准备工作

环境咱们还是使用 Python3.8 和 pycharm2021 即可

实现原理

将视频分为音频和画面；
画面中出现人脸和目标比对，相应人脸进行打码；
处理后的视频添加声音；

模块

手动安装一下 cv2 模块，pip install opencv-python 安装

素材工具

我们需要安装一下 ffmpeg 音视频转码工具
Python实现视频自动打码功能

代码解析

导入需要使用的模块

import cv2  
import face_recognition  # 人脸识别库  99.7%    cmake  dlib  face_recognition
import subprocess

# python学习资料、本项目素材加群 815624229 免费领取

将视频转为音频

def video2mp3(file_name):
    """
    :param file_name: 视频文件路径
    :return:
    """
    outfile_name = file_name.split('.')[0] + '.mp3'
    cmd = 'ffmpeg -i ' + file_name + ' -f mp3 ' + outfile_name
    print(cmd)
    subprocess.call(cmd, shell=False)

打码

def mask_video(input_video, output_video, mask_path='mask.jpg'):
    """
    :param input_video: 需打码的视频
    :param output_video: 打码后的视频
    :param mask_path: 打码图片
    :return:
    """
    # 读取图片
    mask = cv2.imread(mask_path)
    # 读取视频
    cap = cv2.VideoCapture(input_video)
    # 视频  fps  width  height
    v_fps = cap.get(5)
    v_width = cap.get(3)
    v_height = cap.get(4)

    # 设置写入视频参数  格式MP4
    # 画面大小
    size = (int(v_width), int(v_height))
    fourcc = cv2.VideoWriter_fourcc('m', 'p', '4', 'v')

    # 输出视频
    out = cv2.VideoWriter(output_video, fourcc, v_fps, size)

    # 已知人脸
    known_image = face_recognition.load_image_file('tmr.jpg')
    biden_encoding = face_recognition.face_encodings(known_image)[0]

    cap = cv2.VideoCapture(input_video)

    while (cap.isOpened()):
        ret, frame = cap.read()
        if ret:
            # 检测人脸
            # 人脸区域
            face_locations = face_recognition.face_locations(frame)

            for (top_right_y, top_right_x, left_bottom_y, left_bottom_x) in face_locations:
                print((top_right_y, top_right_x, left_bottom_y, left_bottom_x))
                unknown_image = frame[top_right_y - 50:left_bottom_y + 50, left_bottom_x - 50:top_right_x + 50]
                if face_recognition.face_encodings(unknown_image) != []:
                    unknown_encoding = face_recognition.face_encodings(unknown_image)[0]

                    # 对比人脸
                    results = face_recognition.compare_faces([biden_encoding], unknown_encoding)
                    # [True]
                    # 贴图
                    if results == [True]:
                        mask = cv2.resize(mask, (top_right_x - left_bottom_x, left_bottom_y - top_right_y))
                        frame[top_right_y:left_bottom_y, left_bottom_x:top_right_x] = mask
            out.write(frame)


        else:
            break

音频添加到画面

def video_add_mp3(file_name, mp3_file):
    """
    :param file_name: 视频画面文件
    :param mp3_file:  视频音频文件
    :return:
    """
    outfile_name = file_name.split('.')[0] + '-f.mp4'
    subprocess.call('ffmpeg -i ' + file_name + ' -i ' + mp3_file + ' -strict -2 -f mp4 ' + outfile_name, shell=False)

完整代码

import cv2 
import face_recognition  # 人脸识别库  99.7%    cmake  dlib  face_recognition
import subprocess

def video2mp3(file_name):

    outfile_name = file_name.split('.')[0] + '.mp3'
    cmd = 'ffmpeg -i ' + file_name + ' -f mp3 ' + outfile_name
    print(cmd)
    subprocess.call(cmd, shell=False)


def mask_video(input_video, output_video, mask_path='mask.jpg'):

    # 读取图片
    mask = cv2.imread(mask_path)
    # 读取视频
    cap = cv2.VideoCapture(input_video)
    # 视频  fps  width  height
    v_fps = cap.get(5)
    v_width = cap.get(3)
    v_height = cap.get(4)

    # 设置写入视频参数  格式MP4
    # 画面大小
    size = (int(v_width), int(v_height))
    fourcc = cv2.VideoWriter_fourcc('m', 'p', '4', 'v')

    # 输出视频
    out = cv2.VideoWriter(output_video, fourcc, v_fps, size)

    # 已知人脸
    known_image = face_recognition.load_image_file('tmr.jpg')
    biden_encoding = face_recognition.face_encodings(known_image)[0]

    cap = cv2.VideoCapture(input_video)

    while (cap.isOpened()):
        ret, frame = cap.read()
        if ret:
            # 检测人脸
            # 人脸区域
            face_locations = face_recognition.face_locations(frame)

            for (top_right_y, top_right_x, left_bottom_y, left_bottom_x) in face_locations:
                print((top_right_y, top_right_x, left_bottom_y, left_bottom_x))
                unknown_image = frame[top_right_y - 50:left_bottom_y + 50, left_bottom_x - 50:top_right_x + 50]
                if face_recognition.face_encodings(unknown_image) != []:
                    unknown_encoding = face_recognition.face_encodings(unknown_image)[0]

                    # 对比人脸
                    results = face_recognition.compare_faces([biden_encoding], unknown_encoding)
                    # [True]
                    # 贴图
                    if results == [True]:
                        mask = cv2.resize(mask, (top_right_x - left_bottom_x, left_bottom_y - top_right_y))
                        frame[top_right_y:left_bottom_y, left_bottom_x:top_right_x] = mask
            out.write(frame)


        else:
            break


def video_add_mp3(file_name, mp3_file):

    outfile_name = file_name.split('.')[0] + '-f.mp4'
    subprocess.call('ffmpeg -i ' + file_name + ' -i ' + mp3_file + ' -strict -2 -f mp4 ' + outfile_name, shell=False)


if __name__ == '__main__':
    # 1.
    video2mp3('cut.mp4')
    # 2.
    mask_video(input_video='cut.mp4',output_video='output.mp4')
    # 3.
    video_add_mp3(file_name='output.mp4',mp3_file='cut.mp3')