* Avoid removing nodes from STM when using CreateIntermediateNodes
* Clean up the data of the intermediate node when IntermediateNodeDataKept is false
* Save intermediate node input features
* Don't delete intermediate nodes
* comment typo
* Updated parameter description, bump patch for DBReader API chang
* When ID is not udpated to new one, don't disable words because that signature is still used in the bayes filter
* updated comment
* rtabmap-reprocess: support Rtabmap/DetectionRate when intermediate nodes are created and -skip option is not used
* Fixed smallMovement when rehearsal is not trigdered at the same time
* GUI: visualize rehearsal darkblue when intermediate nodes are used
* Fixed RtabmapThread not transfering input features on intermediate nodes
* Do not override smallMovement and fastMovement when converting to intermediate nodes to keep GUI visualization color working
---------
Co-authored-by: Borong Yuan <yuanborong@hotmail.com>
* Auto graph repairing: detect and remove bad loop closures accepted in the past that block new good loop closures to be accepted
* Added parameter RGBD/OptimizeMaxErrorRepairRadius: added approach to repair the graph (remove bad loop closures added in the past). Refactored max optimization error function, split repairGraph to its own function.
* Handling optimization from graph end in repair
* MaxGraphErrors: changed link's pointers to copy directly to avoid seg fault down the road if not used properly
* Dont publish optimization statistics if it was not computed.
* Support intermediate nodes with memory management enabled
* Adding fix from #1686
* Fixed odomCache overlay shown in mapping mode
* Fixed local retrieval with intermediate nodes
* bumping version for updating API. Fixed getting graph error.
* Disable planning if intermediate nodes are there. Updated usage of Kp/BadSignRatio to support intermediate nodes.
* Added liosam odometry integration
* Add kXYZIRT scan format with per-point ring channel for LIO-SAM integration
Introduce PointXYZIRT point type and kXYZIRT LaserScan format (x,y,z,
intensity,ring,time) so that ring indices survive the scan pipeline.
Update OdometryLIOSAM to require kXYZIRT and properly split ring/time
into the parallel buffers LIO-SAM expects. Extend deskewing to preserve
ring data and disable base-class deskew in OdometryLIOSAM since LIO-SAM
handles it internally.
* Address PR review: config file, deferred init, and GUI panel for LIO-SAM
- Add OdomLIOSAM/ConfigPath parameter to load LIO-SAM settings from a
YAML file. When set, individual params are ignored. Extrinsics from
sensor local transforms always override config file values.
- Defer LioSamCore initialization until both IMU and lidar local
transforms are available, computing T_lidar_imu from sensor data.
IMU samples are buffered and replayed after init.
- Fix deferred init for scan-only messages that arrive after IMU
local transform is already cached.
- Add LIO-SAM entry to odometry strategy combo box (index 14) with
full PreferencesDialog panel including config path browse button
and all parameter widgets.
* OdometryLIOSAM: propagate deskewed scan to SensorData
Capture the deskewed cloud produced by LIO-SAM's image projection
stage and replace the raw scan on SensorData with it, so loop closure
registration and other downstream stages operate on the motion-
compensated points instead of the raw pre-deskew input.
* Minor updates for rtabmap_ros
---------
Co-authored-by: matlabbe <matlabbe@gmail.com>
* New rtabmap-reduceGraph CLI tool
* fixed some edge cases
* Regenerating optimized map if there was one before reducing the graph
* addMoreLoopClosures: refactored how ctrl-c is handled to stop faster when no loop closures are added
* Added kilted status
* Make offline tool always propagate neighbor merged links
* removed a parameter
* fixed disconnected graph
* fixed --help
* Added error log on Kp/NNStrategy not compatible with huge vocabulary. ReduceGraph/DetectMoreLoopClosures: Make sure original parameters are saved back on closing. g2o: fixing optimizer to Levenberg for SBA to avoid [SetJac] infinite jac fatal error.
