Data acquisition optimization method and system

Abstract

The invention discloses a data acquisition optimization method and system, which is applied to the field of power system data acquisition, and comprises the following steps: generating an optimization instruction based on a target data acquisition demand, and loading a corresponding reference acquisition parameter vector; collecting data according to the vector, determining a real-time load value, and dividing load grades according to the load value; visual data quality, a data acquisition period and the adjustment amount of a sensing scheduling strategy are obtained by dynamically adjusting a mapping table, an optimized acquisition parameter vector is obtained through reconstruction, and finally data acquisition is executed by the vector. The data acquisition optimization method provided by the invention gets rid of dependence on terminal hardware indexes, realizes dynamic adaptation of acquisition parameters through load feedback, not only avoids resource overload and acquisition interruption under a high load, but also can improve data acquisition quality under a low load, accurately matches differentiated requirements of different business scenes of a power grid, and improves data acquisition efficiency. And the data acquisition efficiency and reliability of the power system are greatly improved.

Description

Technical Field

[0001] This invention relates to the field of power system data acquisition technology, and in particular to an optimized method and system for data acquisition. Background Technology

[0002] In the field of power system operation and maintenance and asset management, in order to digitize business operations such as equipment condition monitoring, asset inventory and inspection management, it is usually necessary to deploy dedicated data acquisition terminals or rely on manual on-site data collection. These collection activities aim to obtain multi-dimensional data, including equipment appearance images, operating environment temperature and humidity, asset location information, etc., to provide decision-making basis for back-end business systems.

[0003] However, existing data acquisition methods for power grid equipment select a fixed acquisition strategy based on the acquisition scenario before data collection begins. This static configuration lacks dynamic adaptability in actual operation. For example, when the terminal temporarily switches to a high-load task, processor resources are already highly strained, yet high-load data acquisition continues, easily leading to system lag, data loss, or even abnormal program exits. Conversely, when the system load is low, it cannot automatically improve acquisition quality to obtain richer data details. This acquisition strategy not only affects the smoothness and data integrity of field operations but also wastes resource efficiency, severely restricting the automation, intelligence, and overall operational efficiency of mobile operation terminals in power grid fields. Summary of the Invention

[0004] This invention provides an optimized method and system for data acquisition to solve the technical problem that data acquisition strategies cannot be dynamically adjusted according to the real-time load of the terminal system, so as to achieve the best balance between acquisition quality and system load.

[0005] To address the aforementioned technical problems, embodiments of the present invention provide an optimized method for data acquisition, comprising: Based on the target data acquisition requirements, generate optimized instructions for the data acquisition strategy; In response to the optimization instruction, a baseline acquisition parameter vector corresponding to the target data acquisition requirement is loaded. The baseline acquisition parameter vector includes at least visual data quality, data acquisition cycle, and sensor resource scheduling strategy. Data is acquired based on the benchmark acquisition parameter vector, and the corresponding real-time load value is determined based on the data acquisition results. Based on the real-time load value, determine the load level under the current data acquisition strategy; Based on the load level, query the pre-built dynamic adjustment mapping table to obtain the first adjustment amount for the visual data quality, the second adjustment amount for the data acquisition cycle, and the third adjustment amount for the sensor resource scheduling strategy. Based on the first adjustment amount, the second adjustment amount, and the third adjustment amount, the benchmark acquisition parameter vector is reconstructed to obtain the optimized acquisition parameter vector; Data acquisition is performed using at least the optimized acquisition parameter vector.

[0006] As one preferred embodiment, the step of loading the baseline acquisition parameter vector corresponding to the target data acquisition requirement in response to the optimization instruction includes: Based on the optimization instructions, the corresponding power grid business scenario is determined; According to the power grid business scenario, the scenario parameter configuration library is invoked, which pre-stores the mapping relationship between each power grid business scenario and the benchmark acquisition parameter vector; For different power grid business scenarios, the corresponding reference acquisition parameter vector is loaded.

[0007] As one preferred embodiment, the step of acquiring data based on the benchmark acquisition parameter vector and determining the corresponding real-time load value based on the data acquisition results includes: Data acquisition is performed according to the sensor resource scheduling strategy based on the benchmark acquisition parameter vector; During the data acquisition process, spatiotemporal alignment of multi-source data is performed synchronously. After the data acquisition is completed, the acquired data undergoes preliminary preprocessing to obtain a valid dataset; Based on the effective data set, extract the data load feature parameter set; Based on the set of data load characteristic parameters, and in accordance with the preset load evaluation criteria, the corresponding real-time load value is determined.

[0008] As one preferred embodiment, determining the load level under the current data acquisition strategy based on the real-time load value includes: Predefine a set of corresponding load thresholds for different data acquisition strategies; The real-time load value is compared with the set of load thresholds corresponding to the current data acquisition strategy to determine the load level.

[0009] As a preferred embodiment, the reconstruction of the reference acquisition parameter vector based on the first adjustment amount, the second adjustment amount, and the third adjustment amount includes: Based on the first adjustment amount, the resolution of the visual data quality in the benchmark acquisition parameter vector is adjusted. Based on the second adjustment amount, the duration of the data acquisition period in the benchmark acquisition parameter vector is adjusted. Based on the third adjustment amount, the state of the sensor resource scheduling strategy in the benchmark acquisition parameters is adjusted.

[0010] Another embodiment of the present invention provides an optimized data acquisition system, comprising: The instruction generation module is used to generate optimized instructions for data acquisition strategies based on the target data acquisition requirements. A vector loading module is used to load a baseline acquisition parameter vector corresponding to the target data acquisition requirements in response to the optimization instruction. The baseline acquisition parameter vector includes at least visual data quality, data acquisition cycle and sensor resource scheduling strategy. The load determination module is used to collect data based on the benchmark acquisition parameter vector and determine the corresponding real-time load value based on the data acquisition results. The load level determination module is used to determine the load level under the current data acquisition strategy based on the real-time load value. The query module is used to query a pre-built dynamic adjustment mapping table according to the load level to obtain a first adjustment amount for the visual data quality, a second adjustment amount for the data acquisition cycle, and a third adjustment amount for the sensor resource scheduling strategy. The parameter reconstruction module is used to reconstruct the benchmark acquisition parameter vector based on the first adjustment amount, the second adjustment amount, and the third adjustment amount to obtain an optimized acquisition parameter vector. A data acquisition module is used to acquire data using at least the optimized acquisition parameter vector.

