boofcv.alg.transform.pyramid

Class PyramidDiscreteSampleBlur<T extends ImageGray>

java.lang.Object

boofcv.struct.pyramid.ImagePyramidBase<T>

boofcv.struct.pyramid.PyramidDiscrete<T>

boofcv.alg.transform.pyramid.PyramidDiscreteSampleBlur<T>

All Implemented Interfaces:

ImagePyramid<T>

public class PyramidDiscreteSampleBlur<T extends ImageGray>
extends PyramidDiscrete<T>

Convolves a re-normalizable blur kernel across the image before down sampling. This is useful for creating a Gaussian pyramid as well as other standard pyramids.

NOTE: This pyramid cannot be configured such that blur will be applied to the input image. It can be done by the user before the image is passed in.

NOTE: This blur magnitude is constant for each level in the pyramid. In general it is desirable to have it dependent on each level's scale factor.

Field Summary

Fields inherited from class boofcv.struct.pyramid.PyramidDiscrete

scale

Fields inherited from class boofcv.struct.pyramid.ImagePyramidBase

layers

Constructor Summary

Constructors

Constructor and Description
PyramidDiscreteSampleBlur(Kernel1D kernel, double sigma, java.lang.Class<T> imageType, boolean saveOriginalReference, int... scaleFactors)

Method Summary

Modifier and Type	Method and Description
double	getSampleOffset(int layer) There is no offset since a symmetric kernel is applied starting at pixel (0,0)
double	getSigma(int layer) Returns the scale-space scale for the specified layer.
void	process(T input) Constructs the image pyramid given the input image.

Methods inherited from class boofcv.struct.pyramid.PyramidDiscrete

getNumLayers, getScale, getScales, setScaleFactors

Methods inherited from class boofcv.struct.pyramid.ImagePyramidBase

getHeight, getImageType, getInputHeight, getInputWidth, getLayer, getWidth, initialize, isSaveOriginalReference, setFirstLayer, setTo

Methods inherited from class java.lang.Object

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

Constructor Detail

PyramidDiscreteSampleBlur

public PyramidDiscreteSampleBlur(Kernel1D kernel,
                                 double sigma,
                                 java.lang.Class<T> imageType,
                                 boolean saveOriginalReference,
                                 int... scaleFactors)

Parameters:

kernel - A blur kernel

sigma - The effective amount of Gaussian blur the kernel applies

imageType - Type of image processed

saveOriginalReference - If a reference to the full resolution image should be saved instead of copied. Set to false if you don't know what you are doing.

scaleFactors - Scale factor for each layer in the pyramid relative to the input layer

Method Detail

process

public void process(T input)

Description copied from interface: ImagePyramid

Constructs the image pyramid given the input image.

getSampleOffset

public double getSampleOffset(int layer)

There is no offset since a symmetric kernel is applied starting at pixel (0,0)

Parameters:

layer - Layer in the pyramid

Returns:

offset

getSigma

public double getSigma(int layer)

Description copied from interface: ImagePyramid

Returns the scale-space scale for the specified layer. This scale is equivalent amount of Gaussian blur applied to the input image. If Gaussian blur is not applied to each layer then an approximation should be returned.

Parameters:

layer - Layer in the pyramid

Returns:

Equivalent sigma for Gaussian blur.