* exposing neighbor merged ratio parameter to the tool
* show param in log
* refactored detectMoreLoopClosures to ignore too close nodes in terms of neighbor links based on Mem/STMSize parameter. Reduce graph: added direction parameter.
* Simplified: removed ratio parameter, removed recursive reduction. Just don't reduce if a NM link is longer than maxDistance.
* Removed NNStrategy override, as it was still done on closing when we changed back to original params
* DBViewer: show missing links when showing OptimizedPoses in GraphView, fixed clicking on landmark links
* DetectMoreLoopClosures: Added support for min graph distance option in MainWindow and DbViewer
* slight renaming of ROS jobs
* reprocess: added option --params_last
* CudaSIFT: filter doubles
* removed fixed threshold
* SSC can be used with multicameras. Refactored CudaSIFT to support SSC. Add new parameter SIFT/MaxGaussianThreshold. DbViewer: show negative features with gray color (so that we can know which features are in the vocabulary)
* Added SIFT/MaxGaussianThreshold parameter
* Updated parameter description
* Detect more loop closures from/to specific map ID only option
* default value -1
* Added elapsed time
* Avoid loading ALL signatures in RAM, only load them when necessary
* Avoid creating flann index when initializing with Kp/NNStrategy>=3
* Setting opt params by default
* Read-only localization mode
* Localization: don't update map's node timestamps if statistic after run are not saved
* Updated Mem/LocalizationReadOnly description
* Updated description of Kp/FlannIndexSaved
* rtabmap-rgbd_camera: added width/height options,
* using CameraViewer for visualization by default (use -pcl option for old visualization).
* CameraViewer: added capture and render rate.
* Note that non-TOF orbbec cameras could not be tested.
* update cuvslam api call to adhere to new v4.0 changes
* update comment
* parse cuvslam major version from the header and throw an error if theversion doesn't match the expected
* moving average filter + extensive debug logging
* lots more logs
* improvement by turning off use_motion_model in cuvslam config. lots of debug logs.
* slight improvements to velocity guard logic. lots of debug logs
* removing debug logs. working.
* let find_package_handle_standard_args handle the version comparison logic. Custom logic to handle newer then requested major version
* proper warning message if the major version doesn't match
* add back debug logs. Handle no guess transform case.
* adding warnings when no guess transform is provided
* clean up comments
* lowering average window to 5
* moderate odometry mode, 5 value averaging window
* adding comment
* add reset logs to investigate reset issue
* try returning guess or identity on init to avoid reset cascade
* disable motion model again
* big log when tracking
* remove some debug logs
* flag to use original covariance from cuVSLAM
* use raw covariance
* data collection for revised lost detection
* compare guess to estimated transform for lost detection
* re-enable motion model, it appears to reduce false-positives for the new lost detection
* run pose estimation afer init on first frame
* don't initialize if we don't have enough features
* dont warm up GPU everytime we reset
* low estimate velocity guard. Passes all challenges. Debug comments still everywhere.
* Remove debug logs. Add thresholds to header in a config section.
* Cleanup logs
* exposing multi-cam mode as a rtabmap param
* remove max frame delta
* change to use assertion instead of conditional
* replaced some UERROR by UWARN
---------
Co-authored-by: Felix Toft <felix@robust.ai>
Co-authored-by: matlabbe <matlabbe@gmail.com>
* Working IMU_RGBD and IMU_STEREO. Changing Inter IMU base frame is now possible.
* Output error msg when vocabulary path is wrong.
* Showing full local feature map
* Prioritize neighbor links in Optimizer::getConnectedGraph() to avoid odometry jumps.
