Hand-drawn style map generation system and method based on artificial intelligence and geographic information

Abstract

The invention discloses a hand-drawn style map generation system and method based on artificial intelligence and geographic information, and belongs to the technical field of map data processing.The method comprises the steps that geographic area information input by a user is obtained, and geographic element data of an area is obtained from geographic information service; the geographic elements are converted into a semantic map with element types distinguished by different colors, a control chart is generated based on the semantic map, a hand-drawn style map is generated under the constraint of the control chart by using a pre-trained deep learning image model, and finally a result is returned and displayed to a user; the system comprises a front-end interaction module, a rear-end processing module, a GIS data acquisition and processing module, a semantic map generation module, an AI image generation module and a result return and display module. The problems that a traditional digital map lacks artistry, the manual hand drawing efficiency is low, and a general AI generation model is geographically distorted are solved.

Description

Technical Field

[0001] This invention relates to the field of map data processing technology, and more specifically, to a system and method for generating hand-drawn style maps based on artificial intelligence and geographic information. Background Technology

[0002] Traditional digital map services, such as Tianditu, Google Maps, OpenStreetMap, Gaode Maps, and Baidu Maps, are built on standardized Geographic Information System (GIS) technology. They accurately and efficiently display geographic information and are widely used in navigation, location queries, and spatial analysis. However, these maps are highly uniform in visual presentation and functional, lacking individuality and artistic appeal, making it difficult to meet the diverse and emotional needs of tourism guides, cultural displays, and commercial promotions for map visual styles. Currently, obtaining artistically styled maps mainly relies on manual hand-drawing. Designers or artists manually draw maps based on actual geographical layouts, creating map works with unique aesthetic value. However, this method has significant drawbacks: long production cycles, high labor costs, strong dependence on the creator's professional skills, difficulty in rapid modification or batch generation, and a lack of flexibility and scalability. In recent years, with the development of deep learning technology, general-purpose artificial intelligence image generation models such as StableDiffusion, Midjourney, and DALL-E have emerged, capable of generating high-quality artistic images based on text descriptions. These models have serious shortcomings in generating map-like images. The generated results lack accuracy in geographical structure, and the road networks, building outlines, water features, etc., are seriously inconsistent with the real world, making them unusable as maps with practical value.

[0003] In summary, traditional digital maps lack artistic expression, hand-drawn maps are inefficient, and general AI generation models cannot guarantee geographical accuracy. Currently, there is no technical solution that can automatically and efficiently generate maps with a hand-drawn artistic style while ensuring the accuracy of geographical information. Summary of the Invention

[0004] In response to the problems in related technologies, this invention proposes a hand-drawn style map generation system and method based on artificial intelligence and geographic information to overcome the aforementioned technical problems existing in the existing related technologies.

[0005] Therefore, the specific technical solution adopted by the present invention is as follows:

[0006] A method for generating hand-drawn style maps based on artificial intelligence and geographic information includes the following steps:

[0007] Obtain the geographic region selection information input by the user;

[0008] Geographic element data is obtained from geographic information data services based on the geographic region selection information.

[0009] The geographic feature data is converted into a semantic map, and different geographic features are assigned different preset colors;

[0010] Based on the semantic map, a control map is generated, and a hand-drawn style map image is generated under the constraints of the control map using a pre-trained deep learning image model.

[0011] The generated hand-drawn style map image is returned and displayed to the user.

[0012] By constructing an automated conversion from geographic information data to semantic maps and then to control charts, and using a controllable artificial intelligence generation model to perform stylized rendering under the strong constraints of control charts, high-quality hand-drawn artistic map images are generated efficiently and automatically while strictly maintaining the accuracy of the real-world geographic spatial layout. This significantly reduces the production threshold and application cost of professional artistic maps, allowing non-professional users to quickly obtain personalized style maps that combine practical navigation value with visual aesthetic value through simple operations.

[0013] Furthermore, the step of generating a control map based on the semantic map includes performing edge detection on the semantic map, generating an edge map as a control map, and using this control map as a rigid constraint to apply to the deep learning image generation model to ensure the accuracy of the geospatial layout of the model output.

