Spectral Image

This example demonstrates how to load and inspect a spectral image using the SiaPy library:

• Load a spectral image.
• Access various properties and metadata.

from pathlib import Path

import spectral as sp

from siapy.entities import SpectralImage

# Set the path to the directory containing the data
# !! ADJUST THIS PATH TO YOUR DATA DIRECTORY !!
data_dir = "./docs/examples/data"

# Find all header and image files in the data directory
header_paths = sorted(Path(data_dir).rglob("*.hdr"))
image_paths = sorted(Path(data_dir).rglob("*.img"))

header_path_img0 = header_paths[0]
image_path_img0 = image_paths[0]

# Load the image using spectral library and then wrap over SpectralImage object
sp_file = sp.envi.open(file=header_path_img0, image=image_path_img0)
assert not isinstance(sp_file, sp.io.envi.SpectralLibrary)
image = SpectralImage(sp_file)

# or you can do the same just by running
image = SpectralImage.envi_open(
    header_path=header_path_img0,
    image_path=image_path_img0,
)

# Now you can easily use various property and util functions of the SpectralImage object
# Get the shape of the image
print("Image shape:", image.shape)

# Get the number of bands
print("Number of bands:", image.bands)

# Get the wavelength information
print("Wavelengths:", image.wavelengths)

# Get the file path
print("File path:", image.filepath)

# Get the metadata
print("Metadata:", image.metadata)

# Get the number of rows
print("Number of rows:", image.rows)

# Get the number of columns
print("Number of columns:", image.cols)

# Get the default bands
print("Default bands:", image.default_bands)

# Get the description
print("Description:", image.description)

# Get the camera ID
print("Camera ID:", image.camera_id)

# Get the geometric shapes
print("Geometric shapes:", image.geometric_shapes)

Source: spectral_image_01.py

---

This example demonstrates how to perform various operations on a spectral image:

• Convert a spectral image to a NumPy array.
• Calculate the mean value per band.
• Create a Pixels object from pixel coordinates.
• Extract spectral signatures and subarrays.
• Display the spectral image.

from pathlib import Path

import matplotlib.pyplot as plt

from siapy.entities import Pixels, SpectralImage

# Set the path to the directory containing the data
# !! ADJUST THIS PATH TO YOUR DATA DIRECTORY !!
data_dir = "./docs/examples/data"

# Get first image
header_path_img0 = sorted(Path(data_dir).rglob("*.hdr"))[1]
image_path_img0 = sorted(Path(data_dir).rglob("*.img"))[1]

# Load spectral image
image = SpectralImage.envi_open(
    header_path=header_path_img0,
    image_path=image_path_img0,
)

# Convert to numpy
image_np = image.to_numpy(nan_value=0.0)
print("Image shape:", image_np.shape)

# Calculate mean
mean_val = image.mean(axis=(0, 1))
print("Mean value per band:", mean_val)

# Create a Pixels object from an iterable with pixels coordinates
# The iterable should be a list of tuples representing (x, y) coordinates
# iterable == [(x1, y1), (x2, y2), ...] -> list of pixels
iterable = [(1, 2), (3, 4), (5, 6)]
pixels = Pixels.from_iterable(iterable)

# Convert the pixel coordinates to spectral signatures
signatures = image.to_signatures(pixels)
print("Signatures:", signatures)

# Extract a subarray from the image using the pixel coordinates
subarray = image.to_subarray(pixels)
print("Subarray shape:", subarray.shape)

# Convert to displayable image
display_image = image.to_display(equalize=True)

# Display the image using matplotlib
plt.figure()
plt.imshow(display_image)
plt.axis("off")  # Hide axes for better visualization
plt.show()

Source: spectral_image_02.py