-
Notifications
You must be signed in to change notification settings - Fork 0
/
fitting_circle.py
425 lines (358 loc) · 17.3 KB
/
fitting_circle.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
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
import os
import sys
import time
import math
import argparse
import copy
import json
from pathlib import Path
import multiprocessing as mp
import cv2
import numpy as np
import shape_based_matching_py
DEBUG_MODE = True
prefix = "/home/harry/yafei_temp/shape_based_matching-python_binding/test/"
def angle_train(src, output_dir=None, class_id="yuanzi", angle_range=[-60, 60], use_rot=True):
if isinstance(src, np.ndarray):
img = src
elif isinstance(src, str):
img = cv2.imread(src)
else:
print('please passing valid image path or image in np.format.')
return
if output_dir is not None and not Path(output_dir).exists():
Path(output_dir).mkdir()
detector = shape_based_matching_py.Detector(128, [4])
# order of ny is row col
#img = img[110:380, 130:400] ROI
mask = np.ones((img.shape[0], img.shape[1]), np.uint8)
mask *= 255
padding = 100
padded_img = np.zeros((img.shape[0]+2*padding,
img.shape[1]+2*padding, img.shape[2]), np.uint8)
padded_mask = np.zeros((padded_img.shape[0], padded_img.shape[1]), np.uint8)
padded_img[padding:padded_img.shape[0]-padding, padding:padded_img.shape[1]-padding, :] = \
img[:, :, :]
padded_mask[padding:padded_img.shape[0]-padding, padding:padded_img.shape[1]-padding] = \
mask[:, :]
# cv2.imshow("padded_img", padded_img)
# cv2.imshow("padded_mask", padded_mask)
# cv2.waitKey()
shapes = shape_based_matching_py.shapeInfo_producer(padded_img, padded_mask)
shapes.angle_range = angle_range
shapes.angle_step = 1
shapes.scale_range = [1]
shapes.produce_infos()
infos_have_templ = []
is_first = True
first_id = 0
first_angle = 0
for info in shapes.infos:
to_show = shapes.src_of(info)
templ_id = 0
if is_first:
templ_id = detector.addTemplate(shapes.src_of(info), class_id, shapes.mask_of(info))
first_id = templ_id
first_angle = info.angle
if use_rot:
is_first = False
else:
templ_id = detector.addTemplate_rotate(class_id, first_id,
info.angle-first_angle,
shape_based_matching_py.CV_Point2f(padded_img.shape[1]/2.0, padded_img.shape[0]/2.0))
templ = detector.getTemplates(class_id, templ_id)
for feat in templ[0].features:
to_show = cv2.circle(to_show, (feat.x+templ[0].tl_x, feat.y+templ[0].tl_y), 3, (0, 0, 255), -1)
#cv2.imshow("show templ", to_show)
#cv2.waitKey(1)
cv2.imwrite("test/template.png", to_show)
if templ_id != -1:
infos_have_templ.append(info)
if output_dir is None:
detector.writeClasses(prefix+"%s_templ.yaml")
shapes.save_infos(infos_have_templ, prefix + f"{class_id}_info.yaml")
else:
detector.writeClasses(os.path.join(output_dir, "%s_templ.yaml"))
shapes.save_infos(infos_have_templ, os.path.join(output_dir, f"{class_id}_info.yaml"))
def angle_test(src, class_id, similarity, use_rot, templ_dir='templ_info'):
if isinstance(src, np.ndarray):
test_img = copy.deepcopy(src)
elif isinstance(src, str):
test_img = cv2.imread(src)
else:
print('please passing valid image path or image in np.format.')
return None
print('start locating...')
detector = shape_based_matching_py.Detector(128, [4])
ids = []
ids.append(class_id)
detector.readClasses(ids, os.path.join(templ_dir, "%s_templ.yaml"))
producer = shape_based_matching_py.shapeInfo_producer()
infos = producer.load_infos(os.path.join(templ_dir, f"{class_id}_info.yaml"))
#test_img = cv2.imread(prefix+"case1/test.png")
padding = 250
padded_img = np.zeros((test_img.shape[0]+2*padding,
test_img.shape[1]+2*padding, test_img.shape[2]), np.uint8)
padded_img[padding:padded_img.shape[0]-padding, padding:padded_img.shape[1]-padding, :] = \
test_img[:, :, :]
stride = 16
img_rows = int(padded_img.shape[0] / stride) * stride
img_cols = int(padded_img.shape[1] / stride) * stride
img = np.zeros((img_rows, img_cols, padded_img.shape[2]), np.uint8)
img[:, :, :] = padded_img[0:img_rows, 0:img_cols, :]
matches = detector.match(img, similarity, ids)
if len(matches) == 0:
print('Template not found in test image.')
