminal according to the first preset algorithm;

[0086] This step S202 aims to transform the macroscopic parameters of the task into a unified, quantitative complexity metric to measure its inherent processing difficulty.

[0087] The first preset algorithm mentioned above can be implemented as the following formula (1):

[0088] (1) Where C_task is the normalized task complexity index, which is dimensionless; the higher the value of C_task, the more demanding the task. This represents the preset reference monitoring area, which is a constant that can be preset by the system. It is the area normalization ratio; This represents the reference monitoring accuracy and is a preset constant of the system, such as 1 m / pixel. It is the accuracy normalization ratio; This is a reference time window, a constant that can be preset by the system; It is the time normalization ratio.

[0089] S203: Based on the normalized task complexity index, the surface characteristics of the monitoring area, and

[0090] Based on illumination conditions, the dynamic workload of the monitoring terminal for fusion of surface and illumination sensing is predicted according to the preset fusion algorithm model.

[0091] The actual computational load at the monitoring end depends not only on the task itself, but is also profoundly affected by the surface characteristics and lighting conditions of the monitoring area. Step 3 in this section uses a multi-level fusion algorithm model to achieve forward-looking prediction of the workload.

[0092] like Figure 6 As shown, step S203 may specifically include the following steps:

[0093] S2031: Based on the normalized task complexity index, the baseline computing power required by the monitoring terminal to process one unit of normalized task complexity, and the land cover complexity index, calculate the baseline load weighted by the land cover characteristics according to the second preset algorithm.

[0094] The second preset algorithm can be implemented, for example, as the following formula (2):

[0095] (2) Where W_base' represents the baseline load weighted by surface characteristics, in units of GOPS. This is a static load that takes into account surface complexity on the original baseline. C_surface represents the surface cover complexity index, which is dimensionless and ranges from [0,1]. Its value can be obtained from land cover classification maps or baseline databases formed by historical monitoring. For example, water surface is 0.1, farmland is 0.3, forest is 0.6, and urban area is 0.9. This represents the baseline coefficient for calculating efficiency, and indicates the normalized task complexity (i.e., the time complexity of processing one unit at the monitoring end) The benchmark computing power required for a task is measured in GOPS. It is an inherent performance parameter of the monitoring terminal hardware system and built-in processing algorithm. Its standard value can be obtained through the factory benchmark test of this model of monitoring terminal. The surface complexity influence coefficient has a value range of [0.1, 0.5] and can be obtained through regression analysis of historical data. It is used to quantify the amplification intensity of surface complexity on computational load.

[0096] This step S2031 refines the spatial dimension, acknowledging that monitoring a city and monitoring a lake, even if tasks A, P, and T are the same, have significantly different basic computational requirements.

[0097] S2032: Determine the light efficiency attenuation factor according to the third preset algorithm based on the solar altitude angle, the preset effective light threshold angle, the preset low light indication function, and the preset light attenuation weight coefficient.

[0098] The third preset algorithm can be implemented as the following formula (3):

[0099] (3)

[0100] Here, K_light(t) represents the light efficiency attenuation factor relative to time t within the task, which simulates the effect of insufficient light causing data quality degradation, thus requiring additional computing resources for compensation; where t ranges from 0 to the monitoring time window parameter T, representing the relative time calculated from the start of the monitoring task. θ_sun(t) represents the solar altitude angle relative to time t within the task, which can be calculated in real time using astronomical algorithms based on the latitude, longitude, date, and time of the monitoring point; θ_threshold represents the preset effective light threshold angle, for example, 20°. When the solar altitude angle is below this value, insufficient light is considered to begin to significantly affect data quality. This represents the preset low-light indicator function, which takes the value 0 or 1. When θ_sun(t) < θ_threshold, I = 1, otherwise it is 0. It is used to activate the compensation mechanism.

[0101] The preset illumination attenuation weighting coefficient has a value range of [0.2, 0.8] and is used to control the intensity of illumination compensation.

[0102] This step S2032 performs environmental perception in the time dimension, which can predict the periodic additional computing needs caused by insufficient light in the early morning, at dusk, or in high-latitude winter.

