A method for integrated management and intelligent rendering of multi-source geological map data

Abstract

This invention belongs to the field of artificial intelligence technology and discloses a method for integrated management and intelligent rendering of multi-source geological map data. This invention utilizes a multi-hypothesis generation method to proactively generate multiple competing but all reasonable geological interpretation schemes for areas with uncertain data, leveraging generative AI and geological knowledge graphs. Subsequently, through an interactive interpretive rendering mode, when a user queries geological elements in any scheme, the system can dynamically redraw the map, highlighting key evidence, darkening irrelevant backgrounds, and linking knowledge graph rules to visually reveal the AI's decision-making basis. This invention elevates AI mapping from an automated tool to an interactive and interpretable scientific verification partner, improving the scientific rigor and reliability of geological decision-making.

Description

Technical Field

[0001] This invention belongs to the field of artificial intelligence technology, specifically relating to a method for integrated management and intelligent rendering of multi-source geological map data. Background Technology

[0002] With the development of artificial intelligence technology, AI-assisted geological mapping has become a research hotspot. Existing technologies, such as AI mapping functions and research systems in commercial GIS software, can automatically generate geological maps based on user instructions or style references. These technologies analyze multi-source data (remote sensing, geophysical and geochemical exploration, etc.) through deep learning models, achieving, to a certain extent, the automatic generation and rendering of geological map content and improving mapping efficiency.

[0003] However, these first-generation AI mapping technologies contain a deep contradiction with geological science practice:

[0004] 1. The decision-making logic behind AI-generated geological boundaries or lithological determinations is opaque to users. Users cannot know what data the AI ​​relied on or what rules it followed to reach its conclusions. This black-box nature makes it difficult for geologists to trust, verify, and correct AI results, severely limiting its application in serious scientific research and high-risk resource exploration decision-making.

[0005] 2. Geological interpretation is essentially a multifaceted scientific reasoning process based on incomplete information. For the same region, different theoretical models or different emphases on the data often lead to multiple reasonable interpretation schemes (i.e., multiple working hypotheses). However, existing AI mapping technology aims to output a single, optimal, and deterministic result. This not only eliminates the multiple possibilities in scientific exploration but also fails to provide geologists with a platform to compare and verify different interpretation schemes. Summary of the Invention

[0006] The present invention aims to at least partially solve the aforementioned technical problems. Therefore, the objective of this invention is to provide a method for integrated management and intelligent rendering of multi-source geological map data.

[0007] The technical solution adopted in this invention is as follows:

[0008] A method for integrated management and intelligent rendering of multi-source geological map data is proposed. Its core lies in realizing intelligent geological map generation and verification through a series of steps, including (A) integrated data management; (B) construction and application of geological knowledge graphs; (C) multiple hypothesis generation; (D) interactive interpretive rendering; and (E) rendering execution and output.

[0009] In step (C) of multiple hypothesis generation, the system first identifies target areas with interpretative ambiguity in multi-source geological data. This interpretative ambiguity can be automatically identified by analyzing areas with data conflicts (e.g., contradictions between geophysical and remote sensing interpretations), missing data (e.g., sparse boreholes), or a data signal-to-noise ratio below a preset threshold. Subsequently, using a generative AI model and based on different subsets of interpretation rules in the geological knowledge graph, at least two sets of hypothetical geological primitives are generated for this area. In a preferred embodiment, these different subsets of interpretation rules can correspond to different advanced geological models. For example, a tensional basin model can be applied as a constraint when generating the first set of hypotheses, while a compressional thrust-nappe model can be applied as a constraint when generating the second set of hypotheses, thus ensuring that the generated hypotheses have a clear theoretical background. To assist user decision-making, this method preferably includes a scoring and ranking step after hypothesis generation. The system automatically calculates a comprehensive score for each hypothetical primitive set, which comprehensively considers the degree of fit between the hypothesis and all source data, as well as the confidence level of the knowledge graph rules used in the generation process. The hypothesis with the highest score is then recommended to the user first.

