Geo-InkGAN: An Adaptive Generative Framework for Topographically Faithful Ink-Wash Style Transfer in Terrain Mapping
Open Access
- 21 July 2026
- journal article
- research article
- Published by MDPI AG in ISPRS International Journal of Geo-Information
- Vol. 15 (7) , 335
Abstract
The compelling visualization of Digital Elevation Models (DEMs) constitutes a vital intersection between Geographic Information Science (GIS) and the digital humanities. Nevertheless, traditional Generative Adversarial Networks (GANs) frequently demonstrate a “geography-blind” characteristic, resulting in structural “topographic drift” by dissociating geomorphic complexity from cartographic constraints. To overcome this limitation, we propose Geo-InkGAN, a geo-heuristic framework that integrates geographic principles with generative processes to achieve high-fidelity ink-wash style synthesis. A key component of our approach is an adaptive optimization strategy grounded in the Slope Standard Deviation (SSD). By establishing a quantitative relationship between geomorphological entropy and the cycle-consistency loss weight (), we effectively address the Pareto trade-off between geomorphic accuracy and esthetic representation. Our results indicate that alluvial plains benefit from low-intensity constraints to facilitate fluid ink diffusion, whereas rugged terrains require high-intensity constraints to maintain the integrity of the topological framework. Additionally, the HCEG-SE mechanism (Hillshade-Contour Edge-Guided Stroke Enhancement) narrows the semantic divide between terrain skeletons and artistic textures by combining multi-directional non-photorealistic rendering with precise edge extraction techniques. Evaluated across five geomorphologically diverse regions—from karst towers to loess plateaus—Geo-InkGAN demonstrably surpasses existing benchmarks in Geomorphological Structure Correlation (GSC). This geomorphology-aware approach advances the scientific rigor of AI-driven cartography and offers a refined methodology for the cultural representation of digital twin landscapes.
Keywords
Funding Information
- National Natural Science Foundation of China (42571515)
This publication has 49 references indexed in Scilit:
- Automated extraction of hill boundaries in Gobi deserts based on DEM using integrated positive–negative terrain segmentation and multi-directional hillshadingGeocarto International, 2026
- An enhanced CycleGAN approach for landscape design: Style transfer and color harmonizationAlexandria Engineering Journal, 2025
- MapGPT: an autonomous framework for mapping by integrating large language model and cartographic toolsCartography and Geographic Information Science, 2024
- Integrating terrain structure characteristics into generative adversarial nets for hillshade generationInternational Journal of Geographical Information Science, 2024
- Artificial intelligence studies in cartography: a review and synthesis of methods, applications, and ethicsCartography and Geographic Information Science, 2024
- Rendering and presentation of 3D digital ink landscape paintingComputer Animation and Virtual Worlds, 2023
- Physics-informed machine learningNature Reviews Physics, 2021
- Improving AI System Awareness of Geoscience Knowledge: Symbiotic Integration of Physical Approaches and Deep LearningGeophysical Research Letters, 2020
- Virtual Geographic Environments (VGEs): A New Generation of Geographic Analysis ToolEarth-Science Reviews, 2013
- Fluid-based hatching for tone mapping in line illustrationsThe Visual Computer, 2009