An ecological environment dynamic monitoring method and system based on remote sensing and geographic information system

Abstract

The present application relates to a kind of ecological environment dynamic monitoring method and system based on remote sensing and geographic information system, comprising: the optical remote sensing image data and SAR radar data of the region to be detected are preprocessed, identify the change region of the region to be detected in current time and last time, obtain target change region;Extract all environmental monitoring information of target change region in current time and last time, and the environmental policy information related to target change region;Text analysis is carried out on environmental policy information using natural language processing technology, and the control area corresponding to environmental policy information is extracted;According to all environmental monitoring information of control area in current time and last time, the environmental change region in control area is identified;The matching region is obtained by position matching of control area and target change region, and all environmental monitoring information between current time and last time of matching region is extracted;All environmental monitoring information between current time and last time of target change region, control area, environmental change region and matching region is spliced, generates to be measured sequence input AI big model and is predicted, obtains whether the environment of the region to be detected is abnormal, if yes, then early warning.

Description

A method and system for dynamic monitoring of the ecological environment based on remote sensing and geographic information systems Technical Field

[0001] This invention belongs to the field of geological monitoring, and in particular relates to a method and system for dynamic monitoring of the ecological environment based on remote sensing and geographic information systems. Background Technology

[0002] With the increasing impact of human activities on the natural environment, ecological and environmental monitoring has become a crucial link in maintaining ecological balance and sustainable development. In existing ecological and environmental monitoring systems, traditional monitoring equipment generally relies on multiple sets of sensors for data collection. This model not only leads to a geometric increase in the amount of monitoring data, increasing the burden of data storage and transmission, but also significantly increases the computational complexity of data preprocessing and feature extraction, directly resulting in a longer overall data processing cycle. At the same time, due to the lack of efficient dynamic data integration and spatial analysis mechanisms, existing monitoring systems struggle to track and provide real-time warnings of ecological and environmental changes. When facing sudden environmental events or dynamic adjustments to policy-controlled areas, they often suffer from regulatory lag and slow response, which in turn significantly reduces the accuracy and timeliness of regional ecological and environmental control measures, failing to fully meet the actual needs for refined and intelligent environmental monitoring in the context of current ecological civilization construction. Summary of the Invention

[0003] To address the problems existing in the background art, one aspect of the present invention provides a method for dynamic monitoring of the ecological environment based on remote sensing and geographic information systems, comprising:

[0004] S1: Preprocess the optical remote sensing image data and SAR radar data of the area to be detected, identify the changed areas of the area to be detected at the current time and the previous time, and obtain the target changed areas;

[0005] S2: Extract all environmental monitoring information of the target change area at the current time and the previous time, as well as environmental policy information related to the target change area;

[0006] S3: Use natural language processing technology to parse environmental policy information and extract the corresponding control areas;

[0007] S4: Identify areas of environmental change within the controlled area based on all environmental monitoring information of the controlled area at the current and previous times;

[0008] S5: Match the controlled area with the target change area to obtain the matching area, and extract all environmental monitoring information of the matching area between the current time and the previous time.

[0009] S6: Combine all environmental monitoring information of the target change area, control area, environmental change area and matching area between the current time and the previous time to generate a test sequence. Input the test sequence into the AI ​​big model for prediction to obtain whether the environment of the test area is abnormal. If so, issue an early warning.

[0010] Another aspect of the present invention provides an ecological environment dynamic monitoring system based on remote sensing and geographic information system, comprising: a memory and a processor; the memory is used to store an application program; the processor is used to run the application program and execute the ecological environment dynamic monitoring method based on remote sensing and geographic information system.

[0011] Another aspect of the present invention provides a computer storage medium storing a program that, when executed by a processor, implements the aforementioned method for dynamic monitoring of the ecological environment based on remote sensing and geographic information systems.

