Data acquisition method and system and storage medium

Abstract

The invention discloses a data acquisition method and system and a storage medium, and relates to the technical field of data acquisition and processing.According to the technical scheme, multiple sensor devices and data interfaces are integrated, multi-source data are acquired in real time, format standardization processing, time synchronization correction and spatial information alignment are carried out through a multi-mode fusion module, and the data acquisition efficiency is improved. Calculating to obtain a data consistency factor Tyhz and evaluating the data consistency factor Tyhz; when the data consistency factor Tyhz does not reach the standard, a data optimization module performs noise filtering and abnormal value elimination on a multi-source data set, and an intelligent analysis module calculates a risk prediction parameter Fcyz by using a convolutional neural network; the early warning evaluation module compares the Fcyz with a risk evaluation threshold Fth, calculates a risk early warning index Gyzs, compares the risk early warning index Gyzs with a risk early warning threshold E, and dynamically generates an early warning execution scheme, so that information pushing and emergency resource scheduling are realized, the problems of low multi-source data fusion efficiency and insufficient early warning precision are solved, and the safety of the system is improved. And the risk identification and emergency response capabilities of the urban emergency management system are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data acquisition and processing, and in particular to a data acquisition method, system and storage medium. Background Art

[0002] The data collection system within the IoT (Internet of Things) system for urban emergency management originated from sensor network technology in the late 20th century. With the rapid development of informatization and intelligentization, it has gradually evolved into a comprehensive system integrating multiple sensors, communication technologies, and cloud computing. Early data collection focused on single-point monitoring, with limited functionality and insufficient real-time performance. Since the beginning of the 21st century, the widespread adoption of IoT technology has driven the scale and diversification of data collection, encompassing a wide range of areas, including environmental monitoring, traffic management, and disaster warning. In recent years, the introduction of 5G communications, edge computing, and artificial intelligence technologies has enabled data collection systems to possess enhanced real-time processing capabilities and intelligence. In urban emergency management, modern data collection systems can efficiently integrate multi-source data, enabling real-time monitoring, accurate early warning, and scientific decision-making support, providing a crucial guarantee for enhancing urban resilience and mitigating the impact of emergencies.

[0003] However, in the urban emergency management IoT system, the existing data collection system has the following technical shortcomings:

[0004] 1. Insufficient multi-source data fusion capabilities: Existing systems generally have difficulty achieving efficient fusion of multiple sensor data, especially in complex environments such as hazardous chemical production, storage, and transportation. Data from different sources come in diverse formats and have poor consistency, resulting in inefficient information integration and affecting the accuracy of decision support.

[0005] 2. The accuracy of intelligent early warning needs to be improved: Although artificial intelligence algorithms are widely used in data collection and analysis, in the early warning of hazardous source events, due to the limitations of historical data quality and algorithm model accuracy, there are still problems of false alarms or missed alarms, which cannot fully meet the precise command needs of emergency management. Summary of the Invention

[0006] In view of the deficiencies in the prior art, the present invention provides a data acquisition method, system and storage medium, which solve the technical shortcomings mentioned in the background technology.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a data acquisition system, including a data source module, a multi-mode fusion module, a data optimization module, an intelligent analysis module and an early warning assessment module;

[0008] The data source module acquires multi-source data by integrating various sensor devices and data interfaces, including environmental monitoring data, equipment operation data, video image data, and historical event data.

[0009] The multimodal fusion module is used to perform cross-modal fusion of multi-source data. It uses a dynamic weight allocation mechanism to assign fusion weights to sensor data based on scenario characteristics. It then outputs a fusion confidence matrix. It calculates the data consistency factor Tyhz based on fuzzy logic evaluation and annotates the data source credibility. Finally, it constructs a multi-source data fusion set and calculates the fusion quality assessment coefficient Rzhl for dynamic optimization.

[0010] The data optimization module filters noise and removes outliers from multi-source data sets based on the evaluation content of the data consistency factor Tyhz; then obtains and evaluates the data optimization index Ujhz through calculation;

[0011] The intelligent analysis module uses deep learning technology to extract event features, match risk patterns, and predict trends from optimized multi-source data sets. It also constructs an analysis data set based on the characteristic information in the analysis results. It generates the risk prediction parameter Fcyz by extracting the analysis data set and fitting it with the consistency factor Tyhz.

[0012] The early warning assessment module is used to evaluate the risk prediction parameter Fcyz; then calculate and evaluate the risk warning index Gyzs, and finally generate an early warning execution plan, including information push and emergency resource scheduling.

[0013] Preferably, the data source module acquires and aggregates multi-source data from different sources in real time by integrating various types of sensor devices and diversified data interfaces. Video image-related data uses high-resolution cameras and infrared imaging devices to collect on-site images and video data in real time, and historical event-related data is called from local databases and cloud platforms, including accident records, emergency response plans, and equipment maintenance records. The collected multi-source data is then preliminarily screened and format converted, and the data is labeled using a data classification algorithm, and grouped and organized according to data type, timestamp, and spatial location information. The multi-source data is then transmitted to the multi-mode fusion module in real time through a built-in data synchronization mechanism.

