Cross-regional environmental data sharing and analysis method and system based on big data deep learning

Abstract

The invention discloses a cross-regional environment data sharing and analysis method and system based on big data deep learning, and relates to the technical field of data processing. Comprising the following steps: determining a linkage range of sharing and analyzing environmental data by target equipment, so as to avoid invalid coverage and resource waste of cross-regional data transmission; acquiring the acquisition frequency of environmental data in the linkage range; and dynamically dividing a big data deep learning region in the region according to the linkage range and the acquisition frequency, so as to realize spatial matching of computing resources and data values and improve the data processing efficiency of a high-priority region. According to the method, the linkage range of the target equipment is defined, the data sharing boundary is precisely limited, redundant coverage of cross-regional data transmission is avoided, invalid data storage and network bandwidth consumption are remarkably reduced, the data collection density is further optimized in combination with dynamic adjustment of the collection frequency, generation of redundant data is reduced, and therefore overall resource waste is reduced.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and specifically to a cross-regional environmental data sharing and analysis method and system based on big data deep learning. Background Art

[0002] In the existing technical field, cross-regional environmental data sharing is a technical means to solve the problem of regional environmental monitoring data islands. In order to further enhance the collaborative analysis capabilities of multi-source heterogeneous data, a cross-regional environmental data sharing and analysis method and system based on big data deep learning is needed.

[0003] After searching, the Chinese invention patent with announcement number "CN111680793A" discloses "a blockchain consensus method and system based on deep learning model training". By introducing the excess computing power of the blockchain into deep learning model training, this application can guide funds, computing power and energy to be invested in more meaningful work, solving the problem of insufficient computing power and high costs.

[0004] In addition, the Chinese invention patent with announcement number "CN113469370A" discloses "an industrial Internet of Things data sharing method based on federated incremental learning". The application proposes a federated incremental learning algorithm to solve the problem of integrating a large amount of new data from the factory sub-end with the original industry joint model. The purpose is to quickly integrate the new status data with the original industry joint model by calculating the incremental weighting of the factory sub-end, thereby realizing effective incremental learning of the new status data.

[0005] However, in actual use, the above-mentioned disclosed patents and similar patent methods do not take into account the heterogeneity and dynamic change characteristics of the spatial distribution of cross-regional environmental data, but instead use a fixed threshold to set the linkage range. This method easily leads to redundancy in inter-regional data transmission and mismatch of edge computing resources. To this end, the applicant proposes a cross-regional environmental data sharing and analysis method and system based on big data deep learning to solve the above-mentioned problems. Summary of the Invention

[0006] The purpose of the present invention is to provide a cross-regional environmental data sharing and analysis method and system based on big data deep learning to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] First, a cross-regional environmental data sharing and analysis method based on big data deep learning was proposed, including:

[0009] Determine the linkage scope of target device sharing and analysis environment data to avoid ineffective coverage and resource waste of cross-regional data transmission;

[0010] Obtain the collection frequency of environmental data within the linkage range;

[0011] Dynamically divide big data deep learning areas within the region based on linkage scope and collection frequency to achieve spatial matching of computing resources and data value, and improve data processing efficiency in high-priority areas;

[0012] Building a shared network based on big data deep learning areas;

[0013] After accessing the shared network, target devices in different areas obtain the required environmental data and analyze and predict the environmental data based on big data deep learning algorithms.

[0014] As a further preferred embodiment of the present technical solution, the method for obtaining the linkage range includes:

[0015] Determine the latitude and longitude coordinates of the center point of each area;

[0016] Calculate the Euclidean distance between the center points of two adjacent regions based on their latitude and longitude coordinates;

[0017] The linkage range is set according to the obtained Euclidean distance.

[0018] As a further preferred embodiment of the present technical solution, the following distance calculation formula is used for calculating the Euclidean distance:

[0019] Radians = degrees × (π / 180), used to convert degrees into radians;

[0020] , used to determine the actual distance between the center points of two adjacent areas through longitude and latitude coordinates, so as to set the linkage range more accurately;

[0021] in is the Euclidean distance between the center points of the two regions, is the latitude difference between the center points of the two regions, in radians, is the longitude difference between the two area center points, is the average latitude of the center points of the two regions, is the radius of the Earth, which is 6371 km;

[0022] and as well as and Respectively represent the latitude and longitude coordinates of the center points of two adjacent areas.