[0011] As one preferred embodiment, the instruction generation module includes: The scenario determination unit is used to determine the corresponding power grid business scenario based on the optimization instructions; The parameter calling unit is used to call the scenario parameter configuration library according to the power grid business scenario. The scenario parameter configuration library pre-stores the mapping relationship between each power grid business scenario and the benchmark acquisition parameter vector. The loading unit is used to load the corresponding reference acquisition parameter vector for different power grid service scenarios.

[0012] As one preferred embodiment, the load determination module includes: The execution unit is used to perform data acquisition according to the sensor resource scheduling strategy of the benchmark acquisition parameter vector; The spatiotemporal alignment processing unit is used to synchronously perform spatiotemporal alignment processing of multi-source data during the data acquisition process; The data preprocessing unit is used to perform preliminary preprocessing on the collected data after the data acquisition is completed, so as to obtain an effective data set. The extraction unit is used to extract a set of data load feature parameters based on the effective data set; The comparison unit is used to determine the corresponding real-time load value based on the set of data load characteristic parameters and a preset load evaluation standard.

[0013] As one preferred embodiment, the level determination module includes: The threshold decision unit is used to predefine a set of corresponding load thresholds for different data acquisition modes; The load assessment unit is used to compare the real-time load value with the set of load thresholds corresponding to the current data acquisition mode to determine the load level.

[0014] As one preferred embodiment, the parameter reconstruction module includes: The first adjustment unit is used to adjust the resolution of the visual data quality in the reference acquisition parameter vector based on the first adjustment amount. The second adjustment unit is used to adjust the duration of the data acquisition period in the reference acquisition parameter vector based on the second adjustment amount. The third adjustment unit is used to adjust the state of the sensor resource scheduling strategy in the benchmark acquisition parameters based on the third adjustment amount.

[0015] Compared with the prior art, the beneficial effects of the embodiments of the present invention are at least one of the following: (1) This invention first generates optimization instructions based on the actual data acquisition needs of the power grid, and calls a benchmark acquisition parameter vector that matches the scenario. This vector covers key dimensions such as visual data quality and acquisition cycle, ensuring that the initial parameter configuration accurately matches the business requirements. During the data acquisition process, the real-time load is determined based on the acquisition results. By matching the corresponding parameter adjustment strategy with the load level, the benchmark parameter vector is dynamically reconstructed and optimized. This optimization logic centered on business data enables the acquisition parameters to be flexibly adjusted according to load changes, avoiding the resource waste or data loss problems caused by traditional fixed parameter acquisition. It significantly improves the targeting and efficiency of power grid data acquisition, and achieves the efficient acquisition goal of "on-demand acquisition and dynamic adaptation".

[0016] (2) In terms of technical practicality, this invention addresses the pain points of existing power grid data acquisition equipment, namely "multiple separate devices, asynchronous data, and low processing efficiency." By combining an integrated terminal with optimization methods, it reduces the number of devices carried and lowers manual operation costs. Regarding scenario adaptability, it addresses the needs of various scenarios such as power grid production and material management by flexibly adjusting parameter vectors to achieve differentiated data acquisition, avoiding the scenario limitations of traditional single acquisition modes and adapting to the data needs of the entire power grid business chain. This invention improves the accuracy and real-time performance of data acquisition, providing reliable data support for core businesses such as power grid equipment defect diagnosis and asset inventory, promoting the digital transformation of the power grid towards a more efficient and flexible direction, and possessing broad application prospects. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating an optimized data acquisition method in one embodiment of the present invention; Figure 2 This is a schematic diagram of an optimized data acquisition system in one embodiment of the present invention.

[0018] Figure label: Among them, 11. Instruction generation module, 12. Vector loading module, 13. Load determination module, 14. Level determination module, 15. Query module, 16. Parameter reconstruction module, and 17. Data acquisition module. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0020] In the description of this application, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first," "second," "third," etc., may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0021] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal communication between two components. The terms "vertical," "horizontal," "left," "right," "upper," "lower," and similar expressions used herein are for illustrative purposes only and do not indicate or imply that the device or component referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as limiting the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0022] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this specification is for the purpose of describing specific embodiments only and is not intended to limit the invention. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0023] One embodiment of the present invention provides an optimization method for data acquisition; for details, please refer to [link to relevant documentation]. Figure 1 , Figure 1 The diagram shown illustrates a flowchart of an optimized data acquisition method according to one embodiment of the present invention, which includes steps S1 to S7: S1: Generate optimized instructions for the data acquisition strategy based on the target data acquisition requirements; Step S1, which generates optimized instructions for the data acquisition strategy based on the target data acquisition requirements, is the starting point and core driving link of the entire data acquisition process. In existing power grid data acquisition scenarios, the acquisition requirements for different business scenarios, such as production inspection, material management, financial verification, and asset inventory, vary significantly. If a fixed acquisition strategy is adopted, data redundancy or missing key data can easily occur. Therefore, it is necessary to generate optimized instructions specifically based on the target data acquisition requirements. This lays the foundation for the subsequent configuration and adjustment of acquisition parameters, ensuring that the acquisition strategy is highly matched with actual business needs.

[0024] Among them, the target data acquisition requirements refer to the specific data acquisition requirements under different power grid business scenarios. These requirements encompass core elements such as the acquisition object, data type (e.g., video, temperature and humidity, asset tags), acquisition accuracy (e.g., video resolution, sensor error range), and acquisition frequency, serving as the core basis for driving the entire acquisition process. The optimized instructions for the data acquisition strategy transform the target data acquisition requirements into standardized instructions that the acquisition terminal can recognize and execute. These instructions include key information such as requirement identifiers, trigger types, and target device information, and are used to instruct the acquisition terminal to load matching acquisition parameters and initiate the acquisition process.