[0012] The present invention has at least the following beneficial effects

[0013] This invention addresses the problems of excessive reliance on multiple sensors in existing ecological and environmental monitoring, resulting in large data volumes, time-consuming processing, inconvenient dynamic supervision, and insufficient policy linkage. It achieves dynamic capture of large-area environmental elements through optical remote sensing and SAR radar data, significantly reducing reliance on dense sensors. Sensor data is extracted only in targeted areas of target change and control. Natural language processing is used to parse environmental policy text and match it with remote sensing data locations. Multi-source data is then stitched together and input into an AI model to predict environmental anomalies and issue early warnings, forming a highly efficient monitoring model of "macroscopic identification - microscopic verification - policy linkage." This reduces data collection and processing pressure and hardware costs while improving the accuracy and timeliness of dynamic supervision, achieving a technological leap from passive data collection to proactive anomaly early warning. It solves the pain points of traditional solutions, such as single data dimensions, discontinuous spatial coverage, and difficulty in tracing the effects of policy implementation. Attached Figure Description

[0014] Figure 1 is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0015] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0016] Please refer to Figure 1. This invention provides a method for dynamic monitoring of the ecological environment based on remote sensing and geographic information systems, including:

[0017] S1: Preprocess the optical remote sensing image data and SAR radar data of the area to be detected, identify the changed areas of the area to be detected at the current time and the previous time, and obtain the target changed areas;

[0018] Preferably, step S1 includes:

[0019] S11: Radiometric calibration of optical remote sensing image data is performed using ENVI software, converting the raw values ​​of the remote sensing image into radiance values.

[0020] S12: After applying precise orbital files to correct the position of SAR radar data using SNAP software, the raw values ​​of SAR radar data are converted into backscattering coefficients through radiometric calibration.

[0021] S13: Convert optical remote sensing image data and SAR radar data into a universal map coordinate system, and use the optical remote sensing image as a reference to select ground control points to perform geometric correction on the SAR radar data.

[0022] S14: Open the radiometrically calibrated optical remote sensing image data in ENVI software, calculate the NDVI index of the area to be detected through NDVI, and calculate the change in the NDVI index of the area to be detected between the current time and the previous time.

[0023] S15: Open the radiometrically calibrated SAR radar data in SNAP software and calculate the change in the backscattering coefficient of the area to be detected at the current time and the previous time.

[0024] S16: Set threshold indicators, compare the changes in NDVI index and backscattering coefficient of the area to be detected at the current time and the previous time with the set threshold indicators, identify the feature abnormal pixels of the area to be detected, and aggregate the identified feature abnormal pixels into polygons to obtain the target change area of ​​the area to be detected.

[0025] In this embodiment, step S1 achieves high-precision spatial positioning and quantitative analysis of ecological and environmental changes through preprocessing of optical remote sensing images and SAR radar data and identification of change areas. Specifically, by using ENVI and SNAP software to perform radiometric calibration and geometric correction on remote sensing data and SAR radar data, the influence of sensor errors and terrain deformation is eliminated, improving data accuracy to the centimeter level, which meets the needs of ecological monitoring. By calculating the changes in NDVI index and backscattering coefficient and combining them with threshold analysis, the spatiotemporal variations of environmental elements such as vegetation cover and surface roughness can be accurately captured. The process of aggregating characteristic anomalous pixels into polygonal target change areas realizes the semantic conversion from pixel-level anomalies to regional changes, providing a clear spatial range for subsequent environmental monitoring information extraction.

[0026] For example, in a forest ecological monitoring scenario in a certain watershed, the S1 step is used to process Landsat optical images and Sentinel-1 SAR data of the area. The image values ​​are converted into radiance values ​​through ENVI radiometric calibration, and the backscattering coefficient is obtained after SNAP correction. The difference between the NDVI index at the current time and the previous time (e.g., from 0.6 to 0.4) and the change in SAR backscattering coefficient (e.g., -3dB) are calculated. After setting a threshold to filter out normal fluctuations, an abnormal pixel aggregation area within a 200-hectare range along the riverbank is identified. Finally, this area is determined to be the target change area caused by recent deforestation, providing an accurate spatial benchmark for the targeted extraction of subsequent water quality and vegetation cover data.