[0023] As a further preferred embodiment of the present technical solution, a method for establishing a priority evaluation standard includes:

[0024] The data collection frequency is divided into several gradient levels;

[0025] Establish the spatial mapping relationship between the acquisition frequency and the visual environment feature matrix;

[0026] Obtain environmental feature data of the target location and construct a frequency-feature space correlation map based on the corresponding frequency levels;

[0027] The priority ranking is determined based on the product of the distribution weight of the frequency parameter and the environmental characteristics in the linkage area.

[0028] As a further preferred embodiment of the present technical solution, the method for constructing a shared network includes:

[0029] Deploy network node clusters in the target area based on linkage range parameters to achieve rapid transmission and sharing of cross-regional environmental data;

[0030] Establish an environmental data sharing network based on spatial location encryption, which uses the longitude and latitude coordinates of the regional center point as the initial vector and combines it with real-time timestamp parameters to generate an irreversible 128-bit session key to prevent data from being intercepted or tampered with during cross-regional transmission.

[0031] As a further preferred embodiment of this technical solution, the implementation method of regionalized deep learning operation optimization includes:

[0032] Dynamically allocate computing resources based on the priority of the divided areas, and allocate computing resources based on the priority ranking of each divided area, so as to ensure that deep learning operations on environmental data in different areas can be carried out efficiently;

[0033] Implement multimodal data fusion processing and feature extraction to improve the comprehensive analysis capabilities of cross-regional environmental data.

[0034] As a further preferred embodiment of the present technical solution, the method for dividing the big data deep learning area includes:

[0035] Construct a visual environment feature matrix of the target area;

[0036] Constructing a priority evaluation criterion for the visual environment feature matrix based on acquisition frequency;

[0037] Dynamically divide the corresponding positions of the linkage area according to the order of priority;

[0038] Implement regionalized deep learning operation optimization based on dynamic partitioning results.

[0039] Secondly, to ensure the integrity of the above technical solutions, a cross-regional environmental data sharing and analysis system based on big data deep learning is proposed. This system uses the above cross-regional environmental data sharing and analysis method based on big data deep learning, and includes:

[0040] A linkage scope determination module is used to determine the linkage scope of target device sharing and analysis of environmental data;

[0041] The data collection frequency acquisition module is used to obtain the collection frequency of environmental data within the linkage range;

[0042] Dynamic region division engine, used to dynamically divide big data deep learning regions within a region based on linkage range and acquisition frequency;

[0043] A secure shared network building unit used to implement dynamic encryption of data channels by combining geographic coordinates and timestamps;

[0044] Distributed deep learning computing cluster, used to perform distributed computing tasks on major data deep learning areas;

[0045] The multimodal data fusion processor adopts a collaborative processing architecture for heterogeneous data sources, integrates a multi-dimensional parameter spatial feature analysis unit for meteorological factors, pollutant concentrations, and ecological indices, and realizes the joint representation of time series data and spatial distribution data by constructing a multi-channel parallel feature extraction module, and establishes a spatiotemporal coupling characteristic analysis channel for cross-modal parameters;

[0046] Visual decision support interface, used to provide a visual interface.

[0047] As a further preferred embodiment of the present technical solution, the linkage range determination module includes: an IoT gateway, a GIS server and an SDN controller, wherein the IoT gateway is used to manage the device connection range, the GIS server is used to determine the linkage area through geo-fencing technology, the SDN controller is used to dynamically configure the network device collaboration range, the data acquisition frequency acquisition module includes: wireless sensor nodes, data collectors and edge gateways, the dynamic area division engine includes: stream processing servers, edge computing nodes and cloud resource managers, the secure shared network construction unit includes: hardware security modules, dynamic encryption routers and blockchain nodes, the distributed deep learning computing cluster includes: GPU clusters, TPU clusters and AI servers, the multimodal data fusion processor includes: FPGA acceleration servers, multi-sensor fusion platforms and distributed storage systems, and the visual decision support interface includes: a large data screen, an interactive touch terminal and a Web visualization server.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] A cross-regional environmental data sharing and analysis method and system based on big data deep learning defines the linkage range of target devices and accurately defines the data sharing boundaries, avoiding redundant coverage of cross-regional data transmission, significantly reducing invalid data storage and network bandwidth consumption. Combined with dynamic adjustment of collection frequency, it further optimizes data collection density, reduces the generation of redundant data, and thus reduces overall resource waste.