[0103] S2033: Based on the baseline load weighted by surface characteristics, the inherent load fluctuation coefficient, and the light efficiency attenuation factor, the comprehensive dynamic workload of the monitoring terminal is determined according to the fourth preset algorithm.

[0104] The fourth preset algorithm can be implemented as the following formula (4):

[0105] (4) Where W(t) represents the comprehensive dynamic workload within the task relative to time t, in units of

[0106] For GOPS;

[0107] A_task(t) represents the task activity factor of the monitoring terminal within a task relative to time t. It is a dimensionless coefficient used to characterize the busyness of data processing activities of the monitoring terminal in different task stages. Its value assignment logic is as follows: According to the standard workflow of the monitoring task, the task is divided into three stages: startup self-check, stable monitoring and data processing, and data packaging and uploading. The task activity factor A_task(t) is assigned a value based on the task stage within the task relative to time t. Specifically, in the stable monitoring and data processing stage, A_task(t) takes a baseline value of 1.0; in the startup self-check stage, A_task(t) takes a value less than 1.0; and in the data packaging and uploading stage, A_task(t) takes a value greater than 1.0. The acquisition method is as follows: the specific values ​​of each stage and the stage boundary time points are determined by statistical analysis and regression analysis of the CPU load data of this model of monitoring terminal when executing historical tasks.

[0108] The principle of the fourth preset algorithm is: the final load (comprehensive dynamic workload) is coupled by the static benchmark weighted by the ground surface, the inherent task execution rhythm and the performance compensation determined by the natural environment lighting conditions. max(1, K_light(t)) ensures that the load is amplified only when the lighting is insufficient (K_light(t) > 1).

[0109] S204: Based on the comprehensive dynamic workload of the monitoring terminal, the single-core processing capability of the cloud server, the weather complexity and terrain undulation of the monitoring area, determine the resource allocation for the monitoring terminal according to the preset coupling model.

[0110] This step further couples the predicted fine-grained dynamic workload with real-time weather and macro-topographical factors, completing the process from demand forecasting to resource allocation.

[0111] The preset coupling model can be implemented as follows (5):

[0112] (5)

[0113] Where R_cpu(t) represents the number of CPU cores allocated to the monitoring terminal at relative time t within the task; E_cpu represents the processing power of a single physical CPU core of the cloud server, in GOPS / core, determined by the cloud server specifications; C_weather represents the weather complexity factor at relative time t within the task, dimensionless, obtained and quantified from the meteorological API, such as: sunny = 0, thin clouds = 1, rain = 3; α is the weather influence weighting coefficient at relative time t within the task, obtained through regression of historical data; ΔTerrain is the topographic relief of the monitoring area, which can be calculated from the digital elevation model (DEM) (such as the normalized value of altitude range); β is the topographic influence weighting coefficient of the monitoring area, which can be obtained through regression of historical data; R_buffer is the preset number of buffer resources, in cores, used to cope with unforeseen emergencies.

[0114] This formula serves as the decision point for resource allocation, ensuring that the allocated resources simultaneously meet the influence of five key factors: task requirements, surface characteristics, lighting conditions, real-time weather, and macro-topography, thus achieving intelligent and flexible allocation.

[0115] The steps S201 to S204 above, through an interlocking decision chain, realize intelligent decision-making for resource allocation, enabling the land surface dynamic monitoring system to have situational awareness and to predictively allocate sufficient resources for complex scenarios and adverse conditions. This solves the problem of missing or degraded monitoring data caused by rigid resource allocation in traditional methods, and greatly improves the reliability, accuracy and intelligence of the entire system.

[0116] In another embodiment, a benchmark database for land surface monitoring is constructed, specifically including:

[0117] A set of historical monitoring data on the land surface is obtained. Based on the label of each historical monitoring data in the set, all historical monitoring data in the set are preprocessed and indexed. The corresponding historical monitoring data are then used to construct a benchmark database. The preprocessing includes standardization preprocessing, and the indexing includes adding unique index tags to the historical monitoring data.