[0025] Specifically, the first step is to obtain the triggering source of the target data collection requirement, which includes three types: manual triggering, timed triggering, and event triggering.

[0026] Manual trigger: Maintenance personnel can initiate the process directly through the physical buttons or touch interface of the terminal, which is suitable for temporary asset verification or equipment spot checks.

[0027] Scheduled triggering: The system automatically generates periodic data collection tasks based on a preset work plan (such as once every 30 minutes), which is suitable for routine scenarios such as material warehouse environmental monitoring.

[0028] Event triggering: When the terminal's built-in sensors (such as temperature and humidity, audio unit) detect that the device temperature exceeds the threshold or abnormal sound signature in real time, an emergency data collection request is automatically generated, realizing a closed-loop linkage from status perception to data collection.

[0029] To improve the accuracy and relevance of data collection, the system queries the power grid asset ledger system (such as PMS2.0) in real time via its communication module (4G / 5G / Wi-Fi) before generating instructions to obtain the unique identifier (ID), asset number, historical defect records, and business priority of the target equipment. For example, for the data collection needs of a main transformer, the instruction will explicitly link its rated parameters and indicate that infrared thermal imaging data and partial discharge audio should be collected as the key data.

[0030] Subsequently, the system encapsulates the above information into a standardized "data acquisition strategy optimization instruction." This instruction strictly adheres to the power grid communication protocol and includes structured fields such as demand identifier, trigger type, target equipment information, and service priority. Before the instruction is issued, the core processing module performs a CRC check on it; if the check fails, it is regenerated to ensure the integrity and reliability of the instruction during transmission and parsing.

[0031] Step S1 generates optimized instructions for the data acquisition strategy, which can accurately target the differentiated acquisition needs of different business scenarios in the power grid, avoiding the blindness of traditional fixed acquisition strategies. By combining power grid asset ledger information to refine acquisition needs, the optimized instructions can accurately target specific acquisition objectives and data dimensions, ensuring that the subsequent acquisition process can obtain the key data required by the business.

[0032] S2: In response to the optimization command, load the baseline acquisition parameter vector corresponding to the target data acquisition requirements. The baseline acquisition parameter vector includes at least visual data quality, data acquisition cycle and sensor resource scheduling strategy. Step S2, which loads the corresponding baseline acquisition parameter vector in response to the optimization command, is a key step in connecting acquisition requirements with actual acquisition operations. Its core purpose is to transform the abstract target data acquisition requirements into specific parameter configurations that the acquisition terminal can directly execute.

[0033] Among them, the benchmark acquisition parameter vector is a set of parameters related to power grid acquisition services, which together define the execution standard of an acquisition task; visual data quality mainly refers to indicators such as resolution and frame rate of acquired video or images that affect the clarity of visual information; data acquisition cycle refers to the time interval between two acquisition operations of the same type of data, which can be set to continuous acquisition or fixed interval acquisition according to requirements; sensor resource scheduling strategy refers to the rules for starting, stopping and prioritizing various sensor modules such as video acquisition units, audio acquisition units, temperature and humidity sensors on the terminal.

[0034] Preferably, in one embodiment of the present invention, loading the baseline acquisition parameter vector corresponding to the target data acquisition requirements in response to the optimization instruction includes: Based on the optimization instructions, the corresponding power grid business scenario is determined; Based on the power grid business scenario, the scenario parameter configuration library is invoked. The scenario parameter configuration library pre-stores the mapping relationship between each power grid business scenario and the benchmark acquisition parameter vector. For different power grid business scenarios, load the corresponding reference acquisition parameter vector.

[0035] The scenario parameter configuration library is used to store the association between different power grid business scenarios and corresponding benchmark acquisition parameter vectors. Its data can be updated and maintained through the terminal backend or the cloud.

[0036] Specifically, the first step is to determine the corresponding power grid business scenario based on the optimization instructions. The core processing module of the data acquisition terminal parses the demand identifiers, target equipment information, and other content contained in the optimization instructions, and then identifies the business scenario through keyword matching. For example, when the instructions contain keywords such as "main transformer," "temperature," and "continuous acquisition," it can be determined as a power grid production scenario; when the instructions contain content such as "material label" and "inventory," it matches as a material management scenario. These keyword matching rules are pre-stored in the terminal system and can be customized and added according to the expansion of power grid business.

[0037] After determining the business scenario, the system calls a pre-stored scenario parameter configuration library. This library uses a key-value pair storage format, with the power grid business scenario name as the "key" and the corresponding baseline acquisition parameter vector as the "value," and is stored in the terminal's 64GBeMMC storage module. During the call, the system retrieves the corresponding parameter vector by scenario name. If a match is found, it is directly extracted; otherwise, the default parameter vector is loaded. The default parameter vector is set to a general configuration that meets the basic requirements of most scenarios, such as 1080P video resolution and a 1-minute temperature and humidity acquisition cycle.

[0038] The loading of benchmark acquisition parameter vectors needs to be differentiated for different power grid business scenarios, including: In the power grid production scenario, the core requirement is to accurately capture the operating status of equipment. Therefore, in the benchmark parameter vector loaded in this embodiment, the visual data quality is set to a high-definition mode of 4K resolution and 30 frames per second to ensure that equipment appearance defects and operational details are clearly identifiable. The data acquisition cycle adopts a high-frequency configuration of temperature and humidity once per second and continuous audio and video acquisition to avoid missing moments of equipment abnormality. The sensor resource scheduling strategy is set to prioritize the activation of video acquisition unit, audio acquisition unit, temperature and humidity acquisition unit and Beidou / GPS positioning unit to ensure synchronous acquisition of multi-dimensional data.

[0039] In the materials management scenario, the focus is on the association and collection of materials information and tags. In this embodiment, the visual data quality in the loaded parameter vector is set to 1080P resolution and 15 frames per second standard definition mode to ensure that the appearance of materials can be identified while reducing resource consumption. The data collection cycle is set to temperature and humidity once every 5 minutes, and RFID tags are read in real time. The sensor resource scheduling strategy mainly focuses on starting the video acquisition unit, RFID reader and temperature and humidity acquisition unit to reduce the start of unnecessary modules.

