MapSense is a fast perception algorithm for extracting a map of the world for various applications including robotics and augmented reality. Planar regions are abundantly available in most indoor and urban environment settings. The current project aims to provide a fast planar region extraction algorithm for footstep planning on legged robots. The MapSense algorithm uses OpenCL to perform GPU-accelerated planar region detection and extraction from depth images. It can be used with any off-the-shelve depth image sensor. Some examples include RealSense L515, RealSense D435, Microsoft Kinect, and others.
This project comes with a pre-packaged Docker container with all necessary dependencies for the algorithm to run. This Docker image can be either built locally, or pulled from the Docker Hub.
The only dependencies that need to be manually installed includes nvidia-docker2 and docker itself. Use the following commands to install them on the host.
distribution=$(. /etc/os-release;echo $ID$VERSION_ID) \
&& curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add - \
&& curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list
sudo apt-get update -y
sudo apt-get install -y nvidia-docker2
sudo systemctl restart docker
The next step is to get the docker image for the container either from Docker Hub or built locally. Use the following commands based on preference.
bash Docker/runRemoveMapSense.sh
cd Docker
mkdir -p Shared_Volume #Volume mapping to share ROSBags for testing MapSense inside the container.
docker build -t bmishra/mapsense-nvidia-ros .
For using MapSense with a depth camera, simply direct the camera_info and depth topics to the appropriate input topics.
rosbag play --loop L515_MotionCinderBlocks.bag
/camera/depth/camera_info
/camera/depth/image_rect_raw
/mapsense/planar_regionsd
Input Resolution: (width, height)
Filter, Pack, Merge Kernels: (thread_pool_width, thread_pool_height)
Patch Angular Merge Threshold
Patch Distance Merge Threshold