[0014] Furthermore, the step of generating a hand-drawn style map image includes inputting the control chart, artistic text prompts, and optional reverse text prompts into the deep learning image model.

[0015] Furthermore, the text prompts include natural language statements describing the target art style, and the reverse prompts include statements excluding unwanted styles to optimize the generated effect.

[0016] Furthermore, the method for generating hand-drawn style maps based on artificial intelligence and geographic information also includes the following steps: receiving POI information input by the user, and adjusting the hand-drawn style map image to include the POI information. The front-end page identifies the coordinate points and content type, submits an adjustment request to the back-end, and the back-end combines the previously generated image with the newly added POI information to generate and adjust the final hand-drawn map image with an artistic style, including POI information.

[0017] A hand-drawn style map generation system based on artificial intelligence and geographic information, the system includes:

[0018] The front-end interaction module provides a web-based user interface, receives the user's input of geographical region selection information, and submits a generation request;

[0019] The backend processing module receives requests from the frontend interaction module and coordinates the map generation process.

[0020] The GIS data acquisition and processing module acquires geographic element data from geographic information data services based on geographic region selection information.

[0021] The semantic map generation module converts geographic feature data into semantic maps, with different geographic features assigned different preset colors.

[0022] The AI ​​image generation module generates control maps based on semantic maps and uses a pre-trained deep learning image model to generate hand-drawn style map images under the constraints of the control maps.

[0023] The results return and display module returns and displays the generated hand-drawn style map image to the user.

[0024] By constructing a fully automated map generation system that integrates front-end user interaction with back-end AI generation, the system deeply blends the accuracy of geographic information data with the creativity of AI art. This enables the rapid generation of maps that strictly adhere to real geographic coordinates and feature distribution while possessing a high-quality hand-drawn artistic style, all without human intervention. The system uses a front-end interaction module as its entry point to effectively capture user intent. Through unified scheduling by the back-end processing module, it collaborates with the GIS data acquisition and processing module and the semantic map generation module to transform raw geographic data into a structured semantic map. Then, the AI ​​image generation module completes artistic rendering under the strong constraints of a control chart. Finally, the results are presented intuitively and interactively through the results return and display module.

[0025] Furthermore, the AI ​​image generation module includes:

[0026] The control chart preprocessing unit performs edge detection on the semantic map and generates a control chart.

[0027] The model inference unit inputs the control chart, artistic text prompts, and optional reverse text prompts into the deep learning image model;

[0028] The image generation unit uses a deep learning image model to generate hand-drawn style map images under the strong constraints of a control chart.

[0029] The control chart preprocessing unit extracts precise edge features from the semantic map to generate a control chart that fully preserves the spatial layout of geographic elements. Then, the model inference unit organically integrates the structural constraints of the control chart with the semantic guidance of artistic style text prompts. Finally, the image generation unit achieves a unified output of geographic accuracy and artistic expression.

[0030] Furthermore, the hand-drawn style map generation system based on artificial intelligence and geographic information also includes a POI adjustment and optimization module, which receives POI information input by the user and adjusts the hand-drawn style map image to include the POI information.

[0031] Furthermore, the GIS data acquisition and processing module is also used to simplify the acquired geographic elements to suit the needs of hand-drawn style expression.

[0032] Furthermore, geographic feature data includes at least one of roads, buildings, water bodies, green spaces, and boundaries.

[0033] The beneficial effects of this invention are as follows:

[0034] 1. In practical use, this invention constructs an automated system from front-end interaction to back-end AI generation, generating hand-drawn style maps that combine accuracy, artistry, and interactivity. The various modules of the system work together to complete the entire process from geographic data acquisition, semantic map construction, control chart generation to stylized rendering without human intervention. Users do not need to have GIS professional knowledge or hand-drawn art skills; they only need to select an area on the map and choose a style to obtain a high-quality stylized map.

[0035] 2. By introducing a control chart derived from a semantic map as a strong constraint on the AI ​​generation model, the generated map maintains consistency with the real world in the layout of key geographical elements such as road directions, building outlines, and water features. This solves the structural distortion problem in map generation. The system provides a POI adjustment and optimization module, allowing users to add, modify, or highlight key point information on the generated map and perform local regeneration based on this, giving the artistic map both navigation and spatial reference practical value.