[0040] Financial verification scenario: It is necessary to clearly collect asset nameplate and physical information. In this embodiment, the visual data quality is set to 1080P resolution, and the video is collected in 10-second segments. The positioning unit is started simultaneously to record the asset location. In the asset inventory scenario, a visual configuration of 720P resolution and 5 frames per second is used, combined with UWB positioning and RFID collection to realize the rapid entry of asset information.

[0041] After the parameter vector is loaded, the system will send each parameter to the corresponding execution unit. For example, the visual data quality parameter will be sent to the video acquisition unit, and the acquisition period parameter will be passed to the timing control module. At the same time, start or standby commands will be sent to each sensor module according to the sensor resource scheduling strategy. The process of loading the benchmark acquisition parameter vector in step S2 accurately matches the core needs of different power grid services through differentiated parameter configuration, avoiding the resource waste or data shortage problems caused by traditional general parameters.

[0042] S3: Data is acquired based on the baseline acquisition parameter vector, and the corresponding real-time load value is determined based on the data acquisition results; Performing data acquisition based on the baseline acquisition parameter vector and determining the real-time load value is the core step in realizing dynamic optimization of the acquisition process. Its fundamental purpose is to first complete the data acquisition operation according to the preset scenario parameters, and then evaluate the load under the current acquisition strategy based on the actual acquisition results, so as to provide an objective basis for subsequent parameter optimization.

[0043] Among them, the real-time load value is the core indicator used in this invention to quantitatively evaluate the current data acquisition strategy's occupation of power grid business resources. Its essence is to reflect the comprehensive resource demand of data transmission, processing and storage links in the acquisition process by combining the business characteristics of the effective data set.

[0044] Preferably, in one embodiment of the present invention, data acquisition is performed based on a benchmark acquisition parameter vector, and the corresponding real-time load value is determined based on the data acquisition results, including: Data acquisition is performed based on the sensor resource scheduling strategy according to the baseline acquisition parameter vector; During the data acquisition process, spatiotemporal alignment of multi-source data is performed simultaneously. After data collection is completed, the collected data undergoes preliminary preprocessing to obtain a valid dataset; Based on the effective dataset, extract the set of data load characteristic parameters; Based on the set of data load characteristic parameters and in accordance with the preset load evaluation criteria, the corresponding real-time load value is determined.

[0045] Among them, multi-source data spatiotemporal alignment processing refers to the process of adding unified timestamps and location information to video, audio, temperature and humidity, and positioning data collected by different sensors through technologies such as time synchronization and positioning, so as to maintain the consistency of various types of data in the time and space dimensions; the effective data set refers to the set of raw collected data after preprocessing, removing invalid information and retaining data with power grid business value. The effective data must include core content related to the collection requirements, such as equipment operating status, material information, and asset identification; the data load characteristic parameter set is a combination of parameters extracted from the business dimension of the effective data set that can reflect the collection load situation; the load assessment standard is a load classification basis based on the historical power grid collection business data. By setting thresholds for different business characteristic indicators, the collection load is divided into different levels for quantitative assessment of the load situation corresponding to the collection results.

[0046] Specifically, under the parameter vector of a typical power grid production scenario, the system follows a startup logic of "positioning → environment → image": First, the BeiDou / GPS dual-mode positioning unit is initialized to establish a spatial reference; then, dual redundant temperature and humidity acquisition units (with primary and backup sensors performing self-calibration) and an omnidirectional microphone are started in parallel; finally, after the environmental sensors are ready, the 20-megapixel industrial camera, which has the highest power consumption and computing power requirements, is activated. All units operate according to the quantization parameters set in the vector. For example, the video acquisition unit shoots at 4K resolution and 30 frames per second, the temperature and humidity acquisition unit continuously acquires data at 1-second intervals, and the positioning unit updates location information in real time, thus ensuring the consistency of the acquisition reference.

[0047] During data acquisition, the system simultaneously executes a three-level redundant time synchronization and spatiotemporal alignment mechanism to solve the fundamental problem of heterogeneous data fusion: Main time source: Relying on dual-frequency reception of BeiDou B1I and GPS L1, it provides a master clock signal with an accuracy of ±1ppm when simultaneously locking onto no less than 4 satellites and the signal strength is better than -130dBm.

[0048] Seamless switching mechanism: The built-in "satellite signal quality detection" circuit continuously monitors the signal status. When the above conditions are not met, the system automatically and seamlessly switches to the backup timing source within 10 clock cycles.

[0049] Backup time source: The local RTC is composed of a high-precision temperature-controlled crystal oscillator (TCXO) with a temperature coefficient of ±0.1ppm / ℃. It can independently maintain an accuracy of ≤10ms cumulative error within 1 hour in the working temperature range of -30℃ to 60℃.

[0050] Periodic network calibration: When connected to the network, the terminal synchronizes its time with the unified clock system of the power grid every 30 minutes via the NTP protocol to correct the accumulated error of the local RTC and ensure long-term consistency between the terminal time and the backend system time.

[0051] Through the above mechanism, the system assigns a unified, traceable millisecond-level timestamp to each video frame, each audio sample, and each temperature and humidity reading, and binds it to the latitude and longitude coordinates provided by the positioning unit, ultimately achieving the alignment of multi-source data in the spatiotemporal dimension, with the synchronization error strictly controlled within 10 milliseconds.

[0052] After data collection is complete, the raw data needs to be preprocessed to obtain a valid dataset: Standardized Formatting: Edge processing chips (such as NPUs) encapsulate raw H.265 video streams, WAV audio, and CSV-formatted temperature and humidity data in real time into structured data packets conforming to the power grid standard data format. This data packet is required to include four fields: data identifier, millisecond-level timestamp, terminal device ID, and data body, eliminating format barriers for subsequent system integration.

[0053] Invalid data cleaning includes video filtering and data smoothing: Video filtering: A lightweight model (such as a pruned and optimized MobileNetV3) is used to analyze the video stream frame by frame. The inference time of the model on the NPU is less than 5ms / frame. Its decision logic is not a simple identification of "present / absent" targets, but a comprehensive judgment based on image entropy and the characteristics of specific power grid equipment (such as insulators and transformers). The filtering accuracy for empty scenes and blurred images is no less than 95%.