[0118] The beneficial effects of the above embodiments are that dynamic monitoring of the land surface generates various types of monitoring data. In order to unify the identification of all monitoring data and improve the traceability of monitoring data, a set of historical monitoring data of the land surface is obtained. Based on the tags of each historical monitoring data in the historical monitoring data set (such as the format tag and the location tag of the historical monitoring data), all historical monitoring data in the historical monitoring data set are preprocessed (format standardization preprocessing) and indexed (location indexing processing). The corresponding historical monitoring data are then used to construct a benchmark database to ensure the data standardization of the benchmark database and provide a data foundation for subsequent determination of monitoring data offset.

[0119] In another embodiment, such as Figure 7 As shown, the baseline database is validated based on multi-source data generated during the monitoring task, including the following steps:

[0120] S310: Divide the multi-source data generated during the monitoring task performed by the monitoring terminal into several single-modal data, and perform standardized preprocessing on the single-modal data;

[0121] S320: Compare historical monitoring data with the same type of attributes as the single-modal data in the benchmark database with the single-modal data to determine the deviation between the single-modal data and the historical monitoring data;

[0122] S330: Based on the deviation, verify and correct the historical monitoring data of the benchmark database.

[0123] The beneficial effects of the above embodiments are considered in light of the inevitable deviations in monitoring data generated during real-time monitoring tasks due to external factors, which render the monitoring data unusable. Considering the massive amounts of similar monitoring data generated by the monitoring terminal itself during historical monitoring, these similar monitoring data generally exhibit a strong normal trend, providing significant reference value for subsequent judgments regarding whether deviations exist in the real-time generated monitoring data. To accurately identify and correct deviations in the monitoring data, the multi-source data generated during the monitoring tasks are first categorized into several single-modal data sets, and these single-modal data sets undergo standardization preprocessing. The single-modal data refers to monitoring data generated during the monitoring process related to any of the land's weather, soil, or topography. The standardization preprocessing refers to standardizing and transforming the single-modal data. Then, historical monitoring data with the same type of attributes as the single-modal data in the benchmark database are compared with the single-modal data to determine the deviation between the single-modal data and the historical monitoring data. This allows for the verification and correction of the historical monitoring data in the benchmark database, improving the accuracy and reliability of the benchmark database.

[0124] In another embodiment, such as Figure 8As shown, feature identification and filtering are performed on the benchmark database that has completed data verification to obtain several reliable multi-source data sets, including the following steps:

[0125] S410: Perform feature identification on the benchmark database that has completed data verification to obtain the physical quantity characteristics, regional characteristics and temporal characteristics of all monitoring data in the benchmark database that has completed data verification;

[0126] S420: Analyze the physical quantity characteristics, spatial characteristics, and temporal characteristics of the same type of monitoring data with the same physical quantity attributes to obtain the spatial and temporal fluctuations of the same type of monitoring data; based on the spatial and temporal fluctuations, select several reliable data from the same type of monitoring data, and integrate the selected reliable data corresponding to multiple physical quantity attributes into several reliable multi-source data.

[0127] The beneficial effects of the above embodiments are as follows: given the vast land surface area monitored and the long monitoring duration, in order to extract reliable multi-source data from the benchmark database from both temporal and spatial perspectives, the benchmark database that has completed data verification is identified by physical quantity characteristics (corresponding to the physical quantity type of the monitoring data), regional characteristics (corresponding to the spatial location of the monitoring data), and temporal characteristics (corresponding to the time point when the monitoring data was acquired). In this way, the physical quantity characteristics, spatial characteristics, and temporal characteristics of the same type of monitoring data with the same physical quantity attributes are analyzed to obtain the spatial and temporal fluctuations of the same type of monitoring data. The changes and fluctuations of the same type of monitoring data in both the spatial range of the land surface and the temporal range corresponding to the entire monitoring period of the land surface are characterized. In this case, several reliable data are selected from the same type of monitoring data, and the several reliable data after the selection of the same type of monitoring data corresponding to multiple physical quantity attributes are integrated into several reliable multi-source data. In practice, spatial and temporal fluctuations of the same type of monitoring data can be compared using thresholds. Monitoring data with spatial fluctuations exceeding a preset first fluctuation threshold or temporal fluctuations exceeding a preset second fluctuation threshold can be removed. This integrates the selected and retained monitoring data of the same type into reliable multi-source data, providing reliable data support for the subsequent generation of land surface condition results.

