Digital environmental protection information interaction display system

Abstract

The invention discloses a digital environmental protection information interaction display system, which comprises a data acquisition and preprocessing layer, an ecological calculation engine layer, a knowledge graph narrative generation layer and an interaction rendering and feedback layer, and is characterized in that the data acquisition and preprocessing layer is used for acquiring multi-source heterogeneous ecological data and user interaction data; the ecological calculation engine layer calculates a regional ecological health index through a constructed multi-dimensional index fusion model, integrates a geographic information system (GIS) to carry out space-time prediction, and generates an ecological index and future trend early warning data; the knowledge graph narrative generation layer is used for generating an RAG framework by utilizing retrieval enhancement based on an environmental protection knowledge graph, combining the received ecological index with future trend early warning data and user behavior data, and generating controllable narrative content with emotional design; and the interactive rendering and feedback layer is used for visually displaying the controllable narrative content and the real-time ecological data through a dynamic level of detail (LOD) rendering pipeline.

Description

A digital environmental protection information interactive display system Technical Field

[0001] This invention relates to the field of environmental information interaction technology, and in particular to a digital environmental information interaction display system. Background Technology

[0002] A digital environmental information interactive display system, built upon an intelligent platform integrating cutting-edge technologies such as the Internet of Things, big data, cloud computing, and artificial intelligence, aims to transform the environmental information dissemination model from "passive reception" to "active interaction" by real-time collection, dynamic integration, and multi-dimensional visualization of key content including environmental quality data, pollution source monitoring information, ecological restoration progress, public environmental demands, and policy and regulatory interpretations. This improves the efficiency and breadth of environmental information dissemination, enhances public awareness and participation in environmental issues, and provides real-time, comprehensive, and accurate data support for governments to formulate precise environmental governance strategies, enterprises to optimize environmentally friendly production processes, and research institutions to conduct environmental research. Ultimately, it has profound significance in promoting the modernization of environmental governance, driving green and low-carbon development, cultivating public environmental awareness, building a collaborative environmental ecosystem involving government, enterprises, and the public, and achieving a win-win situation for both economic and ecological benefits. It forms a closed-loop management mechanism covering the entire chain from information collection to decision-making and public feedback, effectively addressing the pain points of traditional environmental information interaction such as "information silos," "data barriers," and "lack of participation," and promoting the leapfrog development of environmental protection towards digitalization, intelligence, and public participation.

[0003] In existing technologies, personalized ecological stories are generated through AI and users are guided to explore environmental knowledge through voice interaction. However, relying entirely on large language models to generate stories can easily lead to illusions and spread incorrect environmental knowledge, such as incorrect species habits and fabricated environmental policies. This directly violates the core positioning of the system's scientific monitoring. Therefore, a digital environmental information interactive display system is proposed. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a digital environmentally friendly information interactive display system.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a digital environmental protection information interactive display system, comprising a data acquisition and preprocessing layer, an ecological computing engine layer, a knowledge graph narrative generation layer, and an interactive rendering and feedback layer, wherein: the data acquisition and preprocessing layer is used to acquire multi-source heterogeneous ecological data and user interaction data through an IoT sensor array and a user behavior acquisition module, and to perform cleaning, standardization, and spatiotemporal alignment preprocessing on the ecological data and user interaction data; the ecological computing engine layer is used to receive the preprocessed data, calculate the regional ecological health index through a constructed multi-dimensional index fusion model, and... The system integrates a Geographic Information System (GIS) for spatiotemporal prediction, generating ecological indices and future trend warning data. The knowledge graph narrative generation layer, based on an environmental knowledge graph and utilizing retrieval enhancement to generate a RAG framework, combines the received ecological indices and future trend warning data with user behavior data to generate controllable narrative content with emotional design. The interactive rendering and feedback layer visualizes the controllable narrative content and real-time ecological data through a dynamic level of detail (LOD) rendering pipeline, receives user commands using a multimodal interaction system, and simultaneously feeds user behavior data back to the data acquisition and preprocessing layer.

[0006] The above technical solution further includes: The data acquisition and preprocessing layer specifically includes: an IoT sensor array, which includes air, water quality, noise, and soil monitoring modules, and transmits data via the LoRaWAN protocol; and a user behavior acquisition module, which is integrated into a mobile terminal and used to capture user gestures, cadence, and emotion data in real time via an inertial measurement unit (IMU), a camera, and wearable biosensors.