[0054] Data smoothing: For temperature and humidity data, a strategy combining "3σ criterion outlier removal" and "moving average filtering with a window size of 5" is employed. This algorithm effectively suppresses instantaneous impulse noise caused by strong electromagnetic interference while preserving the true trend of the equipment's temperature rise curve.

[0055] Finally, the system extracts a set of data load feature parameters from the valid dataset, including: Effective data throughput (MB / s): Measures the efficiency of generating valuable data per unit of time.

[0056] Data processing time (ms): The total time from the completion of data acquisition to the end of preprocessing.

[0057] Effective data compression ratio: The ratio of data volume before and after preprocessing, reflecting the data purification effect.

[0058] The system then compares the aforementioned parameter set with a load assessment standard derived from historical power grid operation data mining. This standard is not a fixed threshold but a dynamic lookup table. For example, when the effective data throughput is >50MB / s and the data processing time is >100ms, it is mapped to "high load"; when the throughput is <10MB / s and the compression ratio is >0.5, it is mapped to "low load". Through this mapping, a quantified real-time load value is ultimately output, providing an accurate and reliable basis for subsequent dynamic adjustments.

[0059] Step S3 performs data acquisition and determines the real-time load value through the above process. Its advantages are: First, the acquisition unit is started according to the sensor resource scheduling strategy. Combined with the scenario-specific visual data quality and acquisition cycle parameters, it can ensure that the acquired raw data meets business needs and avoid the excessive acquisition of invalid data.

[0060] S4: Determine the load level under the current data acquisition strategy based on the real-time load value; Step S4 determines the load level under the current data acquisition strategy based on the real-time load value. This is a key transitional step connecting load assessment and parameter optimization. Its core significance lies in transforming the quantified real-time load value into a graded result with a clear adjustment direction, providing a direct basis for subsequent queries of the dynamic adjustment mapping table.

[0061] The load level is a graded assessment result based on real-time load values ​​and corresponding load thresholds, typically categorized into low, medium, and high levels. Its core function is to intuitively reflect whether the resource occupancy status under the current data collection strategy is reasonable. It serves as a crucial bridge connecting load assessment and parameter adjustment, determining the direction and magnitude of subsequent parameter adjustments.

[0062] Preferably, in one embodiment of the present invention, determining the load level under the current data acquisition strategy based on the real-time load value includes: Predefine corresponding load threshold sets for different data acquisition strategies; The real-time load value is compared with the load threshold set corresponding to the current data acquisition strategy to determine the load level.

[0063] The load threshold set is a pre-defined combination of numerical ranges used to classify load levels for each data acquisition strategy. It consists of numerical ranges corresponding to low, medium, and high load levels. Its core function is to serve as a criterion for matching real-time load values ​​with load levels. The load threshold set for different acquisition strategies is set differently according to their own business needs and resource consumption characteristics.

[0064] Specifically, firstly, predefine differentiated load threshold sets for different data acquisition strategies and embed them in the terminal's configuration library: Power grid production strategy: Due to the need to process 4K video streams and high-frequency environmental data, the load baseline is the highest, and its threshold is set as follows: low load (0-40), medium load (41-70), high load (71-100). Materials management strategy: For 1080P video processing, the thresholds are set as follows: low load (0-35), medium load (36-65), and high load (66-100). Financial verification strategy and asset inventory strategy: mainly dealing with low bitrate videos and still images, with the lowest load baseline, so their low load ranges are set to (0-30) and (0-25) respectively to improve sensitivity to load fluctuations.

[0065] The method for determining the load threshold set is as follows: First, under typical operating conditions, system load data for each business scenario is collected within at least one complete working cycle to form a historical load dataset. Second, after preprocessing the historical load dataset, the K-Means clustering algorithm is used to divide the load status into three clusters based on the natural distribution of system resource utilization. Finally, the boundaries of each cluster are analyzed, and the density valley values ​​between clusters are set as level thresholds. For example, for the power grid production scenario, cluster analysis reveals that high-density clusters are formed in the load values ​​within the three intervals of 0-40, 41-70, and 71-100, so 40 and 70 are determined as the load thresholds for this scenario.

[0066] The logic for determining the load level is as follows: First, the system retrieves the corresponding threshold set from the configuration library based on the currently effective collection strategy identifier; second, it matches the calculated real-time load value with the threshold range to obtain the initial load level; finally, to resolve level oscillations caused by instantaneous load fluctuations, when the load value is at the critical point of any level (such as 40 under the production strategy), the system automatically initiates a resampling and calculation, and uses the arithmetic mean of the two calculation results as the final judgment basis to ensure the stability of the level.

[0067] Step S4 customizes a set of differentiated load thresholds for different collection strategies, which fully considers the differences in business characteristics and resource requirements of each scenario, making the determination of load level more in line with actual application scenarios and avoiding misjudgment caused by using a uniform threshold.

[0068] S5: Based on the load level, query the pre-built dynamic adjustment mapping table to obtain the first adjustment amount for visual data quality, the second adjustment amount for data acquisition cycle, and the third adjustment amount for sensor resource scheduling strategy. The core execution step for dynamically optimizing data acquisition strategies involves retrieving parameter adjustment amounts from a dynamic adjustment mapping table based on load levels and then reconstructing the baseline acquisition parameter vector based on these adjustments. Its fundamental purpose is to precisely adjust acquisition parameters for different load states, solving the problem that traditional fixed-parameter acquisition cannot adapt to load fluctuations.

[0069] The dynamic adjustment mapping table is a parameter adjustment rule database pre-stored in the acquisition terminal. It uses load level as the core index and stores adjustment amounts for parameters such as visual data quality and acquisition cycle according to power grid business scenarios. It supports updating rules based on actual application feedback, providing a direct basis for parameter adjustment. The adjustment amount is a quantization command specifically used to adjust each vector in the baseline acquisition parameter vector.

