Smart park multi-source data dynamic monitoring and real-time analysis system and method

Abstract

The invention relates to the technical field of data analysis, in particular to a smart park multi-source data dynamic monitoring and real-time analysis system and method, and the system comprises a data collection module, a data processing module, a data analysis module, a data encryption and storage module, a real-time monitoring and early warning module, a comprehensive management module and a communication module. In order to overcome the defects of low data quality, inaccurate feature extraction and the like due to the fact that original data is directly and simply processed and then analyzed in the prior art, the scheme improves the data quality through the preprocessing steps of data cleaning, format conversion, normalization, standardization and the like; according to the method, data reliability is ensured through integrity, accuracy, consistency and interpretability check, then features are accurately selected and extracted, and finally multi-source sensor data are integrated through a data alignment and fusion technology, so that the accuracy and availability of the data are remarkably improved through the comprehensive processing scheme; and a more reliable basis is provided for subsequent data analysis and decision making.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and in particular to a system and method for dynamic monitoring and real-time analysis of multi-source data in a smart park. Background Art

[0002] With the rapid development of smart parks, the demand for dynamic monitoring and real-time analysis of multi-source data in the park is becoming increasingly urgent. Smart parks usually contain a large number of sensors, cameras and IoT devices, which can collect real-time environmental data, video data, audio data and personnel activity information data in the park. However, existing technologies often have some problems when processing these multi-source data.

[0003] First, the quality of raw data is often not high. Since the acquisition equipment may be subject to various interferences and restrictions, the collected data may contain noise, missing data, or inconsistent formats. If these problems are not effectively handled, they will seriously affect the subsequent data analysis and decision-making results.

[0004] Secondly, inaccurate feature extraction is also a challenge faced by existing technologies. In the process of data processing, it is necessary to accurately extract features related to the analysis target. However, existing technologies often lack effective feature extraction methods, resulting in inaccurate or weakly representative features, which affects subsequent analysis and prediction results.

[0005] In addition, existing technologies often lack effective data integration and fusion methods when processing multi-source data. Different acquisition devices may generate data of different types and formats, and there may be redundancy, conflict or inconsistency between these data. If these data cannot be effectively integrated and fused, it will lead to a waste of data resources and reduced decision-making efficiency. Summary of the invention

[0006] In order to overcome the problems raised in the above background technology, the present invention proposes a system and method for dynamic monitoring and real-time analysis of multi-source data in a smart park.

[0007] The technical solution of the present invention is: a smart park multi-source data dynamic monitoring and real-time analysis system, including: The data collection module is used to collect data in the park using sensors, cameras and IoT devices installed in the park. The collected data includes environmental data, video data, audio data and personnel activity information data; The data processing module is used to perform preliminary processing on the collected raw data, wherein the processing methods adopted include data preprocessing, feature extraction and data quality assessment; Data analysis module, which is used to conduct in-depth mining and analysis of processed data using deep learning algorithms, and to perform behavior analysis, pattern recognition, and trend prediction; Data encryption and storage module, used to encrypt sensitive data and store the encrypted data; The real-time monitoring and early warning module is used to visualize the data for users to observe and monitor abnormal behaviors and potential threats in real time based on the results of the data analysis module, and trigger the early warning mechanism; Comprehensive management module, used to provide a comprehensive management interface and user interaction functions; The communication module is used to build a communication network to realize data communication among the data acquisition module, data processing module, data analysis module, data encryption and storage module, and real-time monitoring and early warning module.

[0008] Preferably, the data acquisition module includes the following data acquisition devices: A11: Environmental sensors, including temperature sensors, humidity sensors, and light intensity sensors. The temperature sensor is used to measure and record the temperature information in the park. The humidity sensor is used to monitor the humidity level in the park. The light intensity sensor is used to monitor the light intensity inside the park. A12: Cameras, including surveillance cameras, infrared cameras, and panoramic cameras, used to capture images in the park in real time in a variety of complex scenarios; A13: Audio device, including microphone and audio sensor, used to capture environmental sound; A14: Smart meters for intelligent measurement of energy usage.

[0009] Preferably, the data processing module includes the following steps when performing preliminary processing on the collected raw data: S11: Data preprocessing, including data cleaning, data format conversion, data normalization and data standardization; S12: Data quality assessment: quality assessment of preprocessed data, including completeness check, accuracy check, consistency check and interpretability check; S13: Feature extraction, first select representative, relevant and interpretable features according to the analysis objectives and data characteristics, then calculate the corresponding statistics and indicators based on the selected features, and transform and combine the features to form preset new features.

