NovaDetect/landslide_training_model.py at main · Pooja29Shree/NovaDetect · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
import os
import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt
from tensorflow.keras.layers import Input, Conv2D, MaxPooling2D, UpSampling2D, concatenate
from tensorflow.keras.models import Model
from tensorflow.keras.preprocessing.image import load_img, img_to_array
from sklearn.model_selection import train_test_split
from PIL import Image

# Optional: Allow large image processing
Image.MAX_IMAGE_PIXELS = None

# ---------------------------
# Config
# ---------------------------
IMG_SIZE = 128
IMG_CHANNELS = 1
IMAGE_DIR = "images/"
MASK_DIR = "masks/"

# ---------------------------
# Load and preprocess images/masks
# ---------------------------
import re

def load_images(img_dir, mask_dir, img_size):
    image_list = []
    mask_list = []

    image_files = [f for f in os.listdir(img_dir) if f.endswith(('.png', '.jpg', '.tif'))]
    mask_files = [f for f in os.listdir(mask_dir) if f.endswith(('.png', '.jpg', '.tif'))]

    for img_file in image_files:
        # Extract timestamp using regex
        match = re.search(r'\d{8}T\d+', img_file) # e.g., 20090706T2122023762
        if not match:
            print(f"⚠ Skipping {img_file} — no timestamp found")
            continue

        timestamp = match.group(0)

        # Find matching mask
        matching_mask_file = None
        for mask_file in mask_files:
            if timestamp in mask_file:
                matching_mask_file = mask_file
                break

        if matching_mask_file is None:
            print(f"⚠ No matching mask for {img_file} (timestamp: {timestamp})")
            continue

        img_path = os.path.join(img_dir, img_file)
        mask_path = os.path.join(mask_dir, matching_mask_file)

        img = load_img(img_path, color_mode='grayscale', target_size=(img_size, img_size))
        mask = load_img(mask_path, color_mode='grayscale', target_size=(img_size, img_size))

        img_array = img_to_array(img) / 255.0
        mask_array = img_to_array(mask)
        mask_array = (mask_array > 127).astype(np.float32)

        image_list.append(img_array)
        mask_list.append(mask_array)

    print(f"✅ Loaded {len(image_list)} matched image-mask pairs.")
    return np.array(image_list), np.array(mask_list)


images, masks = load_images(IMAGE_DIR, MASK_DIR, IMG_SIZE)
print("Images shape:", images.shape)
print("Masks shape:", masks.shape)

# ---------------------------
# U-Net Architecture
# ---------------------------
def unet_model(input_size=(IMG_SIZE, IMG_SIZE, IMG_CHANNELS)):
    inputs = Input(input_size)

    # Encoder
    c1 = Conv2D(16, 3, activation='relu', padding='same')(inputs)
    c1 = Conv2D(16, 3, activation='relu', padding='same')(c1)
    p1 = MaxPooling2D()(c1)

    c2 = Conv2D(32, 3, activation='relu', padding='same')(p1)
    c2 = Conv2D(32, 3, activation='relu', padding='same')(c2)
    p2 = MaxPooling2D()(c2)

    # Bottleneck
    c3 = Conv2D(64, 3, activation='relu', padding='same')(p2)
    c3 = Conv2D(64, 3, activation='relu', padding='same')(c3)

    # Decoder
    u1 = UpSampling2D()(c3)
    m1 = concatenate([u1, c2])
    c4 = Conv2D(32, 3, activation='relu', padding='same')(m1)
    c4 = Conv2D(32, 3, activation='relu', padding='same')(c4)

    u2 = UpSampling2D()(c4)
    m2 = concatenate([u2, c1])
    c5 = Conv2D(16, 3, activation='relu', padding='same')(m2)
    c5 = Conv2D(16, 3, activation='relu', padding='same')(c5)

    outputs = Conv2D(1, 1, activation='sigmoid')(c5)

    return Model(inputs, outputs)

# ---------------------------
# Dice Loss Function
# ---------------------------
def dice_loss(y_true, y_pred):
    smooth = 1e-6
    y_true_f = tf.keras.backend.flatten(y_true)
    y_pred_f = tf.keras.backend.flatten(y_pred)
    intersection = tf.reduce_sum(y_true_f * y_pred_f)
    return 1 - (2. * intersection + smooth) / (tf.reduce_sum(y_true_f) + tf.reduce_sum(y_pred_f) + smooth)

# ---------------------------
# Train/Test Split
# ---------------------------
X_train, X_val, y_train, y_val = train_test_split(images, masks, test_size=0.2, random_state=42)

# ---------------------------
# Compile & Train
# ---------------------------
model = unet_model()
model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=1e-4),
              loss=dice_loss,
              metrics=['accuracy'])

model.fit(X_train, y_train, validation_data=(X_val, y_val), epochs=20, batch_size=1)

# ---------------------------
# Predict and Visualize
# ---------------------------
pred_mask = model.predict(np.expand_dims(X_val[0], axis=0))[0]

fig, axes = plt.subplots(1, 3, figsize=(12, 4))

axes[0].imshow(X_val[0].squeeze(), cmap='gray')
axes[0].set_title("Input Image")
axes[0].axis('off')

axes[1].imshow(y_val[0].squeeze(), cmap='gray')
axes[1].set_title("True Mask")
axes[1].axis('off')

axes[2].imshow(pred_mask.squeeze(), cmap='gray')
axes[2].set_title("Predicted Mask")
axes[2].axis('off')

plt.tight_layout()
plt.savefig("prediction_output.png")

model.save("model.h5")