AI-based environment monitoring data query system and method

Abstract

The invention relates to the technical field of data processing, and particularly discloses an AI-based environment monitoring data query system and method, and the system comprises a data storage module which is used for storing and managing multi-source heterogeneous data, including real-time and historical monitoring data of a water environment and an atmospheric environment, meteorological data, monitoring station data, pollution source data and a historical analysis report; the intelligent engine module is constructed based on a large language model and used for executing the basic analysis task, obtaining a natural language query instruction input by a user, querying the stored data according to the natural language query instruction and generating a query result; the scene sensing module is used for acquiring a corresponding working scene according to the equipment information of the login system; the self-adaptive processing module is used for executing differentiated data acquisition and formatting output according to different working scenes; and the interactive display module is used for displaying the final query and analysis result to the user. By adopting the technical scheme of the invention, the data query threshold can be reduced, and the scene-based self-adaptive data service is realized.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an AI-based environmental monitoring data query system and method. Background Technology

[0002] With the advancement of ecological civilization construction and the continuous improvement of environmental monitoring networks, the volume of environmental monitoring data has experienced explosive growth. The existing environmental monitoring system covers multiple dimensions, including water environment (such as surface water and drinking water sources), atmospheric environment (such as air quality stations and micro-stations), and pollution sources (online monitoring CEMS), accumulating massive amounts of real-time monitoring data and historical analysis reports. To utilize this data for environmental supervision and decision analysis, various environmental monitoring data query and management systems have emerged.

[0003] However, the query interfaces of traditional environmental monitoring systems are typically built upon complex forms, drop-down menus, or sophisticated SQL query logic. When users query specific data (e.g., "query the section of a river where ammonia nitrogen exceeded the standard in July 2025"), they often need to manually switch between multiple menus and select filters such as time range, station name, and monitoring indicators. For business personnel without a technical background, the process is cumbersome and prone to errors. Furthermore, when faced with complex statistical analysis needs, existing systems often cannot directly understand the user's natural language intent, resulting in low data retrieval efficiency.

[0004] Therefore, there is a need for an AI-based environmental monitoring data query system and method that can lower the threshold for data querying and enable scenario-based adaptive data services. Summary of the Invention

[0005] One of the objectives of this invention is to provide an AI-based environmental monitoring data query system that can lower the threshold for data querying and enable scenario-based adaptive data services.

[0006] To solve the above-mentioned technical problems, this application provides the following technical solution: An AI-based environmental monitoring data query system includes: The data storage module is used to store and manage multi-source heterogeneous data, including real-time and historical monitoring data of water and atmospheric environments, meteorological data, monitoring station data, pollution source data, and historical analysis reports; The intelligent engine module, built on a large language model, is used to perform basic analysis tasks, obtain natural language query instructions input by users, query stored data according to the natural language query instructions, and generate query results. The scene perception module is used to obtain the corresponding work scene based on the device information logged into the system. The adaptive processing module is used to perform differentiated data acquisition and formatted output according to different work scenarios; The interactive display module is used to show users the final query and analysis results.

[0007] Furthermore, the scene perception module is used to determine the work scene by detecting the type of user's logged-in device and the current geographical location information in the device information; Work scenarios include office scenarios and field investigation scenarios.

[0008] Furthermore, when the scene perception module determines that it is an office scene or a field investigation scene, the adaptive processing module is used to receive the user's screenshot input, perform image recognition and semantic analysis on the screenshot, obtain the input scene and query requirements, and return the corresponding query results.

[0009] Furthermore, when the scene perception module determines that it is an office scene, the adaptive processing module is used to identify the interface features in the screenshot and determine the user's current specific input scene, including instant messaging software chat scene, system filling scene and report writing scene. When the screenshot contains features of a chat software, it is identified as an instant messaging software chat scenario. The adaptive processing module recognizes the natural language requirements in the chat history, generates a query command to perform the query, and generates chat reply text based on the query results. When the screenshot contains features of the data entry interface of the monitoring system, it is identified as a system data entry scenario. The adaptive processing module identifies the field name to be entered, queries the corresponding value, and generates structured data. When a screenshot contains features of a document editing interface, it is identified as a report writing scenario. The adaptive processing module analyzes the context before and after the cursor and automatically generates a paragraph of text.

