Intro
I'm building a house, which introduced me to gis and drone mapping so
roughly documenting my process here.
Goal of this has been get a 3d model of the land allowing for sunlight
coverage simulation for architecture and solar panel placement.
Click below to run the project in the browser, for some reason works great on mobile, github repo is here
A bunch of geotagged photos from the air can be stitched together into a map. With higher frequency of photos 3d reconstruction of the terrain is possible. I'm working with around 100 photographs taken from 200 meters away of roughly 40.000 square meters of land.
- For stitching photos into a map I had luck with opendronemap
- For 3d height map local photogrametry software like Meshroom will do this, will require a GPU, might need a cleanup with something like MeshLab
Mapsmadeeasy service provides a simple service for both of the things above and it works great. While I did run local workflows eventually I decided that it's easier to offload this to a third party.
GIS processing
I'm using qgis for processing geo information and maps. LLMs turned out great for tips on how to do more complex operations there.
High res drone photo map
Elevation tiff created by mapsmadeeasy
Topographic map for architects is easy once heightmap is available: `Raster` → `Extraction` → `Contour`
3D Reconstruction
Tiff heightmap is just a 2D array, so it's easy to map it onto a 3d
plane.
I'm using GeoTIFF.js for tiff
parsing and three.js for
rendering.
Can slap any texture on top, so I add a drone photo