* Fixed landmark order
* report: added search for first valid id
* Fixed non-neighbor comparison logic
* RGBD/LocalizationPriorError can now be 0 (disabled) to avoid using priors to fix the graph (https://github.com/introlab/rtabmap_ros/issues/1371)
* amend previous commit
* initial python implementation of superpoint rpautrat
* working python implementation of superpoint
* small tweaks to try and work around the GIL issue
* fix missing os import
* begging cpp impl of superpoint python model
* finishing cpp superpoint impl using the same SPDetector interface
* finalizing superpoint cpp impl, working with cpu but needs to be cleaned
* fixing feature matching by reworking nms and filtering logic
* speeding up nms with batched operations and cleaning up
* updating conversion script
* adding args for image dimensions and cuda usage to model tracer
* wiring up UI to parse superpoint params
* adding label to superpoint rpautrat ui
* reverting some unintended ui changes
* typo
* oneline revert
* removing nms and threshold filtering from cpp, this is handled internally by the superpoint model
* using python interface to run the superpoint _to_torchscript.py script at runtime
* generate and load model file on the first incoming frame
* reverting unintented change
* ui changes appear mysteriously again, reverting
* remove topk from model to prevent issue when number of detected keypoints is lower then the k value (scripting fails in this case)
* cleaning up for review
* change dest for model file and remove debug logs
* bump patch and add version comment
* rm unucessary comment
* use resources to load file to ensure it works when rtabmap isn't built from source
* execute with pybind runpy instead of system call
* only build superpoint rpautrat if we have torch and python support
* changing param to accept a path to the weights .pth file directly
* parse the default working directory to save the model file in
* rm extraneous change
* rm setters and re-initialize the detector whenever params change. We cannot support changing params after the model is constructed
* update UI text
* only enable superpoint rpautrat when built with python and torch
* more build information regarding superpoint rpautrat
* generate temporary python script in the working dir
* rm unecessary changes to rtabmap_superpoint.py
* fix comment and only add repo root to sys.path
* introduce a new parameter for the superpoint python model definition
* execute script from string instead of writing to a file
* wrap parse params in a single compiler directive
* remove descriptor spatial matching logic and rely on upstream RTAB-Map processes to take the top-K desc and kpts
* only add python script to resources if built with superpoint rpautrat support. supress warning with type casting
* isolate pybind11 setup so it can't affect any other modules or potential regnerations of the model file
* update about dialog to show superpoint rpautrat
* remove debug logs
---------
Co-authored-by: Felix Toft <felix@robust.ai>
* Orbbec SDK 2 integration (new CameraOrbbecSDK class)
* Added rectification option, fixed the build without orbbec sdk
* Fixed timestamp issue on Windows
* Removed WITH_ORBBEC_SDK=ON by default in case v1 is installed. Enabled inter imu when selecting orbbec sdk.
* requiring orbbec version=2 to avoid confusion
* Finalizing options
* integrating cuvslam, stereo cam initialization
* fixing transformations
* use isaac ros TocuVSLAMPose
* organizing
* handling covariance conversion
* cleaning up config
* moving cmake instructions to FindCuVSLAM.cmake file
* fixing ui integration
* updating software license for cuvslam
* improving memory management
* cleaning up header more
* reverting mistake in GUI
* using opaque pointers for header definitions, compiles
* back to typed definitions using forward-referencing
* fixing dangling pointer
* cleaning up
* fixing variable scope issues
* Using cuda to allocate gpu memory
* fixing transfrom from cuvslam back to ros coordinate frame
* reducing excessive logs
* removing redundant if true in cuvslam cmake
* moving find cuda into cuvslam cmake
* cleaning cmake to use modern target
* adding copyright line
* Finding tf2 and eigen3 in cmake
* simplifying clean up logic
* better logs
* moving cuvslam implementation methods into cpp
* fixing build with implmentation function declarations moved to the cpp
* cleaning transformation logic and adding grayscale support
* typo
* removing unecessary class member variable for processed images
* remove trailing underscore from initialize cuvslam function signature
* locally scoping config params since they are copied by cuvslam
