Make a great-looking 3D landscape visualization!
summary
Builds a 10×10 km model of Bergen from a 1 m-resolution geotiff, producing about 1.5 million triangles simplified to ~10% using the terrain-specific polygonal simplifier, textures aerial imagery, and packages everything as geotiff → OBJ → glb with KTX2 textures for fast browser upload. Workflow uses GDAL/geotiff.js, OpenStreetMap geojson for building outlines extruded into meshes, BVH raycasts to drop buildings onto terrain, gltf-packer, and renders with three.js; demos run in-browser and perform smoothly.
AI-written summary · about
chapters (12)
- 0:00Start
- 6:34Landscape visualization
- 10:30What and where: Coordinate systems
- 12:51What data providers can I use?
- 15:27Quality problems
- 17:43Let's build a terrain surface!
- 24:34Optimization: Mesh simplification
- 32:17Let's add some surface color!
- 35:43What to use? Satellite photos or aerial photos?
- 38:02Optimization: Texture compression
- 45:47LOD = levels of detail (for buildings)
- 56:58Rendering
related talks
30mCreating 3D Maps using Python
23mDigital Earth Models
37mVisualizing Large Datasets with JavaScript Using deck.gl
30mA front-end dev epic journey into geography
01h 30m3D geospatial data visualization using Python and Cesiumjs
20mA semantic terrain engine
24mMaps, Data, and Charts - Bound for Great Things!
27mCartography on the Web