[0027] S2: Extract all environmental monitoring information of the target change area at the current time and the previous time, as well as environmental policy information related to the target change area;

[0028] Preferably, the environmental monitoring information includes: air quality data, water quality data, soil quality data, and vegetation cover data; the environmental policy information includes: environmental planning, pollution prevention and control measures, and ecological protection policies.

[0029] In this embodiment, step S2 systematically extracts environmental monitoring information (including air quality (PM2.5, PM10, sulfur dioxide concentrations, etc.), water quality (pH, chemical oxygen demand, ammonia nitrogen content, etc.), soil quality (soil pH, heavy metal content, etc.), vegetation cover data (Normalized Difference Vegetation Index NDVI), etc.) and related environmental policy information (including environmental planning, pollution prevention and control measures, ecological protection policies, etc.) from the target change area at the current and previous time points, constructing a multi-dimensional data association system of "spatial change - environmental parameters - policy constraints." This step can accurately capture the dynamic changes of environmental elements in key areas and parse policy texts through natural language processing technology to clarify the control requirements of the target area, avoiding the problem of data and policy disconnect in traditional monitoring. At the same time, extracting data only for the target change area identified by remote sensing reduces the amount of data processing compared to the full-area sensor acquisition mode, significantly improving data utilization efficiency.

[0030] For example, in a watershed ecological monitoring scenario, after identifying a 200-hectare degraded vegetation area along the riverbank through step S1, step S2 extracts environmental data such as the current water COD concentration increasing by 10 mg / L compared to the previous moment and the soil cadmium content exceeding the standard by 1.2 times. This data is then matched with the policy requirement in the local "Regulations on Pollution Prevention and Control in Water Source Protection Areas" that "direct discharge of industrial wastewater within 2 kilometers of the riverbank is prohibited." This directly links the abnormal environmental data with the policy-controlled area, quickly locating industrial pollution sources. Compared to the traditional single-sensor monitoring mode, this shortens the time for tracing the source of the problem and provides dual support of data and policy for environmental supervision.

[0031] S3: Use natural language processing technology to parse environmental policy information and extract the corresponding control areas;

[0032] Preferably, step S3 includes:

[0033] S31: Download PDF, Word, or HTML documents related to environmental policies from official websites using web crawling technology, and organize the downloaded documents into plain text TXT format;

[0034] S32: Use NLTK or Jieba word segmentation tools to segment the TXT text and perform part-of-speech tagging;

[0035] S33: Using named entity recognition technology, identify the region name and policy implementation time from the text after word segmentation and part-of-speech tagging;

[0036] S34: Determine whether the policy implementation time is between the current moment and the previous moment. If so, compare the extracted regional names with the standard place name database one by one. Standardize the names that are inconsistent in expression but point to the same region.

[0037] S35: Substitute the standardized area name into the environmental policy and rule template library for matching. If the match is successful, the area is determined as a control area. Based on administrative division data, watershed division data, or topographic data, spatial analysis is used to delineate the geographical boundaries and scope of the control area.

[0038] In this embodiment, step S3 uses internet crawlers to obtain policy documents and performs word segmentation and named entity recognition, which can accurately extract the regional names and implementation times involved in the policies. After standardization and rule template matching, the boundaries of the control area are delineated by combining spatial data such as administrative divisions, thus solving the problem of the disconnect between policy text and geospatial information in traditional monitoring. Its core value lies in: transforming abstract policy requirements into quantifiable and locatable geospatial ranges through semantic parsing, providing a policy compliance reference for subsequent environmental change analysis, and realizing spatial traceability of policy implementation effects, thereby improving the accuracy of ecological management.

