lidar_euclidean_cluster_detect

The purpose of this package is to detect individual objects in pointcloud data. Points are grouped into clusters based on proximity and published as detected objects.

NOTE: A new version of this package is available in autoware.auto.

Process

1. Pointcloud preprocessing
   • Points closer than a distance of remove_points_upto meters are removed from the cloud.
   • Points are then downsampled if the downsample_cloud parameter is set to true.
   • The pointcloud is trimmed to remove points based on height thresholds (clip_min_height and clip_max_height).
   • Points are further trimmed based on their y position to either side of the vehicle if keep_lanes is set to true. The bounds are defined by keep_lane_left_distance and keep_lane_right_distance.
   • A RANSAC-based algorithm is then used to determine a ground plane and remove any points belonging to the ground. This is activated by the remove_ground parameter.
   • The pointcloud is further filtered using Difference-of-Normals to remove any points that belong to a smooth surface. This is activated by the use_diffnormals parameter.
2. Pointcloud Clustering
   • The preprocessed pointcloud is then clustered using Euclidean Cluster Extraction, the cluster tolerance is defined by the clustering_distance parameter. This is the only part of the node that provides the option to use the GPU (activated by the use_gpu parameter).
   • Resulting clusters are then checked against neighboring clusters and any clusters which are less than cluster_merge_threshold apart are combined into a single cluster.
   • Rectangluar bounding boxes and polygonal bounds are then fit to the cluster pointclouds.

References

Voxel-based Downsampling
Pointcloud Surface Normal Estimation
Difference of Normals Segmentation
Euclidean Cluster Extraction

ROS API

Subs

• points_raw (sensor_msgs/PointCloud2)
  Input pointcloud from lidar sensor.

Pubs

• detection/lidar_detector/cloud_clusters (autoware_msgs/CloudClusterArray)
  Array of cloud clusters.
• detection/lidar_detector/objects (autoware_msgs/DetectedObjectArray)
  Array of all detected objects.
• cluster_centroids (autoware_msgs/Centroids)
  Centroids of the clusters.
• points_lanes (sensor_msgs/PointCloud2)
  Pointcloud with all preprocessing performed except Difference-of-Normals filtering.
• points_cluster (sensor_msgs/PointCloud2)
  Pointcloud colored according to cluster.
• points_ground (sensor_msgs/PointCloud2)
  Pointcloud of only ground points.

ROS Parameters

See the yaml file in the config folder for all ROS parameters and their descriptions

Changelog for package lidar_euclidean_cluster_detect

1.11.0 (2019-03-21)

• [fix] Install commands for all the packages (#1861)
  • Initial fixes to detection, sensing, semantics and utils
  • fixing wrong filename on install command
  • Fixes to install commands
  • Hokuyo fix name
  • Fix obj db
  • Obj db include fixes
  • End of final cleaning sweep
  • Incorrect command order in runtime manager
  • Param tempfile not required by runtime_manager
  • • Fixes to runtime manager install commands
  • Remove devel directory from catkin, if any
  • Updated launch files for robosense
  • Updated robosense
  • Fix/add missing install (#1977)
  • Added launch install to lidar_kf_contour_track
  • Added install to op_global_planner
  • Added install to way_planner
  • Added install to op_local_planner
  • Added install to op_simulation_package
  • Added install to op_utilities
  • Added install to sync
  • • Improved installation script for pointgrey packages
  • Fixed nodelet error for gmsl cameras
  • USe install space in catkin as well
  • add install to catkin
  • Fix install directives (#1990)
  • Fixed installation path
  • Fixed params installation path
  • Fixed cfg installation path
  • Delete cache on colcon_release
• Fix license notice in corresponding package.xml
• Adaptation of Object Filter with new perception workflow
• Initial release of object filter
• Contributors: Abraham Monrroy, Abraham Monrroy Cano, amc-nu

1.10.0 (2019-01-17)