[0010] Furthermore, when the scene perception module determines that it is a field investigation scene, the adaptive processing module is used to obtain the location, determine the query items and perform pre-analysis based on the monitoring station closest to the location; In field investigation scenarios, the interactive display module is also used to proactively push pre-analysis results.

[0011] Furthermore, the data storage module also stores personnel organizational structure information, including information about the departments to which employees belong; When the scene perception module determines that it is an office scene and identifies it as a report writing scene, the adaptive processing module is also used to perform semantic recognition on the text content in the screenshot to determine whether the current writing content belongs to the atmospheric environment category. If so, it extracts the query location or the location of the analysis object from the screenshot. It is also used to analyze other text content in the screenshot besides the query data to determine whether the report contains the writer's subjective analysis. If subjective analysis content is detected, the adaptive processing module is also used to call the scene perception module to query the real-time login status information to determine whether there are staff members in the field investigation scene within the range of the above-identified query location. If there are field survey staff, the adaptive processing module is also used to read the pre-stored personnel organizational structure information to determine whether the field survey staff and the current office staff belong to the same department; if they belong to the same department, it further determines whether the current survey content is water environment based on the current location of the field survey staff or the historical query records of its mobile device. If the field investigators are indeed investigating the water environment, the interactive display module is also used to generate a sharing prompt on the PC interface of the office staff. The adaptive processing module is also used to automatically perform OCR recognition and content extraction on the screenshot of the currently completed analysis report provided by the office staff after the office staff confirms the sharing. The extracted atmospheric environment analysis text is then pushed to the mobile device of the field investigation staff by the interactive display module.

[0012] The second objective of this invention is to provide an AI-based method for querying environmental monitoring data, comprising the following steps: S1. Pre-store and manage multi-source heterogeneous data, including real-time and historical monitoring data of water and atmospheric environments, as well as meteorological data, monitoring station data, pollution source data, historical analysis reports, and personnel organizational structure information; S2. When a user logs into the system, the system detects the type of the user's login device, network environment, and current geographical location information to determine the user's current work scenario. If the scenario is determined to be an office scenario, proceed to step S3. If the scenario is determined to be a field investigation scenario, proceed to step S4. S3. Receive screenshot input from the user, perform image recognition and semantic analysis on the screenshot, identify the user's current specific input scenario, and perform differentiated queries and outputs; S4. Automatically obtain the user's current geographical location information, calculate the Euclidean distance between the coordinates and each monitoring station in the system, determine the nearest station within a preset radius, and automatically determine the query items based on the station type: Based on the latest monitoring data retrieved, it is determined whether there are any abnormalities exceeding the standard. If an abnormality is found, an environmental briefing containing the location and the index exceeding the standard will be proactively popped up on the user's mobile device.

[0013] Furthermore, step S3 specifically includes: S301. When chat software features are identified in the screenshot, extract the natural language requirements from the chat history, convert the natural language into query instructions, and generate a conversational text containing conclusions to return to the user after obtaining the data; the natural language requirements include time, location, and metrics. S302. When the data entry interface features are detected in the screenshot, the field name to be filled in is identified, the corresponding value is automatically queried, and structured data is generated. S303. When document editing interface features are detected in the screenshot, the context before and after the cursor is analyzed, and paragraph text is automatically generated.

[0014] Furthermore, in step S4, when determining the query item, if the nearest station is a water quality monitoring station, the query item is determined to be water environment, and real-time water quality data is automatically retrieved. If the nearest station is an air quality micro-station or is located in an industrial area, and the query item is determined to be atmospheric environment, real-time air quality data will be automatically retrieved.