[0014] Preferably, the multi-mode fusion module includes a data verification unit and a fusion evaluation unit;

[0015] The data verification unit first converts the format of multi-source data from different data sources; then synchronizes the multi-source data through timestamps; finally, spatially aligns the multi-source data based on spatial positioning information and constructs a multi-source data set;

[0016] Based on multi-source data sets, using correlation analysis, similarity matching and data deviation detection algorithms, we can obtain the time synchronization deviation value Psj, spatial positioning deviation value Pkj and feature correlation Ptz, and obtain the data consistency factor Tyhz through calculation;

[0017] Then, by presetting the first data consistency threshold Q1 and the second data consistency threshold Q2, a comparative evaluation is performed with the data consistency factor Tyhz, and the credibility of the data source is marked; and the first data consistency threshold Q1 is greater than the second data consistency threshold Q2, the following evaluation content is generated:

[0018] When the data consistency factor Tyhz ≥ the first data consistency threshold Q1, it is determined to be a "high-trusted data source" and is given the highest weight in the fusion calculation, directly participating in the final decision;

[0019] If the second data consistency threshold Q2 ≤ the data consistency factor Tyhz < the first data consistency threshold Q1, the data source is determined to be a "medium-trusted data source" and is given a medium weight in the fusion calculation and cross-validated with other data sources.

[0020] If the data consistency factor Tyhz is less than the second data consistency threshold Q2, it is determined to be a "low-trust data source" and is given the lowest weight in the fusion calculation, and re-collected and verified;

[0021] In addition, when the data consistency factor Tyhz of the current data source is less than the second data consistency threshold Q2 within a fixed number of times, it is marked as "equipment to be repaired" and the current data source is excluded from the fusion calculation.

[0022] Preferably, the fusion evaluation unit uses a dynamic weight allocation mechanism to assign fusion weights to sensor data based on scene characteristics, combines domain knowledge graphs for semantic fusion and outputs a fusion confidence matrix to obtain data reliability Twhz and data timeliness Tvhz;

[0023] Secondly, the data consistency factor Tyhz, data reliability Twhz and data timeliness Tvhz are extracted and dimensionlessly processed, and the fusion quality evaluation coefficient Rzhl is calculated for dynamic optimization.

[0024] Preferably, the data optimization module includes a data cleaning calculation unit and an optimization evaluation unit;

[0025] The data cleaning calculation unit is used to perform noise filtering and outlier removal operations on the multi-source data set; when the data consistency factor Tyhz of the multi-source data set is evaluated as unqualified, firstly, a multidimensional filtering algorithm is used to smooth the random noise in the multi-source data set; secondly, an anomaly detection algorithm is used to identify and remove outliers that deviate from normal data; finally, a mean square error calculation model is constructed and used to analyze the multi-source data set after the noise filtering and outlier removal operations have been performed, to obtain the noise residual rate Nr, the abnormal data ratio Er and the mean square error Ms, and to calculate the data optimization index Ujhz;

[0026] The optimization evaluation unit is used to preset the data optimization threshold W and perform comparative evaluation on the data optimization index Ujhz. The specific evaluation contents are as follows:

[0027] When the data optimization index Ujhz ≥ the data optimization threshold W, it means that after noise filtering and outlier removal, the multi-source data set has met the optimization criteria and the data quality is qualified;

[0028] When the data optimization index Ujhz is less than the data optimization threshold W, it means that after noise filtering and outlier removal, the multi-source data set still does not meet the optimization criteria. At this time, further optimization processing is performed, including starting the data correction mechanism.

[0029] Preferably, the intelligent analysis module is used to construct and utilize a convolutional neural network model to automatically extract key feature information from a multi-source data set, including abnormal environmental changes, abnormal equipment operation patterns, and historical event correlation features; secondly, to match potential risk patterns in the multi-source data set through multi-dimensional feature analysis and pattern recognition technology; finally, to perform predictive analysis on future potential risk trends in combination with a time series data prediction algorithm; and finally, to construct the extracted feature information into a structured analysis data set;

[0030] By extracting and analyzing the environmental fluctuation intensity Hbd, equipment abnormality frequency Spl, and historical event similarity Lxs in the data set, and fitting them with the consistency factor Tyhz, the risk prediction parameter Fcyz is calculated.

[0031] Preferably, the early warning assessment module includes a risk comparison calculation unit and an early warning strategy assessment unit;

[0032] The risk comparison calculation unit is used to preset the risk assessment threshold Fth and perform comparative evaluation on the risk prediction parameter Fcyz. The specific evaluation contents are as follows:

[0033] When the risk prediction parameter Fcyz ≥ the risk assessment threshold Fth, it means that the currently monitored risk has reached the preset safety threshold and there is a potential danger. At this time, the corresponding risk warning and emergency response mechanism will be immediately activated.

[0034] When the risk prediction parameter Fcyz is less than the risk assessment threshold Fth, it means that the current risk level is within the controllable range, and the system continues to monitor without triggering an early warning.

