torchbp.grid module

class torchbp.grid.CartesianGrid(x_range, y_range, nx, ny)[source]

Bases: Grid

Cartesian grid.

Parameters:

x_range (Tuple[float, float]) – X extent (x0, x1) in meters. x1 must be > x0.
y_range (Tuple[float, float]) – Y extent (y0, y1) in meters. y1 must be > y0.
nx (int) – Number of X samples. Must be positive.
ny (int) – Number of Y samples. Must be positive.

Examples

>>> grid = CartesianGrid(x_range=(-50, 50), y_range=(-50, 50), nx=100, ny=100)
>>> grid.dx
1.0
>>> grid.dy
1.0
>>> grid.shape()
(100, 100)

property dx: float: X spacing in meters (cached).

property dy: float: Y spacing in meters (cached).

classmethod from_dict(d)[source]

Create CartesianGrid from dict format.

Parameters:: d (dict) – Dict with keys: “x”, “y”, “nx”, “ny”
Returns:: Grid object
Return type:: CartesianGrid

resize(nx=None, ny=None)[source]

Return new grid with different resolution.

Parameters:

nx (int, optional) – New number of X samples. If None, keep current nx.
ny (int, optional) – New number of Y samples. If None, keep current ny.

Returns:

New grid with updated resolution

Return type:

CartesianGrid

shape()[source]

Return grid dimensions (nx, ny).

Return type:: Tuple[int, int]

spacing()[source]

Return grid spacing (dx, dy).

Return type:: Tuple[float, float]

to_dict()[source]

Convert to dict format.

Returns:: Grid dict with keys: “x”, “y”, “nx”, “ny”
Return type:: dict

property x0: float: Minimum X coordinate in meters.

property x1: float: Maximum X coordinate in meters.

property y0: float: Minimum Y coordinate in meters.

property y1: float: Maximum Y coordinate in meters.

class torchbp.grid.Grid[source]

Bases: ABC

Base class for all grid types.

abstractmethod shape()[source]

Return grid dimensions.

Return type:: Tuple[int, …]

abstractmethod spacing()[source]

Return grid spacing.

Return type:: Tuple[float, …]

abstractmethod to_dict()[source]

Convert to dict format for backward compatibility.

Return type:: dict

class torchbp.grid.PolarGrid(r_range, theta_range, nr, ntheta)[source]

Bases: Grid

Pseudo-polar grid (theta stored as sin(angle)).

Parameters:

r_range (Tuple[float, float]) – Range extent (r0, r1) in meters. r1 must be > r0.
theta_range (Tuple[float, float]) – Azimuth extent (theta0, theta1) as sin(angle). Must be in [-1, 1].
nr (int) – Number of range bins. Must be positive.
ntheta (int) – Number of azimuth bins. Must be positive.

Examples

>>> grid = PolarGrid(r_range=(50, 100), theta_range=(-1, 1), nr=100, ntheta=200)
>>> grid.dr  # Range spacing (cached)
0.5
>>> grid.dtheta  # Azimuth spacing (cached)
0.01
>>> grid.shape()
(100, 200)

property dr: float: Range spacing in meters (cached).

property dtheta: float: Azimuth spacing (cached).

classmethod from_dict(d)[source]

Create PolarGrid from dict format.

Parameters:: d (dict) – Dict with keys: “r”, “theta”, “nr”, “ntheta”
Returns:: Grid object
Return type:: PolarGrid

property r0: float: Minimum range in meters.

property r1: float: Maximum range in meters.

resize(nr=None, ntheta=None)[source]

Return new grid with different resolution.

Parameters:

nr (int, optional) – New number of range bins. If None, keep current nr.
ntheta (int, optional) – New number of azimuth bins. If None, keep current ntheta.

Returns:

New grid with updated resolution

Return type:

PolarGrid

shape()[source]

Return grid dimensions (nr, ntheta).

Return type:: Tuple[int, int]

spacing()[source]

Return grid spacing (dr, dtheta).

Return type:: Tuple[float, float]

property theta0: float: Minimum azimuth (sin of angle).

property theta1: float: Maximum azimuth (sin of angle).

to_dict()[source]

Convert to dict format.

Returns:: Grid dict with keys: “r”, “theta”, “nr”, “ntheta”
Return type:: dict

with_doubled_azimuth()[source]

Create grid with 2x azimuth resolution.

Used in FFBP algorithm for azimuth upsampling.

Returns:: New grid with ntheta *= 2
Return type:: PolarGrid

torchbp.grid.unpack_cartesian_grid(grid)[source]

Unpack cartesian grid to (x0, x1, y0, y1, nx, ny, dx, dy).

Accepts both CartesianGrid objects and dicts.

Parameters:: grid (CartesianGrid or dict) – Cartesian grid specification
Returns:: (x0, x1, y0, y1, nx, ny, dx, dy)
Return type:: tuple

torchbp.grid.unpack_polar_grid(grid)[source]

Unpack polar grid to (r0, r1, theta0, theta1, nr, ntheta, dr, dtheta).

Accepts both PolarGrid objects and dicts.

Parameters:: grid (PolarGrid or dict) – Polar grid specification
Returns:: (r0, r1, theta0, theta1, nr, ntheta, dr, dtheta)
Return type:: tuple