boofcv.alg.distort

Class LensDistortionOps

java.lang.Object

boofcv.alg.distort.LensDistortionOps

public class LensDistortionOps
extends Object

Operations related to manipulating lens distortion in images

Author:

Peter Abeles

Constructor Summary

Constructors

Constructor and Description
LensDistortionOps()

Method Summary

Methods

Modifier and Type	Method and Description
static PointTransform_F32	allInside(IntrinsicParameters param, IntrinsicParameters paramAdj) Adjusts the view such that each pixel has a correspondence to the original image while maximizing the view area.
static georegression.struct.shapes.Rectangle2D_F32	boundBoxInside(int srcWidth, int srcHeight, PixelTransform_F32 transform) Finds the maximum area axis-aligned rectangle contained inside the transformed image which does not include any pixels outside the sources border.
static PointTransform_F32	fullView(IntrinsicParameters param, IntrinsicParameters paramAdj) Transforms the view such that the entire original image is visible after lens distortion has been removed.
static <T extends ImageSingleBand> ImageDistort<T>	removeRadialImage(IntrinsicParameters param, BorderType borderType, Class<T> imageType) Creates an which removes radial distortion.
static void	roundInside(georegression.struct.shapes.Rectangle2D_F32 bound) Adjust bound to ensure the entire image is contained inside, otherwise there might be single pixel wide black regions
static PointTransform_F64	transformNormToRadial_F64(IntrinsicParameters param) Converts normalized image coordinates into distorted pixel coordinates.
static PointTransform_F32	transformPixelToRadial_F32(IntrinsicParameters param) Transform from undistorted pixel coordinates to distorted with radial pixel coordinates
static PointTransform_F64	transformRadialToNorm_F64(IntrinsicParameters param) Removes radial distortion from the image in pixel coordinates and converts it into normalized image coordinates
static PointTransform_F64	transformRadialToPixel_F64(IntrinsicParameters param) Removes radial distortion from the pixel coordinate.

Methods inherited from class java.lang.Object

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

Constructor Detail

LensDistortionOps

public LensDistortionOps()

Method Detail

fullView

public static PointTransform_F32 fullView(IntrinsicParameters param,
                          IntrinsicParameters paramAdj)

Transforms the view such that the entire original image is visible after lens distortion has been removed. The appropriate PointTransform_F32 is returned and a new set of intrinsic camera parameters for the "virtual" camera that is associated with the returned transformed.

The original image coordinate system is maintained even if the intrinsic parameter flipY is true.

Parameters:

param - Intrinsic camera parameters.

paramAdj - If not null, the new camera parameters are stored here.

Returns:

New transform that adjusts the view and removes lens distortion..

allInside

public static PointTransform_F32 allInside(IntrinsicParameters param,
                           IntrinsicParameters paramAdj)

Adjusts the view such that each pixel has a correspondence to the original image while maximizing the view area. In other words no black regions which can cause problems for some image processing algorithms.

The original image coordinate system is maintained even if the intrinsic parameter flipY is true.

Parameters:

param - Intrinsic camera parameters.

paramAdj - If not null, the new camera parameters are stored here.

Returns:

New transform that adjusts the view and removes lens distortion..

transformRadialToNorm_F64

public static PointTransform_F64 transformRadialToNorm_F64(IntrinsicParameters param)

Removes radial distortion from the image in pixel coordinates and converts it into normalized image coordinates

Parameters:

param - Intrinsic camera parameters

Returns:

Distorted pixel to normalized image coordinates

transformRadialToPixel_F64

public static PointTransform_F64 transformRadialToPixel_F64(IntrinsicParameters param)

Removes radial distortion from the pixel coordinate.

Parameters:

param - Intrinsic camera parameters

Returns:

Transformation into undistorted pixel coordinates

transformNormToRadial_F64

public static PointTransform_F64 transformNormToRadial_F64(IntrinsicParameters param)

Converts normalized image coordinates into distorted pixel coordinates.

Parameters:

param - Intrinsic camera parameters

Returns:

Transform from normalized image coordinates into distorted pixel coordinates

transformPixelToRadial_F32

public static PointTransform_F32 transformPixelToRadial_F32(IntrinsicParameters param)

Transform from undistorted pixel coordinates to distorted with radial pixel coordinates

NOTE: The original image coordinate system is maintained even if the intrinsic parameter flipY is true.

Parameters:

param - Intrinsic camera parameters

Returns:

Transform from undistorted to distorted image.

removeRadialImage

public static <T extends ImageSingleBand> ImageDistort<T> removeRadialImage(IntrinsicParameters param,
                                                            BorderType borderType,
                                                            Class<T> imageType)

Creates an which removes radial distortion. How pixels outside the image are handled is specified by the BorderType. If BorderType.VALUE then pixels outside the image will be filled in with a value of 0. For viewing purposes it is recommended that BorderType.VALUE be used and BorderType.EXTENDED in computer vision applications. VALUE creates harsh edges which can cause false positives when detecting features, which EXTENDED minimizes.

Parameters:

param - Intrinsic camera parameters

imageType - Type of single band image being processed

borderType - Specifies how the image border is handled.

Returns:

Image distort that removes radial distortion

boundBoxInside

public static georegression.struct.shapes.Rectangle2D_F32 boundBoxInside(int srcWidth,
                                                         int srcHeight,
                                                         PixelTransform_F32 transform)

Finds the maximum area axis-aligned rectangle contained inside the transformed image which does not include any pixels outside the sources border. Assumes that the coordinates are not flipped and some other stuff too.

Parameters:

srcWidth - Width of the source image

srcHeight - Height of the source image

transform - Transform being applied to the image

Returns:

Bounding box

roundInside

public static void roundInside(georegression.struct.shapes.Rectangle2D_F32 bound)

Adjust bound to ensure the entire image is contained inside, otherwise there might be single pixel wide black regions