[0015] Furthermore, in step S303, if the scenario is identified as a report writing scenario, a cross-scenario collaborative linkage step S5 is also executed, specifically including: S501. Recognize the text content in the screenshot. If the content is determined to belong to the atmospheric environment category, extract the relevant query location or analysis object location. At the same time, analyze whether the text contains the author's subjective analysis logic to confirm the author's analysis status of the location. S502. If subjective analysis content is detected, the system queries the real-time login status and determines whether there are other staff members in the field investigation scenario within the set range of the query location. S503. If there are field investigators, read the personnel organizational structure information and determine whether the field investigators and the current office staff belong to the same department; if so, determine whether their current investigation content is water environment based on their location or historical query records. S504. If it is confirmed that the field investigators are conducting a water environment survey, a sharing prompt will be generated on the office staff interface. After the office staff confirms the sharing, the atmospheric environment analysis text in the screenshot provided by the office staff will be automatically extracted and pushed to the field investigators' mobile devices.

[0016] Existing environmental monitoring systems typically only provide standardized table outputs, failing to meet the needs of diverse work scenarios. This solution intelligently identifies whether the user is chatting, filling out a form, or writing a report. In chat scenarios, it generates conversational conclusions for rapid communication; in form-filling scenarios, it generates structured data for easy data migration; and in report scenarios, it generates written paragraphs to aid in document drafting. This differentiated output seamlessly integrates into the user's actual workflow, enhancing the user experience. For field investigation scenarios, the system changes the traditional passive mode of manually searching for data. Through the linkage between GPS positioning and the interactive display module, it automatically calculates the nearest site and determines the environmental category, proactively pushing environmental briefings and warnings of exceeding standards. This allows field personnel to immediately grasp key information upon arrival at the site, improving the efficiency of on-site investigations. Attached Figure Description

[0017] Figure 1 This is a logical block diagram of an AI-based environmental monitoring data query system. Detailed Implementation

[0018] The following detailed description illustrates the specific implementation method: Example 1 like Figure 1 As shown, the AI-based environmental monitoring data query system of this embodiment includes: a data storage module, an intelligent engine module, a scene perception module, an adaptive processing module, and an interactive display module.

[0019] The data storage module is used to store and manage multi-source heterogeneous data, including real-time and historical monitoring data of the water and atmospheric environments, as well as meteorological data, monitoring station data, pollution source data, and historical analysis reports. Water environment data includes parameters such as ammonia nitrogen and total phosphorus; atmospheric environment data includes parameters such as PM2.5 and carbon dioxide.

[0020] The intelligent engine module provides basic data query and analysis capabilities. In this embodiment, the intelligent engine module is built upon a large language model. For example, it employs a large language model similar to DeepSeek and a retrieval enhancement generation architecture similar to RagFlow, integrating traditional AI algorithms and using Text-to-SQL technology to convert natural language into database query commands.

[0021] The intelligent engine module is also used to perform basic analysis tasks, obtain natural language query commands input by the user, query stored data based on the natural language query commands, and generate query results. For example, based on the user's input command "Query the ammonia nitrogen exceeding the standard section of the xx river in July 2025", the module sequentially performs the following steps: filtering the water environment monitoring data of the xx river section in July 2025, checking data integrity, obtaining ammonia nitrogen monitoring standards, judging whether the ammonia nitrogen exceeds the standard in the water environment monitoring data, filtering out the sections that exceed the standard, and generating analysis conclusions.

[0022] The scene perception module is used to obtain the corresponding work scene based on the device information of the logged-in system. Work scenes include office scenes and field investigation scenes. In this embodiment, the scene perception module determines the work scene by detecting the type of user's logged-in device (e.g., PC or mobile), network environment, and current geographical location information (e.g., GPS coordinates). In this embodiment, for PC login, the IP address and office address consistently indicate an office scene; for mobile login, the GPS coordinates do not match the pre-stored office GPS coordinates, indicating a field investigation scene.

[0023] The adaptive processing module is used to execute differentiated data acquisition and formatted output logic according to different work scenarios.

[0024] When the scene perception module determines that it is an office scene, the adaptive processing module receives the user's screenshot input, performs image recognition and semantic analysis on the screenshot, obtains the input scene and query requirements, and returns the corresponding query results.

[0025] In this embodiment, the adaptive processing module has a built-in multimodal visual analysis unit, which is used to identify interface features in the screenshot (such as chat bubbles, form input boxes, and document paragraphs) to determine the user's current specific input scenario, including instant messaging software chat scenario, system filling scenario, and report writing scenario.

