luminouspath / terra

A large scale terrain rendering library written in Rust

Terra

Terra is work in progress large scale terrain rendering library built on top of wgpu.

Overview

Terra supports rendering an entire planet with details ranging in scale from thousands of kilometers down to centimeters. In Terra, terrain is treated as a heightmap along with a collection of texture maps storing the surface normal, albedo, etc.

All of this information can take quite a bit of space, so it isn't included in this repository. Instead, the necessary files are streamed from the internet at runtime and cached locally in a subdirectory with the current user's cache directory (which for instance defaults to ~/.cache/terra on Linux).

Level of detail (LOD)

To ensure smooth frame rates and avoid noticable "LOD popping", Terra internally uses sphere mapped version of the Continuous Distance-Dependent Level of Detail algorithm.

Incremental Generation

Terra works by streaming coarse grained tiles containing terrain attributes and then adding fractal details using wgpu compute shaders.

Getting Started

Install Rust and then the other needed dependencies:

$ sudo apt-get install libegl-dev libudev-dev build-essential libssl-dev cmake

Running should be as simple as:

git clone [email protected]:fintelia/terra && cd terra
cargo run --release

The first time you run Terra, it may take a minute or two to stream the necessary files. Don't worry if you have to kill the process part way through, on subsequent runs it will resume where it left off.

Once that step is done, you should see the main Terra window. You can navigate with the arrow keeps, and increase/decrease your altitude via the Space and Z keys respectively. Joystick controls are also supported if one is detected. To exit, press Escape.

You can also pass --help to see some other command line options.

System Requirements

• Windows or Linux operating system (Terra may work on MacOS with the 'soft-float64' feature enabled, but this hasn't been tested)
• A fast internet connection
• GPU with 2+ GB of VRAM

Data Sources / Credits

During operation, this library downloads and merges datasets from a variety of sources. If you integrate it into your own project, please be sure to give proper credit to all of the following as applicable.

Kind	Source
Elevation	ETOPO1 Global Relief Model
Elevation	NASA Digital Elevation Model
Orthoimagery	Blue Marble Next Generation
Stars	Yale Bright Star Catalog
Treecover	Global Forest Change
Trees	SpeedTree
Ground Textures	FreePBR