Urban Vegetation Landscape Monitoring

URBAN VEGETATION LANDSCAPE MONITORING

An added feature of NZAM's large format digital aerial cameras is that they photograph near infrared (NIR) images at the same time as photographing natural colour (RGB) images. A number of New Zealand territorial local authorities have been taking up the opportunity to have NIR imagery orthorectified concurrently or following the delivery of their RGB orthophotos.

The NIR reflections of vegetation and man-made objects are vastly different from one another. This lends NIR imagery to be a valuable source for impervious area mapping. Also, vegetation reflects varying amounts of NIR light depending on leaf structure and chlorophyll production.

Jean-Pierre van Oudenhove from Auckland City Council's Data Services group used his forward thinking manner to devise a methodology for mapping tree canopies within an urban environment that made use of not only the NIR content of digital camera photography but also their LiDAR point cloud dataset.

The NIR images were classified into areas of vegetation and non-vegetation. Within the vegetated areas the non-ground LiDAR point cloud data was used to distinguish grass areas from tree canopy. The tree canopy was then split into height classes and this information was used in combination with the NIR classification results to generate tree canopy area polygons. A success rate of 90+% was observed when comparing the results against tree canopies manually digitised from the aerial photography.

print this page

email this page