Package boofcv.alg.sfm.d3

Class VisOdomMonoOverheadMotion2D<T extends ImageBase<T>>

java.lang.Object
boofcv.alg.sfm.d3.VisOdomMonoOverheadMotion2D<T>
public class VisOdomMonoOverheadMotion2D<T extends ImageBase<T>> extends Object
Estimates the motion of a monocular camera using the known transform between the camera and the ground plane. The camera's image is converted into an orthogonal overhead view. Features are tracked inside the overhead image and the 2D rigid body motion found. The output can be either the 2D motion or the 3D motion of the camera. There is no scale ambiguity since the transform from the plane to camera is known. The advantage of tracking inside the overhead view instead of the camera image is that the overhead view lacks perspective distortion and is crops the image which does not contain the plane. Features without perspective distortion are easier to track and many false positives are removed by removing many features not on the plane. The plane which is being viewed is defined by the 'planeToCamera' transform. In the plane's reference frame the plane lies along the x-z axis and contains point (0,0,0). See CreateSyntheticOverheadView for more information about the ground plane coordinates and overhead image..

  • Constructor Details

    • VisOdomMonoOverheadMotion2D

      public VisOdomMonoOverheadMotion2D(double cellSize, double maxCellsPerPixel, double mapHeightFraction, ImageMotion2D<T,Se2_F64> motion2D, ImageType<T> imageType)
      Configures motion estimation algorithm.
      Parameters:
      cellSize - Size of cells in plane in world units
      maxCellsPerPixel - Specifies minimum resolution of a region in overhead image. A pixel in the camera can't overlap more than this number of map cells. Higher values allow lower resolution regions. Try 4.
      mapHeightFraction - Reduce the map height by this fraction to avoid excessive unusable image space. Set to 1.0 to maximize the viewing area and any value less than one to crop it.
      motion2D - Estimates motion inside the overhead image.

  • Method Details

    • configureCamera

      public void configureCamera(CameraPinholeBrown intrinsic, Se3_F64 planeToCamera)
      Camera the camera's intrinsic and extrinsic parameters. Can be called at any time.
      Parameters:
      intrinsic - Intrinsic camera parameters
      planeToCamera - Transform from the plane to camera.

    • reset

      public void reset()
      Resets the algorithm into its initial state

    • process

      public boolean process(T image)
      Estimates the motion which the camera undergoes relative to the first frame processed.
      Parameters:
      image - Most recent camera image.
      Returns:
      true if motion was estimated or false if a fault occurred. Should reset after a fault.

    • getWorldToCurr2D

      public Se2_F64 getWorldToCurr2D()
      2D motion.
      Returns:
      from world to current frame.

    • getWorldToCurr3D

      public Se3_F64 getWorldToCurr3D()
      3D motion.
      Returns:
      from world to current frame.

    • getOverhead

      public OverheadView<T> getOverhead()
      Overhead image view

    • getMotion2D

      public ImageMotion2D<T,Se2_F64> getMotion2D()
      2D motion algorithm