[0007] Furthermore, in the ecological computing engine layer, the formula for the multi-dimensional index fusion model is: Regional ecological health index = α * air quality sub-index + β * water quality sub-index + γ * biodiversity sub-index + δ * carbon sequestration capacity sub-index, where α, β, γ, and δ are the weights of each item; the biodiversity sub-index is calculated by weighting species richness, Shannon-Wiener index, and endemic species ratio.

[0008] Furthermore, the ecological computing engine layer integrates a GIS spatiotemporal prediction module, which is used to generate continuous environmental data heat maps through the Kriging interpolation algorithm and to perform time series prediction using a Long Short-Term Memory (LSTM) network to output the pollution diffusion trend for a specified future time period.

[0009] Furthermore, in the knowledge graph narrative generation layer, the execution of the retrieval enhancement generation RAG framework includes: retrieving entities and relationships related to the current ecological data and user context from the environmental knowledge graph to form fact triples; using the fact triples as constraints, inputting them into a fine-tuned domain large language model LLM, and generating personalized ecological story content based on a preset narrative template.

[0010] Furthermore, the knowledge graph narrative generation layer receives real-time user emotion data from biosensors and dynamically adjusts the presentation rhythm and associated visual expression intensity of the generated narrative content based on the real-time user emotion data.

[0011] Furthermore, the dynamic LOD rendering pipeline in the interactive rendering and feedback layer adaptively clusters high-density data points based on quadtree spatial indexing, dynamically aggregates or expands data point icons according to the user's perspective distance, and uses GPU instantiation technology and dynamic batch processing technology to efficiently render homogeneous user behavior trajectory points.

[0012] Furthermore, the multimodal interaction system in the interactive rendering and feedback layer includes: a gesture-voice fusion module for combining the Leap Motion gesture recognition device, the mobile phone's built-in IMU, and camera visual recognition; when the gesture recognition signal-to-noise ratio is low, the camera visual recognition performs weighted compensation; and a haptic feedback module for generating haptic feedback signals synchronized with screen visual events according to preset haptic semantic encoding rules.

[0013] Furthermore, the interactive rendering and feedback layer includes an intelligent resource scheduling module, which is used to monitor the real-time rendering load of the system and user interaction scenarios. When complex augmented reality (AR) interactions are detected, the update frequency of the background ecosystem data is automatically reduced.

[0014] Furthermore, through the interactive rendering and feedback layer, the controllable narrative content is visualized and interactively presented in the form of dynamic graphics, interactive installation art, or sound ecological design. The visualization includes: converting environmental monitoring data into particle fluid animation or digital aurora dynamic images; presenting 3D digital images of endangered species through augmented reality (AR) technology; and constructing a virtual zoo ecological education scene.

[0015] This invention offers the following advantages: It designs a hybrid generation framework combining knowledge graph, RAG (Real Language Acquisition), and template constraints. First, the knowledge graph ensures factual accuracy. Then, RAG technology injects correct facts into the generation process. Finally, templates and domain-specific LLM (Limited Language Modeling) fine-tuning control the logic and style of the generated content. This combines the advantages of symbolic AI and connectionist AI, leveraging the powerful natural language generation capabilities of LLM while using symbolic knowledge to constrain its generation boundaries. This controllable generation architecture, specifically designed to ensure scientific accuracy in the field of environmental information interaction, fundamentally solves the illusion problem that large language models may produce in specialized domains. Each generated ecological story is based on authoritative knowledge graphs and real-time monitoring data, ensuring the content's authenticity and credibility. Attached Figure Description

[0016] Figure 1 is a system block diagram of a digital environmental protection information interactive display system proposed in this invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please refer to Figure 1. This invention is a digital environmental protection information interactive display system, comprising a data acquisition and preprocessing layer, an ecological computing engine layer, a knowledge graph narrative generation layer, and an interactive rendering and feedback layer. Specifically: the data acquisition and preprocessing layer acquires multi-source heterogeneous ecological data and user interaction data through an IoT sensor array and a user behavior acquisition module, and performs cleaning, standardization, and spatiotemporal alignment preprocessing on the ecological data and user interaction data; the ecological computing engine layer receives the preprocessed data, calculates the regional ecological health index through a constructed multi-dimensional index fusion model, and integrates a Geographic Information System (GIS) for spatiotemporal prediction, generating ecological index and future trend warning data; the knowledge graph narrative generation layer, based on an environmental protection knowledge graph and utilizing a retrieval-enhanced RAG framework, combines the received ecological index and future trend warning data with user behavior data to generate controllable narrative content with emotional design; the interactive rendering and feedback layer visualizes the controllable narrative content and real-time ecological data through a dynamic level of detail (LOD) rendering pipeline, receives user commands using a multimodal interaction system, and simultaneously feeds user behavior data back to the data acquisition and preprocessing layer.

