import os
import numpy as np
from polsartools.utils.proc_utils import process_chunks_parallel
from polsartools.utils.utils import conv2d,time_it
from .cp_infiles import cpc2files
[docs]
@time_it
def m_delta(in_dir, chi=45, psi=0, win=1, fmt="tif", cog=False,
ovr = [2, 4, 8, 16], comp=False,
max_workers=None,block_size=(512, 512),
progress_callback=None # for QGIS plugin
):
"""Perform M-Delta Decomposition for compact-pol SAR data.
This function implements the M-Delta Decomposition for
compact-polarimetric SAR data, decomposing the total backscattered power into
surface (Ps), double-bounce (Pd), and volume (Pv) scattering components.
Examples
--------
>>> # Basic usage with default parameters
>>> m_delta("/path/to/cp_data")
>>> # Advanced usage with custom parameters
>>> m_delta(
... in_dir="/path/to/cp_data",
... chi=-45,
... win=5,
... fmt="tif",
... cog=True,
... block_size=(1024, 1024)
... )
Parameters
----------
in_dir : str
Path to the input folder containing compact-pol C2 matrix files.
chi : float, default=45
Ellipticity angle chi of the transmitted wave in degrees.
For circular polarization, chi = 45° (right circular) or -45° (left circular).
psi : float, default=0
Orientation angle psi of the transmitted wave in degrees.
For circular polarization, typically 0°.
win : int, default=1
Size of the spatial averaging window. Larger windows reduce speckle noise
but decrease spatial resolution.
fmt : {'tif', 'bin'}, default='tif'
Output file format:
- 'tif': GeoTIFF format with georeferencing information
- 'bin': Raw binary format
cog : bool, default=False
If True, creates Cloud Optimized GeoTIFF (COG) outputs with internal tiling
and overviews for efficient web access.
ovr : list[int], default=[2, 4, 8, 16]
Overview levels for COG creation. Each number represents the
decimation factor for that overview level.
comp : bool, default=False
If True, applies LZW compression to the output GeoTIFF files.
max_workers : int | None, default=None
Maximum number of parallel processing workers. If None, uses
CPU count - 1 workers.
block_size : tuple[int, int], default=(512, 512)
Size of processing blocks (rows, cols) for parallel computation.
Larger blocks use more memory but may be more efficient.
Returns
-------
Writes four output files to disk:
- Ps_m_delta: Surface scattering power component
- Pd_m_delta: Double-bounce scattering power component
- Pv_m_delta: Volume scattering power component
- m_cp: Degree of polarization
- chi_cp: chi angle in degrees
"""
write_flag=True
input_filepaths = cpc2files(in_dir)
output_filepaths = []
if fmt == "bin":
output_filepaths.append(os.path.join(in_dir, "Ps_m_delta.bin"))
output_filepaths.append(os.path.join(in_dir, "Pd_m_delta.bin"))
output_filepaths.append(os.path.join(in_dir, "Pv_m_delta.bin"))
output_filepaths.append(os.path.join(in_dir, "m_cp.bin"))
output_filepaths.append(os.path.join(in_dir, "delta_cp.bin"))
else:
output_filepaths.append(os.path.join(in_dir, "Ps_m_delta.tif"))
output_filepaths.append(os.path.join(in_dir, "Pd_m_delta.tif"))
output_filepaths.append(os.path.join(in_dir, "Pv_m_delta.tif"))
output_filepaths.append(os.path.join(in_dir, "m_cp.tif"))
output_filepaths.append(os.path.join(in_dir, "delta_cp.tif"))
process_chunks_parallel(input_filepaths, list(output_filepaths),
win,
write_flag,
process_chunk_mdelta,
*[chi, psi],
block_size=block_size,
max_workers=max_workers,
num_outputs=len(output_filepaths),
cog=cog, ovr=ovr, comp=comp,
progress_callback=progress_callback
)
def process_chunk_mdelta(chunks, window_size, *args, **kwargs):
chi=args[-2]
psi=args[-1]
# print(chi,psi):
kernel = np.ones((window_size,window_size),np.float32)/(window_size*window_size)
c11_T1 = np.array(chunks[0])
c12_T1 = np.array(chunks[1])+1j*np.array(chunks[2])
c21_T1 = np.conj(c12_T1)
c22_T1 = np.array(chunks[3])
ncols,nrows = np.shape(c11_T1)
if window_size>1:
c11_T1 = conv2d(np.real(c11_T1),kernel)+1j*conv2d(np.imag(c11_T1),kernel)
c12_T1 = conv2d(np.real(c12_T1),kernel)+1j*conv2d(np.imag(c12_T1),kernel)
c21_T1 = conv2d(np.real(c21_T1),kernel)+1j*conv2d(np.imag(c21_T1),kernel)
c22_T1 = conv2d(np.real(c22_T1),kernel)+1j*conv2d(np.imag(c22_T1),kernel)
# Compute Stokes parameters
s0 = np.real(c11_T1 + c22_T1)
s1 = np.real(c11_T1 - c22_T1)
s2 = np.real(c12_T1 + c21_T1)
# s3 = np.where(chi >= 0, 1j * (c12_T1 - c21_T1), -1j * (c12_T1 - c21_T1))
s3 = -2*np.imag(c12_T1)
m = np.sqrt(s1**2 + s2**2 + s3**2) / (s0)
delta = np.arctan2(s3, s2)
Ps_CP= np.sqrt((m * s0 * (1 + np.sin(delta))) / 2)
Pd_CP= np.sqrt((m * s0 * (1 - np.sin(delta))) / 2)
Pv_CP= np.sqrt((s0 * (1 - m)) / 2)
delta = delta * 180 * np.pi
return Ps_CP.astype(np.float32), Pd_CP.astype(np.float32), Pv_CP.astype(np.float32),m.astype(np.float32),delta.astype(np.float32)
