标签： 小工具

比较笨的方法，用来提取PPT课程视频画面，并生成对应的PPT，代码检测黑屏但没有检测白屏，没有检测重复画面（因为有些人讲课会来回翻PPT），因此还有优化空间。内存占用会逐渐增多，不过测试没有出现崩溃的情况。

PS：做完发现可以直接问讲课人要PPT原件，我，，，

import cv2
import os
import numpy as np
from pptx import Presentation
from pptx.util import Inches
from skimage.metrics import structural_similarity as ssim
import tkinter as tk
from tkinter import filedialog, messagebox

# 选择视频和输出目录
def select_video_and_output():
    video_path = filedialog.askopenfilename(title="选择视频文件", filetypes=[("MP4 files", "*.mp4")])
    if not video_path:
        messagebox.showwarning("选择视频", "未选择视频文件")
        return None, None
    
    output_dir = filedialog.askdirectory(title="选择输出目录")
    if not output_dir:
        messagebox.showwarning("选择输出目录", "未选择输出目录")
        return None, None

    pptx_path = os.path.join(output_dir, "output_presentation.pptx")
    return video_path, pptx_path

# 处理视频并生成 PPT
def process_video_to_ppt(video_path, pptx_path):
    os.makedirs("ppt_images", exist_ok=True)
    
    cap = cv2.VideoCapture(video_path)
    _, prev_frame = cap.read()
    prev_gray = cv2.cvtColor(prev_frame, cv2.COLOR_BGR2GRAY)

    frame_count = 0
    slide_count = 0
    images = []
    similarity_threshold = 0.95  # 提高 SSIM 阈值，减少相似图片
    brightness_threshold = 10  # 黑屏检测（平均亮度 < 10 认为是黑屏）

    def process_frame(frame):
        """ 计算 SSIM 相似度，判断是否保存该帧 """
        nonlocal prev_gray, slide_count
        gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
        score = ssim(prev_gray, gray)

        # 计算平均亮度，过滤黑屏
        avg_brightness = np.mean(gray)
        if avg_brightness < brightness_threshold:
            return  # 跳过黑屏帧

        if score < similarity_threshold:  
            img_path = os.path.join("ppt_images", f"slide_{slide_count}.jpg")

            # 确保不同的幻灯片才保存
            if len(images) == 0 or images[-1] != img_path:  
                cv2.imwrite(img_path, frame)
                images.append(img_path)
                slide_count += 1
                prev_gray = gray  # 只在确认变化时更新参考帧

    while cap.isOpened():
        ret, frame = cap.read()
        if not ret:
            break

        # 仅每隔 15 帧处理一次
        if frame_count % 15 == 0:
            process_frame(frame)

        frame_count += 1

    cap.release()
    # cv2.destroyAllWindows()

    # 创建 PPT
    prs = Presentation()
    for img in images:
        slide = prs.slides.add_slide(prs.slide_layouts[5])  # 空白幻灯片
        left, top, width, height = Inches(0), Inches(0), Inches(10), Inches(7.5)
        slide.shapes.add_picture(img, left, top, width, height)

    prs.save(pptx_path)
    messagebox.showinfo("完成", f"PPTX 生成完成: {pptx_path}")

# 主函数
def main():
    root = tk.Tk()
    root.withdraw()  # 隐藏主窗口
    video_path, pptx_path = select_video_and_output()
    if video_path and pptx_path:
        process_video_to_ppt(video_path, pptx_path)

if __name__ == "__main__":
    main()

此前写了个使用cwebp、gif2webp的脚本，但是由于cwebp目前在win的元数据提取存在问题，因此我们可以使用已经支持exif提取和写入的exiftool进行最后一步的转换，这样我们的图片压缩、转码都在官方库得以实现。

前置条件：

cwebp、gif2webp、exiftool三个组件都注册到系统环境变量。python则使用pil库用于分辨率获取。

实现效果：

使用pil库对分辨率进行获取，但是不介入压缩过程，因为cwebp目前没法获取图片分辨率，使用pil库进行是否执行resize的判断。

使用cwebp处理静态png、jpg，使用gif2webp处理gif图，启用mt多线程，压缩质量85，resize到2560最长、宽边，exiftool采用”-overwrite_original”来避免生成两个图片。