[0019] In one embodiment, the data acquisition and preprocessing layer specifically includes: an IoT sensor array, which includes air, water quality, noise, and soil monitoring modules and transmits data via the LoRaWAN protocol; and a user behavior acquisition module, which is integrated into a mobile terminal and is used to capture user gestures, cadence, and emotion data in real time through an inertial measurement unit (IMU), a camera, and wearable biosensors.

[0020] In this embodiment: An IoT sensor array: Multiple sensor nodes integrating PM2.5 / NO2 monitoring, pH / dissolved oxygen monitoring, noise level monitoring, and soil moisture monitoring modules were deployed along the park's lakes, woodlands, lawns, and main roads. These nodes transmit data to the park management's edge computing gateway every 5 minutes via the LoRaWAN protocol.

[0021] User behavior collection module: The APP on the tourist's mobile phone serves as the main interaction terminal, calling the phone's built-in IMU to record walking trajectory and cadence; it captures user gestures through the camera (in combination with ARKit / ARCore); and it encourages users to wear smartwatches to obtain heart rate variability (HRV) data through their biosensors to indirectly infer emotional state (such as excitement or calmness).

[0022] Data preprocessing: A lightweight Python Pandas script runs on the edge gateway to perform linear imputation of missing values ​​and outlier detection based on the 3σ criterion on the received sensor data. Simultaneously, the GeoPandas library is used to uniformly convert the WGS-84 coordinates of all data points to the local planar coordinate system of the park area, ensuring precise spatial alignment between visitor locations and ecological data.

[0023] In one embodiment, in the ecological computing engine layer, the formula of the multi-dimensional index fusion model is: Regional ecological health index = α * air quality sub-index + β * water quality sub-index + γ * biodiversity sub-index + δ * carbon sequestration capacity sub-index, where α, β, γ, and δ are the weights of each item; the biodiversity sub-index is calculated by weighting species richness, Shannon-Wiener index, and endemic species ratio, and the biodiversity sub-index = species richness (40% weight) + Shannon-Wiener index (40% weight) + endemic species ratio (20% weight).

[0024] In one embodiment, the ecological computing engine layer integrates a GIS spatiotemporal prediction module, which is used to generate continuous environmental data heat maps through the Kriging interpolation algorithm and to perform time series prediction using a Long Short-Term Memory (LSTM) network to output the pollution diffusion trend for a specified future time period.

[0025] In one embodiment, the execution of the retrieval-enhanced generation (RAG) framework in the knowledge graph narrative generation layer includes: retrieving entities and relationships related to current ecological data and user context from the environmental knowledge graph to form fact triples; inputting the fact triples as constraints into a fine-tuned domain large language model (LLM) to generate personalized ecological story content based on a preset narrative template.

[0026] In one embodiment, the knowledge graph narrative generation layer receives real-time user emotion data from biosensors and dynamically adjusts the presentation rhythm and associated visual expression intensity of the generated narrative content based on the real-time user emotion data.

[0027] In this embodiment: Environmental knowledge graph construction: Based on the global biodiversity information agency database, an ontology centered on the flora and fauna of the park was constructed. For example, the entity "Black-crowned Night Heron" is associated with entities such as "lake water quality," "fish resources," and "habitat protection policies."

[0028] The Retrieval Augmentation (RAG) framework works as follows: When a visitor clicks on an AR tag for a park lake in the app, the system triggers narrative generation.

[0029] Search: The system retrieves entities related to "lake" from the knowledge graph, such as "dissolved oxygen in water: 8.5 mg / L", "night heron", and "recent risk of algal proliferation".

[0030] Generate: Fill these facts into the preset template.

