boofcv.alg.sfm

Class DepthSparse3D<T extends boofcv.struct.image.ImageGray<T>>

java.lang.Object

boofcv.alg.sfm.DepthSparse3D<T>

Direct Known Subclasses:

DepthSparse3D.F32, DepthSparse3D.I

public abstract class DepthSparse3D<T extends boofcv.struct.image.ImageGray<T>>
extends java.lang.Object

Computes the 3D coordinate a point in a visual camera given a depth image. The visual camera is a standard camera while the depth camera contains the depth (value along z-axis) of objects inside its field of view. The Kinect (structured light) and flash ladar (time of flight) are examples of sensors which could use this class. The z-axis is defined to be pointing straight out of the visual camera and both depth and visual cameras are assumed to be parallel with identical pointing vectors for the z-axis. A mapping is provided for converting between pixels in the visual camera and the depth camera. This mapping is assumed to be fixed with time.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	DepthSparse3D.F32 Implementation for GrayF32.
static class	DepthSparse3D.I<T extends boofcv.struct.image.GrayI<T>> Implementation for GrayI.

Field Summary

Fields

Modifier and Type	Field and Description
protected T	depthImage

Constructor Summary

Constructors

Constructor and Description
DepthSparse3D(double depthScale) Configures parameters

Method Summary

Modifier and Type	Method and Description
void	configure(boofcv.alg.distort.LensDistortionNarrowFOV model, boofcv.struct.distort.PixelTransform<georegression.struct.point.Point2D_F32> visualToDepth) Configures intrinsic camera parameters
georegression.struct.point.Point3D_F64	getWorldPt() The found 3D coordinate of the point in the visual camera coordinate system.
protected abstract double	lookupDepth(int depthX, int depthY) Internal function which looks up the pixel's depth.
boolean	process(int x, int y) Given a pixel coordinate in the visual camera, compute the 3D coordinate of that point.
void	setDepthImage(T depthImage) Sets the depth image.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

depthImage

protected T extends boofcv.struct.image.ImageGray<T> depthImage

Constructor Detail

DepthSparse3D

public DepthSparse3D(double depthScale)

Configures parameters

Parameters:

depthScale - Used to change units found in the depth image.

Method Detail

configure

public void configure(boofcv.alg.distort.LensDistortionNarrowFOV model,
                      boofcv.struct.distort.PixelTransform<georegression.struct.point.Point2D_F32> visualToDepth)

Configures intrinsic camera parameters

Parameters:

model - Model for narrow FOV cameras

visualToDepth - Transform from visual to depth camera pixel coordinate systems.

setDepthImage

public void setDepthImage(T depthImage)

Sets the depth image. A reference is saved internally.

Parameters:

depthImage - Image containing depth information.

process

public boolean process(int x,
                       int y)

Given a pixel coordinate in the visual camera, compute the 3D coordinate of that point.

Parameters:

x - x-coordinate of point in visual camera

y - y-coordinate of point in visual camera

Returns:

true if a 3D point could be computed and false if not

getWorldPt

public georegression.struct.point.Point3D_F64 getWorldPt()

The found 3D coordinate of the point in the visual camera coordinate system. Is only valid when process(int, int) returns true.

Returns:

3D coordinate of point in visual camera coordinate system

lookupDepth

protected abstract double lookupDepth(int depthX,
                                      int depthY)

Internal function which looks up the pixel's depth. Depth is defined as the value of the z-coordinate which is pointing out of the camera. If there is no depth measurement at this location return 0.

Parameters:

depthX - x-coordinate of pixel in depth camera

depthY - y-coordinate of pixel in depth camera

Returns:

depth at the specified coordinate