• Fixes for catkin_make
• Use colcon as the build tool (#1704)
  • Switch to colcon as the build tool instead of catkin
  • Added cmake-target
  • Added note about the second colcon call
  • Added warning about catkin* scripts being deprecated
  • Fix COLCON_OPTS
  • Added install targets
  • Update Docker image tags
  • Message packages fixes
  • Fix missing dependency
• Feature/perception visualization cleanup (#1648)
  • • Initial commit for visualization package
  • Removal of all visualization messages from perception nodes
  • Visualization dependency removal
  • Launch file modification
  • • Fixes to visualization
  • Error on Clustering CPU
  • Reduce verbosity on markers
  • intial commit
  • • Changed to 2 spaces indentation
  • Added README
  • Fixed README messages type
  • 2 space indenting
  • ros clang format
  • Publish acceleration and velocity from ukf tracker
  • Remove hardcoded path
  • Updated README
  • updated prototype
  • Prototype update for header and usage
  • Removed unknown label from being reported
  • Updated publishing orientation to match develop
  • • Published all the trackers
  • Added valid field for visualization and future compatibility with ADAS ROI filtering
  • Add simple functions
  • Refacor code
  • • Reversed back UKF node to develop
  • Formatted speed
  • Refactor codes
  • Refactor codes
  • Refactor codes
  • Refacor codes
  • Make tracking visualization work
  • Relay class info in tracker node
  • Remove dependency to jskbbox and rosmarker in ukf tracker
  • apply rosclang to ukf tracker
  • Refactor codes
  • Refactor codes
  • add comment
  • refactor codes

  * Revert "Refactor codes" This reverts commit 135aaac46e49cb18d9b76611576747efab3caf9c. * Revert "apply rosclang to ukf tracker" This reverts commit 4f8d1cb5c8263a491f92ae5321e5080cb34b7b9c. * Revert "Remove dependency to jskbbox and rosmarker in ukf tracker" This reverts commit 4fa1dd40ba58065f7afacc5e478001078925b27d. * Revert "Relay class info in tracker node" This reverts commit 1637baac44c8d3d414cc069f3af12a79770439ae.

  • delete dependency to jsk and remove pointcloud_frame
  • get direction nis
  • set velocity_reliable true in tracker node
  • Add divided function
  • add function
  • Sanity checks
  • Relay all the data from input DetectedObject
  • Divided function work both for immukf and sukf
  • Add comment
  • Refactor codes
  • Pass immukf test
  • make direction assisted tracking work
  • Visualization fixes
  • Refacor codes
  • Refactor codes
  • Refactor codes
  • refactor codes
  • refactor codes
  • Refactor codes
  • refactor codes
  • Tracker Merging step added
  • Added launch file support for merging phase
  • lane assisted with sukf
  • Refactor codes
  • Refactor codes
  • • change only static objects
  • keep label of the oldest tracker
  • Static Object discrimination
  • Non rotating bouding box
  • no disappear if detector works
  • Modify removeRedundant a bit
  • Replacement of JSK visualization for RViz Native Markers
  • Added Models namespace to visualization
  • Naming change for matching the perception component graph
  • • Added 3D Models for different classes in visualization
  • 2D Rect node visualize_rects added to visualization_package
• Drop support for ROS Indigo
• Added notes about Indigo being deprecated
• Fix Ros/ROS naming convention
• Contributors: Abraham Monrroy Cano, Esteve Fernandez, amc-nu

1.9.1 (2018-11-06)

1.9.0 (2018-10-31)

• [fix] PascalCase messages (#1408)
  • Switch message files to pascal case
  • Switch message names to pascal case in Runtime Manager
  • Switch message names to pascal case in *.yaml
  • Rename brake_cmd and steer_cmd to BrakeCmd and SteerCmd in main.yaml
• • Fixed typo
  • Changed default values
• Fix multiple definition
• Fix condition for pcl::PCA, issue:#1604
• Feature/beyond pixel tracker (#1473)
  • Add beyond_pixel node
  • Update prototype of beyond pixel (#1430)
  • Add parser of DetectedObjectArray for beyond tracker(#1430)
  • • Adaptations to the original code
  • Added README
  • Added Runtime Manager entry
  • Added Video link
  • Added install commands for cmake
  • • Add ID only to tracked objects
  • Display valid IDs on the 3D labels
  • Display only objects with image coords
  • • Added Minimum dimensions
  • Register angle from the vision tracker if available
  • Keep message publishing rate continuous
  • Revert platform_automation_msgs (#1498)
  • Code cleanup
  • Fixed a crash when the dimensions are outside of the image
  • Fix annoying catkin_make causing to run twice the Cmake generation
• Contributors: Abraham Monrroy, Akihito OHSATO, Esteve Fernandez, amc-nu

