CNN变化检测代码下载

以下是一个基于TensorFlow的CNN变化检测代码示例，可以用于检测遥感图像中的变化：

import tensorflow as tf
import numpy as np
import os
import argparse
from PIL import Image

parser = argparse.ArgumentParser()
parser.add_argument("--train_data_dir", default="./train_data", help="Directory containing the train data.")
parser.add_argument("--test_data_dir", default="./test_data", help="Directory containing the test data.")
parser.add_argument("--epochs", default=10, type=int, help="Number of training epochs.")
parser.add_argument("--batch_size", default=32, type=int, help="Batch size.")
parser.add_argument("--lr", default=0.001, type=float, help="Learning rate.")
parser.add_argument("--model_dir", default="./model", help="Directory to save the trained model.")
args = parser.parse_args()

def load_data(data_dir):
    image_dir = os.path.join(data_dir, "images")
    label_dir = os.path.join(data_dir, "labels")
    image_files = os.listdir(image_dir)
    label_files = os.listdir(label_dir)
    image_files.sort()
    label_files.sort()
    images = []
    labels = []
    for i in range(len(image_files)):
        image_path = os.path.join(image_dir, image_files[i])
        label_path = os.path.join(label_dir, label_files[i])
        image = np.array(Image.open(image_path).convert("RGB"))
        label = np.array(Image.open(label_path).convert("L")) // 255
        images.append(image)
        labels.append(label)
    images = np.array(images)
    labels = np.array(labels)
    return images, labels

def create_model():
    inputs = tf.keras.layers.Input(shape=(256, 256, 3))
    x = tf.keras.layers.Conv2D(32, 3, activation="relu")(inputs)
    x = tf.keras.layers.MaxPooling2D()(x)
    x = tf.keras.layers.Conv2D(64, 3, activation="relu")(x)
    x = tf.keras.layers.MaxPooling2D()(x)
    x = tf.keras.layers.Flatten()(x)
    x = tf.keras.layers.Dense(128, activation="relu")(x)
    outputs = tf.keras.layers.Dense(1, activation="sigmoid")(x)
    model = tf.keras.models.Model(inputs=inputs, outputs=outputs)
    return model

def train():
    train_images, train_labels = load_data(args.train_data_dir)
    test_images, test_labels = load_data(args.test_data_dir)
    model = create_model()
    model.compile(optimizer=tf.keras.optimizers.Adam(lr=args.lr),
                  loss="binary_crossentropy",
                  metrics=["accuracy"])
    model.fit(train_images, train_labels,
              epochs=args.epochs,
              batch_size=args.batch_size,
              validation_data=(test_images, test_labels))
    model.save(args.model_dir)

def predict(image_path, model_path):
    model = tf.keras.models.load_model(model_path)
    image = np.array(Image.open(image_path).convert("RGB"))
    image = np.expand_dims(image, axis=0)
    prediction = model.predict(image)
    return prediction[0][0]

if __name__ == "__main__":
    train()

使用方法：

1. 将训练数据放在`--train_data_dir`指定的目录下，其中包含两个子目录：`images`和`labels`，分别存放原始图像和变化检测标签。
2. 将测试数据放在`--test_data_dir`指定的目录下，格式同训练数据。
3. 运行`python cnn_change_detection.py`进行训练，训练完成后模型会被保存在`--model_dir`指定的目录下。
4. 使用`predict`函数可以对单张图像进行预测，其中`image_path`为待预测图像的路径，`model_path`为训练好的模型路径。返回值为预测的变化概率值（0到1之间）。