boofcv.alg.feature.orientation.impl

Class ImplOrientationAverageGradientIntegral<T extends ImageSingleBand,G extends GradientValue>

java.lang.Object

boofcv.alg.feature.orientation.OrientationIntegralBase<T,G>

boofcv.alg.feature.orientation.impl.ImplOrientationAverageGradientIntegral<T,G>

All Implemented Interfaces:

OrientationIntegral<T>, RegionOrientation

public class ImplOrientationAverageGradientIntegral<T extends ImageSingleBand,G extends GradientValue>
extends OrientationIntegralBase<T,G>

Estimates the orientation of a region by computing the image derivative from an integral image. The derivative along each axis is summed up and the angle computed from that.

Author:

Peter Abeles

Field Summary

Fields inherited from class boofcv.alg.feature.orientation.OrientationIntegralBase

g, ii, period, radius, sampleWidth, scale, weights, width

Constructor Summary

Constructors

Constructor and Description
ImplOrientationAverageGradientIntegral(int radius, double period, int sampleWidth, double weightSigma, Class<T> imageType)

Method Summary

Methods

Modifier and Type	Method and Description
double	compute(double c_x, double c_y) Computes the orientation of a region about its center.
protected double	computeUnweighted(double tl_x, double tl_y, double samplePeriod, SparseImageGradient<T,G> g) Compute the gradient while checking for border conditions
protected double	computeWeighted(double tl_x, double tl_y, double samplePeriod, SparseImageGradient<T,G> g) Compute the gradient while checking for border conditions

Methods inherited from class boofcv.alg.feature.orientation.OrientationIntegralBase

getImageType, setImage, setScale

Methods inherited from class java.lang.Object

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

Constructor Detail

ImplOrientationAverageGradientIntegral

public ImplOrientationAverageGradientIntegral(int radius,
                                      double period,
                                      int sampleWidth,
                                      double weightSigma,
                                      Class<T> imageType)

Parameters:

radius - Radius of the region being considered in terms of Wavelet samples. Typically 6.

weightSigma - Sigma for weighting distribution. Zero for unweighted.

Method Detail

compute

public double compute(double c_x,
             double c_y)

Description copied from interface: RegionOrientation

Computes the orientation of a region about its center.

Parameters:

c_x - Center of the region in image pixels.

c_y - Center of the region in image pixels.

Returns:

Orientation in radians. Angle zero points along x-axis and pi/2 along y-axis.

computeUnweighted

protected double computeUnweighted(double tl_x,
                       double tl_y,
                       double samplePeriod,
                       SparseImageGradient<T,G> g)

Compute the gradient while checking for border conditions

computeWeighted

protected double computeWeighted(double tl_x,
                     double tl_y,
                     double samplePeriod,
                     SparseImageGradient<T,G> g)

Compute the gradient while checking for border conditions