Landscape Visualization

The music I wrote in 2008 in part thinking of Koyaanisqatsi. The footage I collected this year - all from the rooftop of my workplace. Marrying the two together came by chance during the editing process - and it worked.

Shot on a Panasonic HVX200, in 1080 mode using the interval recording function. Rate of recording was 1 shot in every 4 or 8 fps. The soft focus look I added afterwards - it ain´t genuine tilt shift - but I did shoot with that in mind and have applied a filter made for that look.

This is the visualization of the final outuput of my diplomma written in:

Warsaw University of Technology
(Department of Photogrammetry and Remote Sensing)
&
Vienna University of Technology
(Institute of Photogrammetry and Remote Sensing)

Basic data was provided by the Surveying Department of the City of Vienna, Austria:
~ ALS (LiDAR) data: 21pts/m2
~ Buildings outlines
~ Orthophoto with 10 cm pixel.

I have collected also ~300 terrestial images on my own.

Model accuracy:
XY: 0,3 m
Z: 0,2 m
Every detail of elevation bigger then 0,5m was included
City GML LoD 3
DTM: 2 m grid size.

Software used:
# Adobe After Effects CS4 (intro)
# DTMaster (filtering data, postprocessing)
# SCOP++ (DTM & nDSM generatiing)
# GML Undistoretr (undistorion of pictures)
# GVE (supporting modelling)
# Ruby Scripts (import point cloud to SketchUp)
# Google SkecthUp 7 (modeling, texturing, animating)
# Gimp (2D graphic processing)
# Windows Movie Maker (final sync & animation)

Visit my blogs:
miasta3d.blox.pl/html [PL]
city3d.blogspot.com/ [ENG - soon full translated]

This 3d visualization is based on real-world geodata and vegetation maps describing an unique moor landscape in Entlebuch in central Switzerland. In co-operation with the Lenné3D project (www.lenne3d.de), a realistic distribution of plants was calculated on basis of the underlying data. The plant models were produced in Xfrog (www.greenworks.de) using scans of samples from the case study site. The final real-time visualization with 4 million polygons took place in the Lenné3D Player software. The successor of the Lenné3D Player is the OpenSource Software Biosphere3D (www.biosphere3d.org).

The video shows a river renaturation measure 30 years in the future, with the focus on a realistic vegetation representation. The terrain, all buildings and the course of the river are based on CAD and GIS data. The vegetation is based on ecological data and it was a key objective to distinguish the different vegetation zones visually: The alluvial forest, the sandbanks and sparsely covered gravel. The production took place in Autodesk Autocad, Autodesk 3d Studio Max and with the Bionatics EasyNat plugin for the vegetation.

Unfortunately, this version of the movie still lacks from a crude error in the ground texture. However, it was decided to use only still images for the final production so that there is no updated version of this movie.