在城市规划、房地产展示或游戏开发中，我们经常需要将某个真实区域“模型化”。传统的做法是耗费数周进行手工建模，而现在，通过**“微缩城市数字孪生重建 (V1.0)”**，我们可以利用 AI 强大的空间推理能力，直接将一张卫星截图“升维”成一张精美的 3D 移轴摄影图。
AI城市建模
核心特征说明
🛸 维度升维：扔给它一张扁平的谷歌地图，它还你一个立体的《模拟城市》。 📷 移轴美学：强制开启 Tilt-shift 特效。巨大的城市瞬间变成了精致的桌面模型。 🌍 全球适配：AI 能根据卫星图自动识别区域风格（如东京 vs 巴黎）并生成细节。 🔍 细节狂魔：屋顶的水箱、空调外机，路面上的斑马线，AI 都能自动填充。
关键技术拆解
🔑 核心逻辑：Dimensional Up-Scaling (维度升维) 这套 Prompt 的难点在于“无中生有”。卫星图只有 X 和 Y 轴，AI 需要根据阴影和常识推导出 Z 轴（高度）。我们使用 direct 3D reconstruction 和 Isometric projection（等轴测投影）的指令，引导 AI 将平面色块转化为立方体建筑。
🔑 视觉魔法：The Diorama Effect (透视盒效应) 为了让画面具有迷人的“玩具感”，我们使用了 Tilt-shift photography effect（移轴摄影效果）。这种技术通过模糊画面的上下边缘（浅景深），让人眼产生错觉，以为自己在看一个微缩模型。这种风格非常讨喜，能够降低复杂城市的压迫感，增加可玩味性。
🔑 扩展玩法：Time Travel (时空穿越) 这套逻辑不仅适用于空间，也适用于时间。
-
Historical Regression (历史重现)：保持道路布局不变，将建筑风格改为
Medieval（中世纪）或1920s Industrial。你会看到同一个街区在一百年前的样子。 -
Post-Apocalyptic (末世模拟)：在自然环境中加入
Overgrown vines, cracked roads（藤蔓丛生，道路龟裂）。瞬间生成《最后生还者》风格的废土场景。
实际使用场景
这套系统是空间设计师的灵感库：
场景一：城市规划提案。快速生成项目建成后的概念图，展示街区氛围。 场景二：游戏关卡设计。将真实地图转化为游戏场景概念图（Concept Art），辅助关卡搭建。 场景三：历史/地理教育。直观地展示不同地域、不同时期的城市形态演变。 场景四：房地产营销。生成项目周边的配套环境分析图，美观又直观。
AI 不仅能理解语言，也能理解空间。这套 Prompt 让你拥有了上帝的视角，去俯瞰、去重建、去想象我们的城市。以下是完整的 JSON 协议，复制它，上传地图，开始你的造城游戏。
| { “template_name”: “Geospatial_Reality_Architect_V1”, “template_version”: “1.0.0”, “template_purpose”: “Converts 2D satellite imagery into a hyper-realistic, high-density 3D isometric digital twin. Functions as a ‘current state snapshot’ generator without UI or map elements.”, “applicable_models”: [ “Midjourney v6”, “Stable Diffusion XL”, “Unreal Engine 5 Concepting”, “DALL-E 3” ], “input_assumptions”: “The system acts as a ‘Geospatial Reality Architect’. It analyzes a user-provided satellite image to infer layout, density, and style, creating a frameless, immersive ‘SimCity-like’ high-end model.”, “editable_fields”: [ { “field_key”: “detected_region_style”, “label_cn”: “区域识别 (Region Style)”, “description_cn”: “The overall vibe of the area detected from the map (e.g., Tokyo residential, US suburb).”, “example_values”: [ “Dense Tokyo Residential District”, “American Suburban Cul-de-sacs”, “European Old Town Center”, “Industrial Port Zone” ] }, { “field_key”: “road_pattern”, “label_cn”: “道路布局 (Road Layout)”, “description_cn”: “Describes the specific road network observed.”, “example_values”: [ “Strict grid-like roads”, “Winding cul-de-sacs”, “Organic riverbank roads”, “Chaotic alleyways” ] }, { “field_key”: “building_density”, “label_cn”: “建筑密度 (Density)”, “description_cn”: “Describes the verticality and spacing of buildings.”, “example_values”: [ “Dense slum-like structures”, “Sparse villas with pools”, “Soaring reinforced concrete jungle”, “Mid-rise commercial blocks” ] }, { “field_key”: “architectural_style”, “label_cn”: “建筑风格 (Arch Style)”, “description_cn”: “Inferred architectural style based on visual cues.”, “example_values”: [ “Showa-era Japanese houses”, “Haussmann-style Parisian blocks”, “Brutalist Soviet apartments”, “Modern Glass Facades” ] }, { “field_key”: “key_landmarks”, “label_cn”: “关键地标 (Key Landmarks)”, “description_cn”: “Specific unique structures visible in the satellite shot.”, “example_values”: [ “A school with a running track”, “A large roundabout”, “A specific bridge”, “A central park area” ] }, { “field_key”: “natural_environment”, “label_cn”: “自然环境 (Nature)”, “description_cn”: “Vegetation color and type.”, “example_values”: [ “Tropical emerald green vegetation”, “Temperate autumn lushness”, “Arid desert landscaping”, “Manicured urban gardens” ] } ], “generation_constraints”: { “quality”: [ “Isometric Projection”, “Diorama Aesthetic”, “Hyper-realistic scale”, “Tilt-shift photography”, “Unreal Engine 5 Render”, “8k Resolution” ], “textures”: [ “Claymation textures combined with photorealism”, “Fully textured ground”, “No abstract baseplates” ], “negative”: [ “User Interface”, “Map markers”, “Flat maps”, “Text”, “Borders”, “Frames”, “Pins” ] }, “final_image_prompt”: “Frameless, edge-to-edge high-angle isometric view, {{detected_region_style}}. This composition is a direct 3D reconstruction of the provided satellite terrain. The view reveals a