forked from JunwookHeo/YOLO-OT
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathCreateGT.py
247 lines (216 loc) · 8.85 KB
/
CreateGT.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
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
import argparse
import cv2
import numpy as np
import os
import glob
class CreateGT(object):
class CGT:
gt = None
pos = 0
img_size = (0, 0)
img_name = None
frame = None
mode = None
def parse_config(self):
ap = argparse.ArgumentParser()
ap.add_argument("--mode", default="cgt_yot", help="vtoi:video to image files, vyolo:view yolo gt, \
vrolo: view rolo rt, cgt_yolo:create gt for yolo, cgt_yot:create gt for yot")
ap.add_argument("--image_path", default="../Tennis/Tennis_01/images", help="Path to the image/video file or dir")
args, _ = ap.parse_known_args()
return args
def __init__(self):
opt = self.parse_config()
self.image_path = opt.image_path
self.mode = opt.mode
def check_video_file(self, path):
if not path.lower().endswith(('.avi', '.mp4')):
assert False, 'Error : Unknown file format!!!'
def vtoi(self, path):
self.check_video_file(path)
img_path = os.path.join(os.path.dirname(path), 'images')
if not os.path.exists(img_path):
os.makedirs(img_path)
self.cap = cv2.VideoCapture(self.image_path)
assert self.cap.isOpened(), 'Cannot open source'
while self.cap.isOpened():
pos = int(self.cap.get(cv2.CAP_PROP_POS_FRAMES))
ret, frame = self.cap.read()
if ret:
cv2.imwrite(os.path.join(img_path, '{:04d}.jpg'.format(pos+1)), frame)
print(f'saved {os.path.join(img_path, "{:04d}.jpg".format(pos+1))}')
else:
break;
def draw_yolo_rect(self, img, path):
label_path = path.replace(".png", ".txt").replace(".jpg", ".txt")
if os.path.exists(label_path):
labels = np.loadtxt(label_path).reshape(-1, 5)
for l in labels:
x1 = int(l[1]*img.shape[1] - l[3]*img.shape[1]/2)
y1 = int(l[2]*img.shape[0] - l[4]*img.shape[0]/2)
x2 = int(l[1]*img.shape[1] + l[3]*img.shape[1]/2)
y2 = int(l[2]*img.shape[0] + l[4]*img.shape[0]/2)
print(l)
cv2.rectangle(img, (x1, y1), (x2, y2), (0, 0, 255), 2)
cv2.imshow(path, img)
def vyolo(self, path):
files = [f for f in os.listdir(path) if f.lower().endswith(('.png', '.jpg'))]
pos = 0
end = len(files)
isRun = True
while(isRun):
f = os.path.join(path, files[pos])
cap = cv2.VideoCapture(f)
assert cap.isOpened(), 'Cannot open source'
while cap.isOpened():
ret, frame = cap.read()
if ret:
cv2.imshow(f, frame)
self.draw_yolo_rect(frame, f)
key = cv2.waitKey() & 0xFF
if key == ord('q'):
isRun = False
break
elif key == ord('n'):
pos = min(pos + 1, end - 1)
elif key == ord('p'):
pos = max(pos - 1, 0)
else:
break
cv2.destroyAllWindows()
print(f)
pass
def draw_rolo_rect(self, img, img_path, lbl_path):
if os.path.exists(lbl_path):
try:
labels = np.loadtxt(lbl_path, delimiter='\t', dtype=float).reshape(-1, 4)
except:
labels = np.loadtxt(lbl_path, delimiter=',', dtype=float).reshape(-1, 4)
for l in labels:
x1 = int(l[0])
y1 = int(l[1])
x2 = int(l[0] + l[2])
y2 = int(l[1] + l[3])
print(l)
cv2.rectangle(img, (x1, y1), (x2, y2), (0, 0, 255), 2)
cv2.imshow(img_path, img)
def vrolo(self, path):
img_path = os.path.join(path, 'images')
lbl_path = os.path.join(path, 'labels')
img_files = sorted(glob.glob("%s/*.*" % img_path))