[0031] The preset template and factual data are fed into a finely tuned BERT model (trained on environmental science texts) for language polishing, ultimately generating a natural and fluent voice narration.

[0032] In one embodiment, the dynamic LOD rendering pipeline in the interactive rendering and feedback layer is used to implement the following functions in Unity or Unreal Engine: adaptively clustering high-density data points based on quadtree spatial indexing, dynamically aggregating or expanding data point icons according to the user's view distance; and efficiently rendering homogeneous user behavior trajectory points using GPU instantiation technology and dynamic batch processing technology.

[0033] In one embodiment, the multimodal interaction system in the interactive rendering and feedback layer includes: a gesture-voice fusion module for combining the Leap Motion gesture recognition device, the mobile phone's built-in IMU, and camera visual recognition; when the gesture recognition signal-to-noise ratio is low, the camera visual recognition performs weighted compensation, and the weighted compensation algorithm improves the robustness of gesture recognition in complex environments; and a haptic feedback module for generating haptic feedback signals synchronized with screen visual events according to preset haptic semantic encoding rules.

[0034] In one embodiment, the interactive rendering and feedback layer includes an intelligent resource scheduling module, which is used to monitor the real-time rendering load of the system and user interaction scenarios. When complex augmented reality (AR) interactions are detected, the module automatically reduces the update frequency of the background ecosystem data to prioritize the rendering frame rate.

[0035] In one embodiment, the controllable narrative content is visualized and interactively presented in the form of dynamic graphics, interactive installation art, or sound ecological design through the interactive rendering and feedback layer. The visualization includes: converting environmental monitoring data into particle fluid animation or digital aurora dynamic images; presenting 3D digital images of endangered species through augmented reality (AR) technology; and constructing a virtual zoo ecological education scene.

[0036] In this embodiment: Dynamic LOD rendering pipeline: When visitors view the "Garbage Sorting and Disposal Heatmap" of the entire park in the app, the quadtree spatial indexing begins. When the view zooms out to the entire park, hundreds of disposal points are clustered into a dozen or so area icons with numerical labels. When visitors zoom in to view their own area, the cluster icons dynamically expand, displaying an AR model of a single trash can and its real-time overflow status. All trash can models are rendered once via GPU instantiation, greatly reducing rendering calls.

[0037] Multimodal Interaction System: Gesture-Voice Fusion: A visitor attempts to "grab" a virtual 3D tree model using gestures to view its carbon sequestration data. Under the shade of the tree, the Leap Motion recognition accuracy decreases. The system immediately increases the weight of the phone's camera's visual recognition and compensates for the hand movement trajectory using IMU data, successfully completing the grab command.

[0038] Haptic-visual consistency coding: When a visitor successfully completes a "virtual tree planting" task, the phone not only displays an animation of tree growth but also generates a short, crisp single-pulse vibration (coded as "successful confirmation"); while when the system warns of declining air quality, a red pulsating light effect appears at the edge of the screen, accompanied by a continuous, low vibration (coded as "warning").

[0039] Intelligent resource scheduling and context management: When the system plays an in-depth narrative animation about "bird migration," the context manager switches the system to "narrative mode." At this time, the gesture for quickly swiping the screen to switch scenes is temporarily disabled, replaced by a "swipe to fast forward / rewind narrative" function to prevent accidental interruptions to the immersive experience. Simultaneously, the resource manager detects high GPU load and pauses non-core dynamic simulations of ground vegetation to ensure smooth playback of the narrative animation.

[0040] Visitors' walking activities along the lake (collected by the IMU) contribute "ecological energy," which is fed back to the data collection layer. When the energy is full, it triggers the narrative generation layer to generate a new story: "Your low-carbon travel has protected a habitat for migratory birds," and is awarded to the user in the form of an AR virtual badge in the interactive rendering layer. This process forms a closed loop of "behavior-feedback-incentive," greatly enhancing the user's sense of participation and accomplishment.