[0039] For example, in an ecological monitoring scenario of a provincial nature reserve:

[0040] Policy document acquisition and processing: The PDF document "Water Ecological Environment Protection Plan for XX River Basin (2023-2030)" was downloaded from the official website of the Department of Ecology and Environment by web crawling. After being sorted into TXT text, the Jieba word segmentation tool was used to identify key information such as "within 5 kilometers along the XX section of the Yangtze River tributary" and "implemented from January 1, 2024".

[0041] Standardization of regional names: Compare "XX River section" in the text with the standard place name database to confirm that its corresponding administrative division is "XX Town of Yubei District to XX Township of Banan District, Chongqing", and uniformly express it as the official standard name;

[0042] Delineation of Control Areas: The standardized area names are substituted into the environmental protection policy and rule template library and matched with the clause "Prohibition of Development and Construction in Water Source Conservation Areas". Combined with watershed division data and GIS spatial analysis, a 5-kilometer (120 square kilometers) area along the riverbank is delineated as the control area, and a vector data layer containing geographical boundaries is generated.

[0043] S4: Identify areas of environmental change within the controlled area based on all environmental monitoring information of the controlled area at the current and previous times;

[0044] Preferably, step S4 includes:

[0045] S41: For the environmental indicator data monitored by the environmental monitoring stations within the control area at the current time and the previous time, compare the environmental indicator data monitored at the current time with the environmental indicator data at the previous time, and calculate the change value of each environmental indicator.

[0046] S42: Based on the set environmental indicator change threshold, determine whether the environment in the monitoring area corresponding to each environmental monitoring station has changed; if the change value of the environmental indicator monitored by a certain environmental monitoring station exceeds the set threshold, it is considered that the environment in the monitoring area corresponding to that environmental monitoring station has changed.

[0047] S43: Fit the monitoring areas of all environmental monitoring stations where the environment changes to obtain the environmental change area within the control area.

[0048] In this embodiment, step S4 achieves precise location and quantitative analysis of environmental changes within the policy-controlled area through temporal comparison and spatial fitting of environmental monitoring station data within the control area. This step calculates the changes in environmental indicators (such as air quality, water quality, and soil quality) between the current and previous moments and compares them with set thresholds. This automatically identifies monitoring points where significant environmental changes have occurred. Then, spatial interpolation or polygon fitting techniques are used to transform discrete points into continuous areas of change, solving the problem that traditional manual inspections or single-point monitoring struggle to capture regional environmental changes. Its core value lies in dynamically linking the control area defined by the policy text with real-time environmental monitoring data, forming a closed-loop analysis of "policy constraint - environmental response." This provides precise spatial targeting for environmental supervision and allows for tracing the ecological effects after policy implementation, improving the timeliness and accuracy of dynamic supervision.

[0049] For example, in a drinking water source protection area management scenario, this area is designated as a control zone by the "Water Source Protection Regulations," and five water quality monitoring stations and three air quality monitoring stations are deployed. In step S4, the current water quality data (e.g., ammonia nitrogen concentration increasing from 0.5 mg / L to 1.2 mg / L) is compared with the data from the previous time point, calculating a change of 0.7 mg / L, exceeding the set threshold of 0.5 mg / L. Simultaneously, the COD concentration change at another monitoring station also exceeds the threshold; synchronous analysis of air quality data reveals no significant anomalies. Based on these results, a water quality deterioration area of ​​approximately 2 square kilometers is generated through spatial fitting, pinpointing the area near a company's discharge outlet upstream of the protection zone. This process eliminates the need for manual traversal of the entire area's data, directly identifying the polluted area through threshold judgment and spatial fitting. This shortens the response time compared to traditional manual inspection methods and provides precise evidence of the pollution range for environmental law enforcement.

[0050] S5: Match the controlled area with the target change area to obtain the matching area, and extract all environmental monitoring information of the matching area between the current time and the previous time.

