import os
from io import BytesIO
import PyPDF2
from PIL import Image
def pdf_to_images(file_path, output_folder):
    # 打開PDF文件
    with open(file_path, 'rb') as file:
        reader = PyPDF2.PdfFileReader(file)
        # 遍歷每一頁
        for page_num in range(reader.getNumPages()):
            page = reader.getPage(page_num)
            # 將PDF頁轉(zhuǎn)換為Pillow Image對象
            img = page_to_image(page)
            # 保存圖片到文件夾
            save_image(img, page_num, output_folder)
def page_to_image(page):
    # 獲取PDF頁的尺寸
    page_size = page.mediaBox
    # 創(chuàng)建空白的Pillow Image對象
    img = Image.new('RGB', (int(page_size.getWidth()), int(page_size.getHeight())), 'white')
    # 將PDF頁渲染到Pillow Image對象中
    img_draw = ImageDraw.Draw(img)
    img_draw.rectangle((0, 0, img.size[0], img.size[1]), fill='white')
    img_draw_img = ImageDraw.Draw(img)
    img_draw_img.drawImage(page, (0, 0))
    return img
def save_image(img, page_num, output_folder):
    # 創(chuàng)建輸出文件夾
    if not os.path.exists(output_folder):
        os.makedirs(output_folder, exist_ok=True)
    # 保存圖片到輸出文件夾
    file_path = os.path.join(output_folder, f'{page_num}.png')
    img.save(file_path)
# 使用示例
pdf_to_images('example.pdf', 'output_folder')