[0010] Preferably, when the data processing module performs preliminary processing on the collected raw data, it also includes the following steps: S21: data extraction, firstly extracting the sensor data from the raw data after preprocessing; S22: Data alignment, including time alignment and coordinate alignment, aligning the sensor data in time, and converting the data obtained by different sensors into the same coordinate system; S23: Data fusion, using a preset data fusion algorithm to integrate and fuse the data from multiple sensors; S24: Data output, outputting the fused sensor data for processing in the next step, wherein the processing method in the next step is feature extraction.

[0011] Preferably, when the preset data fusion algorithm is used to integrate and fuse the data of multiple sensors, the data fusion algorithm used includes: A21: Weighted average fusion, which performs weighted average of data from multiple sensors, where the weights can be adjusted according to the performance of the model. The principle formula is: ; in, and They are the data of two groups of sensors. and They are and The weight of ; A22: Statistical fusion uses statistical methods and probability theory to combine data from different sources. The principle formula is: ; in, represents parameters related to the reliability of the data source, the noise characteristics of the data, and the weight of the data, It represents the probability of the observed data set X appearing under the given parameter θ. Represents an observation data set, which contains n data observation points.

[0012] Preferably, the data analysis module includes the following steps when using a deep learning algorithm to perform in-depth mining and analysis on the processed data, and perform behavior analysis, pattern recognition and trend prediction: S31: Model training: first, select a suitable deep learning algorithm according to the type of data, and build a behavior analysis, pattern recognition and trend prediction model based on the selected deep learning algorithm, and train the built model; S32: Model application, inputting the data results of the data processing module into the trained model to perform behavior analysis, pattern recognition and trend prediction; S33: Result interpretation and verification: Analyze the output of the model, identify abnormal behaviors, potential threats, and development trends, and compare the analysis results with actual data and known situations to verify the analysis results; S34: Result output: output the analysis result in the form of a report, wherein the output report includes a behavior analysis report, a pattern recognition report and a trend prediction report, and the report includes the analysis process, key findings and recommended measures.

[0013] Preferably, the data processing module adopts a distributed structure, and the data processing module includes multiple groups of edge computing modules, the multiple groups of edge computing modules are connected through communication modules, and the multiple groups of edge computing modules are connected to the data analysis module through the communication module.

[0014] Preferably, the data analysis module, data encryption and storage module, real-time monitoring and early warning module and comprehensive management module are all arranged on the central server, and the communication module includes multiple data transmission modules, and the multiple data transmission modules are respectively arranged on multiple groups of edge computing modules.

[0015] Preferably, the data analysis module further includes the following steps when using a deep learning algorithm to perform deep mining and analysis on the processed data and perform behavior analysis, pattern recognition and trend prediction: S41: Blind spot judgment: analyze the layout and coverage of all data collection devices to identify the current blind spots; S42: Equipment configuration: configure and adjust other data acquisition equipment according to the location and range of the blind area, including adding new data acquisition equipment and adjusting the location and direction of the data acquisition equipment; S43: importance judgment, after configuring and adjusting other data collection devices, judging the importance of the blind area covered by the adjusted data collection device; S44: Adjust the configuration. According to the importance of the blind area and the type and structure of the data to be collected, configure the corresponding collection rules, including adjusting the update frequency.

[0016] A method for dynamic monitoring and real-time analysis of multi-source data in a smart park, comprising the following steps: S51: Data collection, using the data collection equipment installed in the park to collect environmental data, video data, audio data and personnel activity information data in the park; S52: Data preprocessing: cleaning, format conversion, normalization and standardization of the collected raw data; S53: Data fusion, extracting sensor data from the preprocessed data separately, and performing time alignment and coordinate alignment, and then integrating and fusing the data of multiple sensors using a data fusion algorithm; S54: Feature extraction: select representative, relevant and interpretable features according to the analysis objectives and data characteristics, and calculate the corresponding statistics and indicators to form a new feature set; S55: Data analysis and prediction: input the processed data into the trained model to perform behavior analysis, pattern recognition and trend prediction, and analyze the output of the model to identify abnormal behavior, potential threats and development trends; S56: Data storage, encrypting sensitive data and storing the encrypted data in a secure location; S57: Real-time monitoring and early warning, visual display of data to facilitate user observation, and based on the results of the data analysis module, real-time monitoring of abnormal behaviors and potential threats, and triggering of early warning mechanisms.