lbl_files = sorted(glob.glob("%s/*.*" % lbl_path))
pos = 0
end = min(len(img_files), len(lbl_files))
isRun = True
while(isRun):
img_f = img_files[pos]
lbl_f = lbl_files[pos]
cap = cv2.VideoCapture(img_f)
assert cap.isOpened(), 'Cannot open source'
while cap.isOpened():
ret, frame = cap.read()
if ret:
cv2.imshow(img_f, frame)
self.draw_rolo_rect(frame, img_f, lbl_f)
key = cv2.waitKey() & 0xFF
if key == ord('q'):
isRun = False
break
elif key == ord('n'):
pos = min(pos + 1, end - 1)
elif key == ord('p'):
pos = max(pos - 1, 0)
else:
break
cv2.destroyAllWindows()
print(img_f, lbl_f)
pass
def cgt(self, path):
files = [f for f in os.listdir(path) if f.lower().endswith(('.png', '.jpg'))]
files.sort()
pos = 0
end = len(files)
isRun = True
gt = self.CGT()
if self.mode == 'cgt_yot':
name = os.path.join(os.path.dirname(path), 'groundtruth_rect.txt')
if os.path.exists(name):
gt.gt = np.loadtxt(name, delimiter=',')
else:
gt.gt = np.zeros((end, 4), dtype=float)
gt.img_name = self.image_path
while(isRun):
f = os.path.join(path, files[pos])
cap = cv2.VideoCapture(f)
assert cap.isOpened(), 'Cannot open source'
while cap.isOpened():
ret, frame = cap.read()
if ret:
gt.pos = pos
gt.frame = frame
gt.mode = self.mode
print(pos, gt.gt[pos])
self.shape = frame.shape
gt.img_size = (frame.shape[1], frame.shape[0])
cv2.imshow(self.image_path, frame)
cv2.setMouseCallback(self.image_path, CreateGT.mouse_event_handler, gt)
key = cv2.waitKey() & 0xFF
if key == ord('q'):
isRun = False
break
elif key == ord('n'):
pos = min(pos + 1, end - 1)
elif key == ord('p'):
pos = max(pos - 1, 0)
else:
break
cv2.destroyAllWindows()
if self.mode == 'cgt_yot':
name = os.path.join(path, 'groundtruth_rect.txt')
np.savetxt(name, gt.gt.astype(int), delimiter=',', fmt='%d')
elif self.mode == 'cgt_yolo':
for i, f in enumerate(files):
name = f.replace(".png", ".txt").replace(".jpg", ".txt")
name = os.path.join(path, name)
np.savetxt(name, [gt.gt[i]], delimiter=',', fmt='%.06f')
def run(self):
if self.mode is 'vtoi':
self.vtoi(self.image_path)
elif self.mode is 'vyolo':
self.vyolo(self.image_path)
elif self.mode is 'vrolo':
self.vrolo(self.image_path)
elif self.mode is 'cgt_yot' or self.mode is 'cgt_yolo':
self.cgt(self.image_path)
else:
assert False, 'Error : Unknown mode!!!'
@staticmethod
def mouse_event_handler(event, x, y, flags, param):
gt = param
rect = gt.gt[gt.pos]
x = max(0, min(x, gt.img_size[0] - 1))
y = max(0, min(y, gt.img_size[1] - 1))
if event == cv2.EVENT_LBUTTONDOWN:
rect[0] = x
rect[1] = y
elif event == cv2.EVENT_LBUTTONUP:
rect[2] = abs(rect[0] - x)
rect[3] = abs(rect[1] - y)
if rect[0] > x:
rect[0] = x
if rect[1] > y:
rect[1] = y
img = gt.frame.copy()
cv2.rectangle(img, (int(rect[0]), int(rect[1])), (int(rect[2]+rect[0]), int(rect[3]+rect[1])), (0, 255, 0), 1)
cv2.imshow(gt.img_name, img)
cv2.displayStatusBar(gt.img_name, f'{gt.pos} : {rect}')
print(gt.pos, rect)
if gt.mode == 'cgt_yolo': # # (X1, Y1, W, H) ==> (Cx, Cy, W, H)
rect[0] = (rect[0] + rect[2]/2)/gt.img_size[0]
rect[1] = (rect[1] + rect[3]/2)/gt.img_size[1]
rect[2] = rect[2]/gt.img_size[0]
rect[3] = rect[3]/gt.img_size[1]
def main(argvs):
cgt = CreateGT()
cgt.run()
if __name__=='__main__':
main('')