A review of digital terrain modeling

Eric Galin, Eric Guérin, Adrien Peytavie, Guillaume Cordonnier, Marie-Paule Cani, Bedrich Benes, and James Gain

This repository contains Digital Terrain Models collected by the authors of the state-of-the-art or obtained by the re-implementation of methods. It is meant to be public and maintained throughout time.

Authoring Landscapes by Combining Ecosystem and Terrain Erosion Simulation

Description

300 steps have been applied of Cordonnier 2017 method

Format

16 bits greyscale Tiff Cell size: 10m/pixel Alt min: 1057.71 Alt max: 2061.54 Latitude (°): 45 Compass (°); 90

Resolution

256x256

Source

Provided by authors

Files

[Cordonnier-2017.tiff] : main DEM

[Cordonnier-2017-0iteration.tiff] : initial terrain before simulation

Computer rendering of stochastic models

Description

8 iterations of diamond square algorithm applied on an initial 8x8 map.

Format

PNG 16 bits unsigned int greyscale.

Resolution

1793x1793

Source

Re-implemented

Files

[Fournier-1982.png] : final DEM (8 subdivisions)

[Fournier-1982-0.png] : initial 8x8 DEM (no subdivision)

[Fournier-1982-1.png] : first subdivision

[Fournier-1982-2.png] : second subdivision

[Fournier-1982-3.png] : third subdivision

[Fournier-1982-4.png] : fourth subdivision

[Fournier-1982-5.png] : fifth subdivision

[Fournier-1982-6.png] : sixth subdivision

[Fournier-1982-7.png] : seventh subdivision

Faulting/Random faults

Description

Faulting algorithm (also known as random faults).

Format

PNG 16 bits unsigned int greyscale.

Resolution

1024x1024

Source

Re-implemented

Files

[Faulting.png] : 400 iterations of the faulting algorithm (i.e. 400 faults are present)

[Faulting-50.png] : first 50 iterations

[Faulting-100.png] : first 100 iterations

[Faulting-200.png] : first 200 iterations

First Person Sketch-based Terrain Editing

Description

Digital elevation models from Tasse et al 2014. These are real DEMs modified using the first-person technique described in the paper.

Format

Tiff 16 bits greyscale unsigned int. No special information on the spatial range.

Resolution

512x512

Source

Provided by authors

Files

[Tasse-2014.tif] : DEM

[Tasse-2014-2.tif] : another example

[Tasse-2014-3.tif] : another example

Hydraulic erosion

Description

Hydraulic erosion performed on a noise

Format

16 bits greyscale png

Resolution

512x512

Source

Re-implemented

Files

[Hydraulic.png] : eroded DEM

[Hydraulic-original.png] : initial DEM

[Hydraulic-sand.png] : amount of sand

Interactive example-based terrain authoring with conditional generative adversarial networks

Description

Digital elevation model from Guerin et al 2017

Format

PNG 16 bits greyscale unsigned int. 30m per pixel

Resolution

2048x2048

Source

Provided by authors

Files

[Guerin-2017.png] : final DEM amplified x4 with a sparse method

[Guerin-2017-input.png] : the input sketch at 512x512 resolution

[Guerin-2017-river.png] : the river footprint

[Guerin-2017-riverbed.png] : the riverbed footprint

Interactive terrain modeling using hydraulic erosion

Description

Digital elevation model from Stava et al 2008

Format

grayscale PNG 16 bits

Resolution

1024x1024

Source

Re-implemented

Files

[Stava-2008.png] : main DEM

Large scale terrain generation from tectonic uplift and fluvial erosion

Description

Terrain fully generated by combining Stream Power erosion law and procedural tectonic uplift

Format

16 bits greyscale Tiff Cell size: 100m/pixel Alt min: 0 Alt max: 5749.82

Resolution

1000x1000

Source

Provided by authors

Files

[Cordonnier-2016.png] : 512x512 crop of the dem in a 16 bits greyscale format

[Cordonnier-2016-full.tiff] : main DEM

[Cordonnier-2016-uplift.tiff] : uplift rate (0 - 3e-3 m/year)

Modeling Landscapes with Ridges and Rivers: bottom up approach

Description

Digital elevation model from Belhadj et al 2005

Format

color PNG 16 bits

Resolution

513x513

Source

Provided by authors

Files

[Belhadj-2005.png] : main DEM

[Belhadj-2005-skeleton.png] : the input skeleton that gives the constraints

Multifractal sum of ridged noises

Description

Sum of 8 octaves of ridged noises, with a lacunarity of 1/0.467 and a persistance of 0.5. The amplitude of the next octave is attenuated with respect to the previously computed octaves by using a cubic smoothstep, in a way similar to Musgrave 1989.