[0051] Preferably, step S5 includes: using the spatial analysis function of a Geographic Information System (GIS), importing the vector boundary data of the controlled area and the target change area into the GIS platform, and using the overlay analysis tool of GIS to calculate the overlapping part of the two areas, thereby obtaining the matching area.

[0052] In this embodiment, step S5 achieves precise overlay analysis of the controlled area and the target change area through GIS spatial analysis. Its core value lies in spatially matching the policy-defined control scope with the actual environmental change area identified by remote sensing, thereby identifying the key overlapping area of ​​"policy constraints and environmental response" and providing data targets for subsequent refined monitoring. Specifically, this step uses the vector overlay analysis tool of the GIS platform to perform spatial calculations on the boundary data of the controlled area (such as the policy-defined water source protection area) and the target change area (such as the vegetation degradation area identified by remote sensing), automatically calculating the intersection of the two (i.e., the matching area), and simultaneously extracting the time-series data of environmental monitoring stations within this area. This process solves the problem of "disconnect between policy area and actual change area" in traditional monitoring. Through precise spatial matching, environmental monitoring data can be directly mapped to the key areas of policy control, providing high-value input data for AI large models and significantly improving the targeting and accuracy of anomaly warnings.

[0053] S6: Combine all environmental monitoring information of the target change area, control area, environmental change area and matching area between the current time and the previous time to generate a test sequence. Input the test sequence into the AI ​​big model for prediction to obtain whether the environment of the test area is abnormal. If so, issue an early warning.

[0054] Preferably, step S6 includes:

[0055] S61: Clean and standardize the spliced ​​test sequences to remove monitoring data with too many duplicates and missing values, normalize different types of environmental indicator data, and unify the data format and units.

[0056] S62: Input the preprocessed test sequence into the trained AI model. The AI ​​model learns the data features of the test sequence and predicts the probability value of environmental anomalies.

[0057] S63: When the probability value of an abnormal environment is higher than the set threshold, an early warning will be activated.

[0058] In this embodiment, step S6 integrates multi-dimensional spatial data with environmental monitoring information to construct a closed-loop mechanism of "data fusion - intelligent prediction - automatic early warning," achieving intelligent identification and rapid response to ecological and environmental anomalies. This step structurally splices multi-source spatial data, such as target change areas and control areas, with time-series data from environmental monitoring stations. After preprocessing such as cleaning and normalization, the data is input into an AI large-scale model. By mining the implicit correlations between data features, anomaly probability prediction is achieved, solving the problems of low efficiency in manual analysis and weak correlation between multi-source data in traditional monitoring. Its core value lies in transforming scattered environmental information into a knowledge graph with early warning value through collaborative analysis of multi-dimensional data, shifting monitoring from "post-event tracing" to "pre-event prediction," and significantly improving the level of intelligence in ecological and environmental supervision.

[0059] For example, in an ecological monitoring scenario within a certain urban agglomeration, steps S1-S5 identified a target change area of ​​30 hectares of farmland on the urban fringe that was converted to construction land. This area also falls within the control zone defined by the "Soil Pollution Prevention and Control Action Plan." Step S4 further identified 5 hectares of this area showing an increase in soil heavy metal concentrations. Step S6 then stitched together time-series data from three soil monitoring stations and two air quality monitoring stations within this area (e.g., cadmium concentration increasing from 0.3 mg / kg to 0.5 mg / kg, and PM2.5 concentration increasing from 25 μg / m³ to 40 μg / m³). After data cleaning to remove outliers and normalization, the data was input into a trained AI model. The model analyzed the synergistic changes in soil heavy metals and air quality data, predicting an environmental anomaly probability of 82% (exceeding the set threshold of 70%). It then activated an early warning mechanism, indicating a risk of industrial pollution spread in the area.

