-
Notifications
You must be signed in to change notification settings - Fork 1
/
remove_logo.py
202 lines (164 loc) · 6.77 KB
/
remove_logo.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
import cv2
import os
import sys
import numpy as np
import math
import time
# 捕获音频
from moviepy.editor import *
from pydub import AudioSegment
from get_crop_time import get_time
'''
1.读取视频
2.读取视频的音频
3.找到视频的切分点
4.对视频进行切分
5.对音频进行切分
6.对视频调用去水印函数
7.去完水印的视频和音频合成
'''
def get_img_mask(im,file):
shape = im.shape
w = shape[1]
h = shape[0]
# print('shape[0]',shape[0])
# print('shape[1]',shape[1])
if int(h) > int(w):
rect = (math.ceil(0.086 * h), math.ceil(0.41 * w)) # (H, W) 水印框大。 图像高宽为720,1280时为(110,300)
# 左上角的位置 (15,15)
pos1 = (math.ceil(0.02 * w), math.ceil(0.011 * h))
# 右下角的位置 (405,1150)
pos2 = (math.ceil(0.5625 * w), math.ceil(0.9 * h))
nim = np.zeros((shape[0], shape[1]), dtype=np.uint8)
st_lt = 0
# x先遍历高，y遍历宽
for x in range(pos1[1], pos1[1] + rect[0]):
for y in range(pos1[0], pos1[0] + rect[1]):
px = im[x, y] # im[0,1] 0代表图像的高，1代表图像的宽
if sum(px) > 640:
nim[x, y] = 255
st_lt = st_lt + 1 # 像素超过一定数量就判定为右下角， 否则水印在左上角
# st_lt = st_lt + 1 # 像素超过一定数量就判定为右下角， 否则水印在左上角
#print('st_lt', st_lt)
# 左上角logo的像素点为三千多，当低于1000个像素的时候，就认为logo移动到右下角了
if st_lt < 1000:
nim = np.zeros((shape[0], shape[1]), dtype=np.uint8) # clear
# 先遍历宽，高
for x in range(pos2[0], pos2[0] + rect[1]):
for y in range(pos2[1], pos2[1] + rect[0] + 3):
px = im[y, x]
if sum(px) > 700:
nim[y, x] = 255
# for x in range(400, 720):
else:
nim = np.zeros((shape[0], shape[1]), dtype=np.uint8)
st_lt = 0
'''
对于1280*720的图片，左上角坐标为（10，10）（350,125）
右下角的坐标为（940,580）（1270,710）
'''
# x先遍历高，y遍历宽
for x in range(math.ceil(0.0078125 * h), math.ceil(0.0977 * h)):
for y in range(math.ceil(0.0139 * w), math.ceil(0.486 * w)):
px = im[x, y] # im[0,1] 0代表图像的高，1代表图像的宽
if sum(px) > 640:
nim[x, y] = 255
st_lt = st_lt + 1 # 像素超过一定数量就判定为右下角， 否则水印在左上角
#print('st_lt', st_lt)
# 左上角logo的像素点为三千多，当低于1000个像素的时候，就认为logo移动到右下角了
if st_lt < 1000:
nim = np.zeros((shape[0], shape[1]), dtype=np.uint8) # clear
# 先遍历宽，高
for x in range(math.ceil(0.7344 * w), math.ceil(0.9921 * w)):
for y in range(math.ceil(0.805 * h), math.ceil(0.986 * h)):
px = im[y, x]
if sum(px) > 700:
nim[y, x] = 255
# 这里提取的水印的灰度图
# if not os.path.exists("filter"):
# os.makedirs("filter")
# cv2.imwrite("filter/" + file, nim)
return nim
# 传进来一张
def remove_logo(image,file):
mask = get_img_mask(image,file)
kernel = np.ones((3, 3), np.uint8)
dilate = cv2.dilate(mask, kernel, iterations=3)
kernel_2 = np.ones((5, 5), np.uint8)
dilate = cv2.dilate(dilate, kernel_2, iterations=1)
# cv2.imwrite("delate/" + file, dilate)
sp = cv2.inpaint(image, dilate, 7, flags=cv2.INPAINT_TELEA)
#滤波操作
sp = cv2.bilateralFilter(sp, 5, 280, 50)
return sp
if __name__ == '__main__':
input_pa = r'D:\person_work\github\video-qushuiyin\watermark-master\1\input_path'
print('input_pa',input_pa)
filenames = os.listdir(input_pa)
# 遍历每个文件
for fn in filenames:
t0 =time.time()
# 每个文件的完整路径
input_path = os.path.join(input_pa, fn)
print('input_path',input_path)
cap = cv2.VideoCapture(input_path)
# # 读视频文件
videoclip_1 = VideoFileClip(input_path)
vidio_id = os.path.basename(input_path).split('.')[0]
vidio_id_audio = vidio_id + '.wav'
# c=os.path.splitext(input_path)[0]
# 保存去除logo的视频文件
vidio_id_save = vidio_id + '.mp4'
if not os.path.exists("out_path"):
os.makedirs("out_path")
outpath = os.path.join("out_path", vidio_id_save)
tm = get_time(input_path)
print('tm',tm)
# 对视频画面进行截取
# 获取视频分辨率
size = (int(cap.get(cv2.CAP_PROP_FRAME_WIDTH)), int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT)))
#图片名称
frame_png = vidio_id+'.png'
# 输出文件编码，Linux下可选X264
fourcc = cv2.VideoWriter_fourcc('m', 'p', '4', 'v')
# 视频帧率
fps = cap.get(cv2.CAP_PROP_FPS)
print('fps', fps)
success, image = cap.read()
count = 0
success = True
# 输出
out = cv2.VideoWriter(outpath, fourcc, fps, size)
while (count < tm):
success, image = cap.read()
# 去水印，在写入视频
re_logo = remove_logo(image,frame_png)
out.write(re_logo)
count += 1
cap.release()
print('time_all',time.time()-t0)
# # 提取原始视频文件的音频部分
audio_1 = videoclip_1.audio
if not os.path.exists("audio_1_path"):
os.makedirs("audio_1_path")
save_audio_1 = os.path.join("audio_1_path", vidio_id_audio)
audio_1.write_audiofile(save_audio_1)
# 使用剪切音频的方法来对音频剪辑
# 读取音频文件
print('save_audio_1',save_audio_1)
music = AudioSegment.from_wav(save_audio_1)
t = tm/30
# 截取前20秒,tm是求得的秒数
clip = music[:t * 1000]
if not os.path.exists("crop_music"):
os.makedirs("crop_music")
crop_id = vidio_id + '.mp3'
print('crop_id',crop_id)
crop_path = os.path.join('crop_music',crop_id)
print('crop_path',crop_path)
# 保存文件为clip.mp3，格式为mp3
clip.export(crop_path, format='mp3')
# cmd = "rm -rf %s" % res_video_file
# subprocess.call(cmd, shell=True)