如题，代码如下，可设置多个图片水印及它们的移动位置

功能为：可以添加多个动态移动的水印，还可以设置水印的大小以及移动速度，也可以增加文字水印，重点是这个是批量执行的，可以对目录下的所有视频文件批量添加水印

import cv2
import os
import numpy as np
from moviepy.editor import VideoFileClipdef add_dynamic_watermarks_with_audio(video_path, output_path, watermark_image_paths, watermark_sizes, watermark_speeds, add_text_watermark, text_watermark_text, text_watermark_color):# 读取视频video_clip = VideoFileClip(video_path)audio_clip = video_clip.audio# 获取视频的宽度和高度width, height = video_clip.size# 准备文本水印if add_text_watermark:# 创建用于绘制文本水印的画布text_watermark_canvas = np.zeros((height, width, 3), dtype=np.uint8)# 设置文本水印属性font_face = cv2.FONT_HERSHEY_SIMPLEXfont_scale = 1.5font_thickness = 2# 获取文本水印的大小(text_width, text_height), _ = cv2.getTextSize(text_watermark_text, font_face, font_scale, font_thickness)# 计算文本水印的位置text_x = int((width - text_width) / 2)text_y = int(height - text_height - 10)# 在画布上绘制文本水印cv2.putText(text_watermark_canvas, text_watermark_text, (text_x, text_y), font_face, font_scale, text_watermark_color, font_thickness, cv2.LINE_AA)# 在每一帧上添加水印watermark_positions = [(0, 0)] * len(watermark_image_paths)def process_frame(t, x):nonlocal watermark_positionsframe = video_clip.get_frame(t)# 在帧上添加图片水印for i in range(len(watermark_image_paths)):watermark_image_path = watermark_image_paths[i]watermark_size = watermark_sizes[i]watermark_speed = watermark_speeds[i]# 读取水印图片watermark_image = cv2.imread(watermark_image_path)# 调整水印图片大小watermark_height, watermark_width, _ = watermark_image.shapeif watermark_width > width or watermark_height > height:scale_factor = min(width / watermark_width, height / watermark_height)watermark_image = cv2.resize(watermark_image,(int(watermark_width * scale_factor), int(watermark_height * scale_factor)),interpolation=cv2.INTER_LINEAR,)# 更新水印位置if t % watermark_speed == 0:while True:x = np.random.randint(0, width - watermark_image.shape[1])y = np.random.randint(0, height - watermark_image.shape[0])too_close = False# 检查新位置与已选择水印位置之间的距离for j in range(i):dist = np.sqrt((x - watermark_positions[j][0]) ** 2 + (y - watermark_positions[j][1]) ** 2)if dist < watermark_image.shape[1] or dist < watermark_image.shape[0]:too_close = Truebreakif not too_close:watermark_positions[i] = (x, y)breakelse:x, y = watermark_positions[i]# 在帧上添加水印watermark_resized = cv2.resize(watermark_image, (int(watermark_size * watermark_width), int(watermark_size * watermark_height)))alpha = watermark_resized[:, :, 0] / 255.0for c in range(3):frame_copy = frame.copy()frame_copy[y : y + watermark_resized.shape[0], x : x + watermark_resized.shape[1], c] = (frame[y : y + watermark_resized.shape[0], x : x + watermark_resized.shape[1], c] * (1 - alpha)+ watermark_resized[:, :, c] * alpha)# 在帧上添加文本水印if add_text_watermark:frame_with_text = cv2.addWeighted(frame, 1, text_watermark_canvas, 0.7, 0)frame = frame_with_textcv2.imshow("Watermarked Video", frame)cv2.waitKey(1)return frame# 处理视频每一帧processed_clip = video_clip.fl(lambda gf, t: process_frame(t, gf))# 输出处理后的视频final_clip = processed_clip.set_audio(audio_clip)final_clip.write_videofile(output_path, codec='libx264', audio_codec="aac")return "水印添加完成！"def batch_add_watermarks_in_directory_with_audio(directory, output_directory, watermark_image_paths, watermark_sizes, watermark_speeds, add_text_watermark=False, text_watermark_text=None):# 获取目录中的所有文件file_list = os.listdir(directory)# 遍历文件列表for file_name in file_list:# 检查文件是否为视频文件if file_name.lower().endswith(('.avi', '.mp4', '.mov', '.mkv')):file_path = os.path.join(directory, file_name)# 设置输出文件路径output_file_path = os.path.join(output_directory, file_name)# 添加水印add_dynamic_watermarks_with_audio(file_path, output_file_path, watermark_image_paths, watermark_sizes, watermark_speeds, add_text_watermark, text_watermark_text if add_text_watermark else None, (255, 255, 255))return "批量处理完成！"# 使用示例
if __name__ == '__main__':directory = input("请输入视频文件所在目录路径：")  # 视频文件所在目录output_directory = input("请输入输出视频文件目录路径：")  # 输出视频文件目录# 检查并创建输出目录os.makedirs(output_directory, exist_ok=True)watermark_image_paths = []watermark_sizes = []watermark_speeds = []num_watermarks = int(input("请输入要添加的水印数量："))for i in range(num_watermarks):watermark_image_path = input(f"请输入第{i+1}个水印图片的路径：")watermark_size = float(input(f"请输入第{i+1}个水印图片的大小（0~1之间）："))watermark_speed = int(input(f"请输入第{i+1}个水印图片的速度："))watermark_image_paths.append(watermark_image_path)watermark_sizes.append(watermark_size)watermark_speeds.append(watermark_speed)add_text_watermark_input = input("是否添加文本水印？(y/n) ")if add_text_watermark_input.lower() == 'y':add_text_watermark = Truetext_watermark_text = input("请输入文本水印内容：")else:add_text_watermark = Falsetext_watermark_text = Nonebatch_add_watermarks_in_directory_with_audio(directory, output_directory, watermark_image_paths, watermark_sizes, watermark_speeds, add_text_watermark, text_watermark_text)