Official RETINA Dataset

To view the images in Python, ensure you have "numpy", "rasterio", and "matplotlib" installed. Images are viewable by clipping the bands by minimum and maximum threshold values of the pixels intensity.

Python Code:
# Print the number of bands
print(f'Number of bands: {num_bands}')

# Optionally, print additional information
print(f'Width: {dataset.width}')
print(f'Height: {dataset.height}')
print(f'CRS: {dataset.crs}') # Coordinate Reference System (CRS)
print(f'Bounds: {dataset.bounds}')

# BAND STATISTICS
with rasterio.open(file_path) as dataset:
band1 = dataset.read(1)
band2 = dataset.read(2)

def print_stats(band, band_name):
print(f"{band_name} statistics:")
print(f" Min: {np.min(band)}")
print(f" Max: {np.max(band)}")
print(f" Mean: {np.mean(band)}")
print(f" Std: {np.std(band)}")

print_stats(band1, "Band 1")
print_stats(band2, "Band 2")

# Function to clip pixel values
def clip_array(array, min_percentile=2, max_percentile=98):
min_val = np.percentile(array, min_percentile)
max_val = np.percentile(array, max_percentile)
array_clipped = np.clip(array, min_val, max_val) # We keep the values of the array between min and max percentile
return array_clipped

# Clipping the bands
band1_clipped = clip_array(band1)
band2_clipped = clip_array(band2)

# Normalizing the clipped bands
band1_normalized = (band1_clipped - np.min(band1_clipped)) / (np.max(band1_clipped) - np.min(band1_clipped))
band2_normalized = (band2_clipped - np.min(band2_clipped)) / (np.max(band2_clipped) - np.min(band2_clipped))

plt.figure(figsize=(10, 10))
plt.title("Band 1 (Clipped and Normalized)")
plt.imshow(band1_normalized, cmap='gray')
plt.show()

plt.figure(figsize=(10, 10))
plt.title("Band 2 (Clipped and Normalized)")
plt.imshow(band2_normalized, cmap='gray')
plt.show()