[0050] In addition, deep learning areas are dynamically divided based on the linkage range and acquisition frequency to achieve intelligent matching of computing resources and data value. By building a cross-regional shared network, integrating environmental data, and combining deep learning algorithms, environmental trends can be analyzed and predicted. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 A flowchart of the steps of the method disclosed in the present invention;

[0052] Figure 2 This is a flowchart of the steps of the method for obtaining the linkage range of the present invention;

[0053] Figure 3 A flowchart of a method for establishing priority evaluation criteria according to the present invention;

[0054] Figure 4 It is a system composition diagram of the present invention;

[0055] Figure 5 It is a frequency-feature space correlation diagram disclosed in the present invention;

[0056] Figure 6 It is an auxiliary illustration of step S103 of the present invention. DETAILED DESCRIPTION

[0057] 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.

[0058] Before understanding the technical solution proposed in this application, it should be clear that big data deep learning is a technical means to improve the accuracy and efficiency of environmental data sharing through the mining and analysis of massive data. In the technical solution proposed in this application, it is mainly used for cross-regional data sharing and interaction, and can accurately match and efficiently analyze the corresponding environmental data to ensure seamless connection of data between different regions, thereby greatly improving the timeliness of environmental monitoring and early warning.

[0059] In addition, it should be noted that in this application, environmental data refers to a type of data, and in this application, environmental data includes: CO2 content, amount of garbage and vegetation density.

[0060] like Figure 1 As shown, the present invention provides a technical solution: a cross-regional environmental data sharing and analysis method based on big data deep learning, including; steps S100-S500.

[0061] Step S100: Determine the linkage scope of target device sharing and analysis environment data.

[0062] It should be noted that in this application, the linkage range in step S100 refers to the minimum distance threshold within which environmental data can be shared and analyzed across regions in different areas, ensuring the stability and reliability of data during transmission and avoiding data distortion caused by excessive distance.

[0063] As a preferred implementation method, in the scenario of cross-regional linkage of urban networked devices, the sharing and analysis of environmental data between devices relies on short-distance judgment (1-10 kilometers). Traditional GPS ranging wastes computing resources to a certain extent and causes the response time of embedded devices to be too long, resulting in delays between the push and sharing of environmental data. Therefore, a ranging method for short-distance cross-regional application is proposed, which makes the linkage range judgment more accurate, reduces the computing burden, and improves the response speed of environmental data sharing and analysis.

[0064] It should be noted that, in this embodiment, the method for obtaining the linkage range includes: steps S101 to S103.

[0065] Step S101: Determine the longitude and latitude coordinates of the center point of each area.

[0066] It should be noted that the longitude and latitude coordinates in step S101 are determined by using the GPS positioning technology in the prior art.

[0067] Step S102: Calculate the Euclidean distance between two adjacent area center points based on the latitude and longitude coordinates.

[0068] It should be clear that in step S102, the following distance calculation formula is used to calculate the Euclidean distance, and the angle needs to be converted into radians during the calculation. It should be noted that the method of converting the angle into radians is achieved through the trigonometric function formula in the prior art. The specific conversion formula is: radian = angle × (π / 180).

[0069] Specifically, the distance calculation formula is: ;in is the Euclidean distance between the center points of the two regions, is the latitude difference between the center points of the two regions, in radians, is the longitude difference between the two area center points, is the average latitude of the center points of the two regions, is the radius of the earth, which is 6371 km.

[0070] It should be added that in this application and as well as and Respectively represent the latitude and longitude coordinates of the center points of two adjacent areas.