1.8.0 (2018-08-31)

• Add variable number of params for clustering_ranges and clustering_distances
• load multiple thres params from string: less rosparam
• Naming change from segment to clustering
• Apllying ros clang
• Add optional multiple threshold euclidean clustert
• applying ros clang formant
• Add only one segment distance
• Feature/std perception msg (#1418)
  • New standard message definition for the perception nodes
  • New Detected Object message applied to:
  • SSD
  • Integrated RVIZ viewer
  • External Viewer
  • modified yolo2 and yolo3, compiles but cuda issues, trying different PC
  • Boiler plate for range vision fusion node

  * Added GenColors for Kinetic Typo fixes for yolo2

  • testing colors in Yolo3
  • Completed transformation, projection of 3D boxes
  • Fixed error on negative assignation
  • code clean up
  • removed yolo2 and yolo3, replaced by single darknet node. GUI launches yolo3 for now, to change. Pushing to test code on other PC.
  • Readme updated, added gitignore for data folder.

  * Added Runtime manager UI for yolo2, yolo3.Support tested for TinyYolo v2 and v3 * Fusion Vision Range Icons for viewer

  • Range Vision Fusion node
  • Indigo cv im read
  • Indigo compiation fix
  • Topic renaming according to new spec
  • Try to fix arm64 stuff
  • • Added launch file
  • Added Runtime manager entry
  • • Added Publication of non fused objects
  • Fixed topic names
• [Fix] Moved C++11 flag to autoware_build_flags (#1395)
• [Feature] Makes sure that all binaries have their dependencies linked (#1385)
• • CUDA Fixes for indigo (#1280)
  • CMake fixes for vision packages
• Contributors: Abraham Monrroy, Esteve Fernandez, Kosuke Murakami

1.7.0 (2018-05-16)

