ImageBasisConvolutionKernel

class lsst.images.convolution_kernels.ImageBasisConvolutionKernel(basis: ndarray, spatial: Iterable[Field], center_y: int | None = None, center_x: int | None = None)

Bases: ConvolutionKernel

A convolution kernel formed by a linear combination of images multiplied by BaseField instances.

Parameters

basis

A 3-d array holding the kernel images each basis function, with shape (n, height, width).

spatial

Iterable of fields.BaseField of length basis.shape[0], holding the spatial variation of each basis kernel.

center_y

Center of the basis kernels in the x dimension. Defaults to height//2.

center_x

Center of the basis kernels in the x dimension. Defaults to width//2.

Attributes Summary

basis	The kernel basis functions, as an array with shape (n, h, w) (numpy.ndarray).
bounds	The region where this convolution kernel is valid (Bounds).
kernel_bbox	Bounding box of all images returned by compute_kernel_image (Box).
spatial	The spatial variation of each basis function (Sequence [BaseField]).

Methods Summary

compute_kernel_image(*, x, y)	Evaluate the kernel at a point.
from_legacy(legacy_kernel)	Convert from a legacy lsst.afw.math.LinearCombinationKernel.
serialize(archive)	Serialize the kernel to an output archive.
to_legacy()	Convert to a legacy lsst.afw.math.LinearCombinationKernel.

Attributes Documentation

basis

The kernel basis functions, as an array with shape (n, h, w) (numpy.ndarray).

bounds

kernel_bbox

spatial

The spatial variation of each basis function (Sequence [BaseField]).

Methods Documentation

compute_kernel_image(*, x: int, y: int) → Image

Evaluate the kernel at a point.

Parameters

x

Column position coordinate to evaluate at.

y

Row position coordinate to evaluate at.

Returns

Image

An image of the kernel, centered on the center of the center pixel, which is defined to be (0, 0) by the image’s origin.

static from_legacy(legacy_kernel: LegacyLinearCombinationKernel) → ImageBasisConvolutionKernel

Convert from a legacy lsst.afw.math.LinearCombinationKernel.

Parameters

legacy_kernel

The kernel to convert. Must use Chebyshev polynomials for its spatial variation and lsst.afw.math.FixedKernel objects with a consistent shape and center for its basis functions.

serialize(archive: OutputArchive[Any]) → ImageBasisConvolutionKernelSerializationModel

Serialize the kernel to an output archive.

Parameters

archive

Archive to write to.

to_legacy() → LegacyLinearCombinationKernel

Convert to a legacy lsst.afw.math.LinearCombinationKernel.

This only works if all spatial variation is handled by lsst.images.ChebyshevField.