[0041] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A digital environmental protection information interactive display system, characterized in that, The system comprises a data acquisition and preprocessing layer, an ecological computing engine layer, a knowledge graph narrative generation layer, and an interactive rendering and feedback layer. Specifically: the data acquisition and preprocessing layer acquires multi-source heterogeneous ecological data and user interaction data through an IoT sensor array and user behavior acquisition module, and performs cleaning, standardization, and spatiotemporal alignment preprocessing on the ecological data and user interaction data; the ecological computing engine layer receives the preprocessed data, calculates the regional ecological health index through a constructed multi-dimensional index fusion model, and integrates a Geographic Information System (GIS) for spatiotemporal prediction, generating ecological index and future trend warning data; the knowledge graph narrative generation layer, based on an environmental knowledge graph and utilizing a retrieval-enhanced RAG framework, combines the received ecological index and future trend warning data with user behavior data to generate controllable narrative content with emotional design; and the interactive rendering and feedback layer visualizes the controllable narrative content and real-time ecological data through a dynamic level of detail (LOD) rendering pipeline, receives user commands using a multimodal interaction system, and simultaneously feeds user behavior data back to the data acquisition and preprocessing layer.

2. The digital environmental protection information interactive display system according to claim 1, characterized in that, The data acquisition and preprocessing layer specifically includes: an IoT sensor array, which includes air, water quality, noise, and soil monitoring modules, and transmits data via the LoRaWAN protocol; and a user behavior acquisition module, which is integrated into the mobile terminal and is used to capture user gestures, cadence, and emotion data in real time through an inertial measurement unit (IMU), a camera, and wearable biosensors.

3. The digital environmental protection information interactive display system according to claim 1, characterized in that, In the ecological computing engine layer, the formula of the multi-dimensional index fusion model is: Regional ecological health index = α*air quality sub-index + β*water quality sub-index + γ*biodiversity sub-index + δ*carbon sequestration capacity sub-index, where α, β, γ and δ are the weights of each item; the biodiversity sub-index is calculated by weighting species richness, Shannon-Wiener index and endemic species ratio.

4. The digital environmental protection information interactive display system according to claim 3, characterized in that, The ecological computing engine layer integrates a GIS spatiotemporal prediction module, which is used to generate continuous environmental data heat maps through the Kriging interpolation algorithm and to perform time series prediction using a Long Short-Term Memory (LSTM) network to output the pollution diffusion trend for a specified future time period.

5. The digital environmental protection information interactive display system according to claim 1, characterized in that, In the knowledge graph narrative generation layer, the execution of the retrieval enhancement generation RAG framework includes: retrieving entities and relationships related to the current ecological data and user context from the environmental knowledge graph to form fact triples; using the fact triples as constraints, inputting them into a fine-tuned domain large language model LLM, and generating personalized ecological story content based on a preset narrative template.

6. The digital environmental protection information interactive display system according to claim 5, characterized in that, The knowledge graph narrative generation layer receives real-time user emotion data from biosensors and dynamically adjusts the presentation rhythm and associated visual expression intensity of the generated narrative content based on the real-time user emotion data.

7. The digital environmental protection information interactive display system according to claim 1, characterized in that, The dynamic LOD rendering pipeline in the interactive rendering and feedback layer adaptively clusters high-density data points based on quadtree spatial indexing, dynamically aggregates or expands data point icons according to the user's perspective distance, and uses GPU instantiation technology and dynamic batch processing technology to efficiently render homogeneous user behavior trajectory points.

8. The digital environmental protection information interactive display system according to claim 1, characterized in that, The multimodal interaction system in the interactive rendering and feedback layer includes: a gesture-voice fusion module for combining the Leap Motion gesture recognition device, the mobile phone's built-in IMU, and camera visual recognition; when the gesture recognition signal-to-noise ratio is low, the camera visual recognition performs weighted compensation; and a haptic feedback module for generating haptic feedback signals synchronized with screen visual events according to preset haptic semantic encoding rules.

9. A digital environmental protection information interactive display system according to claim 1, characterized in that, The interactive rendering and feedback layer includes an intelligent resource scheduling module, which monitors the system's real-time rendering load and user interaction scenarios. When complex augmented reality (AR) interactions are detected, the module automatically reduces the update frequency of the background ecosystem data.

10. A digital environmental protection information interactive display system according to claim 1, characterized in that, Through the interactive rendering and feedback layer, the controllable narrative content is visualized and interactively presented in the form of dynamic graphics, interactive installation art, or sound ecological design. The visualization presentation includes: converting environmental monitoring data into particle fluid animation or digital aurora dynamic images; presenting 3D digital images of endangered species through augmented reality (AR) technology; and constructing a virtual zoo ecological education scene.

Images