* removing unecessary member variables and using locally scoped values since cuvslam seems to copy them
* more cleaning and refactoring
* fixing cmake inconsistancies and surpress warning
* fixing up baseline transform
* trying to remove tf2 for transformations
* fixing incorect baseline transform swap
* comment, begin reset logic investigation
* better handling of previous pose and transform during error cases
* hunting for reset bug, not found
* observation export implementatin, currently getting strange errors
* multi cam support
* Memory fixes
* comments
* make cuda stream a member variable so it isn't initialized and destroyed on every frame
* preparing to add ground constraints
* adding ground constraints, working
* adding wheel odom fusion into cuvslam tracker
* parsing parameter to apply planar constraints
* removing planar constraint in config, doesn't do anything
* using proper cuda stream type for member variable
* cleaning up for merge
* only parse params when cuvslam is compiled
* updating error message if using rgb-d instead of stereo
* removing wheel odom testing logs
* reverting bgr to rgb conversion
* fixing initial pose logic
* return null transform on invalid covariance
* debugging covariance
* guard against low velocity situations where cuvslam covariance spikes, but we aren't lost
* cleaning things up
---------
Co-authored-by: Felix Toft <felix@robust.ai>
Co-authored-by: matlabbe <matlabbe@gmail.com>
* On init, rebuild the dictionnary only once
* Fixed first dictionary update to avoid rebuilding multiple times. Commented some very verbose debug logs (should create a new level: UVERBOSE or UTRACE)
* bump version 0.23: added parameters "Mem/LoadVisualLocalFeaturesOnInit", "Mem/FlannIndexSaved", "Kp/SerializeWithChecksum".
* Renamed Mem/FlannIndexSaved to Kp/FlannIndexSaved. Update UI Preferences with new parameters.
* commented some very verbose debug logs
* Warn flann index serialization not implemented on windows
* Warn flann index deserialization not implemented on windows
* Removing some verbose logs
* missing header (win32)
* Added GlobalMap::fullUpdateNeeded() function
* Adding log
* Save flann index even if links changed
* log time to serialize flann index
* Make dictionary modified only after we update
---------
Co-authored-by: Mathieu Labbe <mathieu@robust.ai>
* Added parameter RGBD/LocalizationSecondTryWithoutProximityLinks
* Statistics: added new stat to know when proximity links are removed after a successful second loc try
* changed kLoopProximity_links_cleared for a count instead of bool
* Updating odomCachePoses with optimization for better visualization
* Using shadows on all texts and icons to see them better when background is white. Buffering database previews to show faster the Library. Updated how RAM usage is computed and now show max memory. Added warnings when RAM usage is low.
* Added measuring tool
* fixed map not shown when closing visualization
* Added measure text size setting, changed look and feel of point to point
* Fixed measuring point not showing
* Added fix for #1401
* Addressing #1407
* Export: Added colored OBJ options
* fixed index_t not existing on focal
* Fixed Settings not applied after restoring to all default settings witohut restarting the app (Marker detection not working issue #1455 )
* Marker detection: Fixing wrong depth used when rgb and depth image sizes cannot be divided exactly one from the the other #1455
* Added Marker Max Range option (default 1m)
* Added OBJ texture policy option (keep color on textureless polygons)
* Added texture/color blending option directly in the app. Added LAZ export option.
* Hiding measuring button if not mesh, dont zip if exporting to laz, updating Export XXX button based on the current context, fixed always blending texture/color on not visualization mode
* updated default marker max range to 2m
* bump ios app version
* fixing pcl 1.8.1 build
* fixed android build
* Reorganize DepthAI params
* Advanced Tuning of DepthAI’s SGBM Pipeline
* Update IMU Local Transform for OAK-D LR, OAK-D SR, and PoE Devices
* Reorganiz the usage of DepthAI. Get params before creating the pipeline, and automatically determine the resolution of different models.
* Identify the IMU type to specifically handle BMI270 and BNO085/086
* Added OdomF2M/BundleAdjustmentMinMotion and OdomF2M/BundleAdjustmentMaxKeyFramesPerFeature parameters
* Added new odom stats localBundleMaxKeyFramesForInlier
* Adjusted how localBundleMaxKeyFramesForInlier is computed when OdomF2M/BundleAdjustmentMaxFrames is used.
* optimization: Dont compute average pixel distance if parameter is not used
