pylops.signalprocessing.Bilinear

class pylops.signalprocessing.Bilinear(iava, dims, forceflat=None, dtype='float64', name='B')

Bilinear interpolation operator.

Apply bilinear interpolation onto fractionary positions iava along the first two axes of a n-dimensional array.

Note

The vector iava should contain unique pais. If the same pair is repeated twice an error will be raised.

Parameters:

iavalist or numpy.ndarray

Array of size \([2 \times n_\text{ava}]\) containing pairs of floating indices of locations of available samples for interpolation.

dimslist

Number of samples for each dimension

forceflatbool, optional

Added in version 2.2.0.

Force an array to be flattened after rmatvec. Note that this is only required when len(dims)=2, otherwise pylops will detect whether to return a 1d or nd array.

dtypestr, optional

Type of elements in input array.

namestr, optional

Added in version 2.0.0.

Name of operator (to be used by pylops.utils.describe.describe)

Attributes:

iava_tnumpy.ndarray

Indices of points above iava for bilinear interpolation.

iava_bnumpy.ndarray

Indices of points below iava for bilinear interpolation.

iava_lnumpy.ndarray

Indices of points left of iava for bilinear interpolation.

iava_rnumpy.ndarray

Indices of points right of iava for bilinear interpolation.

weights_tbnumpy.ndarray

Weights for top/bottom points for bilinear interpolation.

weights_lrnumpy.ndarray

Weights for left/right points for bilinear interpolation.

dimstuple

Shape of the array after the adjoint, but before flattening.

For example, x_reshaped = (Op.H * y.ravel()).reshape(Op.dims).

dimsdtuple

Shape of the array after the forward, but before flattening.

For example, y_reshaped = (Op * x.ravel()).reshape(Op.dimsd).

shapetuple

Operator shape.

Raises:

ValueError

If the vector iava contains repeated values.

Notes

Bilinear interpolation of a subset of \(N\) values at locations iava from an input n-dimensional vector \(\mathbf{x}\) of size \([m_1 \times m_2 \times ... \times m_{ndim}]\) can be expressed as:

\[y_{\mathbf{i}} = (1-w^0_{i}) (1-w^1_{i}) x_{l^{l,0}_i, l^{l,1}_i} + w^0_{i} (1-w^1_{i}) x_{l^{r,0}_i, l^{l,1}_i} + (1-w^0_{i}) w^1_{i} x_{l^{l,0}_i, l^{r,1}_i} + w^0_{i} w^1_{i} x_{l^{r,0}_i, l^{r,1}_i} \quad \forall i=1,2,\ldots,M\]

where \(\mathbf{l^{l,0}}=[\lfloor l_1^0 \rfloor, \lfloor l_2^0 \rfloor, ..., \lfloor l_N^0 \rfloor]\), \(\mathbf{l^{l,1}}=[\lfloor l_1^1 \rfloor, \lfloor l_2^1 \rfloor, ..., \lfloor l_N^1 \rfloor]\), \(\mathbf{l^{r,0}}=[\lfloor l_1^0 \rfloor + 1, \lfloor l_2^0 \rfloor + 1, ..., \lfloor l_N^0 \rfloor + 1]\), \(\mathbf{l^{r,1}}=[\lfloor l_1^1 \rfloor + 1, \lfloor l_2^1 \rfloor + 1, ..., \lfloor l_N^1 \rfloor + 1]\), are vectors containing the indices of the original array at which samples are taken, and \(\mathbf{w^j}=[l_1^i - \lfloor l_1^i \rfloor, l_2^i - \lfloor l_2^i \rfloor, ..., l_N^i - \lfloor l_N^i \rfloor]\) (\(\forall j=0,1\)) are the bilinear interpolation weights.

Methods

__init__(iava, dims[, forceflat, dtype, name])
adjoint()
apply_columns(cols)	Apply subset of columns of operator
cond([uselobpcg])	Condition number of linear operator.
conj()	Complex conjugate operator
div(y[, niter, densesolver])	Solve the linear problem \(\mathbf{y}=\mathbf{A}\mathbf{x}\).
dot(x)	Matrix-matrix or matrix-vector multiplication.
eigs([neigs, symmetric, niter, uselobpcg])	Most significant eigenvalues of linear operator.
matmat(X)	Matrix-matrix multiplication.
matvec(x)	Matrix-vector multiplication.
reset_count()	Reset counters
rmatmat(X)	Matrix-matrix multiplication.
rmatvec(x)	Adjoint matrix-vector multiplication.
todense([backend])	Return dense matrix.
toimag([forw, adj])	Imag operator
toreal([forw, adj])	Real operator
tosparse()	Return sparse matrix.
trace([neval, method, backend])	Trace of linear operator.
transpose()

Examples using pylops.signalprocessing.Bilinear

Bilinear Interpolation

Bilinear Interpolation