Support

There may be a few reasons, such as:

• The building may be considered smaller in area than 9m²
• The building was missed by the our data contributors process (generally ~99% of all buildings are captured, with usually only smaller buildings occasionally missed)
• The building didn’t exist when our data contributors captured their imagery.

Where tall features like buildings and trees exist, the terrain has been smoothed and areas in close proximity to some of these high features appear to have higher than expected terrain values. With higher than expected terrain values, tree pixels close to tall buildings or other tall trees may appear to have lower than expected resulting tree heights. We’re working closely with data contributors to fix this issue and expect tree heights will become more accurate in future updates to our Geoscape data.

We have identified an over-classification of buildings with a primary roof material of ‘Tile’. These are more evident in areas of industrial zoning and buildings with iron appearance. We’re working with data contributors keep improving the classification accuracy of roof materials.

Surface cover for both 2-metre and 30-metre resolution is captured and supplied to us in ‘strips’ defined by satellite coverage. Variation in ‘strips’ may be dependent on a number of factors including the classification, atmospheric corrections, weather, and season when the image was captured. We’re working with contributors to improve classification consistency for both datasets.

Maximum Roof Heights and Average Eave Heights for buildings are dependent on the date of capture of source imagery, which is used in the creation of the Digital Surface Model. Where buildings are part of a new development or re-development, these heights are impacted. We’re investigating ways to identify or remediate the accuracy of heights across the dataset and specifically in new development or re-development areas.

Although we’ve tried to capture and represent real world trees as accurately as possible, a number of individual trees have not been captured in the trees dataset. We’re working to improve trees, including improvements to source heights, classifications and improved capture dates.

The misalignment of boundaries has resulted in a building (BUILDING_PID = BLDNSW0001772595) being allocated a Mesh Block in a different state/territory. We’re working to fix some of the issues in the boundary misalignment to better reflect the allocation of Mesh Blocks, Suburbs/Localities and state/territories.

We’ve identified that in some areas of the 2m Surface Cover, specifically bitumen, there is an over-classification of water. Although most roads appear unaffected, areas like airport runways have a clear over-classification. We’re working with contributors to improve the classification of bitumen for future releases.