tile serving polylines?


What type of level of detail approaches are recommended for working with very large networks of polylines? e.g. Is it possible to tile serve vector polyline data to Cesium so when zoomed out you only see major roads but full detail is there when you zoom in?

While we’re on the subject, how about level of detail for labels/billboards like place names etc?



For individual polyline LOD, we may use the Douglas-Peucker algorithm. For large networks of polylines/billboards/labels, we’d probably do a HLOD approach. Google Maps and other mapping libraries are moving away from storing vector data in rasters, and instead store it in vector format. A lot of the work will be in tools to create the HLOD tree.