Python读取含url图片链接的txt文档方法小结

import os
import random
import string

# 读取face.txt文件中的URL
with open('face.txt', 'r') as f:
    urls = f.read().splitlines()

# 生成额外URL的函数
def generate_extra_urls(base_urls, num_extra):
    extra_urls = []
    for _ in range(num_extra):
        base_url = random.choice(base_urls)
        # 这里我们简单地通过添加随机字符串来模拟新的URL
        # 在实际应用中，你可能需要更复杂的策略来生成有意义的URL
        random_str = ''.join(random.choices(string.ascii_uppercase + string.digits, k=10))
        extra_url = f"{base_url}?extra={random_str}"
        extra_urls.append(extra_url)
    return extra_urls

# 生成200个额外的URL
extra_urls = generate_extra_urls(urls, 200)

# 将原始URL和额外生成的URL合并
all_urls = urls + extra_urls

# 将所有URL写入新的文件
with open('all_faces.txt', 'w') as f:
    for url in all_urls:
        f.write(f"{url}\n")

print("数据增强完成，所有URL已写入all_faces.txt文件。")

import os
import requests
from PIL import Image, ImageOps, ImageEnhance
import numpy as np
import random

# 下载图像的函数
def download_image(url, save_path):
    response = requests.get(url, stream=True)
    if response.status_code == 200:
        with open(save_path, 'wb') as f:
            for chunk in response:
                f.write(chunk)
    else:
        print(f"Failed to download image from {url}")

# 图像增强的函数
def enhance_image(image_path, output_dir):
    # 读取图像
    image = Image.open(image_path)
    
    # 几何变换示例：随机翻转
    if random.choice([True, False]):
        image = ImageOps.flip(image)
    
    # 像素变换示例：添加噪声
    if random.choice([True, False]):
        noise = np.random.normal(0, 10, image.size)
        noisy_image = Image.fromarray(np.clip(image + noise, 0, 255).astype(np.uint8))
        image = noisy_image
    
    # 保存增强后的图像
    output_path = os.path.join(output_dir, f"enhanced_{os.path.basename(image_path)}")
    image.save(output_path)

# 读取face.txt文件中的URL
with open('face.txt', 'r') as f:
    urls = f.read().splitlines()

# 下载并增强图像的目录
download_dir = 'downloaded_images'
enhanced_dir = 'enhanced_images'

# 如果目录不存在，则创建它们
if not os.path.exists(download_dir):
    os.makedirs(download_dir)
if not os.path.exists(enhanced_dir):
    os.makedirs(enhanced_dir)

# 下载并增强图像
for url in urls:
    image_path = os.path.join(download_dir, f"face_{os.path.basename(url.split('/')[-1])}")
    download_image(url, image_path)
    enhance_image(image_path, enhanced_dir)

print("数据增强完成，增强后的图像已保存在", enhanced_dir)

import os
import requests
from PIL import Image
import imgaug as ia
from imgaug import augmenters as iaa

# 下载图像的函数
def download_image(url, save_path):
    response = requests.get(url, stream=True)
    if response.status_code == 200:
        with open(save_path, 'wb') as f:
            for chunk in response:
                f.write(chunk)
    else:
        print(f"Failed to download image from {url}")

# 定义数据增强序列
def create_augmentation_seq():
    seq = iaa.Sequential([
        iaa.Fliplr(0.5),  # 水平翻转，50%的概率
        iaa.Flipud(0.5),  # 垂直翻转，50%的概率
        iaa.Affine(
            rotate=(-20, 20),  # 随机旋转，范围在-20到20度之间
            scale=(0.8, 1.2),  # 随机缩放，范围在0.8到1.2倍之间
            translate_percent={"x": (-0.1, 0.1), "y": (-0.1, 0.1)},  # 随机平移，范围在-10%到10%之间
            mode="constant",  # 使用常数填充新的像素
            cval=(0, 255)    # 填充像素的值范围
        ),
        iaa.AdditiveGaussianNoise(loc=0, scale=(0.0, 0.05*255), per_channel=0.5),  # 添加高斯噪声
        iaa.ContrastNormalization((0.7, 1.3)),  # 对比度归一化
    ], random_order=True)  # 每次应用增强时的顺序是随机的
    return seq

# 主程序
def main():
    # 读取face.txt文件中的URL
    with open('face.txt', 'r') as f:
        urls = f.read().splitlines()

    # 下载图像的目录
    download_dir = 'downloaded_images'
    # 增强后图像的目录
    augmented_dir = 'augmented_images'

    # 创建目录
    if not os.path.exists(download_dir):
        os.makedirs(download_dir)
    if not os.path.exists(augmented_dir):
        os.makedirs(augmented_dir)

    # 创建数据增强序列
    seq = create_augmentation_seq()

    # 下载并增强图像
    for idx, url in enumerate(urls):
        image_path = os.path.join(download_dir, f"face_{idx}.jpg")
        download_image(url, image_path)

        # 读取图像
        image = Image.open(image_path)
        # 将PIL图像转换为numpy数组
        image_array = np.array(image)

        # 应用数据增强
        augmented_images = seq(images=[image_array])

        # 保存增强后的图像
        for i, aug_image in enumerate(augmented_images):
            aug_image_path = os.path.join(augmented_dir, f"face_{idx}_aug_{i}.jpg")
            Image.fromarray(aug_image).save(aug_image_path)

    print("数据增强完成，增强后的图像已保存在", augmented_dir)

# 运行主程序
if __name__ == "__main__":
    main()