[0035] When the risk prediction parameter Fcyz has reached or exceeded the risk assessment threshold Fth, the risk warning index Gyzs is calculated using normalization and quantitative analysis methods.

[0036] Preferably, the early warning strategy evaluation unit is used to evaluate the risk early warning index Gyzs; compare the risk early warning index Gyzs with the preset risk early warning threshold E, and generate an early warning execution plan; the specific contents are as follows:

[0037] When the risk warning index Gyzs ≥ the preset risk warning threshold E: the current risk is marked as a priority warning. The information push method adopted at this time is: a priority warning notification is sent to the emergency command center, management departments at all levels, and emergency teams in real time;

[0038] The specific content of emergency resource dispatch is: immediately launch a comprehensive emergency response, deploy all emergency supplies, including firefighting, medical and security equipment, assemble emergency personnel, coordinate multi-department linkage, and comprehensively deal with the incident;

[0039] When the risk warning index Gyzs is less than the preset risk warning threshold E: the current risk is marked as a secondary warning. The information push method adopted at this time is to send risk reminders to relevant personnel to inform them of the existing safety hazards;

[0040] The specific contents of emergency resource dispatch are: strengthen monitoring of high-risk areas and key equipment, pre-position some emergency resources, and be on standby for response.

[0041] A data collection method comprises the following steps:

[0042] Step 1: Integrate multiple sensor devices and data interfaces to obtain multi-source data, including environmental monitoring data, equipment operation data, video image data, and historical event data;

[0043] Step 2: Perform cross-modal fusion on multi-source data, including format standardization, time synchronization correction, and spatial information alignment; use a dynamic weight allocation mechanism to assign fusion weights to sensor data based on scene characteristics; combine domain knowledge graphs for semantic fusion and output a fusion confidence matrix; calculate the data consistency factor Tyhz based on fuzzy logic evaluation and annotate the data source credibility; finally, construct a multi-source data fusion set and calculate the fusion quality assessment coefficient Rzhl for dynamic optimization;

[0044] Step 3: Based on the evaluation content of the data consistency factor Tyhz, the multi-source data set is filtered for noise and outliers are eliminated; then the data optimization index Ujhz is obtained by calculation and evaluated;

[0045] Step 4: Use deep learning technology to extract event features, match risk patterns, and predict trends on the optimized multi-source data set, and construct an analysis data set based on the feature information in the analysis results; generate the risk prediction parameter Fcyz by extracting the analysis data set and fitting it with the consistency factor Tyhz;

[0046] Step 5: Preset the risk assessment threshold Fth and compare it with the risk prediction parameter Fcyz; based on the evaluation content of the risk prediction parameter Fcyz, calculate and evaluate the risk warning index Gyzs, and finally generate a warning execution plan, including information push and emergency resource scheduling.

[0047] A data acquisition storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the process of a data acquisition system as described in any one of claims 1 to 8.

[0048] The present invention provides a data acquisition method, system, and storage medium. It has the following beneficial effects:

[0049] (1) A data acquisition method, system and storage medium, by integrating multiple sensor devices and diversified data interfaces, acquires environmental monitoring related data, equipment operation related data, video image related data and historical event related data in real time, uses a multi-mode fusion module to perform format standardization, time synchronization correction and spatial information alignment on multi-source data, further obtains time synchronization deviation value Psj, spatial positioning deviation value Pkj and feature correlation degree Ptz by calculation, and uses data consistency factor Tyhz to evaluate the multi-source data fusion effect; when the data consistency factor Tyhz does not meet the standard, the data optimization module uses the noise residual rate Nr, abnormal data ratio Er and mean square error Ms are calculated, and multidimensional filtering algorithm and anomaly detection algorithm are executed to filter noise and eliminate outliers in the data set. The optimized data is automatically extracted from the environmental fluctuation intensity Hbd, equipment abnormality frequency Spl and historical event similarity Lxs through the intelligent analysis module based on the convolutional neural network model, and the risk prediction parameter Fcyz is obtained by fitting with the data consistency factor Tyhz. This effectively solves the problem of low data integration efficiency caused by diverse formats and poor consistency of multi-source data; through multidimensional feature analysis, pattern recognition and time series data prediction, the accuracy and stability of data fusion are comprehensively improved, providing high-quality data support for subsequent risk analysis and decision-making.

[0050] (2) A data collection method, system and storage medium. In the early warning assessment link, the early warning assessment module calculates the risk warning index Gyzs by comparing the risk prediction parameter Fcyz with the preset risk assessment threshold Fth, using normalization and quantitative analysis methods, and dynamically generates early warning execution plans of different levels by comparing it with the preset risk warning threshold E, including real-time push of priority early warning notifications to the emergency command center, management departments at all levels and emergency teams, and automatic allocation of emergency resources such as fire fighting, medical and security equipment according to the risk level. Combined with multi-source data real-time monitoring and deep learning technology, the system's risk identification and accurate early warning capabilities in complex environments are significantly improved; the false alarm and missed alarm problems caused by low historical data quality and insufficient model accuracy are effectively reduced, and the urban emergency management system's prevention and control capabilities for hazardous source events and emergency response efficiency are comprehensively improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a schematic diagram of the framework structure of a data acquisition system of the present invention;

[0052] Figure 2 The figure is a flow chart of the steps of a data acquisition method of the present invention. DETAILED DESCRIPTION

[0053] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0054] Example 1

[0055] See also Figure 1 ,A data acquisition system includes a data source module, a multi-mode fusion module, a data optimization module, an intelligent analysis module and an early warning assessment module;

[0056] The data source module acquires multi-source data by integrating multiple sensor devices and data interfaces, including environmental monitoring related data, equipment operation related data, video image related data and historical event related data;

[0057] The data source module acquires multi-source data by integrating various sensor devices and data interfaces, including environmental monitoring data, equipment operation data, video image data, and historical event data.