[0036] 3. The system supports controlling the artistic style of the generated map through natural language text prompts. Users can freely specify various visual styles such as watercolor, sketch, and cartoon, and can combine them with reverse prompts to exclude unwanted visual elements. This transforms the traditional map-making process that relies on professional designers and artists into a task that can be completed through a simple web interface. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 This is a flowchart of a hand-drawn style map generation method based on artificial intelligence and geographic information according to an embodiment of the present invention. Detailed Implementation

[0039] 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.

[0040] Example 1

[0041] According to embodiments of the present invention, a method for generating hand-drawn style maps based on artificial intelligence and geographic information is provided. The method includes: acquiring geographic region selection information input by a user; acquiring geographic element data from a geographic information data service based on the geographic region selection information; converting the geographic element data into a semantic map, assigning different preset colors to different geographic elements; performing edge detection on the semantic map to generate an edge map as a control map; and using a pre-trained deep learning image model, under the constraints of the control map, inputting the control map, artistic style text prompts, and optional reverse text prompts into the deep learning image model to generate a hand-drawn style map image. The text prompts include natural language statements describing the target artistic style, and the reverse prompts include statements excluding unwanted styles to optimize the generation effect. In addition to generating the hand-drawn style map image, the method can also receive POI information input by the user, adjust the hand-drawn style map image to include the POI information, and return and display the generated hand-drawn style map image to the user.

[0042] The aforementioned technical solution, through ControlNet technology and using semantic maps as precise constraints, ensures that the generated map remains consistent with the real world in terms of road networks, building outlines, and water morphology. Users can freely select any geographical area and change the artistic style of the generated map by adjusting text prompts, such as watercolor, sketch, or cartoon, offering extremely high flexibility and customization potential. Users do not need any professional GIS knowledge or artistic drawing skills; they can obtain professional-level hand-drawn style maps simply by selecting areas on the map, greatly lowering the barrier to entry. It can be widely used in various fields such as smart tourism, cultural guidance, real estate project display, personalized decorative map production, and game scene design, possessing enormous commercial and cultural value.

[0043] Example 2

[0044] According to an embodiment of the present invention, a hand-drawn style map generation system based on artificial intelligence and geographic information is provided, comprising:

[0045] The front-end interaction module provides a web-based user interface, receives the user's input of geographical region selection information, and submits a generation request;

[0046] The backend processing module receives requests from the frontend interaction module and coordinates the map generation process.

[0047] The GIS data acquisition and processing module obtains geographic element data from geographic information data services based on geographic region selection information. The geographic elements obtained from the GIS database are first converted into a semantic map, and then a control map is generated from the semantic map. This control map is used as a rigid constraint and applied to a deep learning image generation model to ensure the accuracy of the geospatial layout of the model output.

[0048] The semantic map generation module converts geographic feature data into semantic maps, with different geographic features assigned different preset colors.

[0049] The AI ​​image generation module generates control maps based on semantic maps and uses a pre-trained deep learning image model to generate hand-drawn style map images under the constraints of the control maps.

[0050] The results return and display module returns and displays the generated hand-drawn style map image to the user.

[0051] The AI ​​image generation module includes:

[0052] The control chart preprocessing unit performs edge detection on the semantic map and generates a control chart.

[0053] The model inference unit inputs control charts, artistic text prompts, and optional reverse text prompts into the deep learning image model;

[0054] The image generation unit uses a deep learning image model to generate hand-drawn style map images under the strong constraints of a control chart.

[0055] This generation system implements a decoupled control of the AI ​​generation process. In terms of structure, it applies precise, pixel-level hard constraints to the geometric layout of the generated image through control maps generated from GIS data to ensure the accuracy of geographic information. In terms of style, it provides flexible guidance on the artistic style, color, texture and other aesthetic features of the generated image through natural language text prompts.

[0056] The hand-drawn style map generation system based on artificial intelligence and geographic information also includes a POI adjustment and optimization module, which receives POI information input by the user and adjusts the hand-drawn style map image to include POI information.

