High-resolution digital imagery (satellite, aerial or drone-based) and high-resolution elevation data (typically derived from LiDAR point clouds) are increasingly used in local and regional environmental studies. But it is still challenging to extract high-resolution feature attributes (e.g. building footprints from imagery and terrain metrics from digital elevation models) for deeper geospatial analysis and mapping. Over the past few years we have invested heavily and developed capabilities and workflows to extract those feature attributes more cost-effectively, and in this blog we showcase studies for Newcastle and Wollongong coastal areas, New South Wales, Australia.

For underlying methods and techniques, this blog (link) summarised our work of creating (generalised) building footprints or outlines using new AI and deep learning. We have to stress a point which is not very clear in the user community that the building footprints extracted (automatically or semi-automatically) only show approximate shapes of buildings, and in map production one has to balance the level of cartographic generalisation and the true representation of features on the ground. In our analysis, we make efforts to achieve >99% locational accuracy (in terms of building positions) and ~95% areal accuracy (in terms of areas being extracted and matched). On terrain features, such as detailed contours and modelled surface water flow directions, we published a series of blogs (link) reporting related projects last year.

1.  Web Mapping

Figure 1 shows integrated web mapping where high-resolution imagery, extracted building footprints, detailed contours and modelled surface water flow directions are displayed together, for coastal areas in Newcastle City Council and Wollongong City Council. In this way important location information and insights can be communicated more effectively. With the measured context it is easier to compare two contrasting coastal stretches (one with steep cliffs and the other with gentle beaches) and to identify buildings vulnerable to different coastal hazards (e.g. erosion, storm surge and sea level rise).

We carry out the extraction of building footprints on demand, with a focus on major population centres and hazard-prone areas. For terrain and hydrology features, we have already prepared ready-to-use, API-enabled tiled web maps covering the majority of populated coastal areas in the country (please follow this link for more information).

Figure 1: Integrated web mapping where extracted building footprints, contours and modelled surface water flow directions are superimposed. Web App is developed using Google Maps APIs. Locations: Newcastle (top), and Wollongong (bottom).

2.  Desktop Mapping

Figure 2 displays the same set of feature layers in desktop mapping software (e.g. QGIS). API-enabled web maps on terrain ad hydrology can be directly linked and vector-based building footprint data mapped in versatile ways in desktop GIS.

Figure 2: Integrated desktop mapping (in QGIS) where extracted building footprints, contours and modelled surface water flow directions are superimposed. Locations: Newcastle (top) and Wollongong (bottom).

3.  Ongoing Work

Extracting detailed and accurate location attributes or metrics from high-resolution data sources (e.g. imagery and elevation) is an important part of our ongoing work. The newly-created geospatial feature layers can be very useful for a wide range of applications, including those in risk analysis, emergency management, local government, catchment management, and urban development & planning.


– Mapping the Coastal Settlement in Wamberal, NSW: Two Approaches link

– Extracting Features from High-resolution Imagery: An AI-based Processing Service link

– Building-level Geocoding of G-NAF (Address Database) for Improved Flood & Bushfire Risk Analysis in Australia link

– Products: High-resolution Web Maps on Terrain & Hydrology link