[0070] Specifically, a dynamically adjusted mapping table needs to be pre-built. This table uses "load level - acquisition strategy" as a dual index, and associates and stores adjustment rules for visual data quality, data acquisition cycle, and sensor resource scheduling strategy. The mapping table is configured with content for four scenarios, including power grid production and material management. For example, under the power grid production strategy: High load: The first adjustment to visual data quality is "4K reduced to 1080P, 30fps reduced to 15fps", the second adjustment to data acquisition cycle is "temperature and humidity sampling interval extended from 1s to 5s", and the third adjustment to sensor resource scheduling strategy is "UWB positioning unit turned off". Medium load: The first adjustment is "maintain 4K resolution, reduce frame rate to 25fps", the second adjustment is "extend sampling interval to 2s", and the third adjustment is "retain core sensing unit". Low load: The first adjustment is to "enable wide dynamic range function", the second adjustment is to "shorten the sampling interval to 0.5s", and the third adjustment is to "activate RFID extension module".

[0071] After determining the load level, the core processing module calls the dynamic adjustment mapping table, first matching the current acquisition strategy, and then extracting the corresponding adjustment amount according to the load level. For example, when executing the material management strategy, it is determined to be a high load. After querying the mapping table, the first adjustment amount is obtained: "reduce 1080P to 720P", the second adjustment amount is "extend the temperature and humidity sampling interval from 5 minutes to 10 minutes", and the third adjustment amount is "turn off the audio acquisition unit".

[0072] The rules in the dynamic adjustment mapping table are comprehensively customized based on the actual needs and application conditions of power grid data acquisition. The core basis includes the priority characteristics of various power grid business scenarios, the hardware performance boundaries of the acquisition terminal, historically acquired load feedback data, and the resource and environmental characteristics of the power grid site.

[0073] The priority characteristics of business scenarios determine the core direction of parameter adjustments; hardware performance boundaries set basic constraints for the rules, ensuring that the adjustment amount will not exceed the hardware capacity limits of the terminal video acquisition unit, such as 4K resolution and 30fps frame rate; historically collected load feedback data provides a quantitative basis for the adjustment amount. For example, historical data shows that when the 4K resolution is reduced to 1080P, the transmission load can be reduced by about 60%, and this value will be directly incorporated into the adjustment rules for visual data quality; and the resource and environmental characteristics of the power grid site enable the rules to be adaptable to different scenarios. Different adjustment strategies will be customized for different environments such as narrow bandwidth in remote distribution areas and strong electromagnetic interference in substations.

[0074] The advantage of the above process is that dynamically adjusting the mapping table enables standardized output of adjustment schemes, avoids the subjectivity of manual intervention, and allows for customized rules for different scenarios to ensure that the adjustment amount matches business needs.

[0075] S6: Based on the first adjustment amount, the second adjustment amount, and the third adjustment amount, the baseline acquisition parameter vector is reconstructed to obtain the optimized acquisition parameter vector; Reconstructing the baseline acquisition parameter vector based on each adjustment amount to obtain the optimized acquisition parameter vector is the core implementation step for realizing dynamic optimization of the acquisition strategy. Its purpose is to transform the abstract adjustment amount into a new executable parameter configuration and solve the problem of mismatch between the baseline parameter vector and the actual load.

[0076] The optimized acquisition parameter vector is a new set of parameters formed after the baseline acquisition parameter vector has been corrected by various adjustments. It integrates the adjusted visual data quality, acquisition cycle, and sensor scheduling strategy, and serves as the direct basis for the acquisition terminal to execute subsequent acquisition tasks. Its parameter configuration is adapted to the real-time load status. Parameter reconstruction refers to the process of modifying and integrating specific parameters in the baseline acquisition parameter vector according to the adjustments. It covers adjustments to dimensions such as resolution, duration, and module status. The core is to transform the adjustments into executable parameter configurations.

[0077] Preferably, in one embodiment of the present invention, the reconstruction of the reference acquisition parameter vector based on the first adjustment amount, the second adjustment amount, and the third adjustment amount includes: Based on the first adjustment amount, the resolution of the visual data quality in the benchmark acquisition parameter vector is adjusted. Based on the second adjustment amount, the duration of the data acquisition period in the benchmark acquisition parameter vector is adjusted. Based on the third adjustment, the state of the sensor resource scheduling strategy in the benchmark acquisition parameters is adjusted.

[0078] Specifically, the system first extracts three core parameters from the baseline acquisition parameter vector: visual data quality, data acquisition cycle, and sensor resource scheduling strategy. Then, it reconstructs the data based on the corresponding adjustment amounts. For visual data quality, resolution adjustment is performed based on the first adjustment amount. If the adjustment amount is to reduce the resolution, the 4K resolution in the baseline parameter is downgraded to 1080P or 720P. If the adjustment amount is to improve the quality, wide dynamic range is enabled or the frame rate is increased. The adjustment range strictly matches the hardware support capability of the terminal's 20-megapixel industrial camera. For the data acquisition cycle, the duration is adjusted based on the second adjustment amount. Under high load, the sampling interval for data such as temperature, humidity, and video is extended, while under low load, the interval is shortened. The adjustment duration is set within the business requirement range of 1 second to 300 seconds. For the sensor resource scheduling strategy, status adjustment is implemented according to the third adjustment amount. Under high load, non-core modules such as UWB positioning and audio acquisition are shut down, while under low load, extended modules such as RFID readers are activated. Status adjustment commands are sent to each sensor unit through the terminal's internal bus to ensure timely module response.

[0079] S7: Data acquisition should be performed with at least an optimized acquisition parameter vector.

[0080] At least optimizing the acquisition parameter vector for data acquisition is the final execution step of the entire data acquisition optimization method. Its core purpose is to implement the reconstructed adaptability parameter configuration into actual acquisition operations, so that the data acquisition strategy can dynamically iterate according to the load status and solve the problem that traditional fixed parameter acquisition cannot adapt to environmental changes.

