Tags

v2.2 for ROS2. 
See https://gitlab.kitware.com/keu-computervision/slam/-/tags/v2.2 for more details.

This release mainly brings available features from the library to the wrappers and improves the wrappers interfaces.

Features available in this version : 
* Loop closure in ROS
* Loop closure detection supervision (PV/ROS)
* IMU and wheel encoder in ROS
* Automatic calibration with GPS
* Generic conversion from any PointCloud2 in ROS
* Calibration estimation between 2 trajectories
* Handling of Livox sensors in ROS
* Pose graph optimization with external poses is allowed

In the library, the keypoints extractor is deeply refactored and a new extractor is added to be able to handle dense LiDAR sensors in the future.
This new version also solves a lot of bugs and is more stable.

This new version mainly brings fixes to the SLAM on pose graph optimization and GPS handling and on the build and install cmake processes. 

Features available in this version : 
* Loop closure automatic detection + loading of various detections from the disk
* Failure detection and recovery mode
* Calibration estimation between 2 trajectories
* Handling of Livox sensors in ROS
* Pose graph optimization with external poses is allowed

Performances are also improved modifying the keypoints extraction and the submap extraction. The ROS interface is also improved.

The main changes are related to a postprocess use of the SLAM. 

Features available in this version : 
* Trajectory upload
* Loop closure
* IMU integration
* RGB camera integration
* External cross-platform superbuild

The library is cmakified and its install has been cleaned so that it can be easily integrated in an external project.

Features available in this version : 
* Landmarks handling (locally and in pose graph optimization)
* Wheel odometer handling
* External pose integration
* Localization/Mapping modes
* Modular keypoints extraction 
* MultiLidar
* GPS for pose graph optimization
* Moving objects rejection
* Confidence estimators available
* Ouster/Velodyne ROS support
* Superbuild

Version 1.2

This release brings important improvements (in terms of processing speed as well as precision) such as *undistortion* or *motion extrapolation*.
It also greatly improves user interface for easier parameters settings.

Version 1.1

Add several functionalities to v1.0, such as compressed pointclouds logging, latency compensation, multi-threading or ParaView/LidarView plugin.

First release of LiDAR SLAM as an independent project.

As this is the first 'official' version, most changes are not reported since **v0.0**, and only a small subset of useful or major changes is listed below.

### Core lib

* Numerous misc bug fixes and improvements
* Major code cleaning and refactoring
* Add CI for core SLAM lib
* Add pose graph optimization (PGO)
* Add optional logging of keypoints and trajectory
* Add verbosity modes to display state, steps durations and results
* Replace 6 DoF state vector by `Eigen::Isometry3d`
* Major acceleration of `RollingGrid`
* Add documentation for dependencies and installation

### ROS wrapping

* First version of ROS package `lidar_slam`, supporting all core SLAM lib functionalities
* Add SLAM parameters setting from ROS parameter server
* Optional SLAM/GPS global calibration from trajectories
* Add ROS package `gps_conversions` to manage conversions to standard `gps_common::GPSFix` message and process UTM/WGS84 transformations
* Compute GPS heading from movement when it is not available.
* Add documentation for usage

Slam code extracted from LidarView without any modification.