[0057] The GIS data acquisition and processing module is also used to simplify the acquired geographic features to meet the needs of hand-drawn style expression. It establishes a technical path to convert structured geographic information system (GIS) vector data into rasterized visual instructions that can be understood and executed by artificial intelligence image generation models. The geographic feature data includes at least one of roads, buildings, water bodies, green spaces and boundaries.

[0058] In summary, by utilizing the above-mentioned technical solutions of this invention, in practical use, this invention constructs an automated system from data acquisition to final generation through the collaborative work of modules such as front-end interaction, back-end processing, GIS data acquisition, semantic map generation, AI image generation, and result display. It converts GIS vector data into semantic maps and further generates control charts as strong constraints for the AI ​​image generation model, ensuring that the generated map maintains consistency with the real world in the layout of key geographical elements such as road directions, building outlines, and water morphology. At the same time, it controls the artistic style through text prompts, solving the structural distortion problem existing in general text-based map models in map generation, and achieving a perfect combination of practicality and artistry.

[0059] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for generating hand-drawn style maps based on artificial intelligence and geographic information, characterized in that: Includes the following steps: Obtain the geographic region selection information input by the user; Geographic element data is obtained from geographic information data services based on the geographic region selection information. The geographic feature data is converted into a semantic map, and different geographic features are assigned different preset colors; Based on the semantic map, a control map is generated, and a hand-drawn style map image is generated under the constraints of the control map using a pre-trained deep learning image model. The generated hand-drawn style map image is returned and displayed to the user.

2. The method for generating hand-drawn style maps based on artificial intelligence and geographic information according to claim 1, characterized in that, The step of generating a control map based on the semantic map includes performing edge detection on the semantic map and generating an edge map as a control map.

3. The method for generating hand-drawn style maps based on artificial intelligence and geographic information according to claim 1, characterized in that, The steps for generating hand-drawn style map images include inputting the control chart, artistic text hints, and optional reverse text hints into the deep learning image model.

4. The method for generating hand-drawn style maps based on artificial intelligence and geographic information as described in claim 3, characterized in that, The text prompts include natural language statements describing the target art style, and the reverse prompts include statements excluding unwanted styles to optimize the generated effect.

5. The method for generating hand-drawn style maps based on artificial intelligence and geographic information according to claim 1, characterized in that, It also includes the steps of: receiving POI information input by the user and adjusting the hand-drawn style map image to include the POI information.

6. A hand-drawn style map generation system based on artificial intelligence and geographic information, characterized in that, The system, applicable to the method of any one of claims 1-5, comprises: The front-end interaction module provides a web-based user interface, receives the user's input of geographical region selection information, and submits a generation request; The backend processing module receives requests from the frontend interaction module and coordinates the map generation process. The GIS data acquisition and processing module acquires geographic element data from geographic information data services based on the geographic region selection information. The semantic map generation module converts the geographic element data into a semantic map, with different geographic elements being assigned different preset colors. The AI ​​image generation module generates a control map based on the semantic map, and uses a pre-trained deep learning image model to generate a hand-drawn style map image under the constraints of the control map. The results return and display module returns and displays the generated hand-drawn style map image to the user.

7. The hand-drawn style map generation system based on artificial intelligence and geographic information according to claim 6, characterized in that, The AI ​​image generation module includes: The control graph preprocessing unit performs edge detection on the semantic map and generates a control graph. The model inference unit inputs the control chart, artistic text prompts, and optional reverse text prompts into the deep learning image model; The image generation unit uses a deep learning image model to generate the hand-drawn style map image under the strong constraints of the control chart.

8. The hand-drawn style map generation system based on artificial intelligence and geographic information according to claim 6, characterized in that, It also includes a POI adjustment and optimization module, which receives POI information input by the user and adjusts the hand-drawn style map image to include the POI information.

9. A hand-drawn style map generation system based on artificial intelligence and geographic information according to claim 6, characterized in that, The GIS data acquisition and processing module is also used to simplify the acquired geographic elements to suit the needs of hand-drawn style expression.

10. A hand-drawn style map generation system based on artificial intelligence and geographic information according to claim 6, characterized in that, The geographic feature data includes at least one of roads, buildings, water bodies, green spaces, and boundaries.

Images