[0058] The multimodal fusion module is used to perform cross-modal fusion of multi-source data. It uses a dynamic weight allocation mechanism to assign fusion weights to sensor data based on scenario characteristics. It then outputs a fusion confidence matrix. It calculates the data consistency factor Tyhz based on fuzzy logic evaluation and annotates the data source credibility. Finally, it constructs a multi-source data fusion set and calculates the fusion quality assessment coefficient Rzhl for dynamic optimization.

[0059] The data optimization module filters noise and removes outliers from multi-source data sets based on the evaluation content of the data consistency factor Tyhz; then obtains and evaluates the data optimization index Ujhz through calculation;

[0060] The intelligent analysis module uses deep learning technology to extract event features, match risk patterns, and predict trends from optimized multi-source data sets. It also constructs an analysis data set based on the characteristic information in the analysis results. It generates the risk prediction parameter Fcyz by extracting the analysis data set and fitting it with the consistency factor Tyhz.

[0061] The early warning assessment module is used to evaluate the risk prediction parameter Fcyz; then calculate and evaluate the risk warning index Gyzs, and finally generate an early warning execution plan, including information push and emergency resource scheduling.

[0062] In this embodiment, the data source module integrates multiple sensor devices and data interfaces to acquire real-time data related to environmental monitoring, equipment operation, video images, and historical events. This fully covers multi-dimensional data sources and ensures comprehensive and real-time data collection.

[0063] The multi-mode fusion module calculates and evaluates the data consistency factor Tyhz through format standardization, time synchronization correction, and spatial information alignment. This can effectively solve the problem of low data integration efficiency caused by diverse multi-source data formats, time asynchrony, and spatial deviation, and improve the quality of data fusion.

[0064] The data optimization module uses noise filtering and outlier removal operations based on the evaluation content of the data consistency factor Tyhz, and calculates the data optimization index Ujhz, which can effectively remove noise and abnormal data and improve the accuracy and reliability of the data;

[0065] The intelligent analysis module extracts event feature information from the optimized multi-source data set based on deep learning technology, performs risk pattern matching and trend prediction, constructs an analysis data set, and generates a risk prediction parameter Fcyz by fitting with the data consistency factor Tyhz. This can accurately identify potential risks and predict risk development trends, thereby improving the accuracy of risk identification.

[0066] The early warning assessment module calculates the risk warning index Gyzs by comparing the risk prediction parameter Fcyz with the preset risk assessment threshold Fth, and generates an early warning implementation plan based on the assessment results, implements information push and emergency resource scheduling, and can quickly respond to potential risks and link relevant departments in real time, thereby improving the early warning response speed and emergency handling capabilities of the urban emergency management system.

[0067] Example 2

[0068] The data source module acquires and aggregates multi-source data from different sources in real time by integrating various types of sensor devices and diversified data interfaces. Video image-related data uses high-resolution cameras and infrared imaging devices to collect on-site images and video data in real time, while historical event-related data is called from local databases and cloud platforms, including accident records, emergency response plans, and equipment maintenance records. The collected multi-source data is then preliminarily screened and formatted, and the data is labeled using a data classification algorithm. The data is then grouped and organized according to data type, timestamp, and spatial location information. The multi-source data is then transmitted to the multi-mode fusion module in real time through a built-in data synchronization mechanism.

[0069] The multi-mode fusion module includes a data verification unit and a fusion evaluation unit;

[0070] The data verification unit first converts the format of multi-source data from different data sources; then synchronizes the multi-source data through timestamps; finally, spatially aligns the multi-source data based on spatial positioning information and constructs a multi-source data set;

[0071] Based on multi-source data sets, we use correlation analysis, similarity matching, and data deviation detection algorithms to obtain the time synchronization deviation value Psj, spatial positioning deviation value Pkj, and feature correlation Ptz. We then calculate the data consistency factor Tyhz using the following formula:

[0072]

[0073] Then, by presetting the first data consistency threshold Q1 and the second data consistency threshold Q2, a comparative evaluation is performed with the data consistency factor Tyhz, and the credibility of the data source is marked; and the first data consistency threshold Q1 is greater than the second data consistency threshold Q2, the following evaluation content is generated:

[0074] When the data consistency factor Tyhz ≥ the first data consistency threshold Q1, it is determined to be a "high-trusted data source" and is given the highest weight in the fusion calculation, directly participating in the final decision;

[0075] If the second data consistency threshold Q2 ≤ the data consistency factor Tyhz < the first data consistency threshold Q1, the data source is determined to be a "medium-trusted data source" and is given a medium weight in the fusion calculation and cross-validated with other data sources.

