Documentation of 'boofcv.alg.sfm.d3.VisOdomDualTrackPnP' Java class
VisOdomDualTrackPnP
boofcv.alg.sfm.d3

Class VisOdomDualTrackPnP<T extends ImageBase,Desc extends TupleDesc>



  • public class VisOdomDualTrackPnP<T extends ImageBase,Desc extends TupleDesc>
    extends java.lang.Object
    Stereo visual odometry algorithm which relies on tracking features independently in the left and right images and then matching those tracks together. The idea behind this tracker is that the expensive task of association features between left and right cameras only needs to be done once eat time a track is spawned. Triangulation is used to estimate each feature's 3D location. Motion is estimated robustly using a RANSAC type algorithm provided by the user which internally uses PnP type algorithm. Estimated motion is relative to left camera.
    • Constructor Detail

      • VisOdomDualTrackPnP

        public VisOdomDualTrackPnP(int thresholdAdd,
                                   int thresholdRetire,
                                   double epilolarTol,
                                   PointTracker<T> trackerLeft,
                                   PointTracker<T> trackerRight,
                                   DescribeRegionPoint<T,Desc> describe,
                                   AssociateDescription2D<Desc> assocL2R,
                                   TriangulateTwoViewsCalibrated triangulate,
                                   ModelMatcher<Se3_F64,Stereo2D3D> matcher,
                                   ModelFitter<Se3_F64,Stereo2D3D> modelRefiner)
        Specifies internal algorithms and parameters
        Parameters:
        thresholdAdd - When the number of inliers is below this number new features are detected
        thresholdRetire - When a feature has not been in the inlier list for this many ticks it is dropped
        epilolarTol - Tolerance in pixels for enforcing the epipolar constraint
        trackerLeft - Tracker used for left camera
        trackerRight - Tracker used for right camera
        describe - Describes features in tracks
        assocL2R - Assocation for left to right
        triangulate - Triangulation for estimating 3D location from stereo pair
        matcher - Robust motion model estimation with outlier rejection
        modelRefiner - Non-linear refinement of motion model
    • Method Detail

      • reset

        public void reset()
        Resets the algorithm into its original state
      • process

        public boolean process(T left,
                               T right)
        Updates motion estimate using the stereo pair.
        Parameters:
        left - Image from left camera
        right - Image from right camera
        Returns:
        true if motion estimate was updated and false if not
      • getCurrToWorld

        public Se3_F64 getCurrToWorld()
      • getTick

        public int getTick()
      • isFault

        public boolean isFault()
        If there are no candidates then a fault happened.
        Returns:
        true if fault. false is no fault
      • getCandidates

        public java.util.List<PointTrack> getCandidates()
        Returns a list of active tracks that passed geometric constraints

DataMelt 3.0 © DataMelt by jWork.ORG

You see the box below because you did not login.