-
Notifications
You must be signed in to change notification settings - Fork 0
/
detect_tears.py
159 lines (144 loc) · 6.48 KB
/
detect_tears.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
import numpy as np
import matplotlib.pyplot as plt
from matplotlib import patches
from skimage import io
from scipy import ndimage as ndi
from skimage.filters import gaussian
import pickle
from skimage.feature import register_translation
from skimage.feature import canny
import os
import time
from glob import glob
# target_folder = 'F:/yankai_levitation_transfer_highspeed_videos/10Hz_300V_sinewave_1/whole_video_1/'
# raw_frames_suffix = 'FASTS4_000000/'
def process_one_folder(target_folder,
raw_frames_suffix = '',
thresh_1 = 15,
start_frame = 1,
do_canny=False,
do_figures = False,
low_threshold=30,
high_threshold=150,
radius=False,
center=False,
sum_0 = 31476558.0):
video_folder = target_folder + raw_frames_suffix
frames_folder = target_folder + 'frames_1/'
events_folder = target_folder + 'events_pickle/'
misc_folder = target_folder + 'misc/'
for my_folder in [frames_folder, events_folder, misc_folder]:
if not os.path.exists(my_folder):
os.makedirs(my_folder)
filelist = []
# find all "imgNNN" subfolders
img_subfolders = sorted(glob(video_folder + "img*/"))
for video_subfolder in img_subfolders:
for file in sorted(os.listdir(video_subfolder)):
if file.endswith(".jpg"):
filelist.append(os.path.join(video_subfolder, file))
# filelist = sorted(filelist)
print(filelist)
print('Found {0} jpg frames.'.format(len(filelist)))
frame_id_max = len(filelist) - 1
def load_image(frame_id):
# return np.sum(io.imread('{0}{1:07d}.jpg'.format(video_folder, frame_id)).astype(float), axis=2)
return np.sum(io.imread(filelist[frame_id]).astype(float), axis=2)
# sum_0 = np.sum(load_image(0))
def load_normed(frame_id):
image = load_image(frame_id)
return image*sum_0/np.sum(image)
# plt.imshow(load_normed(1), cmap='plasma', vmax=256*3)
# plt.show()
# Load picture and detect edges
image = load_normed(1)
edges = canny(image, sigma=3, low_threshold=low_threshold, high_threshold=high_threshold)
fill_coins = ndi.binary_fill_holes(edges)
if not center:
center = ndi.center_of_mass(fill_coins)
if not radius:
radius = np.sqrt((np.sum(fill_coins)/np.pi))
print(radius)
np.save(misc_folder + 'radius', radius)
np.save(misc_folder + 'center', center)
np.save(misc_folder + 'first_frame', image)
plt.scatter(center[1], center[0], color='red')
image[edges]=1000
plt.imshow(fill_coins)
plt.show()
int_rad = int(round(radius*1.1))
ref_patch = image[int(round(center[0]))-int_rad:int(round(center[0]))+int_rad,
int(round(center[1]))-int_rad:int(round(center[1]))+int_rad]
plt.imshow(ref_patch)
plt.show()
prev_image = load_normed(0)
# center=(200, 200)
t0 = time.time()
for frame_id in range(start_frame,frame_id_max):
print(frame_id)
prev_image = load_normed(frame_id-1)
image = load_normed(frame_id)
target_patch = image[int(round(center[0])) - int_rad:int(round(center[0])) + int_rad,
int(round(center[1])) - int_rad:int(round(center[1])) + int_rad]
shift, error, diffphase = register_translation(src_image=ref_patch, target_image=target_patch,
upsample_factor=4)
if do_canny:
edges = canny(image, sigma=3, low_threshold=20, high_threshold=110)
fill_coins = ndi.binary_fill_holes(edges)#, origin=[int(round(center[0])), int(round(center[1]))])
center_canny = ndi.center_of_mass(fill_coins)
center -= np.array(shift)
diffimg = image-prev_image
diffimg2 = gaussian(diffimg, sigma=(2,2))
if do_figures:
fig, axarr = plt.subplots(ncols=4, nrows=1, figsize=(14, 4))
plt.subplots_adjust(wspace=0, hspace=0)
for ax in axarr:
ax.set_axis_off()
img_for_show = np.copy(image)
axarr[0].imshow(img_for_show, cmap='gray')
axarr[0].scatter(center[1], center[0], color='red', alpha=0.5)
if do_canny:
axarr[1].imshow(fill_coins)
axarr[1].scatter(center[1], center[0], color='blue', alpha=0.5)
axarr[1].scatter(center_canny[1], center_canny[0], s=40, marker='x', color='black', alpha=0.5)
axarr[2].imshow(diffimg2, cmap='seismic', vmin=-120, vmax=120)
diffimg3 = gaussian(np.abs(diffimg), sigma=(2,2))
diffimg3[:int(round(center[0])), :] = 0
if do_figures:
im = axarr[3].imshow(diffimg3, vmin=0, vmax=150)
loc_max = np.unravel_index(np.argmax(diffimg3, axis=None), diffimg3.shape)
if np.max(diffimg3) > thresh_1:
for_pickle = [frame_id, np.max(diffimg3), loc_max, center]
pickle.dump(for_pickle, open(target_folder + 'events_pickle/{0:07d}.p'.format(frame_id), "wb"))
if do_figures:
for ax in axarr:
e1 = patches.Ellipse((loc_max[1], loc_max[0]), 50, 50,
angle=0, linewidth=2, fill=False, zorder=2, color='red')
ax.add_patch(e1)
if do_figures:
axarr[3].scatter(center[1], center[0], color='red')
plt.colorbar(im)
fig.savefig('{0}frames_1/{1:07d}'.format(target_folder, frame_id))
# plt.show()
plt.close(fig)
prev_image = np.copy(image)
print('Speed was {0:.2f} frames/min'.format((frame_id_max - start_frame)/(time.time()-t0)*60))
plt.show()
if __name__ == '__main__':
# process_one_folder('H:/FASTS4_2020-10-15_1409_000000/img0000/',
# raw_frames_suffix='',
# thresh_1=25,
# start_frame=1,
# radius=50,
# center=(90,150),
# do_figures=False
# )
process_one_folder('E:/fastcam/2020_nov_10/FASTS4_11-10_1000rpm_000000/',
raw_frames_suffix='',
thresh_1=20,
start_frame=1,
radius=35,
center=(50,110),
do_figures=False,
sum_0 = 31476558.0/8
)