[0026] When a screenshot contains chat software features (such as chat bubbles or avatars), it is identified as an instant messaging chat scenario. The adaptive processing module recognizes the natural language requirements in the chat history. For example, if the other party in the screenshot asks, "Send me the total phosphorus data for Yutan Reservoir last month," the module extracts the query conditions (time: last month, location: Yutan Reservoir, indicator: total phosphorus), generates a query command, and produces a conversational chat reply text containing a conclusion based on the query results. For example, "The average total phosphorus concentration in Yutan Reservoir last month was 0.12 mg / L, and the overall water quality was stable." The user can directly copy and paste this text to reply, avoiding the tedious data filtering process.

[0027] When a screenshot contains features of the monitoring system's data entry interface (such as tables, input boxes, or "Fill in" buttons), it is identified as a system data entry scenario. The adaptive processing module then identifies the field names to be entered. For example, if fields such as "COD" or "NH3-N" are identified, the module automatically queries the corresponding values ​​and generates structured data in pure numerical or JSON format, without any decorative text, making it convenient for users to directly copy and enter or for automatic filling via plugins.

[0028] When a screenshot contains features of a document editing interface (such as a Word toolbar or cursor position), it is identified as a report writing scenario, and the adaptive processing module analyzes the context before and after the cursor. For example, if the cursor is after the text "Water Quality Analysis of Yutan Reservoir," the module automatically generates a professional, formal, and logically sound paragraph. For example: "Monitoring data shows that the overall water quality of Yutan Reservoir remained stable this month, with the concentration of major pollutants decreasing by 5.2% compared to the same period last year, and the impact of rainfall and runoff was minimal." When the scene perception module determines that it is a field investigation scene, the adaptive processing module is used to automatically obtain the location, determine the query items and perform pre-analysis based on the monitoring station closest to the location.

[0029] The adaptive processing module is specifically used to calculate the Euclidean distance between the current GPS coordinates and the coordinates of each monitoring station in the system, and to determine the nearest station within a preset radius. The adaptive processing module is also used to automatically determine the query items according to the station type: in this embodiment, the preset radius is 1 kilometer. If the nearest station is a water quality monitoring station or is located within the buffer zone of a river or lake, the query item is determined to be water environment, and real-time water quality data (such as dissolved oxygen and permanganate index) of the nearest section or water quality monitoring station is automatically retrieved. If the nearest station is an air quality micro-station or is located near an industrial park or road, the query item will be determined as atmospheric environment, and the real-time air quality data (such as AQI, PM2.5) of that station will be automatically retrieved.

[0030] The interactive display module is used to show users the final query and analysis results. In field investigation scenarios, the interactive display module is also used to proactively push pre-analysis results. In this embodiment, after determining the query item, the adaptive processing module will judge whether there are any abnormalities exceeding the standard based on the latest monitoring data. If an abnormality is found, the interactive display module will pop up an environmental briefing on the mobile homepage. For example, when a user arrives at a river, the phone will automatically pop up: "You are currently located at section xx of the Liangtan River. The ammonia nitrogen concentration at the nearest section xx (200 meters) exceeds the standard, and the trend is continuously rising. It is recommended to focus on investigating the sewage outlet 1km upstream." Existing environmental monitoring systems typically only provide standardized table outputs, failing to meet the needs of diverse work scenarios. This embodiment, through the collaboration of a scenario-aware module and an adaptive processing module, intelligently identifies whether the user is chatting, filling out a form, or writing a report. In chat scenarios, it generates conversational conclusions for rapid communication; in form-filling scenarios, it generates pure numerical values / JSON for easy data migration; and in report scenarios, it generates formalized paragraphs to aid in document writing. This differentiated output seamlessly integrates into the user's actual workflow, enhancing the user experience. For field investigation scenarios, the system changes the traditional passive mode of manually searching for data. Through the linkage of GPS positioning and interactive display modules, it automatically calculates the nearest site and determines the environmental category, proactively pushing environmental briefings and warnings of exceeding standards. This allows field personnel to immediately grasp key information upon arrival at the site, improving the efficiency of on-site investigations.