[0010] In step (D), interactive interpretive rendering, when a user queries a specific geological feature in a hypothetical map, the system enters interpretive rendering mode. The specific geological feature can be a geological boundary, fault, or rock mass boundary, while the source data includes borehole data points, geophysical survey lines, geochemical sampling points, or remote sensing imagery. The system dynamically reconstructs the map's visual appearance, specifically by increasing the brightness and saturation of the source data layers that play a decisive role in generating the specific feature (such as several key seismic profiles), or adding a flashing effect to attract the user's attention; simultaneously, all other irrelevant data layers and features are semi-transparent or grayscaled to reduce visual interference. Furthermore, while visually highlighting, the system clearly displays the interpretation rules in the knowledge graph triggered by this judgment (e.g., "IF remote sensing imagery shows linear characteristics AND gravity gradient zone matches THEN, determined as a fault structure") on the interface (e.g., through pop-ups or sidebars) using text or graphical flowcharts. To achieve deeper interaction, the interpretability rendering mode of this invention preferably also supports user adjustment of interpretability parameters. For example, users can drag a "confidence threshold" slider, and the system will update the highlighted evidence data in real time according to the new threshold, allowing users to intuitively see which evidence is robust and which is secondary under different stringent standards. Users can also adjust the weight of specific data sources (such as geophysical or geochemical exploration) to observe how stable the interpretation results are when more emphasis is placed on a certain type of data.

[0011] To facilitate comparison of different hypotheses, this method preferably includes a differential contrast rendering step. When multiple sets of hypothetical primitives are displayed on the screen simultaneously, the system can automatically calculate the spatial differences between them and render these contradictory primitives (such as faults with different locations or attributes) with special colors or line types, allowing users to easily locate the focus of the dispute.

[0012] To enable the system to learn and evolve, this method also includes a feedback learning step. The system records the hypothetical geological primitive set that the user ultimately adopts or edits to confirm after comparison and verification, and uses it as a high-quality positive sample. These samples can be used for reinforcement learning or fine-tuning of the generative AI model, enabling the model to generate hypotheses that are more consistent with expert experience and of higher quality in future tasks, and to optimize its recommendation ranking.

[0013] The beneficial effects of this invention are as follows:

[0014] This invention, through interactive interpretive rendering, for the first time realizes the visualization, traceability, and auditability of the AI ​​geological mapping decision-making process, enhancing the credibility of the results; through multiple hypothesis generation, it transforms AI from an answer provider into a thought-provoking agent, providing a platform for scientific research to evaluate multiple possibilities. Attached Figure Description

[0015] Figure 1 This is an overall flowchart of the method provided in the embodiments of the present invention.

[0016] Figure 2 This is a schematic diagram illustrating the working principle of the multiple hypothesis generation module in this embodiment of the invention.

[0017] Figure 3 This is a schematic diagram illustrating the effect of the interactive interpretive rendering mode in an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0019] It should be understood that, and also noted, in the embodiments, the functions / actions may appear in a different order than those shown in the figures. For example, depending on the functions / actions involved, they may actually be performed substantially concurrently, or sometimes the two figures shown consecutively may be performed in reverse order.

[0020] like Figure 1 and Figure 2As shown, the present invention provides a method for integrated management and intelligent rendering of multi-source geological map data, implemented within an integrated intelligent GIS platform, specifically including:

[0021] Step S101: Integrated Data Management

[0022] Through an integrated management interface, seamless access and unified visual management of multi-source heterogeneous geological data, such as MapGIS format data, Shapefile, remote sensing imagery, geophysical and geochemical grid data, borehole database, etc., can be achieved.

[0023] Step S102: Constructing and applying geological knowledge maps

[0024] Construct a rich geological knowledge map. This map not only includes geological entities (such as granites and Permian strata) and geological relationships (such as intrusions and unconformities), but more importantly, it also contains a large number of interpretation rules. These rules are categorized into different subsets of interpretation rules, such as the {compressional structure interpretation rule set}, which includes rules such as "recurrence of strata - thrust faults" and "old core, new wing - anticlines"; and the {extensional structure interpretation rule set}, which includes rules such as "absence of strata - normal faults" and "graben-horst combination".