[0017] Beneficial effects of the present invention: 1. Compared with the existing technology that directly performs simple processing on the raw data and then performs analysis, which has the disadvantages of low data quality and inaccurate feature extraction, this solution adopts a comprehensive processing solution that first pre-processes the data, conducts quality assessment, extracts features, and further performs data extraction, alignment, fusion and output. This solution improves data quality through pre-processing steps such as data cleaning, format conversion, normalization and standardization, ensures data reliability through integrity, accuracy, consistency and interpretability checks, then accurately selects and extracts features, and finally integrates multi-source sensor data through data alignment and fusion technology. This comprehensive processing solution significantly improves the accuracy and availability of data, and provides a more reliable basis for subsequent data analysis and decision-making; 2. Compared with the centralized data processing scheme adopted in the prior art, there may be disadvantages such as low processing efficiency, heavy equipment burden, and the whole system may be paralyzed once the central equipment is damaged. This scheme adopts a distributed data processing module, which includes multiple groups of edge computing modules and realizes efficient interconnection through communication modules. The edge computing module can process data nearby, reduce the pressure of the central server, and improve data processing efficiency. In addition, since multiple edge computing modules coexist, even if a module fails, it will not affect the operation of the whole system, effectively improving the stability and reliability of the system. This technical solution that is biased towards edge computing and the Internet of Everything not only realizes the efficient processing of large amounts of data, but also effectively avoids the risk of system paralysis caused by equipment damage; 3. Compared with the existing technology that adopts static data collection and analysis solutions, there may be shortcomings such as blind spots in data collection and incomplete analysis. This solution adds blind spot judgment, equipment configuration, importance judgment and adjustment configuration steps in the data analysis module. By dynamically analyzing the layout and coverage of data collection equipment, blind spots are identified and equipment configuration and adjustment are carried out in a targeted manner. At the same time, the collection rules are optimized according to the importance of blind spots and data types. This solution not only effectively eliminates data collection blind spots and improves the comprehensiveness and accuracy of data, but also meets the collection needs of different regions and types of data through flexible collection rule configuration, providing strong support for the refined management and decision-making of smart parks. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 Shown is a schematic diagram of the structure of the smart park multi-source data dynamic monitoring and real-time analysis system of the present invention; Figure 2 What is shown is a schematic diagram of the workflow of the method for dynamic monitoring and real-time analysis of multi-source data in a smart park of the present invention. DETAILED DESCRIPTION

[0019] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0020] See also Figure 1 The present invention provides an embodiment: a smart park multi-source data dynamic monitoring and real-time analysis system, including: The data collection module is used to collect data in the park using sensors, cameras and IoT devices installed in the park. The collected data includes environmental data, video data, audio data and personnel activity information data; The data processing module is used to perform preliminary processing on the collected raw data, wherein the processing methods adopted include data preprocessing, feature extraction and data quality assessment; Data analysis module, which is used to conduct in-depth mining and analysis of processed data using deep learning algorithms, and to perform behavior analysis, pattern recognition, and trend prediction; Data encryption and storage module, used to encrypt sensitive data and store the encrypted data; The real-time monitoring and early warning module is used to visualize the data for users to observe and monitor abnormal behaviors and potential threats in real time based on the results of the data analysis module, and trigger the early warning mechanism; Comprehensive management module, used to provide a comprehensive management interface and user interaction functions; The communication module is used to build a communication network to realize data communication among the data acquisition module, data processing module, data analysis module, data encryption and storage module, and real-time monitoring and early warning module.

[0021] As described above, the present invention collects environmental, video, audio and personnel activity information data through sensors, cameras and IoT devices in the park. After preprocessing, feature extraction and quality assessment by the data processing module, the data analysis module uses a deep learning algorithm to perform deep mining, behavior analysis, pattern recognition and trend prediction. Sensitive data is encrypted and stored by the encryption and storage module, and the real-time monitoring and early warning module realizes data visualization, abnormal behavior monitoring and early warning. In addition, the comprehensive management module provides a user interaction interface, and the communication module ensures smooth data communication between modules.

