### 图像归一化代码示例 以下是使用Python对文件夹中的图像进行归一化的代码示例。此代码利用了Pillow库来处理图像，并通过调整像素值范围实现归一化。 ```python from PIL import Image import numpy as np import os def normalize_image(image_array): """Normalize an image array to range [0, 1].""" min_val = np.min(image_array) max_val = np.max(image_array) normalized_image = (image_array - min_val) / (max_val - min_val) return normalized_image def save_normalized_image(normalized_image, output_path): """Save a normalized image back to disk.""" normalized_image_uint8 = (normalized_image * 255).astype(np.uint8) img = Image.fromarray(normalized_image_uint8) img.save(output_path) def normalize_images_in_folder(input_folder, output_folder): """ Normalize all images in the input folder and save them to the output folder. Parameters: input_folder (str): Path to the folder containing original images. output_folder (str): Path to the folder where normalized images will be saved. """ if not os.path.exists(output_folder): os.makedirs(output_folder) for filename in os.listdir(input_folder): if filename.lower().endswith(('.png', '.jpg', '.jpeg', '.bmp', '.gif')): input_path = os.path.join(input_folder, filename) output_path = os.path.join(output_folder, filename) with Image.open(input_path).convert('L') as img: # Convert to grayscale if needed image_array = np.array(img) normalized_image = normalize_image(image_array) save_normalized_image(normalized_image, output_path) ``` #### 解释 - 函数`normalize_image`用于将图像数组的像素值归一化到 `[0, 1]` 的范围内[^1]。 - 函数`save_normalized_image`负责将归一化后的图像保存回磁盘，同时将其转换为适合存储的数据类型 (`uint8`)。 - 主函数`normalize_images_in_folder`遍历输入文件夹中的所有图像文件，调用上述辅助函数完成归一化操作并将结果保存至输出文件夹[^3]。