heightcraft 2.0.0

Heightmap Generation & AI Upscaling Tool

Heightcraft is a high-performance CLI tool designed for Game Engineers, GIS Professionals, and Researchers. It automates the pipeline of converting 3D data into high-precision heightmaps, featuring accessible AI Upscaling.

🚀 Key Features

• 🤖 Heightmap AI Upscaling: Transform low-res inputs into crisp, hi-res/high-detail heightmaps. Also supporting bit depth increase (8-bit → 16/32-bit) with AI hallucination.
• 📡 LiDAR Support: Stream process massive .las and .laz point clouds into Digital Elevation Models (DEMs).
• 🏔️ Mesh to Heightmap: Bake .gltf, .glb, .obj, .stl, and .ply meshes into heightmaps with automated chunking.
• 🌊 Feature Masks: Optionally export water masks with sea level thresholds and texture masks for slope and curvature.
• 🎯 High Precision: Native support for 32-bit Float TIFF, 16-bit PNG, and RAW formats.
• ⚡ Performance: GPU acceleration and memory-efficient streaming for gigabyte-scale datasets.

📖 Documentation

Full documentation is available in the Docs.

• Getting Started
• LiDAR Processing
• Mesh Processing
• AI Upscaling Deep Dive

📦 Installation

pipx install heightcraft

⚡ Quick Start

Upscale an image (8-bit → 16-bit and 3x upscale):

heightcraft input.png --upscale --upscale-factor 3 --bit_depth 16 --output_path high_res.png

Process LiDAR data into an high-res 32-bit heightmap:

heightcraft scan.laz --bit_depth 32 --chunk_size 600000 --max_resolution 24576

Convert a 3D Mesh to an upscaled heightmap:

heightcraft terrain.obj --large-model --chunk_size 200000 --max_resolution 8192 --upscale

Development

To install development dependencies:

pipx install uv
uv sync --group dev
uv run main.py

🧪 Testing

Run the test suite:

pytest tests