[0076] If the data consistency factor Tyhz is less than the second data consistency threshold Q2, it is determined to be a "low-trust data source" and is given the lowest weight in the fusion calculation, and re-collected and verified;

[0077] In addition, when the data consistency factor Tyhz of the current data source is less than the second data consistency threshold Q2 within a fixed number of times, it is marked as "equipment to be repaired" and the current data source is excluded from the fusion calculation.

[0078] The fusion evaluation unit uses a dynamic weight allocation mechanism to assign fusion weights to sensor data based on scene characteristics, combines domain knowledge graphs for semantic fusion and outputs a fusion confidence matrix to obtain data reliability Twhz and data timeliness Tvhz;

[0079] Secondly, the data consistency factor Tyhz, data reliability Twhz, and data timeliness Tvhz are extracted and dimensionlessly processed to calculate the fusion quality evaluation coefficient Rzhl. The specific calculation formula is as follows:

[0080]

[0081] Perform dynamic optimization;

[0082] The data optimization module includes a data cleaning calculation unit and an optimization evaluation unit;

[0083] The data cleaning calculation unit is used to perform noise filtering and outlier removal operations on the multi-source data set; when the data consistency factor Tyhz of the multi-source data set is evaluated as unqualified, firstly, a multidimensional filtering algorithm is used to smooth the random noise in the multi-source data set; secondly, an anomaly detection algorithm is used to identify and remove outliers that deviate from normal data; finally, a mean square error calculation model is constructed and used to analyze the multi-source data set after the noise filtering and outlier removal operations have been performed, and the noise residual rate Nr, the abnormal data ratio Er and the mean square error Ms are obtained, and the data optimization index Ujhz is calculated using the following formula:

[0084]

[0085] The optimization evaluation unit is used to preset the data optimization threshold W and perform comparative evaluation on the data optimization index Ujhz. The specific evaluation contents are as follows:

[0086] When the data optimization index Ujhz ≥ the data optimization threshold W, it means that after noise filtering and outlier removal, the multi-source data set has met the optimization criteria and the data quality is qualified;

[0087] When the data optimization index Ujhz is less than the data optimization threshold W, it means that after noise filtering and outlier removal, the multi-source data set still does not meet the optimization criteria. At this time, further optimization processing is performed, including starting the data correction mechanism.

[0088] In this embodiment, various types of sensor devices and diversified data interfaces are integrated through the data source module to obtain real-time environmental monitoring related data, equipment operation related data, video image related data and historical event related data. The video image data is collected in real time by high-resolution cameras and infrared imaging devices, and the historical event data is called from the local database and cloud platform, effectively ensuring the comprehensiveness and real-time nature of data collection; the multi-mode fusion module converts the format, time synchronizes and spatially aligns the multi-source data through the data calculation unit to construct a multi-source data set, and uses correlation analysis, similarity matching and data deviation detection algorithms to obtain the time synchronization deviation value Psj, spatial positioning deviation value Pkj and feature correlation degree Ptz, and then calculates the data consistency factor Tyhz to achieve accurate evaluation of the consistency of multi-source data fusion; the consistency evaluation unit compares the preset data consistency threshold Q with the data consistency factor Tyhz to determine whether the data fusion effect meets the standard, thereby ensuring the stability and reliability of the data;

[0089] The data optimization module uses the data cleaning calculation unit to filter noise and remove outliers from the data set, adopts a multidimensional filtering algorithm to smooth random noise, uses an anomaly detection algorithm to remove data that deviates from the normal state, and obtains the noise residual rate Nr, abnormal data ratio Er and mean square error Ms through the mean square error calculation model, calculates the data optimization index Ujhz, and further improves data quality; the optimization evaluation unit compares and evaluates the data optimization index Ujhz using the preset data optimization threshold W to determine whether the data meets the standard. If not, the data correction mechanism is activated to ensure the high consistency and reliability of multi-source data;

[0090] The synergistic effect of various modules enables the system to have powerful data collection, integration, optimization and analysis capabilities, improves data processing accuracy and risk identification capabilities, and provides efficient and accurate data support for urban emergency management and production safety monitoring.

[0091] Example 3

[0092] The intelligent analysis module is used to build and utilize a convolutional neural network model to automatically extract key feature information from multi-source data sets, including abnormal environmental changes, abnormal equipment operation patterns, and historical event correlation features. Secondly, it matches potential risk patterns in the multi-source data sets through multi-dimensional feature analysis and pattern recognition technology. Finally, it combines time series data prediction algorithms to predict and analyze future potential risk trends. Finally, the extracted feature information is constructed into a structured analysis data set.

