Layered Wavemap extends Wavemap's compressed multi-resolution occupancy map with typed environmental attributes. Continuous fields such as reflectivity, signal, near-infrared intensity, and RGB share the wavelet hierarchy with occupancy. Categorical fields such as semantic classes use aligned sparse side layers with lossless majority-with-exceptions compression.
Get started · Run the RGB demo · Original Wavemap documentation
Watch the Layered Wavemap demo on YouTube.
The synthetic forest demo needs no sensor data. On Ubuntu 20.04 with ROS Noetic:
cd ~/catkin_ws
catkin build wavemap_all
source devel/setup.bash
roslaunch wavemap_ros synthetic_rgb_layered_map.launchThe launch file opens RViz and progressively builds a colored layered map. Docker setup, native installation, real Ouster usage, map persistence, and a layer-extension walkthrough are in the Layered Wavemap guide.
The Layered Wavemap guide covers installation, examples, map persistence, layer schemas, update policies, and extension points. Documentation for the inherited Wavemap APIs and tools remains available in the original Wavemap documentation.
Layered Wavemap is built on and derived from Wavemap, originally developed by Victor Reijgwart and contributors at the Autonomous Systems Lab, ETH Zurich. The original project provides the wavelet-based multi-resolution occupancy mapping framework on which this extension is based. This repository adds typed continuous and discrete layers, their integration and storage policies, ROS1 support, persistence, visualization, and sensor examples.
The original Wavemap paper is:
Victor Reijgwart, Cesar Cadena, Roland Siegwart, and Lionel Ott. “Efficient Volumetric Mapping of Multi-Scale Environments using Wavelet-Based Compression.” Robotics: Science and Systems XIX, 2023. https://doi.org/10.15607/RSS.2023.XIX.065
Please cite the original work when using Wavemap or Layered Wavemap in research. The upstream copyright and BSD 3-Clause license are preserved in LICENSE.