Terrain Coverage and Autogen

AVHRR Global Land Cover Classification
by M. Hansen, R. DeFries, J. R. G. Townshend, R. Sohlberg

Entry created by Andreas Jung.

Rough global land classification, based on AVHRR RADAR data; generated by a sophisticated classification algorithm. Resolution: Up to 30" (~1km). Dataset can be download as one big file for entire earth - the file itself is quite small because of the huge compression potential. [IMO] Not breathtaking, but sufficient for large scale sceneries. It's a good start to give the simulated world a basic destinction between crops, forest, and city. It is advisable to combine the data with other data such as culture and climate - this way, one can easily generate more distinctions, e.g. european-forest/brazilian-rainforest. Good: The dataset is available in lat/lon-project (it is not uncommon for these datasets to be available in equal-area-projection=non-lat/lon-projection, which renders it virtually useless for flight simulation).

Dealing with Coastline Disagrees
by Andreas Jung

Entry created by Andreas Jung.

In case you use land classification and vector-costlines at the same time, it is unlikely, that vector-coastlines and classification-transitions from ocean to land are exactly at the same location. In other words: it is possible, that land on the inland-side of the shore is still classified as ocean. To challange this disagree, one should render terrain, which is classified as ocean, using a dune texture (or beach or sand or whatever). Then, the actual water is masked out using a water-mask (e.g. alpha channel), which has been generated from water-body-vector-data. This way, the coastline-disagree-areas are rendered as dunes.

Land Cover in Flight Simulator Project
by Matthieu Laban

Entry created by Andreas Jung.

Some implementation experiences of land cover datasets in a flight simulator project. [IMO] Shares my experiences with this type of dataset. From his blog: "... but for some reason there is an offset in the land cover texture... I suspect this is because the texture mapping uses some sort of odd mapping... which is not the same as the one that's used for the satellite images..." - interesting.

VideoTrace
by A. van den Hengel, A. Dick, T. Thormaehlen, B. Ward, P. H. S. Torr

Entry created by Andreas Jung.

From the website: "VideoTrace is a system for interactively generating realistic 3D models of objects from video—models that might be inserted into a video game, a simulation environment, or another video sequence.". [IMO] Must see the video! This technology could become a key aspect in the creation of a digital globe with actual 3D objects placed on it. Unfortunately, I suspect the technology won't be inexpensive.

NIC Urban Cultivation
by Sam Chui

Entry created by Andreas Jung.

Shows the typical urban cultivation of a newly industrializing country (NIC).

Urban Light Distribution At Night
by Rafael Cordero

Entry created by Andreas Jung.

Low angle shot, hazy night, Mexico City (probably slums). [IMO] Could be your primary source of information when you plan to create the lightmap for an urban landclass texture.

How In The World Textures
by How In The World

Entry created by Andreas Jung.

237 geotypical tileable/blendable textures. [IMO] The textures are good. The offer can be considered fair and it also allows the eventual commercial use. All in all a good starting point, especially for teams with only few artists and non-satellite-imagery-projects. Long-term projects have to consider, that a texture resolution of 1024x1024 may not be state-of-the-art in a few years from now.