[0022] Preferably, the data acquisition module includes the following data acquisition devices: A11: Environmental sensors, including temperature sensors, humidity sensors, and light intensity sensors. The temperature sensor is used to measure and record the temperature information in the park. The humidity sensor is used to monitor the humidity level in the park. The light intensity sensor is used to monitor the light intensity inside the park. A12: Cameras, including surveillance cameras, infrared cameras, and panoramic cameras, used to capture images in the park in real time in a variety of complex scenarios; A13: Audio device, including microphone and audio sensor, used to capture environmental sound; A14: Smart meters for intelligent measurement of energy usage.

[0023] As mentioned above, the data acquisition module is equipped with a variety of devices to comprehensively monitor the smart park: the environmental sensor combines temperature sensors, humidity sensors and light intensity sensors to accurately record the temperature, humidity and light intensity in the park; the camera combines monitoring, infrared and panoramic cameras to ensure that the park dynamics can be captured in real time in various complex scenarios; the audio equipment captures environmental sounds through microphones and audio sensors; in addition, the smart meter is responsible for intelligently measuring the energy usage of the park, and together provide detailed data support for park management.

[0024] Preferably, the data processing module includes the following steps when performing preliminary processing on the collected raw data: S11: Data preprocessing, including data cleaning, data format conversion, data normalization and data standardization; S12: Data quality assessment: quality assessment of preprocessed data, including completeness check, accuracy check, consistency check and interpretability check; S13: Feature extraction, first select representative, relevant and interpretable features according to the analysis objectives and data characteristics, then calculate the corresponding statistics and indicators based on the selected features, and transform and combine the features to form preset new features.

[0025] Preferably, when the data processing module performs preliminary processing on the collected raw data, it also includes the following steps: S21: data extraction, firstly extracting the sensor data from the raw data after preprocessing; S22: Data alignment, including time alignment and coordinate alignment, aligning the sensor data in time, and converting the data obtained by different sensors into the same coordinate system; S23: Data fusion, using a preset data fusion algorithm to integrate and fuse the data from multiple sensors; S24: Data output, outputting the fused sensor data for processing in the next step, wherein the processing method in the next step is feature extraction.

[0026] As described above, compared with the prior art, which directly performs simple processing on the raw data before analysis, the present invention has the disadvantages of low data quality and inaccurate feature extraction. This solution adopts a comprehensive processing solution that first pre-processes the data, conducts quality assessment, extracts features, and further performs data extraction, alignment, fusion and output. This solution improves data quality through pre-processing steps such as data cleaning, format conversion, normalization and standardization, ensures data reliability through integrity, accuracy, consistency and interpretability checks, then accurately selects and extracts features, and finally integrates multi-source sensor data through data alignment and fusion technology. This comprehensive processing solution significantly improves the accuracy and availability of data, providing a more reliable foundation for subsequent data analysis and decision-making.

[0027] Preferably, when the preset data fusion algorithm is used to integrate and fuse the data of multiple sensors, the data fusion algorithm used includes: A21: Weighted average fusion, which performs weighted average of data from multiple sensors, where the weights can be adjusted according to the performance of the model. The principle formula is: ; in, and They are the data of two groups of sensors. and They are and The weight of ; A22: Statistical fusion uses statistical methods and probability theory to combine data from different sources. The principle formula is: ; in, represents parameters related to the reliability of the data source, the noise characteristics of the data, and the weight of the data, It represents the probability of the observed data set X appearing under the given parameter θ. Represents an observation data set, which contains n data observation points.

[0028] As mentioned above, compared with the existing technology that uses single sensor data or simple data combination schemes, which have shortcomings such as one-sided data and error accumulation, this scheme adopts a data fusion scheme that combines weighted average fusion with a statistical fusion algorithm. The weighted average fusion algorithm effectively balances the contribution of different sensor data by dynamically adjusting the weights; while the statistical fusion algorithm makes full use of statistical methods and probability theory, comprehensively considers the reliability, noise characteristics and weights of the data source, and achieves more accurate data integration. This comprehensive data fusion scheme significantly improves the accuracy and comprehensiveness of the data, and provides more reliable and accurate data support for multi-source data analysis and decision-making in smart parks.