• [Fix] rename packages (#1269)
  • rename lidar_tracker
  • Modify pf_lidar_track's cmake file
  • Refactor code
  • Rename from euclidean_lidar_tracker to lidar_euclidean_track
  • Rename from kf_contour_track to lidar_kf_contour_track
  • Rename from kf_lidar_track to lidar_kf_track, but need some modification in euclidean cluster(Cluster.h)
  • Rename from pf_lidar_tarck to lidar_pf_track
  • Rename range_fusion
  • Rename obj_reproj
  • Rename euclidean_cluster to lidar_euclidean_cluster_detect
  • Rename svm_lidar_detect to lidar_svm_detect
  • Rename kf_lidar_track to lidar_kf_track
  • Change version 1.6.3 to 1.7.0 in pacakge.xml
  • Modify CMake so that extrenal header would be loaded
  • Remove obj_reproj from cv_tracker
  • Add interface.yaml
  • Rename road_wizard to trafficlight_recognizer
  • create common directory
  • Add lidar_imm_ukf_pda_track
  • create vision_detector and moved cv
  • Modify interface.yaml and package.xml
  • remove dpm_ocv
  • moved directory
  • Delete unnecessary launch file
  • Delete rcnn related file and code
  • separated dummy_track from cv_tracker
  • separated klt_track from cv_tracker
  • Fix a cmake
  • Remove unnecessary dependency of lidar_euclidean_cluster_detect package
  • Rename image_segmenter to vision_segment_enet_detect
  • Remove unnecessary dependency of lidar_svm_detect package
  • separated kf_track and fix a some compiling issue
  • move viewers
  • merge ndt_localizer and icp_localizer, and rename to lidar_localizer
  • Remove unnecessary dependency of lidar_euclidean_track
  • moved image lib
  • add launch
  • lib move under lidar_tracker
  • Rename dpm_ttic to vision_dpm_ttic_detect
  • rename yolo3detector to vision_yolo3_detect
  • Modify cmake and package.xml in vision_dpm_ttic_detect
  • moved sourcefiles into nodes dir
  • moved sourcefiles into nodes dir
  • Move cv_tracker/data folder and delete cv_tracker/model folder
  • fix a package file and cmake
  • Rename yolo2 -> vision_yolo2_detect
  • fix a package file and cmake
  • Fix package name of launch file
  • Rename ssd to vision_ssd_detect
  • fixed cmake and package for decerese dependencies
  • remove top packages dir for detection
  • fixed cmake for cuda
  • Rename lane_detector to vision_lane_detect
  • Modify package.xml in lidar-related packages
  • Remove unnecessary dependencies in lidar_detector and lidar_tracker
  • Modify computing.yaml for dpm_ttic
  • Modify dpm_ttic launch file
  • Remove/Add dependencies to trafficlight_recognizer
  • Update data folder in dpm_ttic
  • Modified CMake and package file in dpm_ttic.
  • Remove src dir in imm_ukf_pda_track
  • removed unnecessary comments
  • rename lidar_tracker
  • Modify pf_lidar_track's cmake file
  • Refactor code
  • Rename from euclidean_lidar_tracker to lidar_euclidean_track
  • Rename from kf_contour_track to lidar_kf_contour_track
  • Rename from kf_lidar_track to lidar_kf_track, but need some modification in euclidean cluster(Cluster.h)
  • Rename from pf_lidar_tarck to lidar_pf_track
  • Rename range_fusion
  • Rename obj_reproj
  • Rename road_wizard to trafficlight_recognizer
  • Rename euclidean_cluster to lidar_euclidean_cluster_detect
  • Rename svm_lidar_detect to lidar_svm_detect
  • Rename kf_lidar_track to lidar_kf_track
  • Change version 1.6.3 to 1.7.0 in pacakge.xml
  • Modify CMake so that extrenal header would be loaded
  • Remove obj_reproj from cv_tracker
  • Add interface.yaml
  • create common directory
  • Add lidar_imm_ukf_pda_track
  • create vision_detector and moved cv
  • Modify interface.yaml and package.xml
  • remove dpm_ocv
  • moved directory
  • Delete unnecessary launch file
  • Delete rcnn related file and code
  • separated dummy_track from cv_tracker
  • separated klt_track from cv_tracker
  • Fix a cmake
  • Remove unnecessary dependency of lidar_euclidean_cluster_detect package
  • Rename image_segmenter to vision_segment_enet_detect
  • Remove unnecessary dependency of lidar_svm_detect package
  • separated kf_track and fix a some compiling issue
  • move viewers
  • merge ndt_localizer and icp_localizer, and rename to lidar_localizer
  • Remove unnecessary dependency of lidar_euclidean_track
  • moved image lib
  • add launch
  • lib move under lidar_tracker
  • Rename dpm_ttic to vision_dpm_ttic_detect
  • rename yolo3detector to vision_yolo3_detect
  • Modify cmake and package.xml in vision_dpm_ttic_detect
  • moved sourcefiles into nodes dir
  • moved sourcefiles into nodes dir
  • Move cv_tracker/data folder and delete cv_tracker/model folder
  • fix a package file and cmake
  • Rename yolo2 -> vision_yolo2_detect
  • fix a package file and cmake
  • Fix package name of launch file
  • Rename ssd to vision_ssd_detect
  • fixed cmake and package for decerese dependencies
  • remove top packages dir for detection
  • fixed cmake for cuda
  • Rename lane_detector to vision_lane_detect
  • Modify package.xml in lidar-related packages
  • Remove unnecessary dependencies in lidar_detector and lidar_tracker
  • Modify computing.yaml for dpm_ttic
  • Modify dpm_ttic launch file
  • Remove/Add dependencies to trafficlight_recognizer
  • Update data folder in dpm_ttic
  • Modified CMake and package file in dpm_ttic.
  • Remove src dir in imm_ukf_pda_track
  • Fix bug for not starting run time manager
  • Remove invalid dependency
• Contributors: Kenji Funaoka, Kosuke Murakami

1.6.3 (2018-03-06)

1.6.2 (2018-02-27)

1.6.1 (2018-01-20)

1.6.0 (2017-12-11)

1.5.1 (2017-09-25)

1.5.0 (2017-09-21)

1.4.0 (2017-08-04)

1.3.1 (2017-07-16)

1.3.0 (2017-07-14)

1.2.0 (2017-06-07)

1.1.2 (2017-02-27 23:10)

1.1.1 (2017-02-27 22:25)

1.1.0 (2017-02-24)

1.0.1 (2017-01-14)

1.0.0 (2016-12-22)