single, coherent moment in time, capturing the bustling reality of the site. The ground is fully textured, removing all abstract baseplates. [BUILDING SILHOUETTES]: Layout features {{road_pattern}}. Density is characterized by {{building_density}}. Visible landmarks include {{key_landmarks}}. The architectural style is {{architectural_style}}. [LIFE & ATMOSPHERE]: Street level features miniature cars aligned with roads, visible crosswalks, and streetlights. Rooftop level features air conditioning units, roof gardens, water tanks, or solar panels (depending on region). Nature features street-lined trees with {{natural_environment}}. [RENDER SPECS]: Isometric projection, diorama aesthetic but hyper-realistic scale. Tilt-shift photography effect (blurring distant edges), magnificent sunlight casting crisp shadows. Unreal Engine 5 render, combining claymation textures with photorealism, 8K resolution. –no User Interface, map markers, flat maps, text, borders, frames –ar 16:9 –stylize 250 –v 6.0” } |
🖼️更多案例：https://my.feishu.cn/wiki/OzfIw5oTii0C60k9baLcGbdFnie?from=from_copylink
AI Urban Planning: Generating “Miniature Digital Twins” from Satellite Maps
Want to turn a flat satellite map into a stunning 3D miniature city? This article introduces the “Miniature Urban Digital Twin Reconstruction” prompt. It uses AI to perform “Dimensional Up-Scaling,” applying a tilt-shift aesthetic to create hyper-realistic, isometric cityscapes perfect for planning and game design.Keywords #AIUrbanPlanning #DigitalTwin #TiltShift #CityscapeGenerator #PromptEngineering #MidjourneyWorkflow #DALLE3 #GameLevelDesign #ArchitecturalVisualization #IsometricArt #moyuai
In urban planning and game development, visualizing a specific real-world location as a 3D model is often required. Traditional modeling is slow. Today, I’m sharing the “Miniature Urban Digital Twin Reconstruction (V1.0),” a workflow that allows AI to instantly “extrude” a 2D satellite screenshot into a beautiful, tilt-shift 3D diorama.
Core Features
🛸 Dimensional Up-Scaling: Converts 2D map grids into 3D volumes, inferring building heights and styles automatically. 📷 Tilt-Shift Aesthetic: Applies a shallow depth of field effect, making massive city blocks look like exquisite tabletop toys. 🌍 Context Awareness: Whether it’s the dense alleys of Tokyo or the boulevards of Paris, the AI adapts the architectural style to the map’s geography. 🔍 Living Details: Automatically populates the scene with cars, AC units, and trees, breathing life into the static geometry.
Technical Breakdown
🔑 Core Logic: Dimensional Up-Scaling The challenge is inferring the Z-axis (height) from a flat image. The prompt uses direct 3D reconstruction and Isometric projection instructions. This guides the AI to interpret the map’s footprints as foundations for 3D structures, respecting the original road network while inventing vertical details.
🔑 Visual Magic: The Tilt-Shift Effect To achieve that charming “SimCity” look, we enforce a Tilt-shift photography effect. By blurring the foreground and background, we trick the brain into perceiving the subject as miniature. This aesthetic makes complex urban data approachable and visually delightful.
🔑 Expansion: Time Travel Modes This logic can reconstruct the past or predict the future:
-
Historical: Keep the road layout but change the architecture to
Medievalor1920s. See the same streets in a different era. -
Post-Apocalyptic: Add
Overgrown vinesandcracked roadsto visualize the city after nature reclaims it—perfect for game concept art.
Use Cases
This tool serves spatial storytellers:
Case 1: Urban Planning Pitches. Visualizing the “vibe” of a proposed district renovation. Case 2: Game Level Design. Rapidly generating concept art based on real-world locations for open-world games. Case 3: Education. Showing students how cities evolve over time or differ across cultures.
Conclusion & Full Prompt
With this prompt, you hold the power to rebuild the world in miniature. Below is the full JSON protocol. Copy it, upload a map, and become the architect of reality.