[0093] The risk prediction parameter Fcyz is calculated by extracting and analyzing the environmental fluctuation intensity Hbd, the equipment abnormality frequency Spl, and the historical event similarity Lxs in the data set, fitting them with the consistency factor Tyhz. The specific calculation formula is as follows:

[0094]

[0095] The early warning assessment module includes a risk comparison calculation unit and an early warning strategy assessment unit;

[0096] The risk comparison calculation unit is used to preset the risk assessment threshold Fth and perform comparative evaluation on the risk prediction parameter Fcyz. The specific evaluation contents are as follows:

[0097] When the risk prediction parameter Fcyz ≥ the risk assessment threshold Fth, it means that the currently monitored risk has reached the preset safety threshold and there is a potential danger. At this time, the corresponding risk warning and emergency response mechanism will be immediately activated.

[0098] When the risk prediction parameter Fcyz is less than the risk assessment threshold Fth, it means that the current risk level is within the controllable range, and the system continues to monitor without triggering an early warning.

[0099] When the risk prediction parameter Fcyz has reached or exceeded the risk assessment threshold Fth, the risk warning index Gyzs is calculated using the normalization and quantitative analysis method. The specific calculation formula is as follows:

[0100]

[0101] The warning strategy evaluation unit is used to evaluate the risk warning index Gyzs; compare the risk warning index Gyzs with the preset risk warning threshold E, and generate a warning execution plan; the specific contents are as follows:

[0102] When the risk warning index Gyzs ≥ the preset risk warning threshold E: the current risk is marked as a priority warning. The information push method adopted at this time is: a priority warning notification is sent to the emergency command center, management departments at all levels, and emergency teams in real time;

[0103] The specific content of emergency resource dispatch is: immediately launch a comprehensive emergency response, deploy all emergency supplies, including firefighting, medical and security equipment, assemble emergency personnel, coordinate multi-department linkage, and comprehensively deal with the incident;

[0104] When the risk warning index Gyzs is less than the preset risk warning threshold E: the current risk is marked as a secondary warning. The information push method adopted at this time is to send risk reminders to relevant personnel to inform them of the existing safety hazards;

[0105] The specific contents of emergency resource dispatch are: strengthen monitoring of high-risk areas and key equipment, pre-position some emergency resources, and be on standby for response.

[0106] In this embodiment, the data acquisition system is constructed through an intelligent analysis module and utilizes a convolutional neural network model to automatically extract abnormal environmental changes, abnormal equipment operation modes, and historical event correlation features from multi-source data sets to effectively identify potential risks; the risk patterns in the data are accurately matched through multi-dimensional feature analysis and pattern recognition technology, and future risk trends are predicted by combining the time series data prediction algorithm, extracting the environmental fluctuation intensity Hbd, the equipment abnormality frequency Spl, and the historical event similarity Lxs, and fitting the data consistency factor Tyhz to obtain the risk prediction parameter Fcyz, thereby realizing quantitative analysis of the risk status; the early warning assessment module sets the risk assessment through the risk comparison calculation unit The threshold Fth is compared with the risk prediction parameter Fcyz to determine whether the risk meets the warning conditions. The normalization and quantitative analysis methods are used to calculate the risk warning index Gyzs to accurately quantify the risk level. The warning strategy evaluation unit compares the risk warning index Gyzs with the preset risk warning threshold E to generate a graded warning implementation plan, and sends risk reminder information based on the assessment content, strengthens the monitoring of high-risk areas and key equipment, and pre-positions emergency resources on standby response. The synergistic effect of each module ensures that the system can accurately identify, warn in real time, and efficiently respond to potential risks, significantly improving the city's emergency management and safety prevention and control capabilities, and achieving rapid perception, scientific decision-making, and efficient handling of emergencies.

[0107] Example 4

[0108] See also Figure 2 , a data collection method, comprising the following steps:

[0109] Step 1: Integrate multiple sensor devices and data interfaces to obtain multi-source data, including environmental monitoring data, equipment operation data, video image data, and historical event data;

[0110] Step 2: Perform cross-modal fusion on multi-source data, including format standardization, time synchronization correction, and spatial information alignment; use a dynamic weight allocation mechanism to assign fusion weights to sensor data based on scene characteristics; combine domain knowledge graphs for semantic fusion and output a fusion confidence matrix; calculate the data consistency factor Tyhz based on fuzzy logic evaluation and annotate the data source credibility; finally, construct a multi-source data fusion set and calculate the fusion quality assessment coefficient Rzhl for dynamic optimization;

[0111] Step 3: Based on the evaluation content of the data consistency factor Tyhz, the multi-source data set is filtered for noise and outliers are eliminated; then the data optimization index Ujhz is obtained by calculation and evaluated;

[0112] Step 4: Use deep learning technology to extract event features, match risk patterns, and predict trends on the optimized multi-source data set, and construct an analysis data set based on the feature information in the analysis results; generate the risk prediction parameter Fcyz by extracting the analysis data set and fitting it with the consistency factor Tyhz;

[0113] Step 5: Preset the risk assessment threshold Fth and compare it with the risk prediction parameter Fcyz; based on the evaluation content of the risk prediction parameter Fcyz, calculate and evaluate the risk warning index Gyzs, and finally generate a warning execution plan, including information push and emergency resource scheduling.