[0071] It is worth noting that when the distance calculation formula disclosed in step S102 is actually used, assuming that area A and area B are adjacent areas, and the longitude and latitude coordinates of area A and area B are obtained through GPS, the Euclidean distance between the center points of area A and area B can be accurately calculated according to the above formula.

[0072] Specifically, for area A, the longitude and latitude coordinates are (31.2423°, 121.4903°), and area B, the longitude and latitude coordinates are (31.2385°, 121.5020°). At this time, after converting the angles of the longitude and latitude coordinates into interactive, 、 、 as well as The values are: , , , .

[0073] at this time , ,

[0074] ;

[0075] . Through the above calculation, it is concluded that the Euclidean distance between the center points of Area A and Area B is approximately 2.1 kilometers. In addition, it should be added that the units of the values in the above formula are radians.

[0076] Step S103: setting a linkage range according to the obtained Euclidean distance.

[0077] It should be noted that the reference Figure 6The setting of the linkage range threshold in step 103 is determined by adding the average value of the maximum and minimum values within the Euclidean distance of all adjacent areas. After determining the average value, the circular range with the position of the target device in the area as the center and the average value as the radius is set as the specific linkage range. It should be added that this setting method eliminates the impact of extreme distances on the system through extreme value averaging, thereby avoiding response delays caused by excessively large thresholds or false linkages caused by too small thresholds.

[0078] Step S200: Acquire the collection frequency of environmental data within the linkage range.

[0079] It should be noted that in step S200, the frequency of collecting environmental data is used to provide data samples for big data deep learning.

[0080] It should be further explained that in this application, the method for obtaining the acquisition frequency is to connect the current sensor to the environmental monitoring module in the target device to monitor the working status of the target device in real time. When the target device is in working state, the current sensor will detect the current change and convert the current change into a digital signal, thereby determining the acquisition frequency of the environmental data. It should be added that since the use of current sensors to connect to the modules of corresponding functions and then detect the execution frequency of the function is common knowledge in the existing sensor detection field, the applicant will not make too much explanation and description of the specific linkage method between the current sensor and the environmental detection module.

[0081] Step S300: Dynamically divide the big data deep learning area within the region based on the linkage range and acquisition frequency.

[0082] It should be noted that the big data deep learning area in step S300 is used to determine the high-incidence range of environmental data sharing and analysis, ensuring that the deep learning algorithm can perform efficient data processing and analysis for specific areas.

[0083] As a preferred implementation method, in the scenario where devices are linked across regions after the city is connected to the Internet, due to the complexity of the regions within the city, applying the big data deep learning algorithm to the entire region will lead to the dispersion of computing resources and low data processing efficiency. Therefore, a method for dynamically dividing big data deep learning areas is needed.

[0084] It should be noted that, in this embodiment, reference Figure 2 It can be seen that the method for dividing the big data deep learning area includes: steps S301 to S304.

[0085] Step S301: Constructing a visual environment feature matrix of the target area.

[0086] It should be noted that the method for constructing the visual environmental feature matrix in this implementation scheme is as follows: first, the linkage area is mapped to the two-dimensional electronic map of the target area, and then the environmental data in the mapping area is collected, and finally the location features collected by the environmental data are integrated into the linkage area of the electronic map in the form of annotations.

[0087] Step S302: Constructing a priority evaluation standard of the visualization environment feature matrix based on the acquisition frequency.

[0088] It should be noted that the reference Figure 3 In this embodiment, the method for establishing the priority evaluation criteria includes: Step S302A to Step S302D.

[0089] Step S302A: Divide the data acquisition frequency into several gradient levels.

[0090] Specifically, this plan adopts a three-tier classification system: high-frequency collection level (sampling interval ≤ 5 minutes), medium-frequency collection level (sampling interval 6-30 minutes), and low-frequency collection level (sampling interval > 30 minutes). This classification system is based on the magnitude difference of the sampling time threshold.

[0091] Step S302B: establishing a spatial mapping relationship between acquisition frequencies in the visualization environment feature matrix.

[0092] It should be noted that this step aims to analyze the spatial position between the acquisition frequency and the environmental characteristic parameters in the corresponding linkage area.