[0025] Step S103: Multiple Hypothesis Generation

[0026] 1. Identifying Ambiguous Areas: The system automatically scans all incoming data and discovers that in a certain target area, the NE-trending linear structural zone interpreted from the remote sensing imagery forms an angle of approximately 30 degrees with the gradient zone of the regional gravity anomaly. Furthermore, there is only one deep borehole in this area, and the data is insufficient to uniquely determine the structural nature. The system marks this area as an interpretation ambiguity zone.

[0027] 2. Generating Hypothesis A: The system initiates a multi-hypothesis generation engine, invoking a pre-trained generative AI model. The engine first loads the {compressional tectonic interpretation rule set} from the knowledge graph as strong constraints. The model, after analyzing all data in the region and adhering to this rule set, generates the geological primitive set for "Hypothesis A". In this hypothesis, the NE-trending linear tectonic structure is interpreted as a series of thrust faults and associated tight folds.

[0028] 3. Generating Hypothesis B: Subsequently, the engine clears the constraints and reloads the {strike-slip-pull-apart structure interpretation rule set}. The model interprets the same data again under the new rule constraints, generating the geological primitive set of "Hypothesis B". In this hypothesis, the NE-trending structure is interpreted as a large right-lateral strike-slip fault.

[0029] 4. Scoring and Ranking: After generating two sets of hypotheses, the system automatically scores them. The system evaluation found that "Hypothesis A" highly matches the stratigraphic repetition phenomenon revealed by the geophysical profile, with a comprehensive score of 9.2. "Hypothesis B," however, slightly conflicts with the local sequence stratigraphy revealed by the borehole, with a comprehensive score of 8.5. Therefore, the system marks "Hypothesis A" as the preferred recommendation on the interface.

[0030] 5. Differentiated Rendering: When two hypothetical layers are displayed simultaneously on the interface, the system automatically performs spatial overlay analysis on them, identifying fault attributes and locations as the main differences. Therefore, the system renders the thrust fault in "Hypothesis A" as a solid red line with teeth, while rendering the strike-slip fault in "Hypothesis B" as a solid blue line with arrows, making the differences immediately apparent.

[0031] Step S104: Interactive Explanatory Rendering

[0032] The user became interested in a major thrust fault, 301, in "Hypothesis A" and wanted to know why the AI ​​drew it that way.

[0033] 1. Trigger command: The user right-clicks on the fault line and selects "Interpret this element" from the pop-up menu.

[0034] 2. Entering Interpretation Mode: The system immediately enters the interpretation mode as follows. Figure 3 The interactive, interpretive rendering mode shown.

[0035] Evidence highlighted: On the map, the two geophysical profiles 302a and 302b, which play the most crucial role in identifying this fault (and whose phase axes show obvious truncation and uplift), begin to highlight and slowly flash. A borehole 303, located on the hanging wall of the fault and revealing older strata overlying newer strata, also changes its symbol from a regular dot to a striking star. These are the source data that play a decisive role in the identification.

[0036] Darken the background: All other irrelevant layers on the map, such as the surface vegetation cover map and other unqueried elements, are rendered as a semi-transparent gray.

[0037] Rule display: An interactive information box 304 pops up on the right side of the screen, showing the triggered interpretation rules in the form of a flowchart: "Input: Geophysical profile data - Identify phase axis interruption and uplift - Trigger 'Reverse structural features' - Input: Borehole data - Identify stratigraphic inversion - Trigger 'Reverse evidence' - Comprehensive judgment - Output: Thrust fault (confidence level: 0.85)".

[0038] 3. Interactive Parameter Adjustment: The user drags a "Confidence Threshold" slider in information box 304, increasing it from 0.8 to 0.9. The system responds immediately in real time: on the map, the previously highlighted geophysical profile 302b loses its highlight because its individual contribution of confidence is less than 0.9, leaving only profile 302a and the key borehole 303 still highlighted. This interactive process instantly makes it clear to the user that the borehole and profile 302a are the core, irrefutable evidence for this judgment.