[0031] Example 2 The difference between this embodiment and Embodiment 1 is that the data storage module in this embodiment also stores personnel organizational structure information. This personnel organizational structure information includes information about the departments to which employees belong.

[0032] When the scene perception module determines that it is an office scene and identifies it as a report writing scene, the adaptive processing module is also used to perform linked analysis: The adaptive processing module is also used to perform semantic recognition on the text content in the screenshot to determine whether the currently written content belongs to the atmospheric environment category (e.g., recognizing keywords such as "exhaust gas", "PM2.5", and "VOCs"). If so, it further extracts the query location or the location of the analysis object from the screenshot (e.g., recognizing "XX Industrial Park" or "XX Enterprise").

[0033] The adaptive processing module is also used to analyze other text content in the screenshot besides the query data to determine whether the report contains the author's subjective analysis. Specifically, it identifies whether there are logical connectors such as "analysis believes," "the main reason is," and "in conclusion" in the paragraphs, or determines whether there are natural language paragraphs not directly returned by the database before the cursor position, thereby determining whether the author is conducting subjective analysis of the current query location or the location of the analysis object, or has completed the analysis of related locations.

[0034] If subjective analysis content is detected, the adaptive processing module calls the scene perception module to query the real-time login status information and determine whether there are staff members in the field investigation scene within the set range (e.g., within a radius of 2 kilometers) of the above-identified query location (e.g., XX Industrial Park).

[0035] If field survey personnel are present, the adaptive processing module also reads pre-stored personnel organizational structure information to determine whether the field survey personnel and the current office staff belong to the same department. If they belong to the same department, it further determines whether the current survey content is related to the water environment based on the field survey personnel's current location (whether it is near a river or sewage outlet) or their mobile device's historical query records.

[0036] If the field investigator's current investigation is indeed about the water environment, the interactive display module will generate a sharing prompt on the office staff's PC interface. The prompt might read: "Your colleague xx is conducting a water environment patrol near the XX Industrial Park. We suggest sharing your current atmospheric analysis findings for this location with your colleague so they can combine them with the latest atmospheric data for further investigation." Once the office staff confirms the sharing, the adaptive processing module automatically performs OCR recognition and content extraction on the screenshot of the currently completed analysis report provided by the office staff. The extracted atmospheric environment analysis text is then pushed to the mobile device of the field investigation staff by the interactive display module. For example, the extracted atmospheric environment analysis text might suggest an abnormally high VOCs concentration in the area, indicating suspected illegal emissions from a company.

[0037] During the testing and research of the system, it was found that there is a common problem of poor information communication between office analysts and field investigators. Especially when the atmospheric environment and water environment belong to different monitoring categories, this contradiction is particularly prominent. When office personnel analyze air pollution clues in a specific area (such as an industrial park), they often do not actively inquire whether there are field personnel from the same department conducting water environment inspections on site; while field personnel also lack auxiliary clues from real-time analysis in the office during the investigation process. Since these two types of environmental elements are highly correlated in actual scenarios, for example, abnormal exhaust gas and wastewater emissions from the same enterprise often occur simultaneously, the lag in communication will directly lead to the neglect of compound pollution problems or the inability to effectively investigate them in a timely manner at the first time. In response to the above problems found, in this embodiment, the personnel organizational structure information is incorporated into the data storage module, and logics such as environmental category recognition, location extraction, subjective analysis and judgment, and personnel status perception are introduced into the adaptive processing module, realizing an intelligent collaboration mechanism across environmental categories and scenarios. When office personnel write a report, the system can identify that the screenshot content belongs to atmospheric environment analysis, and further extract the location involved and the completion status of the analysis conclusion. On this basis, the adaptive processing module can query the login status information obtained by the scenario perception module in real time, judge whether there are field investigators from the same department performing tasks near the same location, and decide whether to trigger a linkage prompt according to the on-site environmental category (such as water environment inspection). After the linkage is triggered, after the office personnel complete the analysis of air pollution in a certain industrial park, the system can synchronize the analysis clues to colleagues who are conducting water environment inspections on site in a timely manner at key nodes. It can significantly bridge the communication gap between different monitoring categories, enabling air pollution clues to be transformed into references for water environment on-site inspections in a timely manner. Based on this, field personnel can quickly lock in rainwater outlets, hidden pipes, or sewage outlets, achieving precise identification of gas-water co-discharge behaviors, and greatly improving the collaboration efficiency and problem discovery ability.