[0029] Preferably, the data analysis module includes the following steps when using a deep learning algorithm to perform in-depth mining and analysis on the processed data, and perform behavior analysis, pattern recognition and trend prediction: S31: Model training: first, select a suitable deep learning algorithm according to the type of data, and build a behavior analysis, pattern recognition and trend prediction model based on the selected deep learning algorithm, and train the built model; S32: Model application, inputting the data results of the data processing module into the trained model to perform behavior analysis, pattern recognition and trend prediction; S33: Result interpretation and verification: Analyze the output of the model, identify abnormal behaviors, potential threats, and development trends, and compare the analysis results with actual data and known situations to verify the analysis results; S34: Result output: output the analysis result in the form of a report, wherein the output report includes a behavior analysis report, a pattern recognition report and a trend prediction report, and the report includes the analysis process, key findings and recommended measures.

[0030] As mentioned above, compared with the traditional data analysis methods used in existing technologies, there may be shortcomings such as weak model adaptability and shallow analysis results. This solution uses deep learning algorithms for model training and applies them to behavior analysis, pattern recognition, and trend prediction. By building targeted deep learning models, this solution can deeply mine data features, accurately identify abnormal behaviors and potential threats, and effectively predict development trends. At the same time, the result interpretation and verification steps ensure the accuracy and reliability of the analysis results. The final detailed report provides strong support for the decision-making of the smart park, significantly improving the depth and practicality of data analysis.

[0031] Preferably, the data processing module adopts a distributed structure, and the data processing module includes multiple groups of edge computing modules, the multiple groups of edge computing modules are connected through communication modules, and the multiple groups of edge computing modules are connected to the data analysis module through the communication module.

[0032] Preferably, the data analysis module, data encryption and storage module, real-time monitoring and early warning module and comprehensive management module are all arranged on the central server, and the communication module includes multiple data transmission modules, and the multiple data transmission modules are respectively arranged on multiple groups of edge computing modules.

[0033] As described above, compared with the centralized data processing scheme adopted in the prior art, there may be shortcomings such as low processing efficiency, heavy equipment burden, and once the central equipment is damaged, it is easy to cause the entire system to be paralyzed. This scheme adopts a distributed structure data processing module, which includes multiple groups of edge computing modules, and realizes efficient interconnection through communication modules. The edge computing module can process data nearby, reduce the pressure on the central server, and improve data processing efficiency. In addition, since multiple edge computing modules coexist, even if a module fails, it will not affect the operation of the entire system, effectively improving the stability and reliability of the system. This technical solution that is biased towards edge computing and the Internet of Everything not only realizes the efficient processing of large amounts of data, but also effectively avoids the risk of system paralysis caused by equipment damage.

[0034] Preferably, the data analysis module further includes the following steps when using a deep learning algorithm to perform deep mining and analysis on the processed data and perform behavior analysis, pattern recognition and trend prediction: S41: Blind spot judgment: analyze the layout and coverage of all data collection devices to identify the current blind spots; S42: Equipment configuration: configure and adjust other data acquisition equipment according to the location and range of the blind area, including adding new data acquisition equipment and adjusting the location and direction of the data acquisition equipment; S43: importance judgment, after configuring and adjusting other data collection devices, judging the importance of the blind area covered by the adjusted data collection device; S44: Adjust the configuration. According to the importance of the blind area and the type and structure of the data to be collected, configure the corresponding collection rules, including adjusting the update frequency.

[0035] As mentioned above, compared with the static data collection and analysis solutions used in the prior art, there may be shortcomings such as blind spots in data collection and incomplete analysis. This solution adds blind spot judgment, equipment configuration, importance judgment, and configuration adjustment steps to the data analysis module. By dynamically analyzing the layout and coverage of data collection equipment, blind spots are identified and equipment configuration and adjustment are carried out in a targeted manner. At the same time, the collection rules are optimized according to the importance of the blind spots and the data type. This solution not only effectively eliminates data collection blind spots and improves the comprehensiveness and accuracy of data, but also meets the collection needs of different regions and types of data through flexible collection rule configuration, providing strong support for the refined management and decision-making of smart parks.

