Last year saw the release of Enhanced Automatic Terrain Extraction (or eATE), which allows for the creation of point clouds with unparalleled density, derived fully automatic from stereo imagery.
In this past year, Imagem has done extensive testing with this new module and has been able to achieve very nice results with various datasets. When correlation is accurate and the stereo images have an overlap of 60% or more, almost every pixel in the original photographs can be matched and given its proper elevation. This creates a point cloud with density that far exceeds commonly available height models, like AHN2 in the Netherlands.
Tests that we have done on 2,5cm aerial photographs result in DEM's with an elevation point for every 2,5cm. An additional advantage from this way of generating point clouds is the fact that every elevation point is given the correct RGB-value, derived from the input imagery. This results in very realistic depictions of an area and creates a clear picture, with a density you might be familiar with from LIDAR (laser based altimetry).
Because the process is completely automated and done on stereo imagery already available, this way of generating height models is far more cost effective than having to do a separate LIDAR flight.
Below you can see an example of what the general output of eATE can be. If you would like to know more or experience what we can achieve for your own area, please contact Imagem.
Nederlands
English