[0038] Meanwhile, this embodiment adopts a differentiated linkage strategy, only pushing information when the monitoring categories are different (e.g., air versus water), avoiding valueless information interference due to timeliness differences between similar monitoring tasks. From a communication cost perspective, the same monitoring category often has a fixed communication mechanism, and information exchange is relatively smooth; however, different monitoring categories lack stable proactive communication. In actual work, due to the division of labor and different task rhythms, a high information gap often forms, resulting in pollution clues that should be interconnected not being shared in a timely manner. Moreover, similar analysis reports written by the office are usually based on historical data from the previous monitoring cycle, which is usually understood before field inspections. This makes it difficult to provide substantial guidance for field inspections focusing on real-time monitoring of the current cycle. However, the situation is completely different in cross-category scenarios. Although the atmospheric analysis report comes from historical data, the abnormal exhaust gas reflected in it often means that the enterprise is still in an active production state recently. Active production is a key prerequisite for the possible hidden sewage outlets, intermittent emissions, or mixed rainwater and sewage discharge problems on site, and this is usually not proactively understood before field inspections. Therefore, atmospheric clues are actually of high real-time value for personnel conducting on-site water environment inspections, helping them to accurately narrow down the scope of their investigations and avoid blind searches. This embodiment, through a differentiated push strategy, avoids frequent interference from similar invalid information while ensuring that cross-category information is delivered at the most needed time, thereby simplifying user operations and improving pollution source tracing capabilities.

[0039] Example 3 This embodiment provides an AI-based method for querying environmental monitoring data, using the system of Embodiment 1 or Embodiment 2, and includes the following steps: S1. Pre-store and manage multi-source heterogeneous data, including real-time and historical monitoring data of water and atmospheric environments, as well as meteorological data, monitoring station data, pollution source data, historical analysis reports, and personnel organizational structure information; among which, personnel organizational structure information includes information on the departments to which employees belong.

[0040] S2. When a user logs into the system, the system detects the type of the user's login device (PC or mobile), network environment, and current geographical location information (GPS coordinates) to determine the user's current work scenario. If the scenario is determined to be an office scenario, proceed to step S3. If the scenario is determined to be a field investigation scenario, proceed to step S4.

[0041] S3. Receive screenshot input from the user, perform image recognition and semantic analysis on the screenshot, identify the user's current specific input scenario, and execute differentiated queries and outputs, specifically including: S301. When chat software features are identified in the screenshot, extract the natural language requirements (including time, location, and metrics) from the chat history, convert the natural language into query instructions through an intelligent engine, and generate a conversational text containing conclusions after obtaining the data and returning it to the user; the natural language requirements include time, location, and metrics. S302. When the data entry interface features are detected in the screenshot, the field name to be filled in is identified, the corresponding value is automatically queried, and data in pure numerical or JSON format is generated. S303. When document editing interface features are detected in the screenshot, the context before and after the cursor is analyzed, and formalized paragraph text is automatically generated.

[0042] S4. Automatically obtain the user's current geographic location information (GPS coordinates), calculate the Euclidean distance between these coordinates and each monitoring station in the system, determine the nearest station within a preset radius (e.g., 1 kilometer), and automatically determine the query items based on the station type: If the nearest station is a water quality monitoring station or is located in a water buffer zone, and the query item is determined to be water environment, real-time water quality data will be automatically retrieved. If the nearest station is an air micro-station or is located in an industrial area or near a road, and the query item is determined to be atmospheric environment, real-time air quality data will be automatically retrieved. Based on the latest monitoring data retrieved, it is determined whether there are any abnormalities exceeding the standard. If an abnormality is found, an environmental briefing containing the location and the index exceeding the standard will be proactively popped up on the user's mobile device.