[0093] Step S302C: Acquire environmental feature data of the target location, and construct a frequency-feature space association map based on the corresponding frequency levels.

[0094] It should be noted that the reference Figure 5 It can be seen that the frequency-feature space correlation map is a two-dimensional electronic map that contains all target location points and their corresponding environmental feature data. It intuitively displays the distribution of environmental features at different frequency levels through color depth or pattern differences. This map not only helps to quickly identify data-dense or sparse areas, but also provides an important reference for subsequent data processing and regional division.

[0095] Step S302D: Determine the priority ranking based on the product of the frequency parameter and the distribution weight of the environmental characteristics in the linkage area.

[0096] Step S303: Dynamically divide the corresponding positions of the linkage area in order of priority.

[0097] Step S304: Implement regionalized deep learning operation optimization based on the dynamic segmentation results.

[0098] As a preferred embodiment, in this embodiment, in order to allocate computing resources according to the priority ranking status of each divided area, so as to ensure that deep learning operations on environmental data in different regions can be carried out efficiently, an implementation method for regionalized deep learning operation optimization is proposed, which specifically includes: step S304A-step S304B.

[0099] Step S304A: Dynamically allocate computing resources according to the priorities of the divided areas.

[0100] It should be noted that the computing resources in step S304A specifically refer to the computing power allocated for executing big data deep learning.

[0101] Specifically, in step S304A, the high-frequency acquisition area is configured with a real-time stream processing engine and a GPU cluster, the medium-frequency acquisition area is configured with medium-scale computing resources and a memory database, and the low-frequency acquisition area is allocated with basic computing units and local storage.

[0102] Step S304B: Implement multimodal data fusion processing and feature extraction to enhance the comprehensive analysis capability of cross-regional environmental data.

[0103] It should be noted that this step uses the existing convolutional neural network (ST-CNN) to perform feature fusion processing on environmental data. During implementation, lightweight real-time feature extraction is deployed for high-frequency areas, while intermittent feature extraction is performed through cloud computing nodes in medium and low-frequency areas.

[0104] Step S400: Constructing a shared network based on the big data deep learning area.

[0105] It should be noted that the method for constructing a shared network in this application includes: step S401-step S402.

[0106] Step S401: deploying a network node cluster in a target area based on linkage range parameters.

[0107] Specifically, network node clusters are evenly distributed at any location within the linkage range of a two-dimensional electronic map to achieve phased transmission of environmental data. Furthermore, it should be noted that in this application, a network node cluster is a computing node in the prior art that enables rapid data processing and transmission. Each computing node has the functions of receiving, storing, forwarding, and performing preliminary processing on data, ensuring the integrity and accuracy of data during transmission.

[0108] Step S402: Establishing an environmental data sharing network based on spatial location encryption.

[0109] It should be noted that when the network node cluster transmits environmental data, a dynamic hash value is generated based on the latitude and longitude coordinates of the target area as an encryption key, and the data is encrypted using the AES-GCM algorithm in the existing technology. It should be added that the environmental data sharing network uses the latitude and longitude coordinates of the center point of the area (such as the value obtained in step S101) as the initial vector and combines it with the real-time timestamp parameter to generate an irreversible 128-bit session key to prevent data interception or tampering during cross-regional transmission.

[0110] Step S500: After accessing the shared network, target devices in different areas obtain the required environmental data and analyze and predict the environmental data based on big data deep learning algorithms.

[0111] It is worth noting that in step S500, the big data deep learning algorithm includes neural network algorithm, support vector machine algorithm and decision tree algorithm in the existing technology. These algorithms can mine potential correlations and regularities from massive environmental data, providing strong support for the accurate analysis and prediction of environmental data.

[0112] As a preferred embodiment, refer to Figure 4 It is known that a system has been disclosed, which is a cross-regional environmental data sharing and analysis system based on big data deep learning. It uses the public "cross-regional environmental data sharing and analysis method based on big data deep learning" and is used to realize the sharing and analysis of cross-regional environmental data.

