import os
import numpy as np
from polsartools.utils.utils import conv2d,time_it
from osgeo import gdal
gdal.UseExceptions()
from .pauli_rgb import generate_rgb_png, create_pgw, generate_rgb_tif, norm_data, read_bin #create_prj
def ensure_array(x):
# If it's a path-like object (string or Path), read from file
if isinstance(x, (str, os.PathLike)):
return np.array(read_bin(x), dtype=np.float32)
# If it's already a NumPy array, just cast dtype
elif isinstance(x, np.ndarray):
return x.astype(np.float32)
# If it's a list or other sequence, convert to array
else:
return np.array(x, dtype=np.float32)
[docs]
@time_it
def rgb(Rpath,Gpath,Bpath,
norm_each=True,lp=5,gp=95,
norm_span = False,
save_tif=False,
window_size=None):
"""Generate false-color RGB visualization from three input raster files.
Examples
--------
>>> # Generate Pauli RGB from a polarimetric folder
>>> rgb("/path/to/red.tif", "/path/to/green.tif", "/path/to/blue.tif")
>>> # Output both PNG and GeoTIFF versions
>>> rgb("/path/to/red.tif", "/path/to/green.tif", "/path/to/blue.tif", tif_flag=True)
Parameters
----------
Rpath : str
Path to the red channel raster file.
Gpath : str
Path to the green channel raster file.
Bpath : str
Path to the blue channel raster file.
norm_each : bool, default=True
If True, normalize each channel separately.
lp : int, default=5
Lower percentile for normalization (0-100). Only used if `norm_each=True`.
gp : int, default=95
Upper percentile for normalization(0-100). Only used if `norm_each=True`.
norm_span : bool, default=False
If True, normalize each channel with total power.
save_tif : bool, default=False
If True, generates a GeoTIFF (.tif) file alongside the PNG image.
window_size : int, optional
Size of the moving average window for smoothing. If None, no smoothing is applied.
Returns
-------
None
Writes output to:
- RGB.png: RGB composite with world file for georeferencing
- RGB.tif: Optional GeoTIFF output if `tif_flag=True`
"""
# R = read_bin(Rpath)
# G = read_bin(Gpath)
# B = read_bin(Bpath)
R = ensure_array(Rpath)
G = ensure_array(Gpath)
B = ensure_array(Bpath)
# R = np.array(read_bin(Rpath), dtype=np.float32)
# G = np.array(read_bin(Gpath), dtype=np.float32)
# B = np.array(read_bin(Bpath), dtype=np.float32)
if window_size is not None:
kernel = np.ones((window_size, window_size), np.float32) / (window_size * window_size)
R = conv2d(R, kernel).astype(np.float32)
G = conv2d(G, kernel).astype(np.float32)
B = conv2d(B, kernel).astype(np.float32)
if norm_each:
R = norm_data(R, lp, gp, dB_scale=True).astype(np.float32)
G = norm_data(G, lp, gp, dB_scale=True).astype(np.float32)
B = norm_data(B, lp, gp, dB_scale=True).astype(np.float32)
if norm_span:
total = R + G + B
red = R / total
green = G / total
blue = B / total
del R,G,B,total
else:
red = R
green = G
blue = B
if isinstance(Rpath, (str, os.PathLike)):
georef_file = Rpath
infolder = os.path.dirname(georef_file)
output_path = os.path.join(infolder, 'RGB.png')
generate_rgb_png(red, green, blue, georef_file, output_path)
create_pgw(georef_file, output_path)
print(f"RGB image saved as {output_path}")
if save_tif:
tif_path = os.path.join(infolder, f'RGB.tif')
generate_rgb_tif(red, green, blue, georef_file, tif_path)
print(f"RGB GeoTIFF saved as {tif_path}")
elif isinstance(Rpath, np.ndarray):
georef_file = None
infolder = './'
output_path = os.path.join(infolder, 'RGB.png')
generate_rgb_png(red, green, blue, georef_file, output_path)
# create_pgw(georef_file, output_path)
print(f"RGB image saved as {output_path}")
else:
raise('Invalid input!!')
# plt.imshow(RGB_image)
# plt.axis('off') # Hide axes
# if pname:
# plt.savefig(pname, bbox_inches='tight', pad_inches=0,dpi=300)
# plt.show()