[0036] See also Figure 2 The present invention provides an embodiment: a method for dynamic monitoring and real-time analysis of multi-source data in a smart park, comprising the following steps: S51: Data collection, using the data collection equipment installed in the park to collect environmental data, video data, audio data and personnel activity information data in the park; S52: Data preprocessing: cleaning, format conversion, normalization and standardization of the collected raw data; S53: Data fusion, extracting sensor data from the preprocessed data separately, and performing time alignment and coordinate alignment, and then integrating and fusing the data of multiple sensors using a data fusion algorithm; S54: Feature extraction: select representative, relevant and interpretable features according to the analysis objectives and data characteristics, and calculate the corresponding statistics and indicators to form a new feature set; S55: Data analysis and prediction: input the processed data into the trained model to perform behavior analysis, pattern recognition and trend prediction, and analyze the output of the model to identify abnormal behavior, potential threats and development trends; S56: Data storage, encrypting sensitive data and storing the encrypted data in a secure location; S57: Real-time monitoring and early warning, visual display of data to facilitate user observation, and based on the results of the data analysis module, real-time monitoring of abnormal behaviors and potential threats, and triggering of early warning mechanisms.

[0037] As described above, this method of dynamic monitoring and real-time analysis of multi-source data in smart parks covers the entire process from data collection, preprocessing, fusion, feature extraction, analysis and prediction, encrypted storage to real-time monitoring and early warning. Multi-dimensional data is collected through equipment in the park, and the data quality is improved through preprocessing and data fusion. After feature extraction, it is input into the training model for in-depth analysis and prediction, anomalies and threats are identified, and sensitive data is encrypted and stored to ensure security, ultimately achieving visual monitoring and real-time early warning of data, providing comprehensive, accurate, and real-time decision support for park management.

[0038] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of those skilled in the art without departing from the spirit of the present invention.

Claims

1. A smart park multi-source data dynamic monitoring and real-time analysis system; characterized by: Included are: The data collection module is used to collect data in the park using sensors, cameras and IoT devices installed in the park. The collected data includes environmental data, video data, audio data and personnel activity information data; The data processing module is used to perform preliminary processing on the collected raw data, wherein the processing methods adopted include data preprocessing, feature extraction and data quality assessment; Data analysis module, which is used to conduct in-depth mining and analysis of processed data using deep learning algorithms, and to perform behavior analysis, pattern recognition, and trend prediction; Data encryption and storage module, used to encrypt sensitive data and store the encrypted data; The real-time monitoring and early warning module is used to visualize the data for users to observe and monitor abnormal behaviors and potential threats in real time based on the results of the data analysis module, and trigger the early warning mechanism; Comprehensive management module, used to provide a comprehensive management interface and user interaction functions; The communication module is used to build a communication network to realize data communication among the data acquisition module, data processing module, data analysis module, data encryption and storage module, and real-time monitoring and early warning module.

2. According to claim 1, a smart park multi-source data dynamic monitoring and real-time analysis system is characterized by: The data acquisition module includes the following data acquisition devices: A11: Environmental sensors, including temperature sensors, humidity sensors, and light intensity sensors. The temperature sensor is used to measure and record the temperature information in the park. The humidity sensor is used to monitor the humidity level in the park. The light intensity sensor is used to monitor the light intensity inside the park. A12: Cameras, including surveillance cameras, infrared cameras, and panoramic cameras, used to capture images in the park in real time in a variety of complex scenarios; A13: Audio device, including microphone and audio sensor, used to capture environmental sound; A14: Smart meters for intelligent measurement of energy usage.

3. The smart park multi-source data dynamic monitoring and real-time analysis system according to claim 2 is characterized by: The data processing module performs preliminary processing on the collected raw data, including the following steps: S11: Data preprocessing, including data cleaning, data format conversion, data normalization and data standardization; S12: Data quality assessment: quality assessment of preprocessed data, including completeness check, accuracy check, consistency check and interpretability check; S13: Feature extraction, first select representative, relevant and interpretable features according to the analysis objectives and data characteristics, then calculate the corresponding statistics and indicators based on the selected features, and transform and combine the features to form preset new features.

4. The smart park multi-source data dynamic monitoring and real-time analysis system according to claim 3 is characterized by: The data processing module also includes the following steps when performing preliminary processing on the collected raw data: S21: data extraction, firstly extracting the sensor data from the raw data after preprocessing; S22: Data alignment, including time alignment and coordinate alignment, aligning the sensor data in time, and converting the data obtained by different sensors into the same coordinate system; S23: Data fusion, using a preset data fusion algorithm to integrate and fuse the data from multiple sensors; S24: Data output, outputting the fused sensor data for processing in the next step, wherein the processing method in the next step is feature extraction.