测试效果：

该图片原图7M多，压缩质量选择85，可能由于细节较为丰富，压缩到WebP大小仍为1M左右，还是比较大，不过细节保留充分，同时保留了EXIF信息。

import tkinter as tk
from tkinter import filedialog, messagebox
import os
import subprocess
from PIL import Image

def validate_file(input_path):
    input_path = os.path.abspath(input_path)
    if not os.path.exists(input_path):
        raise FileNotFoundError(f"文件 {input_path} 不存在，请检查路径。")
    return input_path

def get_resized_dimensions(width, height, max_size):
    if width > height:
        new_width = max_size
        new_height = int((new_width / width) * height)
    else:
        new_height = max_size
        new_width = int((new_height / height) * width)
    return new_width, new_height

def convert_image(input_path, output_path, new_width=None, new_height=None):
    try:
        file_extension = os.path.splitext(input_path)[1].lower()
        if file_extension == ".gif":
            command = ["gif2webp","mt", input_path, "-o", output_path]
        else:
            if new_width and new_height:
                command = ["cwebp","mt", "-q", "85", "-resize", str(new_width), str(new_height), input_path, "-o", output_path]
            else:
                command = ["cwebp","mt", "-q", "85", input_path, "-o", output_path]
        subprocess.run(command, check=True)
    except subprocess.CalledProcessError as e:
        raise RuntimeError(f"转换工具运行出错: {e}")

def embed_exif(input_path, output_path):
    try:
        command = ["exiftool", "-overwrite_original", "-tagsfromfile", input_path, "-all:all", output_path]
        subprocess.run(command, check=True)
    except subprocess.CalledProcessError as e:
        raise RuntimeError(f"EXIF 数据嵌入失败: {e}")

def convert_to_webp(input_path, max_size=2560):
    try:
        # 验证文件路径
        input_path = validate_file(input_path)
        output_path = os.path.splitext(input_path)[0] + ".webp"

        # 使用 PIL 获取图像分辨率
        with Image.open(input_path) as img:
            width, height = img.size
            if width <= max_size and height <= max_size:
                convert_image(input_path, output_path)
            else:
                new_width, new_height = get_resized_dimensions(width, height, max_size)
                convert_image(input_path, output_path, new_width, new_height)

        # 嵌入 EXIF 数据
        embed_exif(input_path, output_path)

        return f"图片已转换并保存为 {output_path}"

    except (subprocess.CalledProcessError, FileNotFoundError) as e:
        return str(e)
    except Exception as e:
        return f"处理文件时发生错误: {e}"

def select_files():
    file_paths = filedialog.askopenfilenames(
        title="选择图片文件",
        filetypes=[("*所有图片格式", "*.jpg;*.jpeg;*.png;*.gif"),
                   ("JPEG 图片", "*.jpg;*.jpeg"),
                   ("PNG 图片", "*.png"),
                   ("GIF 图片", "*.gif")]
    )
    if file_paths:
        for path in file_paths:
            file_listbox.insert(tk.END, path)

def convert_and_save_batch():
    files = file_listbox.get(0, tk.END)
    if not files:
        messagebox.showerror("错误", "请选择至少一个图片文件！")
        return

    results = [convert_to_webp(file_path) for file_path in files]
    messagebox.showinfo("完成", "\n".join(results))

def clear_list():
    file_listbox.delete(0, tk.END)

root = tk.Tk()
root.title("批量图片转换为 WebP 工具")
root.geometry("600x400")

frame = tk.Frame(root)
frame.pack(pady=10, padx=10, fill=tk.BOTH, expand=True)

scrollbar = tk.Scrollbar(frame, orient=tk.VERTICAL)
file_listbox = tk.Listbox(frame, selectmode=tk.EXTENDED, yscrollcommand=scrollbar.set)
scrollbar.config(command=file_listbox.yview)
scrollbar.pack(side=tk.RIGHT, fill=tk.Y)
file_listbox.pack(side=tk.LEFT, fill=tk.BOTH, expand=True)

button_frame = tk.Frame(root)
button_frame.pack(pady=10)

select_button = tk.Button(button_frame, text="选择文件", command=select_files, width=15)
select_button.grid(row=0, column=0, padx=5)

clear_button = tk.Button(button_frame, text="清空列表", command=clear_list, width=15)
clear_button.grid(row=0, column=1, padx=5)

convert_button = tk.Button(button_frame, text="批量转换", command=convert_and_save_batch, width=15)
convert_button.grid(row=0, column=2, padx=5)

root.mainloop()

此前的代码使用了Pillow库集成的库，这次使用WebP官方库，对GIF、PNG的处理也比较友好。需要添加WebP的库到系统环境变量后使用。

功能实现：

代码使用cwebp、gif2webp两种方式转换不同的格式图片，使用库本身的压缩分辨率方法，压缩图片到2560最长、宽。

对于gif，该代码可以实现对原图动态的保持，此前使用pillow库则可以设定gif的持续或者最后一帧静帧。

遗憾：

没有能够传递exif，win平台中cwebp压根没法有效传递exif信息，显示——Warning: only ICC profile extraction is currently supported on this platform!元数据只有ICC才能支持传递，所以只依靠cwebp是没法很好在win中进行转换的。