[0114] It also includes a data acquisition storage medium on which a computer program is stored, characterized in that when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

[0115] In this embodiment, the above-mentioned data acquisition storage medium can be implemented in whole or in part by software, hardware, firmware, or any other combination thereof; when implemented using software, the above-mentioned embodiment can be implemented in whole or in part in the form of a computer program product. Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution.

[0116] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0117] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A data acquisition system, characterized in that: It includes data source module, multi-mode fusion module, data optimization module, intelligent analysis module and early warning assessment module; The data source module acquires multi-source data by integrating various sensor devices and data interfaces, including environmental monitoring data, equipment operation data, video image data, and historical event data. The multimodal fusion module is used to perform cross-modal fusion of multi-source data; a dynamic weight allocation mechanism is used to assign fusion weights to sensor data based on scenario characteristics; the fusion confidence matrix is ​​then output; the data consistency factor Tyhz is calculated based on fuzzy logic evaluation and the data source credibility is annotated; Finally, a multi-source data fusion set is constructed and the fusion quality evaluation coefficient Rzhl is calculated for dynamic optimization; The data optimization module filters noise and removes outliers from multi-source data sets based on the evaluation content of the data consistency factor Tyhz; then obtains and evaluates the data optimization index Ujhz through calculation; The intelligent analysis module uses deep learning technology to extract event features, match risk patterns, and predict trends from optimized multi-source data sets. It also constructs an analysis data set based on the characteristic information in the analysis results. It generates the risk prediction parameter Fcyz by extracting the analysis data set and fitting it with the consistency factor Tyhz. The early warning assessment module is used to evaluate the risk prediction parameter Fcyz; then calculate and evaluate the risk warning index Gyzs, and finally generate an early warning execution plan, including information push and emergency resource scheduling.

2. A data acquisition system according to claim 1, characterized in that: The data source module acquires and aggregates multi-source data from different sources in real time by integrating various types of sensor devices and diversified data interfaces. Video image-related data uses high-resolution cameras and infrared imaging devices to collect on-site images and video data in real time, while historical event-related data is called from local databases and cloud platforms, including accident records, emergency response plans, and equipment maintenance records. The collected multi-source data is then preliminarily screened and formatted, and the data is labeled using a data classification algorithm. The data is then grouped and organized according to data type, timestamp, and spatial location information. The multi-source data is then transmitted to the multi-mode fusion module in real time through a built-in data synchronization mechanism.

3. A data acquisition system according to claim 1, characterized in that: The multi-mode fusion module includes a data verification unit and a fusion evaluation unit; The data verification unit first converts the format of multi-source data from different data sources; then synchronizes the multi-source data through timestamps; finally, spatially aligns the multi-source data based on spatial positioning information and constructs a multi-source data set; Based on multi-source data sets, using correlation analysis, similarity matching and data deviation detection algorithms, we can obtain the time synchronization deviation value Psj, spatial positioning deviation value Pkj and feature correlation degree Ptz, and obtain the data consistency factor Tyhz through calculation; Then, by presetting the first data consistency threshold Q1 and the second data consistency threshold Q2, a comparative evaluation is performed with the data consistency factor Tyhz, and the credibility of the data source is marked; and the first data consistency threshold Q1 is greater than the second data consistency threshold Q2, the following evaluation content is generated: When the data consistency factor Tyhz ≥ the first data consistency threshold Q1, it is determined to be a "high-trusted data source" and is given the highest weight in the fusion calculation, directly participating in the final decision; If the second data consistency threshold Q2 ≤ the data consistency factor Tyhz < the first data consistency threshold Q1, the data source is determined to be a "medium-trusted data source" and is given a medium weight in the fusion calculation and cross-validated with other data sources. If the data consistency factor Tyhz is less than the second data consistency threshold Q2, it is determined to be a "low-trust data source" and is given the lowest weight in the fusion calculation, and re-collected and verified; In addition, when the data consistency factor Tyhz of the current data source is less than the second data consistency threshold Q2 within a fixed number of times, it is marked as "device to be repaired" and the current data source is excluded from the fusion calculation.

4. A data acquisition system according to claim 3, characterized in that: The fusion evaluation unit uses a dynamic weight allocation mechanism to assign fusion weights to sensor data based on scene characteristics, combines domain knowledge graphs for semantic fusion and outputs a fusion confidence matrix to obtain data reliability Twhz and data timeliness Tvhz; Secondly, the data consistency factor Tyhz, data reliability Twhz and data timeliness Tvhz are extracted and dimensionlessly processed, and the fusion quality evaluation coefficient Rzhl is calculated for dynamic optimization.