[0060] Preferably, the AI ​​large-scale model of this invention can adopt a model architecture capable of handling multi-source heterogeneous data fusion and spatiotemporal feature analysis, such as a hybrid model combining Transformer and Graph Neural Network (GNN), or a spatiotemporal sequence prediction model integrating Convolutional Neural Network (CNN) and Long Short-Term Memory Network (LSTM). Such models can effectively learn the feature correlations between multi-dimensional spatial data such as target change areas and control areas and time-series data from environmental monitoring stations. Through a weighted fusion of remote sensing image features, policy text semantic information, and environmental indicator values ​​using an attention mechanism, they can achieve probabilistic prediction of ecological and environmental anomalies. For example, in urban agglomeration ecological monitoring scenarios, this model can learn the spatiotemporal coupling patterns of soil heavy metal concentration, air quality indicators, and land use change through training. When inputting a spliced ​​multi-source data sequence, it can quickly capture the implicit correlations between data features, output environmental anomaly probability values, and trigger early warnings, meeting the full-process requirements from multi-source data to intelligent prediction.

[0061] Another aspect of the present invention provides an ecological environment dynamic monitoring system based on remote sensing and geographic information system, comprising: a memory and a processor; the memory is used to store an application program; the processor is used to run the application program and execute the ecological environment dynamic monitoring method based on remote sensing and geographic information system.

[0062] Another aspect of the present invention provides a computer storage medium storing a program that, when executed by a processor, implements the aforementioned method for dynamic monitoring of the ecological environment based on remote sensing and geographic information systems.

[0063] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0064] In summary, this invention addresses the problems of excessive reliance on multiple sensors in existing ecological and environmental monitoring, resulting in large data volumes, time-consuming processing, inconvenient dynamic supervision, and insufficient policy linkage. It achieves dynamic capture of environmental elements over large areas using optical remote sensing and SAR radar data, significantly reducing reliance on dense sensors. Sensor data is extracted only in targeted areas of target change and control. Natural language processing is used to parse environmental policy text and match it with remote sensing data locations. Multi-source data is then stitched together and input into a large AI model to predict environmental anomalies and issue early warnings. This forms a highly efficient monitoring model of "macroscopic identification - microscopic verification - policy linkage," reducing data collection and processing pressure and hardware costs while improving the accuracy and timeliness of dynamic supervision. It represents a technological leap from passive data collection to proactive anomaly early warning, solving the pain points of traditional solutions such as single data dimensions, discontinuous spatial coverage, and difficulty in tracing the effects of policy implementation.