这个问题采用exiftool进行了解决，下篇文章可以看到。

import tkinter as tk
from tkinter import filedialog, messagebox
import os
import subprocess
from PIL import Image

def validate_file(input_path):
    input_path = os.path.abspath(input_path)
    if not os.path.exists(input_path):
        raise FileNotFoundError(f"文件 {input_path} 不存在，请检查路径。")
    return input_path

def get_resized_dimensions(width, height, max_size):
    if width > height:
        new_width = max_size
        new_height = int((new_width / width) * height)
    else:
        new_height = max_size
        new_width = int((new_height / height) * width)
    return new_width, new_height

# 使用 cwebp 或 gif2webp 进行转换
def convert_image(input_path, output_path, new_width=None, new_height=None, cwebp_metadata="none", gif2webp_metadata="none"):
    try:
        file_extension = os.path.splitext(input_path)[1].lower()
        if file_extension == ".gif":
            command = ["gif2webp", "-metadata", gif2webp_metadata, "-mt", input_path, "-o", output_path]
        else:
            if new_width and new_height:
                command = ["cwebp", "-q", "85", "-resize", str(new_width), str(new_height), "-metadata", cwebp_metadata, "-mt", input_path, "-o", output_path]
            else:
                command = ["cwebp", "-q", "85", "-metadata", cwebp_metadata, "-mt", input_path, "-o", output_path]
        subprocess.run(command, check=True)
    except subprocess.CalledProcessError as e:
        raise RuntimeError(f"转换工具运行出错: {e}")

def convert_to_webp(input_path, max_size=2560, cwebp_metadata="none", gif2webp_metadata="none"):
    try:
        # 验证文件路径
        input_path = validate_file(input_path)
        output_path = os.path.splitext(input_path)[0] + ".webp"

        # 使用 PIL 获取图像分辨率
        with Image.open(input_path) as img:
            width, height = img.size
            if width <= max_size and height <= max_size:
                convert_image(input_path, output_path, cwebp_metadata=cwebp_metadata, gif2webp_metadata=gif2webp_metadata)
            else:
                new_width, new_height = get_resized_dimensions(width, height, max_size)
                convert_image(input_path, output_path, new_width, new_height, cwebp_metadata=cwebp_metadata, gif2webp_metadata=gif2webp_metadata)

        return f"图片已转换并保存为 {output_path}"

    except (subprocess.CalledProcessError, FileNotFoundError) as e:
        return str(e)
    except Exception as e:
        return f"处理文件时发生错误: {e}"

def select_files():
    file_paths = filedialog.askopenfilenames(
        title="选择图片文件",
        filetypes=[("*所有图片格式", "*.jpg;*.jpeg;*.png;*.gif"),
                   ("JPEG 图片", "*.jpg;*.jpeg"),
                   ("PNG 图片", "*.png"),
                   ("GIF 图片", "*.gif")]
    )
    if file_paths:
        for path in file_paths:
            file_listbox.insert(tk.END, path)

def convert_and_save_batch():
    files = file_listbox.get(0, tk.END)
    if not files:
        messagebox.showerror("错误", "请选择至少一个图片文件！")
        return

    cwebp_metadata = "exif"  # 设置cwebp要复制的元数据类型为exif
    gif2webp_metadata = "xmp"  # 设置gif2webp要复制的元数据类型为xmp
    results = [convert_to_webp(file_path, cwebp_metadata=cwebp_metadata, gif2webp_metadata=gif2webp_metadata) for file_path in files]
    messagebox.showinfo("完成", "\n".join(results))

def clear_list():
    file_listbox.delete(0, tk.END)

# 创建主窗口
root = tk.Tk()
root.title("批量图片转换为 WebP 工具")
root.geometry("600x400")

frame = tk.Frame(root)
frame.pack(pady=10, padx=10, fill=tk.BOTH, expand=True)

scrollbar = tk.Scrollbar(frame, orient=tk.VERTICAL)
file_listbox = tk.Listbox(frame, selectmode=tk.EXTENDED, yscrollcommand=scrollbar.set)
scrollbar.config(command=file_listbox.yview)
scrollbar.pack(side=tk.RIGHT, fill=tk.Y)
file_listbox.pack(side=tk.LEFT, fill=tk.BOTH, expand=True)

button_frame = tk.Frame(root)
button_frame.pack(pady=10)

select_button = tk.Button(button_frame, text="选择文件", command=select_files, width=15)
select_button.grid(row=0, column=0, padx=5)

clear_button = tk.Button(button_frame, text="清空列表", command=clear_list, width=15)
clear_button.grid(row=0, column=1, padx=5)

convert_button = tk.Button(button_frame, text="批量转换", command=convert_and_save_batch, width=15)
convert_button.grid(row=0, column=2, padx=5)

root.mainloop()