[0128] In another embodiment, such as Figure 9 As shown, temporal and spatial correlation processing is performed on several reliable multi-source data to obtain a land surface condition representation map, including the following steps:

[0129] S511: Based on the time points and spatial points of several trusted multi-source data, determine the degree of temporal correlation and spatial correlation between any two trusted multi-source data in the several trusted multi-source data;

[0130] S512: Based on the degree of temporal and spatial correlation, several reliable multi-source data are arranged and mapped to obtain a land surface condition representation map.

[0131] The beneficial effects of the above embodiments are that reliable multi-source data essentially describes the multimodal situation of different locations on the land surface at different times. Therefore, based on the time and spatial points of several reliable multi-source data sets, the temporal and spatial correlation between any two reliable multi-source data sets are determined. The temporal correlation refers to the time difference between the generation times of the two reliable multi-source data sets; the smaller the time difference, the greater the temporal correlation. The spatial correlation refers to the distance difference between the land surface locations corresponding to the two reliable multi-source data sets; the smaller the distance difference, the greater the spatial correlation. Then, based on the temporal and spatial correlation, the reliable multi-source data sets are arranged and mapped in both the temporal and spatial dimensions. Reliable multi-source data sets with temporal correlation greater than a preset first correlation threshold and spatial correlation greater than a preset second correlation threshold are arranged and mapped in both the temporal and spatial domains to obtain a land surface condition representation map. In this way, the land surface condition representation map can accurately represent the multimodal changes of the land surface in the temporal and spatial domains.

[0132] In another embodiment, such as Figure 10 As shown, based on the query request, the elements of the land surface condition representation map are modified to obtain the target representation map and uploaded to the cloud, including the following steps:

[0133] S521: Parse the query request to obtain the query target object; based on the query target object, determine the map elements within the map that represent the actual land surface status and meet the preset related conditions of the query target object;

[0134] S522: Based on the actual state of map elements on the land surface, represent the distribution of the map, change the visualization state of the map elements, obtain the target representation map, and upload it to the cloud.

[0135] The beneficial effects of the above embodiments are as follows: In practical operation, to quickly and accurately meet the query needs for land surface conditions, the query target object is obtained from the query request, and map elements within the land surface condition representation map that meet preset related conditions with the query target object are identified. The query target object may include, but is not limited to, the buildings and / or natural structures to be queried; the map elements that meet the preset related conditions may include, but are not limited to, map object elements that meet preset shape similarity conditions (e.g., shape similarity exceeds a preset similarity threshold). Then, based on the distribution of map elements on the land surface condition representation map, the visualization states of the map elements, such as color or transparency, are changed to form a target representation map with differentiated map elements. This target representation map is then uploaded to the cloud, allowing users to obtain the required target representation map from the cloud in a timely and accurate manner.

[0136] Please see Figure 11 The present invention also provides a multi-source data processing system for dynamic monitoring of land surface corresponding to the above method, comprising:

[0137] The resource pool construction module is used to determine the task allocation for all monitoring terminals based on their working status within the land area; and to construct an elastic resource pool based on the actual operating status of the cloud accessed by all monitoring terminals.

[0138] The resource allocation module is used to parse the monitoring task instructions received by the monitoring terminal and determine the resource allocation of the elastic resource pool to the monitoring terminal.

[0139] The data verification module is used to build a benchmark database for land surface monitoring and to verify the benchmark database based on multi-source data generated during the monitoring tasks performed by the monitoring terminal.

[0140] The identification and filtering module is used to identify and filter features in the benchmark database that has completed data verification, and obtain several reliable multi-source data.

[0141] The data processing module is used to perform temporal and spatial correlation processing on several reliable multi-source data to obtain a land surface condition representation map; according to the query request, the elements of the land surface condition representation map are modified to obtain the target representation map and uploaded to the cloud.