[0113] Specifically, they include:

[0114] A linkage scope determination module is used to determine the linkage scope of target device sharing and analysis of environmental data;

[0115] The data collection frequency acquisition module is used to obtain the collection frequency of environmental data within the linkage range;

[0116] Dynamic region division engine, used to dynamically divide big data deep learning regions within a region based on linkage range and acquisition frequency;

[0117] A secure shared network building unit used to implement dynamic encryption of data channels by combining geographic coordinates and timestamps;

[0118] Distributed deep learning computing cluster, used to perform distributed computing tasks on major data deep learning areas;

[0119] The multimodal data fusion processor adopts a collaborative processing architecture for heterogeneous data sources, integrates a multi-dimensional parameter spatial feature analysis unit for meteorological factors, pollutant concentrations, and ecological indices, and realizes the joint representation of time series data and spatial distribution data by constructing a multi-channel parallel feature extraction module, and establishes a spatiotemporal coupling characteristic analysis channel for cross-modal parameters;

[0120] Visual decision support interface, used to provide a visual interface.

[0121] As a preferred implementation scheme, the system architecture proposed in this embodiment implements hardware entity deployment through a modular mapping method. Specifically, the linkage range determination module, data acquisition frequency acquisition module, dynamic area division engine, secure shared network construction unit, distributed deep learning computing cluster, multimodal data fusion processor, and visual decision support interface can be mapped to the following hardware entities:

[0122] At the hardware implementation level, the linkage range determination module consists of an IoT gateway, a spatial location service (GIS) server, and a software-defined network (SDN) controller. The IoT gateway is responsible for device access range management, the GIS server uses geo-fencing technology to accurately define the linkage area, and the SDN controller implements dynamic configuration of network device collaboration domains. The data acquisition frequency acquisition module integrates wireless sensor nodes, data acquisition terminals, and edge gateways to achieve adaptive adjustment of multi-dimensional data acquisition frequency. The dynamic region partitioning engine architecture includes a streaming data processing server, edge computing nodes, and a cloud resource manager to build an elastic computing resource scheduling system. The secure shared network construction unit integrates a hardware security module (HSM), dynamic encryption routing devices, and blockchain consensus nodes to establish a trusted data exchange mechanism. The distributed deep learning computing cluster deploys a GPU parallel computing array, a TPU dedicated acceleration unit, and an AI inference server to form a hierarchical intelligent computing architecture. The multimodal data fusion processor is equipped with an FPGA programmable accelerator, a multi-source sensor fusion platform, and a distributed object storage system to complete real-time heterogeneous data processing. The visual decision support interface integrates a large-screen display system, a multi-touch interactive terminal, and a WebGL visualization server to build a multi-dimensional data presentation system.

[0123] Although embodiments of the present invention have been shown and described, it will be understood 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 limited by the accompanying embodiments and their equivalents.

Claims

1. A cross-regional environmental data sharing and analysis method based on big data deep learning, characterized by: include: Determine the linkage scope of target device sharing and analysis environment data to avoid ineffective coverage and resource waste of cross-regional data transmission; Obtain the collection frequency of environmental data within the linkage range; Dynamically divide big data deep learning areas within the region based on linkage scope and collection frequency to achieve spatial matching of computing resources and data value, and improve data processing efficiency in high-priority areas; Building a shared network based on big data deep learning areas; After accessing the shared network, target devices in different areas obtain the required environmental data and analyze and predict the environmental data based on big data deep learning algorithms.

2. The cross-regional environmental data sharing and analysis method based on big data deep learning according to claim 1 is characterized by: Methods for obtaining linkage range include: Determine the latitude and longitude coordinates of the center point of each area; Calculate the Euclidean distance between the center points of two adjacent regions based on their latitude and longitude coordinates; The linkage range is set according to the obtained Euclidean distance.

3. The cross-regional environmental data sharing and analysis method based on big data deep learning according to claim 2 is characterized by: The following distance calculation formula is used for the calculation of Euclidean distance include: Radians = degrees × (π / 180), used to convert degrees into radians; , used to determine the actual distance between the center points of two adjacent areas through longitude and latitude coordinates, so as to set the linkage range more accurately; is the Euclidean distance between the center points of the two regions, is the latitude difference between the center points of the two regions, in radians, is the longitude difference between the two area center points, is the average latitude of the center points of the two regions, is the radius of the earth, which is 6371km; and as well as and Respectively represent the latitude and longitude coordinates of the center points of two adjacent areas.