[0043] In this embodiment, if the scenario is identified as a report writing scenario, the cross-scenario collaborative linkage step S5 is further executed, which specifically includes: S501. Recognize the text content in the screenshot. If the content is determined to belong to the atmospheric environment category (e.g., containing the keyword "exhaust gas"), extract the relevant query location or analysis object location. At the same time, analyze whether the text contains the author's subjective analysis logic (e.g., recognizing "analysis believes") to confirm the author's analysis status of the location. S502. If subjective analysis content is detected, the system queries the real-time login status and determines whether there are other staff members in the field investigation scenario within the set range of the query location. S503. If there are field investigators, read the personnel organizational structure information and determine whether the field investigators and the current office staff belong to the same department; if so, determine whether their current investigation content is water environment based on their location or historical query records. S504. If it is confirmed that the field investigators are conducting a water environment survey, a sharing prompt will be generated on the office staff interface. After the office staff confirms the sharing, the atmospheric environment analysis text in the screenshot provided by the office staff will be automatically extracted and pushed to the field investigators' mobile devices to achieve collaborative investigation of air and water discharge.

[0044] The above are merely embodiments of the present invention. The invention is not limited to the fields covered by these embodiments. Commonly known structures and characteristics in the solutions are not described in detail here. Those skilled in the art are aware of all common technical knowledge in the field prior to the application date or priority date, are able to access all existing technologies in that field, and have the ability to apply conventional experimental methods prior to that date. Those skilled in the art can, under the guidance of this application, improve and implement this solution in combination with their own capabilities. Some typical known structures or methods should not be obstacles for those skilled in the art to implement this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the structure of the present invention. These should also be considered within the scope of protection of the present invention, and will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. An AI-based environmental monitoring data query system, characterized in that, include: The data storage module is used to store and manage multi-source heterogeneous data, including real-time and historical monitoring data of water and atmospheric environments, meteorological data, monitoring station data, pollution source data, and historical analysis reports; The intelligent engine module, built on a large language model, is used to perform basic analysis tasks, obtain natural language query instructions input by users, query stored data according to the natural language query instructions, and generate query results. The scene perception module is used to obtain the corresponding work scene based on the device information logged into the system. The adaptive processing module is used to perform differentiated data acquisition and formatted output according to different work scenarios; The interactive display module is used to show users the final query and analysis results.

2. The AI-based environmental monitoring data query system according to claim 1, characterized in that: The scene perception module is used to determine the work scene by detecting the type of user's logged-in device and the current geographical location information in the device information. Work scenarios include office scenarios and field investigation scenarios.

3. The AI-based environmental monitoring data query system according to claim 2, characterized in that: When the scene perception module determines that it is an office scene or a field investigation scene, the adaptive processing module receives the user's screenshot input, performs image recognition and semantic analysis on the screenshot, obtains the input scene and query requirements, and returns the corresponding query results.

4. The AI-based environmental monitoring data query system and method according to claim 3, characterized in that: When the scene perception module determines that it is an office scene, the adaptive processing module is used to identify the interface features in the screenshot and determine the user's current specific input scene, including instant messaging software chat scene, system filling scene and report writing scene. When the screenshot contains features of chat software, it is identified as an instant messaging software chat scenario. The adaptive processing module recognizes the natural language requirements in the chat history, generates a query command to perform the query, and generates chat reply text based on the query results. When the screenshot contains features of the data entry interface of the monitoring system, it is identified as a system data entry scenario. The adaptive processing module identifies the field name to be entered, queries the corresponding value, and generates structured data. When a screenshot contains features of a document editing interface, it is identified as a report writing scenario. The adaptive processing module analyzes the context before and after the cursor and automatically generates a paragraph of text.

5. The AI-based environmental monitoring data query system according to claim 4, characterized in that: When the scene perception module determines that it is a field investigation scene, the adaptive processing module is used to obtain the location, determine the query items and perform pre-analysis based on the monitoring station closest to the location; In field investigation scenarios, the interactive display module is also used to proactively push pre-analysis results.