[0142] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. This disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims. Thus, if these modifications and variations of the invention fall within the scope of the claims of the invention and their equivalents, the invention is also intended to include these modifications and variations.

Claims

1. A multi-source data processing method for dynamic monitoring of land surface, characterized in that, include: Based on the operational status of all monitoring terminals deployed within the land area, determine the task allocation for all monitoring terminals; Based on the actual operating status of all monitoring terminals connected to the cloud, an elastic resource pool is constructed. The monitoring task instructions received by the monitoring terminal are parsed to determine the resource allocation of the elastic resource pool to the monitoring terminal. A benchmark database for land surface monitoring is constructed, and the benchmark database is validated based on multi-source data generated during the monitoring tasks performed by the monitoring terminal. Feature identification and filtering are performed on the benchmark database that has completed the data verification to obtain several reliable multi-source data; Temporal and spatial correlation processing is performed on the aforementioned reliable multi-source data to obtain a land surface condition representation map. Based on the query request, the elements of the land surface condition representation map are changed to obtain the target representation map and uploaded to the cloud. The process of parsing the monitoring task instructions received by the monitoring terminal and determining the resource allocation of the elastic resource pool to the monitoring terminal includes: extracting the following three types of parameters from the monitoring task instructions received by the monitoring terminal: monitoring area spatial range parameters, monitoring accuracy parameters, and monitoring time window parameters; determining the normalized task complexity index of the monitoring terminal according to a first preset algorithm based on the monitoring area spatial range parameters, monitoring accuracy parameters, and monitoring time window parameters; predicting the dynamic workload of the monitoring terminal for fusion of surface and light sensing according to a preset fusion algorithm model based on the normalized task complexity index, surface characteristics and lighting conditions of the monitoring area; and determining the resource allocation to the monitoring terminal according to a preset coupling model based on the dynamic workload of the monitoring terminal, the single-core processing capability of the cloud server, the weather complexity and terrain undulation of the monitoring area.

2. The multi-source data processing method for dynamic monitoring of land surface as described in claim 1, characterized in that: Based on the operational status of all monitoring terminals deployed within the land area, determine the task allocation for all monitoring terminals, including: The historical working status records and actual working status records of all monitoring terminals deployed within the land area are obtained. The monitoring data change trends of the monitoring terminals during the historical monitoring period are extracted from the historical working status records. The monitoring data change trends include the correlation change trends between the monitoring data error rate and monitoring data redundancy of the monitoring terminals during the historical monitoring period and the generated monitoring data types and their data volume. Extract the real-time generated monitoring data type and its volume from the work record; compare the real-time generated monitoring data type and its volume with the monitoring data change trend to estimate the time point in the future when the monitoring terminal will be in an unreliable state; Based on the temporal relationship between the time points when each monitoring terminal is in an untrusted state, the task allocation time order for all monitoring terminals is determined.

3. The multi-source data processing method for dynamic monitoring of land surface as described in claim 2, characterized in that: Based on the actual cloud operation status of all monitoring terminals, an elastic resource pool is constructed, including: Obtain the real-time CPU operating status of all monitoring terminals connected to the cloud, and determine the spare CPU resources of the cloud in maintaining the preset working mode based on the real-time CPU operating status. Based on the available CPU resources, construct an elastic resource pool and adjust the actual resource capacity of the elastic resource pool.

4. The multi-source data processing method for dynamic monitoring of land surface as described in claim 1, characterized in that: Construct a benchmark database for land surface monitoring, including: A set of historical monitoring data on the land surface is obtained. Based on the label of each historical monitoring data in the set, all historical monitoring data in the set are preprocessed and indexed. The corresponding historical monitoring data are then used to construct a benchmark database. The preprocessing includes standardization preprocessing. The indexing includes adding unique index tags to the historical monitoring data.

