JoaoSimoes97
commented
2 years ago Hi Patrick, I don't have the time right now, but I have used ucoslam for a year and I can offer my help next week if you want to. I will try and run your video in my computer in the meantime, but I don't think I can do that before Monday. I will get back to you as soon as I can.
Best regards, João Simões
A terça, 17/05/2022, 23:02, Patrik Vácal @.***> escreveu:
First of all, thanks for such a nice library (and easy build procedure).
I'm currently trying to use ucoslam for tracking a self-driving car.
I have a problem: In the beginning it is tracking correctly, but then it failes both orb tracking and adding new aruco tags.
This is how it looks: https://youtu.be/BlF6uOruMDk
My ucoslam version is 1.1.0. My config is:
%YAML:1.0
kpDescriptorType: orb KPNonMaximaSuppresion: 0 targetFocus: -1. minKFConfidence: 6.0000002384185791e-01 maxNewPoints: 700 enableLoopClosure: 1 targetWidth: 1. maxFeatures: 4000 nthreads_feature_detector: 2 nOctaveLevels: 8 scaleFactor: 1.2000000476837158e+00 runSequential: 1 detectMarkers: 1 # Enable / disable markers detection forceInitializationFromMarkers: 0 aruco_allowOneFrameInitialization: 0 reLocalizationWithKeyPoints: 1 reLocalizationWithMarkers: 1 inPlaneMarkers: 0 saveImageInMap: 0 minBaseLine: 7.0000000298023224e-02 detectKeyPoints: 1 # Enable / Disable keypoint detecion removeKeyPointsIntoMarkers: 1 minDescDistance: 3.4028234663852886e+38 baseline_medianDepth_ratio_min: 9.9999997764825821e-03 aruco_markerSize: 0.187 # TODO projDistThr: 15 maxVisibleFramesPerMarker: 10 minNumProjPoints: 3 keyFrameCullingPercentage: 8.0000001192092896e-01 thRefRatio: 8.9999997615814209e-01 aruco_minerrratio_valid: 3. aruco_minNumFramesRequired: 3 markersOptWeight: 5.0000000000000000e-01 minMarkersForMaxWeight: 5 autoAdjustKpSensitivity: 0 global_optimizer: g2o aruco-dictionary: ARUCO_MIP_36h12 aruco-detectMode: DM_NORMAL aruco-cornerRefinementM: CORNER_SUBPIX aruco-minSize: 0.
kptImageScaleFactor: 0.5 # Scale down the image after Aruco detection, before it goes to ORB features detection
extraParams: ""
It doesn't work even in the GUI with this config: [image: image] https://user-images.githubusercontent.com/11137519/168914708-be74e4fa-e84e-4e3a-b051-2f08eb754d45.png
My camera calibration (I got this one using ROS camera calibrator): https://drive.google.com/file/d/1QjSpXUSDGssILD3-hXZn2GIlBV6hrN5f/view?usp=sharing
Data I tested it on: as mp4: https://drive.google.com/file/d/1k7shXZJbzylLkF99xo_W0awmMfE0Mf48/view?usp=sharing as .bag file: https://drive.google.com/file/d/1GmokkLaOXp5ilVwc89UA7HTPtaIacadi/view?usp=sharing
— Reply to this email directly, view it on GitHub https://github.com/lambdaloop/ucoslam-cv3/issues/4, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEVKXJTJTFNPTAVDQBIU643VKQJPRANCNFSM5WGKUEFA . You are receiving this because you are subscribed to this thread.Message ID: @.***>
gamecraftCZ
ZitongLan
First of all, thanks for such a nice library (and easy build procedure).
I'm currently trying to use ucoslam for tracking a self-driving car.
I have a problem: In the beginning it is tracking correctly, but then it failes both orb tracking and adding new aruco tags.
This is how it looks: https://youtu.be/BlF6uOruMDk
My ucoslam version is 1.1.0. My config is:
It doesn't work even in the GUI with this config:
My camera calibration (I got this one using ROS camera calibrator): https://drive.google.com/file/d/1QjSpXUSDGssILD3-hXZn2GIlBV6hrN5f/view?usp=sharing
Data I tested it on: as mp4: https://drive.google.com/file/d/1k7shXZJbzylLkF99xo_W0awmMfE0Mf48/view?usp=sharing as .bag file: https://drive.google.com/file/d/1GmokkLaOXp5ilVwc89UA7HTPtaIacadi/view?usp=sharing