Python代码实现PDF与TIFF互转

pip install Spire.Pdf
pip install Pillow

from spire.pdf.common import *
from spire.pdf import *
from PIL import Image
from io import BytesIO

def convert_pdf_to_tiff(input_path, output_path):
    # 初始化PDF文档对象
    pdf_manager = PdfDocument()

    try:
        # 加载PDF文件
        pdf_manager.LoadFromFile(input_path)

        # 存储每页图像的列表
        collected_images = []

        # 遍历PDF中的每一页
        for i in range(pdf_manager.Pages.Count):
            # 渲染页面为图像流
            with pdf_manager.SaveAsImage(i) as image_stream:
                # 将图像流转换为PIL能理解的格式
                pil_img = Image.open(BytesIO(image_stream.ToArray()))
                collected_images.append(pil_img)

        # 如果成功提取了图像
        if collected_images:
            # save_all=True是创建多页TIFF的关键
            # append_images从第二页开始添加
            collected_images[0].save(
                output_path, 
                save_all=True, 
                append_images=collected_images[1:], 
                compression="tiff_lzw"  # LZW无损压缩，广泛支持
            )
            print(f"成功：{output_path}已生成，共包含{len(collected_images)}页。")

    except Exception as e:
        print(f"转换过程中发生错误：{e}")
    finally:
        # 释放资源以节省内存
        pdf_manager.Dispose()

# 执行函数
convert_pdf_to_tiff("Annual_Internal_Audit.pdf", "Audit_Archive_Fixed.tiff")

from spire.pdf.common import *
from spire.pdf import *
from PIL import Image
import io
import os

def convert_tiff_to_pdf(input_tiff, output_pdf_path):
    # 创建一个新的PDF文档
    new_pdf = PdfDocument()

    # 设置页面边距为零，确保图像填充整个页面
    new_pdf.PageSettings.SetMargins(0.0)

    try:
        # 打开TIFF文件
        with Image.open(input_tiff) as multi_frame_tiff:
            # 遍历TIFF中的每一帧
            for i in range(multi_frame_tiff.n_frames):
                multi_frame_tiff.seek(i)
                
                # 将每一帧保存到一个临时缓冲区
                frame_buffer = io.BytesIO()
                multi_frame_tiff.save(frame_buffer, format="PNG")
                frame_buffer.seek(0)
                
                # 将缓冲区中的图像加载到PDF文档中
                pdf_img = PdfImage.FromStream(frame_buffer)

                # 获取图像的宽高
                img_width = pdf_img.PhysicalDimension.Width
                img_height = pdf_img.PhysicalDimension.Height

                # 创建一个与图像尺寸相同的PDF页面
                pdf_page = new_pdf.Pages.Add(SizeF(img_width, img_height))

                # 将图像绘制到PDF页面的画布上
                pdf_page.Canvas.DrawImage(pdf_img, 0.0, 0.0, img_width, img_height)

        # 保存PDF文件
        new_pdf.SaveToFile(output_pdf_path, FileFormat.PDF)
        print(f"PDF文件已成功创建：{output_pdf_path}")

    except Exception as e:
        print(f"处理TIFF时发生错误：{e}")
    finally:
        new_pdf.Dispose()

# 执行函数
convert_tiff_to_pdf("Legacy_Records_Vol_1.tiff", "Digital_Library_V1.pdf")

# 动态添加水印的代码
font = PdfTrueTypeFont("Arial", 20.0, PdfFontStyle.Bold, True)
brush = PdfBrushes.get_Red()
pdf_page.Canvas.SetTransparency(0.5)
pdf_page.Canvas.DrawString("CONFIDENTIAL - INTERNAL ONLY", font, brush, 50, 50)

pip install pdf2image pillow

from pdf2image import convert_from_path
from PIL import Image
import os
def pdf_to_tiff(pdf_path, tiff_path, dpi=200):
    """
    将PDF转换为多页TIFF
    """
    # 将PDF每一页转换为PIL图像列表
    # Windows用户需要添加 poppler_path 参数，例如：poppler_path=r'C:\path\to\poppler\bin'
    images = convert_from_path(pdf_path, dpi=dpi)
    if not images:
        print("PDF未产生任何图像，请检查文件。")
        return
    # 将第一张图像保存为TIFF，并附加剩余图像
    images[0].save(
        tiff_path,
        save_all=True,
        append_images=images[1:],
        compression="tiff_lzw"  # 使用 LZW 无损压缩，可节省空间
    )
    print(f"转换成功，文件已保存至: {tiff_path}")
# 使用示例
pdf_to_tiff("输入.pdf", "输出.tiff", dpi=300)

pip install PyMuPDF pillow

import fitz  # PyMuPDF
from PIL import Image
import io
def pdf_to_tiff_fitz(pdf_path, tiff_path, dpi=200):
    # 1. 打开PDF文档
    doc = fitz.open(pdf_path)
    images = []
    # 2. 遍历每一页，渲染为PNG格式的图像字节流
    for page_num in range(len(doc)):
        page = doc[page_num]
        # 设置缩放矩阵：dpi/72 为标准转换因子
        zoom = dpi / 72
        mat = fitz.Matrix(zoom, zoom)
        pix = page.get_pixmap(matrix=mat)  # 渲染为像素图
        img_bytes = pix.tobytes("png")     # 转为PNG字节流
        img = Image.open(io.BytesIO(img_bytes))
        images.append(img)
    doc.close()
    # 3. 保存为多页TIFF
    if images:
        images[0].save(
            tiff_path,
            save_all=True,
            append_images=images[1:],
            compression="tiff_lzw"
        )
        print(f"转换成功，文件已保存至: {tiff_path}")
# 使用示例
pdf_to_tiff_fitz("输入.pdf", "输出.tiff", dpi=300)

pip install imageconvert

from imageconvert import ImageConvert
# 调用库内置方法，一步完成转换
ImageConvert.pdf_to_images("输入.pdf", "输出文件夹", dpi=300)

from spire.pdf import PdfDocument
from PIL import Image
from io import BytesIO
doc = PdfDocument()
doc.LoadFromFile("输入.pdf")
images = []
for i in range(doc.Pages.Count):
    with doc.SaveAsImage(i) as img_data:
        img = Image.open(BytesIO(img_data.ToArray()))
        images.append(img)
if images:
    images[0].save("输出.tiff", save_all=True, append_images=images[1:])
doc.Dispose()

pip install pillow

from PIL import Image
def tiff_to_pdf(tiff_path, pdf_path):
    # 打开一个多页TIFF文件
    tiff_image = Image.open(tiff_path)
    images = []
    # 遍历每一帧，将图像转换为RGB模式（PDF/A格式要求）
    for i in range(getattr(tiff_image, 'n_frames', 1)):
        tiff_image.seek(i)
        # 转换为RGB模式，避免透明度等兼容性问题
        rgb_image = tiff_image.convert("RGB")
        images.append(rgb_image)
    if images:
        # 将第一页保存为PDF，并附加其余页面
        images[0].save(pdf_path, save_all=True, append_images=images[1:])
        print(f"转换成功，PDF已保存至: {pdf_path}")
# 使用示例
tiff_to_pdf("输入.tiff", "输出.pdf")

pip install img2pdf

import img2pdf
def tiff_to_pdf_img2pdf(tiff_path, pdf_path):
    # 直接读取文件并将TIFF数据转换为PDF
    with open(pdf_path, "wb") as f:
        f.write(img2pdf.convert(tiff_path))
    print(f"转换成功，PDF已保存至: {pdf_path}")
# 使用示例
tiff_to_pdf_img2pdf("输入.tiff", "输出.pdf")