Ortho-graphy: city base layers from an aerial photo

Base layers are usually built slowly: buildings in one project, pavements in another, trees perhaps a few years later. Meanwhile a recent aerial photo is already sitting with the authority, and what it shows does not yet exist as a layer anyone can filter, measure or maintain.

We ran machine recognition (AI) over the authority's existing aerial photo to flag features. Interpreters then completed by hand whatever the machine missed and checked every feature, with heights taken from the 3D model. Depending on the brief, the layers cover buildings, drainage inlets, manholes, culverts, fire hydrants, pavements and kerb colour, poles, signs, traffic lights, bus stops, bike lanes, roundabouts, parking, roads, traffic islands, lane lines, trees, shrubs and fences.

The authority gets a full set of base layers for a whole area in a single project and works with it in its own system: counting, measuring, planning maintenance and using the layers as the starting point for future surveys.