[0081] Specifically, the system retrieves the optimized acquisition parameter vector and, according to the updated sensor resource scheduling strategy in the vector, issues start commands to each acquisition unit. Simultaneously, it transmits the adjusted visual data quality parameters and data acquisition cycle parameters to the corresponding execution modules. For example, if the optimized vector represents a high-load configuration for a power grid production scenario, the video acquisition unit will operate at 1080P resolution and 15 frames per second, the temperature and humidity acquisition unit will sample at 5-second intervals, and the UWB positioning unit will be disabled. If the optimized configuration represents a low-load configuration, the video unit will enable 4K wide dynamic range mode, the temperature and humidity sampling interval will be shortened to 0.5 seconds, and the RFID extension module will be activated. During the acquisition process, the terminal continuously monitors the operating status and data output of each module. If a hardware anomaly or data transmission failure occurs during the execution of the optimized parameter vector, it will automatically switch to the baseline acquisition parameter vector or the minimum guaranteed parameter configuration to continue acquisition, ensuring uninterrupted acquisition. The minimum guaranteed parameter configuration is a pre-stored set of basic acquisition parameters that includes the minimum parameter standards required to meet the core data acquisition needs of various power grid business scenarios. When the optimized acquisition parameter vector execution malfunctions, the terminal will automatically switch to this configuration to ensure uninterrupted acquisition of core data.

[0082] The advantage of step S7 lies in achieving dynamic closed-loop optimization of the data acquisition strategy. Acquisition parameters are adjusted in real time according to load changes, reducing acquisition pressure and preventing terminal overload under high load, while improving acquisition quality and enriching data dimensions under low load, effectively balancing resource consumption and data value. Simultaneously, the backup parameter switching mechanism in abnormal situations ensures the continuity of power grid data acquisition, meeting the reliability requirements of power grid production, material management, and other businesses. Furthermore, the scenario-based adaptation of parameters makes the acquired data more closely aligned with the actual analytical needs of various businesses.

[0083] Another embodiment of the present invention provides an optimized data acquisition system; for details, please refer to [link to relevant documentation]. Figure 2 , Figure 2 The diagram shown illustrates the structure of an optimized data acquisition system according to one embodiment of the present invention, comprising: The instruction generation module 11 is used to generate optimized instructions for the data acquisition strategy based on the target data acquisition requirements. The vector loading module 12 is used to load a baseline acquisition parameter vector corresponding to the target data acquisition requirements in response to the optimization command. The baseline acquisition parameter vector includes at least visual data quality, data acquisition cycle and sensor resource scheduling strategy. The load determination module 13 is used to collect data based on the benchmark acquisition parameter vector and determine the corresponding real-time load value based on the data acquisition results. The load level determination module 14 is used to determine the load level under the current data acquisition strategy based on the real-time load value. The query module 15 is used to query a pre-built dynamic adjustment mapping table according to the load level to obtain the first adjustment amount for visual data quality, the second adjustment amount for data acquisition cycle, and the third adjustment amount for sensor resource scheduling strategy. The parameter reconstruction module 16 is used to reconstruct the benchmark acquisition parameter vector based on the first adjustment amount, the second adjustment amount, and the third adjustment amount to obtain the optimized acquisition parameter vector. The data acquisition module 17 is used to acquire data with at least an optimized acquisition parameter vector.

[0084] Preferably, in one embodiment of the present invention, the instruction generation module includes: The scenario determination unit is used to determine the corresponding power grid business scenario based on optimization instructions; The parameter calling unit is used to call the scenario parameter configuration library according to the power grid business scenario. The scenario parameter configuration library pre-stores the mapping relationship between each power grid business scenario and the benchmark acquisition parameter vector. The loading unit is used to load the corresponding reference acquisition parameter vectors for different power grid business scenarios.

[0085] Preferably, in one embodiment of the present invention, the load determination module includes: The execution unit is used to perform data acquisition according to the sensor resource scheduling strategy of the benchmark acquisition parameter vector; The spatiotemporal alignment processing unit is used to synchronously perform spatiotemporal alignment processing of multi-source data during the data acquisition process; The data preprocessing unit is used to perform preliminary preprocessing on the collected data after data acquisition to obtain a valid dataset. The extraction unit is used to extract a set of data load feature parameters based on the effective data set; The comparison unit is used to determine the corresponding real-time load value based on the set of data load characteristic parameters and a preset load evaluation standard.

[0086] Preferably, in one embodiment of the present invention, the level determination module includes: The threshold decision unit is used to predefine the corresponding set of load thresholds for different data acquisition modes; The load assessment unit is used to compare the real-time load value with the load threshold set corresponding to the current data acquisition mode in order to determine the load level.

[0087] Preferably, in one embodiment of the present invention, the parameter reconstruction module includes: The first adjustment unit is used to adjust the resolution of the visual data quality in the reference acquisition parameter vector based on the first adjustment amount. The second adjustment unit is used to adjust the duration of the data acquisition period in the reference acquisition parameter vector based on the second adjustment amount. The third adjustment unit is used to adjust the state of the sensor resource scheduling strategy in the benchmark acquisition parameters based on the third adjustment amount.

[0088] Compared with the prior art, the beneficial effects of the embodiments of the present invention are at least one of the following: (1) This invention first generates optimization instructions based on the actual data acquisition needs of the power grid, and calls a benchmark acquisition parameter vector that matches the scenario. This vector covers key dimensions such as visual data quality and acquisition cycle, ensuring that the initial parameter configuration accurately matches the business requirements. During the data acquisition process, the real-time load is determined based on the acquisition results. By matching the corresponding parameter adjustment strategy with the load level, the benchmark parameter vector is dynamically reconstructed and optimized. This optimization logic centered on business data enables the acquisition parameters to be flexibly adjusted according to load changes, avoiding the resource waste or data loss problems caused by traditional fixed parameter acquisition. It significantly improves the targeting and efficiency of power grid data acquisition, and achieves the efficient acquisition goal of "on-demand acquisition and dynamic adaptation".