5. The data acquisition system according to claim 1, wherein: The data optimization module includes a data cleaning calculation unit and an optimization evaluation unit; The data cleaning calculation unit is used to perform noise filtering and outlier removal operations on the multi-source data set; when the data consistency factor Tyhz of the multi-source data set is evaluated as unqualified, firstly, a multidimensional filtering algorithm is used to smooth the random noise in the multi-source data set; secondly, an anomaly detection algorithm is used to identify and remove outliers that deviate from normal data; finally, a mean square error calculation model is constructed and used to analyze the multi-source data set after the noise filtering and outlier removal operations have been performed, to obtain the noise residual rate Nr, the abnormal data ratio Er and the mean square error Ms, and to calculate the data optimization index Ujhz; The optimization evaluation unit is used to preset the data optimization threshold W and perform comparative evaluation on the data optimization index Ujhz. The specific evaluation contents are as follows: When the data optimization index Ujhz ≥ the data optimization threshold W, it means that after noise filtering and outlier removal, the multi-source data set has met the optimization criteria and the data quality is qualified; When the data optimization index Ujhz is less than the data optimization threshold W, it means that after noise filtering and outlier removal, the multi-source data set still does not meet the optimization criteria. At this time, further optimization processing is performed, including starting the data correction mechanism.

6. A data acquisition system according to claim 1, characterized in that: The intelligent analysis module is used to build and utilize a convolutional neural network model to automatically extract key feature information from multi-source data sets, including abnormal environmental changes, abnormal equipment operation patterns, and historical event correlation features. Secondly, it matches potential risk patterns in the multi-source data sets through multi-dimensional feature analysis and pattern recognition technology. Finally, it combines time series data prediction algorithms to predict and analyze future potential risk trends. Finally, the extracted feature information is constructed into a structured analysis data set. By extracting and analyzing the environmental fluctuation intensity Hbd, equipment abnormality frequency Spl, and historical event similarity Lxs in the data set, and fitting them with the consistency factor Tyhz, the risk prediction parameter Fcyz is calculated.

7. A data acquisition system according to claim 1, characterized in that: The early warning assessment module includes a risk comparison calculation unit and an early warning strategy assessment unit; The risk comparison calculation unit is used to preset the risk assessment threshold Fth and perform comparative evaluation on the risk prediction parameter Fcyz. The specific evaluation contents are as follows: When the risk prediction parameter Fcyz ≥ the risk assessment threshold Fth, it means that the currently monitored risk has reached the preset safety threshold and there is a potential danger. At this time, the corresponding risk warning and emergency response mechanism will be immediately activated. When the risk prediction parameter Fcyz is less than the risk assessment threshold Fth, it means that the current risk level is within the controllable range, and the system continues to monitor without triggering an early warning. When the risk prediction parameter Fcyz has reached or exceeded the risk assessment threshold Fth, the normalization and quantitative analysis method is used to calculate the risk warning index Gyzs.

8. A data acquisition system according to claim 7, characterized in that: The warning strategy evaluation unit is used to evaluate the risk warning index Gyzs; compare the risk warning index Gyzs with the preset risk warning threshold E, and generate a warning execution plan; the specific contents are as follows: When the risk warning index Gyzs ≥ the preset risk warning threshold E: the current risk is marked as a priority warning. The information push method adopted at this time is: a priority warning notification is sent to the emergency command center, management departments at all levels, and emergency teams in real time; The specific content of emergency resource dispatch is: immediately launch a comprehensive emergency response, deploy all emergency supplies, including firefighting, medical and security equipment, assemble emergency personnel, coordinate multi-department linkage, and comprehensively deal with the incident; When the risk warning index Gyzs is less than the preset risk warning threshold E: the current risk is marked as a secondary warning. The information push method adopted at this time is to send risk reminders to relevant personnel to inform them of the existing safety hazards; The specific contents of emergency resource dispatch are: strengthen monitoring of high-risk areas and key equipment, pre-position some emergency resources, and be on standby for response.

9. A data acquisition method, according to any one of claims 1 to 8, characterized in that: The following steps are involved: Step 1: Integrate multiple sensor devices and data interfaces to obtain multi-source data, including environmental monitoring data, equipment operation data, video image data, and historical event data; Step 2: Cross-modal fusion of multi-source data, including format standardization, time synchronization correction, and spatial information alignment; A dynamic weight allocation mechanism is used to assign fusion weights to sensor data based on scene characteristics; Combine domain knowledge graphs for semantic fusion and output a fusion confidence matrix; calculate data consistency factor Tyhz based on fuzzy logic evaluation and annotate data source credibility; Finally, a multi-source data fusion set is constructed and the fusion quality evaluation coefficient Rzhl is calculated for dynamic optimization; Step 3: Based on the evaluation content of the data consistency factor Tyhz, the multi-source data set is filtered for noise and outliers are eliminated; then the data optimization index Ujhz is obtained by calculation and evaluated; Step 4: Use deep learning technology to extract event features, match risk patterns, and predict trends on the optimized multi-source data set, and construct an analysis data set based on the feature information in the analysis results; generate the risk prediction parameter Fcyz by extracting the analysis data set and fitting it with the consistency factor Tyhz; Step 5: Preset the risk assessment threshold Fth and compare it with the risk prediction parameter Fcyz; based on the evaluation content of the risk prediction parameter Fcyz, calculate and evaluate the risk warning index Gyzs, and finally generate a warning execution plan, including information push and emergency resource scheduling.

10. A data acquisition storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the process of the data acquisition system according to any one of claims 1 to 8 is implemented.

Images