5. The multi-source data processing method for dynamic monitoring of land surface as described in claim 4, characterized in that: Based on the multi-source data generated during the monitoring task performed by the monitoring terminal, data verification is performed on the benchmark database, including: The multi-source data generated during the monitoring task performed by the monitoring terminal is divided into several single-modal data, and the single-modal data is standardized and preprocessed. By comparing the historical monitoring data with the same type of attributes as the single-modal data in the benchmark database with the single-modal data, the deviation between the single-modal data and the historical monitoring data is determined. Based on the deviation, the historical monitoring data of the benchmark database are verified and corrected.

6. The multi-source data processing method for dynamic monitoring of land surface as described in claim 1, characterized in that: Feature identification and filtering are performed on the benchmark database that has completed the data verification to obtain several reliable multi-source data, including: Feature identification is performed on the benchmark database that has completed the data verification to obtain the physical quantity characteristics, regional characteristics, and temporal characteristics of all monitoring data in the benchmark database that has completed the data verification. The physical quantity characteristics, spatial characteristics, and temporal characteristics of the same type of monitoring data with the same physical quantity attributes are analyzed to obtain the spatial and temporal fluctuations of the same type of monitoring data. Based on the spatial and temporal fluctuations, several reliable data are selected from the same type of monitoring data, and the several reliable data selected from the same type of monitoring data with multiple physical quantity attributes are integrated into several reliable multi-source data.

7. The multi-source data processing method for dynamic monitoring of land surface as described in claim 1, characterized in that: Temporal and spatial correlation processing is performed on the aforementioned reliable multi-source data to obtain a land surface condition representation map, including: Based on the respective time points and spatial points of the aforementioned trusted multi-source data, determine the degree of temporal correlation and spatial correlation between any two trusted multi-source data among the aforementioned trusted multi-source data. Based on the temporal and spatial correlation, the several reliable multi-source data are arranged and mapped to obtain a land surface condition representation map.

8. The multi-source data processing method for dynamic monitoring of land surface as described in claim 1, characterized in that: Based on the query request, the elements of the land surface condition representation map are modified to obtain the target representation map, which is then uploaded to the cloud, including: The query target object is obtained by parsing the query request; based on the query target object, the map elements in the map that meet the preset related conditions with the query target object are determined to represent the actual state of the land surface. Based on the distribution of the map elements on the land surface in terms of their actual state, the visualization state of the map elements is changed to obtain the target representation map and upload it to the cloud.

9. A multi-source data processing system for dynamic monitoring of land surface, characterized in that, include: The resource pool construction module is used to determine the task allocation for all monitoring terminals based on their working status within the land area. Based on the actual operating status of all monitoring terminals connected to the cloud, an elastic resource pool is constructed. The resource allocation module is used to parse the monitoring task instructions received by the monitoring terminal and determine the resource allocation of the elastic resource pool to the monitoring terminal. The data verification module is used to construct a benchmark database for land surface monitoring and to verify the benchmark database based on multi-source data generated during the monitoring tasks performed by the monitoring terminal. The identification and filtering module is used to perform feature identification and filtering on the benchmark database that has completed the data verification, and to obtain several reliable multi-source data. The data processing module is used to perform temporal and spatial correlation processing on the aforementioned reliable multi-source data to obtain a land surface condition representation map. Based on the query request, the elements of the land surface condition representation map are modified to obtain the target representation map and uploaded to the cloud. The process of parsing the monitoring task instructions received by the monitoring terminal and determining the resource allocation of the elastic resource pool to the monitoring terminal includes: extracting the following three types of parameters from the monitoring task instructions received by the monitoring terminal: monitoring area spatial range parameters, monitoring accuracy parameters, and monitoring time window parameters; determining the normalized task complexity index of the monitoring terminal according to a first preset algorithm based on the monitoring area spatial range parameters, monitoring accuracy parameters, and monitoring time window parameters; predicting the dynamic workload of the monitoring terminal for fusion of surface and light sensing according to a preset fusion algorithm model based on the normalized task complexity index, surface characteristics and lighting conditions of the monitoring area; and determining the resource allocation to the monitoring terminal according to a preset coupling model based on the dynamic workload of the monitoring terminal, the single-core processing capability of the cloud server, the weather complexity and terrain undulation of the monitoring area.

Images