perception_cupoch

This package provides functions that can convert pointclouds from ROS to cupoch and vice-versa.

Dependencies

• Eigen3-(currently 3.3.7)
• cupoch-(currently 0.1.9.1)

System Requirements

• Ubuntu 18.04+: GCC 5+
• cuda10.2+

Installation

eigen3

upgrade eigen to 3.3.7

# 1804
sudo apt install libopenblas-dev freeglut3-dev libglew-dev libadolc-dev libcholmod3 libumfpack5 libmetis-dev libspqr2 libsparsehash-dev libmpfr-dev libmpfrc++-dev libfftw3-dev
# jetson
sudo apt install libopenblas-dev freeglut3-dev libglew-dev libadolc-dev libcholmod3.0.6 libumfpack5.7.1 libmetis-dev libspqr2.0.2 libsparsehash-dev libmpfr-dev libmpfrc++-dev libfftw3-dev

cmake .. -DCMAKE_INSTALL_PREFIX=/usr
sudo make install

cupoch

* Instructions to setup cupoch can be found here.

Because of issues, you can use the custom version of cupoch, naming cupoch-fat.

And follow this guide to install cupoch to the path of /opt/cupoch/.

cupoch_conversions

• In case you are building this package from source, time taken for the conversion functions will be much larger if it is not built in Release mode.

Build

in cupoch_conversions/cmake/ga_build_common.cmake, set the CUPOCH_ROOT correctly.

export GPUAC_COMPILE_WITH_CUDA=1
catkin_make -j5 -DCMAKE_BUILD_TYPE=Release

(Optional) Performance tunning for jetson

NOTE: You should set power management policy for cpu and gpu to get the best performance of cudaMemCpy.

cpu frequency scaling

refer to this guide from stackoverflow to disable cpu scaling.

sudo apt-get install cpufrequtils
# to get the highest freq
sudo cpufreq-set -g performance

# to get the default mode powersave
sudo cpufreq-set -g powersave

gpu frequency scaling

refer to this guide for more details.

# If not, please set CONFIG_DEVFREQ_GOV_PERFORMANCE=y and update the kernel
zcat /proc/config.gz | grep CONFIG_DEVFREQ_GOV_PERFORMANCE
# Run below command on board as root to keep GPU alive (not suspend)
echo 0 > /sys/devices/17000000.gv11b/railgate_enable
# Enable performance governor for GPU
cat /sys/class/devfreq/17000000.gv11b/available_governors
echo performance > /sys/class/devfreq/17000000.gv11b/governor
cat /sys/class/devfreq/17000000.gv11b/governor

test

Unit Test

cd ./devel/lib/cupoch_conversions
./cupoch_test
./cupoch_conversions_test

ROS node test

# t1
roslaunch cupoch_conversions cupoch_conversions_test_node.launch
# t2
rosrun plotjuggler plotjuggler
# plot the topic /cupoch/ros2cupoch

performance benchmark

rmm benchmark.

cd ./devel/lib/cupoch_conversions
./RANDOM_ALLOCATIONS_BENCH  -n 1000 -m 1024

API Usage

There are two functions provided in this library:

void cupochToRos(std::shared_ptr<cupoch::geometry::PointCloud> &pointcloud, sensor_msgs::PointCloud2 &ros_pc2,
                   std::string frame_id = "cupoch_pointcloud");

void rosToCupoch(const sensor_msgs::PointCloud2ConstPtr &ros_pc2, std::shared_ptr<cupoch::geometry::PointCloud> &o3d_pc,
                   bool skip_colors = false);

• As cupoch pointclouds only contain points, colors and normals, the interface currently supports XYZ, XYZRGB pointclouds. XYZI pointclouds are handled by placing the intensity value in the colors_.
• On creating a ROS pointcloud from an cupoch pointcloud, the user is expected to set the timestamp in the header and pass the frame_id to the conversion function.

Example

Please refer to this repo ga_points_downsampler which is a ros node to downsample pointclouds with cupoch.

Documentation

Documentation can be generated using Doxygen and the configuration file by executing doxygen Doxyfile in the package.

Contact

Feel free to contact us for any questions:

• ZhenshengLi