5. The smart park multi-source data dynamic monitoring and real-time analysis system according to claim 4 is characterized by: When the preset data fusion algorithm is used to integrate and fuse the data of multiple sensors, the data fusion algorithms used include: A21: Weighted average fusion, which performs weighted average of data from multiple sensors, where the weights can be adjusted according to the performance of the model. The principle formula is: ； in, and They are the data of two groups of sensors. and They are and The weight of ; A22: Statistical fusion uses statistical methods and probability theory to combine data from different sources. The principle formula is: ； in, represents parameters related to the reliability of the data source, the noise characteristics of the data, and the weight of the data, It represents the probability of the observed data set X appearing under the given parameter θ. Represents an observation data set, which contains n data observation points.

6. The smart park multi-source data dynamic monitoring and real-time analysis system according to claim 5 is characterized by: The data analysis module uses deep learning algorithms to conduct in-depth mining and analysis of processed data, and conducts behavior analysis, pattern recognition, and trend prediction, including the following steps: S31: Model training: first, select a suitable deep learning algorithm according to the type of data, and build a behavior analysis, pattern recognition and trend prediction model based on the selected deep learning algorithm, and train the built model; S32: Model application, inputting the data results of the data processing module into the trained model to perform behavior analysis, pattern recognition and trend prediction; S33: Result interpretation and verification: Analyze the output of the model, identify abnormal behaviors, potential threats, and development trends, and compare the analysis results with actual data and known situations to verify the analysis results; S34: Result output: output the analysis result in the form of a report, wherein the output report includes a behavior analysis report, a pattern recognition report and a trend prediction report, and the report includes the analysis process, key findings and recommended measures.

7. The smart park multi-source data dynamic monitoring and real-time analysis system according to claim 6 is characterized by: The data processing module adopts a distributed structure, and the data processing module includes multiple groups of edge computing modules. The multiple groups of edge computing modules are connected through communication modules, and the multiple groups of edge computing modules are connected to the data analysis module through the communication module.

8. The smart park multi-source data dynamic monitoring and real-time analysis system according to claim 7 is characterized by: The data analysis module, data encryption and storage module, real-time monitoring and early warning module and comprehensive management module are all set on the central server. The communication module includes multiple data transmission modules, and the multiple data transmission modules are respectively set on multiple groups of edge computing modules.

9. A smart park multi-source data dynamic monitoring and real-time analysis system according to claim 8, characterized in that: The data analysis module uses deep learning algorithms to conduct in-depth mining and analysis of processed data, and conducts behavior analysis, pattern recognition, and trend prediction, and also includes the following steps: S41: Blind spot judgment: analyze the layout and coverage of all data collection devices to identify the current blind spots; S42: Equipment configuration: configure and adjust other data acquisition equipment according to the location and range of the blind area, including adding new data acquisition equipment and adjusting the location and direction of the data acquisition equipment; S43: importance judgment, after configuring and adjusting other data collection devices, judging the importance of the blind area covered by the adjusted data collection device; S44: Adjust the configuration. According to the importance of the blind area and the type and structure of the data to be collected, configure the corresponding collection rules, including adjusting the update frequency.

10. A smart park multi-source data dynamic monitoring and real-time analysis system and method according to claim 9, characterized in that: The following steps are included: S51: Data collection, using the data collection equipment installed in the park to collect environmental data, video data, audio data and personnel activity information data in the park; S52: Data preprocessing: cleaning, format conversion, normalization and standardization of the collected raw data; S53: Data fusion, extracting sensor data from the preprocessed data separately, and performing time alignment and coordinate alignment, and then integrating and fusing the data of multiple sensors using a data fusion algorithm; S54: Feature extraction: select representative, relevant and interpretable features according to the analysis objectives and data characteristics, and calculate the corresponding statistics and indicators to form a new feature set; S55: Data analysis and prediction: input the processed data into the trained model to perform behavior analysis, pattern recognition and trend prediction, and analyze the output of the model to identify abnormal behavior, potential threats and development trends; S56: Data storage, encrypting sensitive data and storing the encrypted data in a secure location; S57: Real-time monitoring and early warning, visual display of data to facilitate user observation, and based on the results of the data analysis module, real-time monitoring of abnormal behaviors and potential threats, and triggering of early warning mechanisms.

Images