Landing Area Detector

This project is a C++ implementation of the Generalized Landing Aread Detection algorithm part of a pipeline to estimate terrain landability for field robotics applications such as underwater, areial (drone) and planetary landers. It computes a terrain landability map from the vehicle (AUV/ROV/Drone/Lander) characteristics and a high-resolution terrain evelation data (GeoTIFF or unstructured XYZ files). The approach is based on the algorithm described in:

Autonomous Landing of Underwater Vehicles Using High-Resolution Bathymetry
[https://ieeexplore.ieee.org/document/8897695]

Key Features

• Multi-threaded processing: Utilizes OpenMP for parallel acceleration.
• Optional CUDA support: Enable GPU acceleration with -DUSE_CUDA=ON.
• GeoTIFF & Raster Image Processing: Uses GDAL and OpenCV for image I/O and transformations.
• 3D Geometry Computation: Employs CGAL for plane fitting, convex hulls, and alpha-shapes.
• Configuration via YAML: Easily define vehicle and environment parameters without modifying the code.

Dependencies

• CMake: >= 3.23
• C++ Standard: C++17 or newer
• OpenCV: >= 4.6
• GDAL: Latest recommended
• Boost: >= 1.74
• CGAL: Latest recommended
• OpenMP: For parallel processing, 2.0 on Windows and 4.0 on Linux
• yaml-cpp: For experiment configuration

Optional:

• CUDA: If USE_CUDA=ON is specified during configuration, currently on experimentation.

Supported Platforms

• Linux (e.g. Ubuntu 20.X, 22.X and 24.04): Fully supported.
• Windows (e.g. via Git-Bash): Should be compatible, using standard CMake platform detection and toolchains.

Building

mkdir build
cd build
cmake ..
make

CMake will automatically detect the necessary dependencies and build the project. If CMAKE_BUILD_TYPE is not set, it will default to Release for speed performance. If you want to see the CMake configuration, run:

cmake -L ..

To enable CUDA support (experimental):

cmake -DUSE_CUDA=ON ..
make

Note:

• Ensure all dependencies are available on your system (via package managers or from source).
• You can adjust the installation directory and other parameters using standard CMake variables and options.

Installation (optional)

After building, install the executables to $HOME/bin:

make install

This will copy the executables to $HOME/bin folder. Please add this folder to your system's PATH environment variable.

Executables

The project builds three main executable targets:

• land: Main landability calculation pipeline
• tiff2rugosity: Rugosity calculation from GeoTIFF
• img.resample: Image resampling utility

Each tool integrates versioning information from the git repository (including commit hash) and is configured to handle various input formats as specified via YAML configuration files.

Usage

After building and installing, run the executables from $HOME/bin. For example:

$HOME/bin/land --config config.yaml --input bathymetry.tif
$HOME/bin/tiff2rugosity --input dem.tif --output rugosity_map.tif
$HOME/bin/img.resample --input large_image.tif --output resized_image.tif