How to capture a 3D city model
3D city maps are so much more than the simple evolution of paper maps. 3D city models already play a game-changing role in planning and architecture. Here are 4 most common approaches to the creation of 3D city models - 2 free options, 1 very expensive option and 1 industry-standard option.
Step 1: Digital Surface Model from Lidar
Creating a 3D city model from Lidar
Taking raw Lidar data and turning it into a 3D model with no additional information, such as a building’s polygons, is an essentially automated process. Simply download the data in the raw LAS format to create Digital Surface Model (DSM). The process is very simple, follow this Step-by-step Guide if you are unsure how to Process Lidar Data, Create a DSM and/or Export the DSM into a .STL format.
Pros & Cons of creating a 3D city model from Lidar
+ It is a very cost effective method and the resulting 3D city models have numerous applications. These 3D models can offer quick and inexpensive view of wide city areas.
+ Almost no previous knowledge or software is necessary.
- Finished product is an unseparated digital surface model with untrue vertices, meaning it would provide limited opportunity for close-up analysis. Any work needing clean data will require a considerable amount of editing.
Step 2: Extruded Buildings Footprints
Extruding buildings footprint polygons is usually the fastest, most cost effective way of adding a 3rd dimension to 2D mapping data. In essence the technique is taking an estimate of a building’s height and extending the footprint vertically to that height. OpenStreetMap is a good example of this technique. The good news is - all the hard work has been done by OSM mapping community.
OSM Buildings 3D models
OSM Buildings building data works with web maps, GIS applications, analysis tools and 3d visualisation and non-commercial map applications may use our data tiles free of charge. Visit OSM Buildings to download required 3D data.
Pros & Cons of using Extruded Buildings Footprints
+ Further data such as footprint geometry, height, type, source id's and other properties are usually included.
+ Very quick, usually fairly cheap if background data already exists but lacking in finer detail and a terrain model.
+ These 3D models can give better understanding of the as-built environment than 2D maps.
+ Easy to work Separated Buildings with clean geometry.
- Terrain is usually only flat or flattened.
- 3D Model has low Level of detail and accuracy.
Step 3: Semi-Automated Processes
If you are a hi-tech company (helps if you are Google, Apple or Microsoft) you can commission an earial survey with ultra high shutter speed cameras. You can then use this data in combination with other data (such as extruded footprint polygons) and with help of AI and Machine Learning you can create a highly accurate, visually stunning 3D city model with most of residuals (even cars and people) cleaned from the 3D model.
3D City Models by Google
Structure-from-motion (SfM) pipeline for visual 3D modelling of a large city area using 360° field of view Google Street View images. The core of the pipeline combines the state of the art techniques such as SURF feature detection, tentative matching by an approximate nearest neighbour search, relative camera motion estimation by solving 5-pt minimal camera pose problem, and sparse bundle adjustment.
Pros & Cons of using Semi-Automated Processes
+ Visually attractive and clean 3D geometry.
+ 3D terrain capture & geometry cleaning.
- Very expensive.
Step 4: Manual Aerial Photogrammetry
Tried-and-tested, reliable and accurate, manual stereo photogrammetry is consistently accurate and detailed. The 3D models produced are highly usable and, once the project is set-up, various levels of detail can be captured for very little extra cost. There are lots of ways to collect source data but the downside is that it’s a manual process and can be expensive in terms of time and resources.
Sourcing Imagery Survey Data
Both drone-based and satellite-imagery photogrammetries have huge potential. Drones can produce visually stunning 3D models which can then be combined to create wide-area 3D city models. However, they are yet to be allowed to fly over cities to any meaningful extent, which limits their value for larger-scale 3D city maps. For small sites or local sites-of-change, drones are a great way to build and update 3D city models.
Civilian satellites currently offer very limited use in major world cities but with better and better resolution available, it should be possible to substitute aerial surveys for satellite surveys in the foreseeable future.
Manual Aerial Photogrammetry
3D London model by AccuCities is a good example of a large 3D city model captured using stereo photogrammetry from fixed-wing aircraft. These 3D models can be rapidly captured and delivered as AutoCad, Sketchup or FBX files can be directly imported into almost any 3D editing software and have very high, up to 15cm accuracy.
Pros & Cons of using Manual Aerial Photogrammetry
+ Accurate and detailed 3D model with clean geometry.
+ 3D terrain, Buildings & Trees can be captured and separated into layers and tiles.
+ Rapid & Cost effective capture.
- Requires existing Imagery Survey Data.
- No facade features or details are captured.