6. The AI-based environmental monitoring data query system according to claim 5, characterized in that: The data storage module also stores personnel organizational structure information, which includes information about the departments to which employees belong. When the scene perception module determines that it is an office scene and identifies it as a report writing scene, the adaptive processing module is also used to perform semantic recognition on the text content in the screenshot to determine whether the current writing content belongs to the atmospheric environment category. If so, it extracts the query location or the location of the analysis object from the screenshot. It is also used to analyze other text content in the screenshot besides the query data to determine whether the report contains the writer's subjective analysis. If subjective analysis content is detected, the adaptive processing module is also used to call the scene perception module to query the real-time login status information to determine whether there are staff members in the field investigation scene within the range of the above-identified query location. If there are field survey staff, the adaptive processing module is also used to read the pre-stored personnel organizational structure information to determine whether the field survey staff and the current office staff belong to the same department; if they belong to the same department, it further determines whether the current survey content is water environment based on the current location of the field survey staff or the historical query records of its mobile device. If the field investigators are indeed investigating the water environment, the interactive display module is also used to generate a sharing prompt on the PC interface of the office staff. The adaptive processing module is also used to automatically perform OCR recognition and content extraction on the screenshot of the currently completed analysis report provided by the office staff after the office staff confirms the sharing. The extracted atmospheric environment analysis text is then pushed to the mobile device of the field investigation staff by the interactive display module.

7. An AI-based method for querying environmental monitoring data, characterized in that, Includes the following steps: S1. Pre-store and manage multi-source heterogeneous data, including real-time and historical monitoring data of water and atmospheric environments, as well as meteorological data, monitoring station data, pollution source data, historical analysis reports, and personnel organizational structure information; S2. When a user logs into the system, the system detects the type of the user's login device, network environment, and current geographical location information to determine the user's current work scenario. If the scenario is determined to be an office scenario, proceed to step S3. If the scenario is determined to be a field investigation scenario, proceed to step S4. S3. Receive screenshot input from the user, perform image recognition and semantic analysis on the screenshot, identify the user's current specific input scenario, and perform differentiated queries and outputs; S4. Automatically obtain the user's current geographical location information, calculate the Euclidean distance between the coordinates and each monitoring station in the system, determine the nearest station within a preset radius, and automatically determine the query items based on the station type: Based on the latest monitoring data retrieved, it is determined whether there are any abnormalities exceeding the standard. If an abnormality is found, an environmental briefing containing the location and the index exceeding the standard will be proactively popped up on the user's mobile device.

8. The AI-based environmental monitoring data query method according to claim 7, characterized in that: Step S3 specifically includes: S301. When chat software features are identified in the screenshot, extract the natural language requirements from the chat history, convert the natural language into query instructions, and generate a conversational text containing conclusions to return to the user after obtaining the data; the natural language requirements include time, location, and metrics. S302. When the data entry interface features are detected in the screenshot, the field name to be filled in is identified, the corresponding value is automatically queried, and structured data is generated. S303. When document editing interface features are detected in the screenshot, the context before and after the cursor is analyzed, and paragraph text is automatically generated.

9. The AI-based environmental monitoring data query method according to claim 8, characterized in that: When determining the query item in step S4, if the nearest station is a water quality monitoring station, the query item is determined to be water environment, and real-time water quality data is automatically retrieved. If the nearest station is an air quality micro-station or is located in an industrial area, and the query item is determined to be atmospheric environment, real-time air quality data will be automatically retrieved.

10. The AI-based environmental monitoring data query method according to claim 9, characterized in that: In step S303, if the scenario is identified as a report writing scenario, a cross-scenario collaborative linkage step S5 is also executed, which specifically includes: S501. Recognize the text content in the screenshot. If the content is determined to belong to the atmospheric environment category, extract the relevant query location or analysis object location. At the same time, analyze whether the text contains the author's subjective analysis logic to confirm the author's analysis status of the location. S502. If subjective analysis content is detected, the system queries the real-time login status and determines whether there are other staff members in the field investigation scenario within the set range of the query location. S503. If there are field investigators, read the personnel organizational structure information and determine whether the field investigators and the current office staff belong to the same department; if so, determine whether their current investigation content is water environment based on their location or historical query records. S504. If it is confirmed that the field investigators are conducting a water environment survey, a sharing prompt will be generated on the office staff interface. After the office staff confirms the sharing, the atmospheric environment analysis text in the screenshot provided by the office staff will be automatically extracted and pushed to the field investigators' mobile devices.

Images