[0089] (2) In terms of technical practicality, this invention addresses the pain points of existing power grid data acquisition equipment, namely "multiple separate devices, asynchronous data, and low processing efficiency." By combining an integrated terminal with optimization methods, it reduces the number of devices carried and lowers manual operation costs. Regarding scenario adaptability, it addresses the needs of various scenarios such as power grid production and material management by flexibly adjusting parameter vectors to achieve differentiated data acquisition, avoiding the scenario limitations of traditional single acquisition modes and adapting to the data needs of the entire power grid business chain. This invention improves the accuracy and real-time performance of data acquisition, providing reliable data support for core businesses such as power grid equipment defect diagnosis and asset inventory, promoting the digital transformation of the power grid towards a more efficient and flexible direction, and possessing broad application prospects.

[0090] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method of optimizing data collection, characterized by, The method comprises the following steps: generating an optimization instruction of a data collection strategy based on a target data collection requirement; loading a reference collection parameter vector corresponding to the target data collection requirement in response to the optimization instruction, the reference collection parameter vector at least including visual data quality, a data collection period, and a sensing resource scheduling strategy; performing data collection based on the reference collection parameter vector, and determining a real-time load value corresponding to the data collection result; determining a load level under the current data collection strategy according to the real-time load value; querying a pre-constructed dynamic adjustment mapping table according to the load level to obtain a first adjustment amount of the visual data quality, a second adjustment amount of the data collection period, and a third adjustment amount of the sensing resource scheduling strategy; reconstructing the reference collection parameter vector based on the first adjustment amount, the second adjustment amount, and the third adjustment amount to obtain an optimized collection parameter vector; performing data collection at least with the optimized collection parameter vector.

2. A method of optimizing data collection as claimed in claim 1, wherein, The step of loading the reference collection parameter vector corresponding to the target data collection requirement in response to the optimization instruction comprises the following steps: determining a corresponding power grid business scenario based on the optimization instruction; calling a scenario parameter configuration library according to the power grid business scenario, wherein the scenario parameter configuration library pre-stores a mapping relationship between each power grid business scenario and the reference collection parameter vector; loading the corresponding reference collection parameter vector for different power grid business scenarios.

3. The method of optimizing data collection of claim 1, wherein, The step of performing data collection based on the reference collection parameter vector and determining a real-time load value corresponding to the data collection result comprises the following steps: performing data collection according to the sensing resource scheduling strategy of the reference collection parameter vector; synchronously performing multi-source data space-time alignment processing in the process of data collection; performing preliminary preprocessing on the collected data to obtain an effective data set after completing the data collection; extracting a data load feature parameter set based on the effective data set; determining the corresponding real-time load value by comparing the data load feature parameter set with a preset load evaluation standard.

4. The method of optimizing data collection of claim 1, wherein, The step of determining a load level under the current data collection strategy according to the real-time load value comprises the following steps: predefining a corresponding load threshold set for different data collection strategies; comparing the real-time load value with the load threshold set corresponding to the current data collection strategy to determine the load level.

5. The method of optimizing data collection of claim 1, wherein, The step of reconstructing the reference collection parameter vector based on the first adjustment amount, the second adjustment amount, and the third adjustment amount comprises the following steps: performing resolution adjustment on the visual data quality in the reference collection parameter vector based on the first adjustment amount; performing time length adjustment on the data collection period in the reference collection parameter vector based on the second adjustment amount; performing state adjustment on the sensing resource scheduling strategy in the reference collection parameter vector based on the third adjustment amount.

6. An optimization system for data collection, characterized by, The method comprises the following steps: an instruction generation module is configured to generate an optimization instruction of a data collection strategy based on a target data collection requirement; The vector loading module is configured to load, in response to the optimization instruction, a benchmark collection parameter vector corresponding to the target data collection requirement, the benchmark collection parameter vector including at least visual data quality, a data collection period, and a sensing resource scheduling strategy. The load determination module is configured to determine, based on the benchmark collection parameter vector, a real-time load value corresponding to a data collection result. The level determination module is configured to determine, according to the real-time load value, a load level under the current data collection strategy. The query module is configured to query, according to the load level, a pre-constructed dynamic adjustment mapping table to obtain a first adjustment amount for the visual data quality, a second adjustment amount for the data collection period, and a third adjustment amount for the sensing resource scheduling strategy. The parameter reconstruction module is configured to reconstruct, based on the first adjustment amount, the second adjustment amount, and the third adjustment amount, the benchmark collection parameter vector to obtain an optimized collection parameter vector. The data collection module is configured to perform data collection at least with the optimized collection parameter vector.

7. An optimized system for data collection as defined in claim 6, wherein, The instruction generation module includes: The scene determination unit is configured to determine, based on the optimization instruction, a corresponding power grid business scene. The parameter calling unit is configured to call, according to the power grid business scene, a scene parameter configuration library in which a mapping relationship between each power grid business scene and the benchmark collection parameter vector is pre-stored. The loading unit is configured to load, for different power grid business scenes, corresponding benchmark collection parameter vectors.

8. The system for optimizing data collection of claim 6, wherein, The load determination module includes: The execution unit is configured to perform data collection according to the sensing resource scheduling strategy of the benchmark collection parameter vector. The space-time alignment processing unit is configured to synchronously perform multi-source data space-time alignment processing in the process of data collection. The data preprocessing unit is configured to preliminarily preprocess collected data to obtain an effective data set after the data collection is completed. The extraction unit is configured to extract a data load feature parameter set based on the effective data set. The comparison unit is configured to determine a corresponding real-time load value according to the data load feature parameter set and a preset load evaluation standard.

9. The system for optimizing data collection of claim 6, wherein, The level determination module includes: The threshold decision unit is configured to predefine a corresponding load threshold set for different data collection modes. The load evaluation unit is configured to compare the real-time load value with the load threshold set corresponding to the current data collection mode to determine the load level.

10. The system for optimizing data collection of claim 6, wherein, The parameter reconstruction module includes: The first adjustment unit is configured to perform resolution adjustment on the visual data quality in the benchmark collection parameter vector based on the first adjustment amount. The second adjustment unit is configured to perform time length adjustment on the data collection period in the benchmark collection parameter vector based on the second adjustment amount. The third adjustment unit is configured to perform state adjustment on the sensing resource scheduling strategy in the benchmark collection parameter vector based on the third adjustment amount.

Images