[0039] Step S105: Rendering Execution and Output

[0040] After repeatedly and interactively debating the two hypotheses, the user concluded that the chain of evidence for "Hypothesis A" was stronger. He then made minor manual adjustments to "Hypothesis A" and clicked to confirm the solution.

[0041] 1. Feedback Learning: The system records the geological primitive set finally confirmed by the user, marks it as a high-quality positive sample, packages it with the original input data, and sends it to the model training module for future fine-tuning of the generative AI model.

[0042] 2. Final output: The system applies standard national or industry rendering schemes to perform final symbolization, finishing and map frame configuration on the confirmed primitive set, generating a final geological map that can be printed or published.

[0043] This invention is not limited to the above-described optional embodiments. Anyone can derive other various forms of products under the guidance of this invention. However, regardless of any changes made in their shape or structure, any technical solution that falls within the scope of the claims of this invention shall be protected by this invention.

Claims

1. A geological map multi-source data integrated management and intelligent rendering method, characterized in that, The method comprises the following steps: (A) Data integration management: accessing and uniformly managing multi-source geological data through an integrated management interface; (B) Building and applying a geological knowledge graph: building or loading a geological knowledge graph containing geological entities, geological relationships, interpretation rules, and rendering constraints; (C) Multi-hypothesis generation: identifying target areas with interpretation ambiguity in the multi-source geological data, using a generative AI model, and generating at least two sets of mutually independent, geologically reasonable hypothetical geological map elements for the target areas based on different subsets of interpretation rules in the geological knowledge graph; (D) Interactive interpretive rendering: receiving user query instructions for a specific geological map element in a hypothetical geological map element set, and the system enters an interpretive rendering mode, dynamically adjusting the visual representation of the map to highlight the source data and / or interpretation rules in the geological knowledge graph that played a decisive role in generating the specific geological map element; (E) Rendering execution and output: generating a final geological map based on a set of geological map elements selected or edited by the user.

2. The method of claim 1, wherein, In step (C), the interpretation ambiguity is identified by analyzing areas with data conflicts, data missing, or a signal-to-noise ratio below a preset threshold.

3. The method of claim 1, wherein, In step (C), the generative AI model is applied with different geological structure models or ore-forming models as high-level constraints when generating different sets of hypothetical geological map elements.

4. The method of claim 1, wherein, The method further comprises, after step (C): scoring and ranking the generated at least two sets of hypothetical geological map elements, wherein the score is calculated based on the goodness of fit of the geological map element set to the multi-source geological data and the confidence of the interpretation rules on which the map element set is based.

5. The method of claim 1, wherein, In step (D), the highlighting includes: increasing the brightness, saturation, or adding a flashing effect to the source data layer; and semi-translucency or greying out other irrelevant data layers.

6. The method of claim 1, wherein, In step (D), while dynamically adjusting the visual representation of the map, the triggered interpretation rules are displayed on the interface in text or graphical form.

7. The method according to any one of claims 1, 5 or 6, characterized in that, The interpretive rendering mode in step (D) further includes: receiving user instructions to adjust explainability parameters, and updating the highlighted source data and / or displayed interpretation rules in real time according to the adjusted parameters; the explainability parameters include a confidence threshold or the weight of a specific data source.

8. The method of claim 1, wherein, It further comprises a differential contrast rendering step: when displaying at least two sets of hypothetical geological map elements simultaneously, special symbolization processing is performed on the geological map elements that differ between them to highlight the spatial differences between different hypotheses.

9. The method of claim 1, wherein, The specific geological map element is a geological boundary, fault, or rock body boundary; the source data includes drill data points, geophysical survey lines, geochemical sampling points, or remote sensing images.

10. The method of claim 1, wherein, It further comprises a feedback learning step: recording the user's final selected set of hypothetical geological map elements and using them as positive samples for reinforcement learning or fine-tuning of the generative AI model to optimize the ranking and quality of future hypothesis generation.

Images