[0065] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for dynamic monitoring of the ecological environment based on remote sensing and geographic information systems, characterized in that, include: S1: Preprocess the optical remote sensing image data and SAR radar data of the area to be detected, identify the changed areas of the area to be detected at the current time and the previous time, and obtain the target changed areas; Step S1 includes: S11: Radiometric calibration of the optical remote sensing image data using ENVI software, converting the raw values ​​of the remote sensing image into radiance values; S12: After applying precise orbital file correction to the SAR radar data using SNAP software, radiometric calibration is performed to convert the raw values ​​of the SAR radar data into backscattering coefficients; S13: The optical remote sensing image data and SAR radar data are uniformly converted into a universal map coordinate system, and ground control points are selected to perform geometric correction on the SAR radar data based on the optical remote sensing image; S14: The radiometrically calibrated optical remote sensing image data is opened in ENVI software, and the NDVI index of the area to be detected is calculated using NDVI, and the NDVI index of the area to be detected is calculated. S15: Open the radiometrically calibrated SAR radar data in the SNAP software and calculate the change in backscattering coefficient of the area to be detected between the current and previous times; S16: Set threshold indicators, compare the changes in NDVI and backscattering coefficient of the area to be detected between the current and previous times with the set threshold indicators, identify the characteristic abnormal pixels of the area to be detected, and aggregate the identified characteristic abnormal pixels into polygons to obtain the target change area of ​​the area to be detected; S2: Extract all environmental monitoring information of the target change area between the current and previous times, as well as environmental policy information related to the target change area; The environmental monitoring information includes: air quality data, water quality data, soil quality data, and vegetation cover data; the environmental policy information includes: environmental planning, pollution prevention and control measures, and ecological protection policies; S3: Use natural language processing technology to parse the environmental policy information and extract the control areas corresponding to the environmental policy information; Step S3 includes: S31: Download PDF, Word, or HTML documents related to environmental policies from official websites using web crawling technology, and organize the downloaded documents into plain text TXT format; S32: Use NLTK or Jieba word segmentation tools to segment the TXT text and perform part-of-speech tagging; S33: Use named entity recognition technology to extract the control areas corresponding to the environmental policy information. S34: Identify the region name and policy implementation time from the text after word segmentation and part-of-speech tagging; If the policy implementation time is between the current time and the previous time, compare the extracted region name with the standard place name database one by one, and standardize the names that are inconsistent but point to the same region; S35: Substitute the standardized region name into the environmental policy rule template database for matching. If the match is successful, the region is determined to be a control area; and the geographical boundaries and scope of the control area are delineated through spatial analysis based on administrative division data, watershed division data or topographic data; S4: Identify the environmental change areas within the control area based on all environmental monitoring information of the control area at the current time and the previous time.S5: Match the controlled area with the target change area to obtain the matching area, and extract all environmental monitoring information of the matching area between the current time and the previous time. S6: Concatenate all environmental monitoring information of the target change area, controlled area, environmental change area, and matching area between the current time and the previous time to generate a test sequence. Input this sequence into the AI ​​large model for prediction to determine whether the environment of the area to be detected is abnormal. If so, issue an early warning.

2. The method for dynamic monitoring of the ecological environment based on remote sensing and geographic information systems according to claim 1, characterized in that, Step S4 includes: S41: Comparing the environmental indicator data monitored at the current time within the control area with the environmental indicator data at the previous time, and calculating the change value of each environmental indicator; S42: Determining whether the environment within the monitoring area corresponding to each environmental monitoring station has changed according to the set environmental indicator change threshold; If the change value of the environmental indicator monitored by a certain environmental monitoring station exceeds the set threshold, it is considered that the environment of the monitoring area corresponding to that environmental monitoring station has changed; S43: Fitting the monitoring areas of all environmental monitoring stations where the environment has changed to obtain the environmental change area within the control area.

3. The method for dynamic monitoring of the ecological environment based on remote sensing and geographic information systems according to claim 1, characterized in that, Step S5 includes: using the spatial analysis function of Geographic Information System (GIS), importing the vector boundary data of the controlled area and the target change area into the GIS platform, and using the overlay analysis tool of GIS to calculate the overlapping part of the two areas, thereby obtaining the matching area.

4. The method for dynamic monitoring of the ecological environment based on remote sensing and geographic information systems according to claim 1, characterized in that, Step S6 includes: S61: Cleaning and standardizing the spliced ​​test sequence to remove monitoring data with too many duplicates and missing values, normalizing different types of environmental indicator data, and unifying the data format and units; S62: Inputting the preprocessed test sequence into the trained AI model, which learns the data features of the test sequence to predict the probability value of environmental anomalies; S63: Activating an early warning when the probability value of environmental anomalies is higher than a set threshold.

5. An ecological environment dynamic monitoring system based on remote sensing and geographic information systems, characterized in that, The system includes a memory and a processor; the memory is used to store an application program; the processor is used to run the application program and execute the method for dynamic monitoring of the ecological environment based on remote sensing and geographic information system as described in any one of claims 1 to 4.

6. A computer storage medium, characterized in that, The computer storage medium stores a program, which, when executed by a processor, implements any one of the ecological environment dynamic monitoring methods based on remote sensing and geographic information systems as described in claims 1 to 4.

Images