Update app.py

app.py

@@ -47,10 +47,39 @@ def load_and_convert_images(directory):
                 dataset.append(resized_image)
     return dataset
 
+# Ensure that both datasets contain 2D arrays (grayscale images)
+def check_and_reshape(dataset):
+    """Ensure all images in the dataset are 2D arrays.
+    
+    Args:
+        dataset (list): A list of images.
+        
+    Returns:
+        list: A list of reshaped images to ensure they are all 2D.
+    """
+    reshaped_dataset = []
+    for img in dataset:
+        if img.ndim == 3:  # If the image has 3 dimensions (like RGB), convert it to grayscale
+            gray_img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
+            reshaped_dataset.append(gray_img)
+        elif img.ndim == 2:  # If it's already grayscale (2D), just append it
+            reshaped_dataset.append(img)
+        else:
+            raise ValueError("Unexpected image dimension: {}".format(img.ndim))
+    return reshaped_dataset
+
 # Load datasets using glob
 grass_dataset = load_and_convert_images(grass_dir)
 wood_dataset = load_and_convert_images(wood_dir)
 
+# Apply the reshaping to both datasets
+grass_dataset = check_and_reshape(grass_dataset)
+wood_dataset = check_and_reshape(wood_dataset)
+
+# Now concatenate the datasets
+X = np.concatenate([grass_dataset, wood_dataset])
+y = np.concatenate([np.zeros(len(grass_dataset)), np.ones(len(wood_dataset))])  # Labels: 0 for Grass, 1 for Wood
+
 def calc_glcm_features(images):
     """Calculate GLCM features for a list of images.
 
@@ -90,10 +119,6 @@ def extract_lbp_features(images):
     return features
 
 # Train classifiers (for example purposes, using KNN)
-X = np.concatenate([grass_dataset, wood_dataset])
-y = np.concatenate([np.zeros(len(grass_dataset)), np.ones(len(wood_dataset))])  # Labels: 0 for Grass, 1 for Wood
-
-# Extract GLCM and LBP features
 glcm_features = calc_glcm_features(X)
 lbp_features = extract_lbp_features(X)