4. The cross-regional environmental data sharing and analysis method based on big data deep learning according to claim 1 is characterized by: Methods for establishing priority evaluation criteria include: The data collection frequency is divided into several gradient levels; Establish the spatial mapping relationship between the acquisition frequency and the visual environment feature matrix; Obtain environmental feature data of the target location and construct a frequency-feature space correlation map based on the corresponding frequency levels; The priority ranking is determined based on the product of the distribution weight of the frequency parameter and the environmental characteristics in the linkage area.

5. The cross-regional environmental data sharing and analysis method based on big data deep learning according to claim 1 is characterized by: Methods for building a shared network include: Deploy network node clusters in the target area based on linkage range parameters to achieve rapid transmission and sharing of cross-regional environmental data; Establish an environmental data sharing network based on spatial location encryption, which uses the longitude and latitude coordinates of the regional center point as the initial vector and combines it with real-time timestamp parameters to generate an irreversible 128-bit session key to prevent data from being intercepted or tampered with during cross-regional transmission.

6. The cross-regional environmental data sharing and analysis method based on big data deep learning according to claim 1 is characterized by: Implementation methods for regionalized deep learning operation optimization include: Dynamically allocate computing resources based on the priority of the divided areas, and allocate computing resources based on the priority ranking of each divided area, so as to ensure that deep learning operations on environmental data in different areas can be carried out efficiently; Implement multimodal data fusion processing and feature extraction to improve the comprehensive analysis capabilities of cross-regional environmental data.

7. The cross-regional environmental data sharing and analysis method based on big data deep learning according to claim 1 is characterized by: The method for dividing the big data deep learning area includes: Construct a visual environment feature matrix of the target area; Constructing a priority evaluation criterion for the visual environment feature matrix based on acquisition frequency; Dynamically divide the corresponding positions of the linkage area according to the order of priority; Implement regionalized deep learning operation optimization based on dynamic partitioning results.

8. A cross-regional environmental data sharing and analysis system based on big data deep learning, using the cross-regional environmental data sharing and analysis method based on big data deep learning according to any one of claims 1 to 7, characterized in that: include: A linkage scope determination module is used to determine the linkage scope of target device sharing and analysis of environmental data; The data collection frequency acquisition module is used to obtain the collection frequency of environmental data within the linkage range; Dynamic region division engine, used to dynamically divide big data deep learning regions within a region based on linkage range and acquisition frequency; A secure shared network building unit used to implement dynamic encryption of data channels by combining geographic coordinates and timestamps; Distributed deep learning computing cluster, used to perform distributed computing tasks on major data deep learning areas; The multimodal data fusion processor adopts a collaborative processing architecture for heterogeneous data sources, integrates a multi-dimensional parameter spatial feature analysis unit for meteorological factors, pollutant concentrations, and ecological indices, and realizes the joint representation of time series data and spatial distribution data by constructing a multi-channel parallel feature extraction module, and establishes a spatiotemporal coupling characteristic analysis channel for cross-modal parameters; Visual decision support interface, used to provide a visual interface.

9. The cross-regional environmental data sharing and analysis system based on big data deep learning according to claim 8 is characterized by: The linkage range determination module includes: an IoT gateway, a GIS server and an SDN controller, wherein the IoT gateway is used to manage the device connection range, the GIS server is used to determine the linkage area through geographic fencing technology, and the SDN controller is used to dynamically configure the network device collaboration range. The data collection frequency acquisition module includes: wireless sensor nodes, data collectors and edge gateways. The dynamic area division engine includes: a stream processing server, an edge computing node and a cloud resource manager. The secure shared network construction unit includes: a hardware security module, a dynamic encryption router and a blockchain node. The distributed deep learning computing cluster includes: a GPU cluster, a TPU cluster and an AI server. The multimodal data fusion processor includes: an FPGA acceleration server, a multi-sensor fusion platform and a distributed storage system. The visual decision support interface includes: a large data screen, an interactive touch terminal and a Web visualization server.

Images