update

data/ubc/embedding_test.py

@@ -4,6 +4,7 @@
 python3 embedding_test.py small/G80223_20230513.bin_scale.bin
 
 '''
+
 import os
 import sys
 import umap
@@ -78,16 +79,16 @@ def get_model(model_type='tsne'):
         rgb_values[:, i][np.where(rgb_values[:, i] < 0.)] = 0.
         rgb_values[:, i][np.where(rgb_values[:, i] > 1.)] = 1.
 
-# Create a figure with two subplots (for UMAP/t-SNE and RGB values)
-fig, axs = plt.subplots(1, 2, figsize=(16, 6))  # 1 row, 2 columns
+# Create a figure with three subplots (for UMAP/t-SNE, RGB values with masks, and original image)
+fig, axs = plt.subplots(1, 3, figsize=(18, 6))  # 1 row, 3 columns
 
 # First subplot: UMAP or t-SNE projection colored by RGB values
 scatter = axs[0].scatter(embedding[:, 0], embedding[:, 1], c=rgb_values, s=1, alpha=0.5)  # Scatter plot with RGB coloring
 axs[0].set_title(f'{model_type.upper()} Projection of Raster Data')
 axs[0].set_xlabel(f'{model_type.upper()} Component 1')
 axs[0].set_ylabel(f'{model_type.upper()} Component 2')
 
-# Second subplot: Display RGB values using imshow
+# Second subplot: Display RGB values using imshow with the mask applied
 # Reshape the RGB values back to (height, width, 3) for imshow
 rgb_image = rgb_values.reshape(height, width, 3)  # Recreate the RGB image
 
@@ -96,6 +97,15 @@ def get_model(model_type='tsne'):
 axs[1].axis('off')  # Hide axes for better visualization
 axs[1].set_title('RGB Values of Raster')
 
+# Third subplot: Original image (without any modifications)
+# Display the original image (in this case, we are using RGB bands for visualization)
+original_image = np.moveaxis(data[:3, :, :], 0, -1)  # First 3 bands for the original image
+
+# Display the original image
+axs[2].imshow(original_image)
+axs[2].axis('off')  # Hide axes for better visualization
+axs[2].set_title('Original Image')
+
 # Initialize variables for polygon drawing
 polygon_points = []
 polygon_path = None
@@ -170,3 +180,4 @@ def mask_polygon():
 plt.tight_layout()
 plt.show()
 
+