return None
top5 = 1
if top5 > len(matches):
top5 = 1
for i in range(top5):
match = matches[i]
templ = detector.getTemplates(class_id, match.template_id)
r_scaled = 1062/2.0*infos[match.template_id].scale
train_img_half_width = 1213/2.0 + 100
train_img_half_height = 1208/2.0 + 100
img = cv2.circle(img, (round(match.x), round(match.y)), 3, (0, 200, 255), -1)
x = match.x - templ[0].tl_x + train_img_half_width
y = match.y - templ[0].tl_y + train_img_half_height
print('train half width {}, height {}, tl {} {}'.format(train_img_half_width, train_img_half_height, templ[0].tl_x, templ[0].tl_y))
print('center x, y {} {}'.format(x, y))
img = cv2.circle(img, (round(x), round(y)), 3, (0, 255, 255), -1)
for feat in templ[0].features:
img = cv2.circle(img, (feat.x+match.x, feat.y+match.y), 3, (0, 0, 255), -1)
# cv2 have no RotatedRect constructor?
print('match.template_id: {}'.format(match.template_id))
print('match.similarity: {}'.format(match.similarity))
print('center in test image {} {}'.format(x, y))
print('image shape {}'.format(img.shape))
print('rotating angle {}'.format(infos[match.template_id].angle))
#cv2.imshow("img", img)
#cv2.waitKey(0)
if DEBUG_MODE:
cv2.imwrite("test/result.png", img)
return (round(x-250), round(y-250), infos[match.template_id].angle)
def metrology2D(img, cx, cy, r, perpendicular_len, tangential_len, rotate_angle, num=6):
if img is None:
print('Image is None in metraology2D.')
return None
if len(img.shape) == 3 and img.shape[-1] == 3:
img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
img = cv2.GaussianBlur(img, (3,3), 1.0)
ear_interval = 60
angle = round(rotate_angle + 360) % 360
arc_angle = 2*math.pi/num
first_tl_x = max(0, round(cx - tangential_len))
first_tl_y = max(0, round(cy - r - perpendicular_len))
rect_width = tangential_len * 2
rect_height = perpendicular_len * 2
circle_contour = []
for i in range(num):
projected_mid_x, projected_mid_y, tranformed_x, transformed_y = 0, 0, 0, 0
response_val = np.zeros(rect_height)
max_response_index = 0;
projection_points = []
arc_angle = (angle + 30 + ear_interval * i) / 180 * math.pi
max_x = int(first_tl_x + rect_width)
max_y = int(first_tl_y + rect_height)
img_height, img_width = img.shape[0], img.shape[1]
for row in range(first_tl_y, max_y):
projected_mean = 0
for col in range(first_tl_x, max_x):
transformed_x = min(max(0, round((col - cx) * math.cos(arc_angle) - (cy - row) * math.sin(arc_angle) + cx)), img_width-1)
transformed_y = round((col - cx) * math.sin(arc_angle) + (cy - row) * math.cos(arc_angle) + img.shape[0] - cy)
transformed_y = min(img_height-1, max(0, img_height - transformed_y))
if transformed_x < 0 or transformed_x >= img_width or transformed_y < 0 or transformed_y >= img_height:
print(f'transform point [{col}, {row}] to [{transformed_x} {transformed_y}] which is beyond image boundary. image width and height: {img.shape[1]} {img.shape[0]}')
return None
if col == round(first_tl_x + rect_width/2):
projected_mid_x = transformed_x
projected_mid_y = transformed_y
#cv2.circle(img, (transformed_x, transformed_y), 5, (20, 0, 0))
projected_mean += img[transformed_y, transformed_x]
#print('projected mean: {}, rect width: {}'.format(projected_mean, rect_width))
projected_mean /= rect_width
response_val[row-first_tl_y] = projected_mean
projection_points.append((projected_mid_x, projected_mid_y))
max_response_index = max_response(response_val)
circle_contour.append(projection_points[max_response_index])
cv2.circle(img, projection_points[max_response_index], 5, (20, 0, 0))
if DEBUG_MODE:
color_img = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR)
cv2.imwrite("test/circle_points.png", color_img)
return circle_contour
def max_response(vals):
max_res = 0
index = 0
for i in range(len(vals)-1):
vals[i] = abs(vals[i+1] - vals[i])
if vals[i] > max_res:
max_res = vals[i]
index = i+1
print('max index {}'.format(index))
return index
def defect_gate(src_img, min_ellipse, fitting_radius, fitting_center, rotation_angle):
if len(src_img.shape) == 3:
src_img = cv2.cvtColor(src_img, cv2.COLOR_BGR2GRAY)
min_ellipse_center = (round(min_ellipse[0][0]), round(min_ellipse[0][1]))
# otsu binarization
ret, src_img = cv2.threshold(src_img, 0, 255, cv2.THRESH_BINARY+cv2.THRESH_OTSU)
edges = cv2.Canny(src_img, 60, 60*2)
cv2.circle(edges, min_ellipse_center, fitting_radius-50, 0, -1)
if DEBUG_MODE:
cv2.imwrite("test/edges.png", edges)
conts_indices = []
conts, _ = cv2.findContours(edges, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
for i,cont in enumerate(conts):
rect = cv2.minAreaRect(cont)
area = rect[1][0] * rect[1][1]
if area > 40000:
conts_indices.append(i)
edge_circle = np.zeros_like(src_img)
for i in conts_indices:
cv2.drawContours(edge_circle, conts, i, (255))
if DEBUG_MODE:
cv2.imwrite("test/edge_cleaned.png", edge_circle)
portion_angle = 60
angle_arc = 0
tolerant_offset = 10
measurement_rects = []
left_corner, right_corner, tmp_point, mid_point = {}, {}, {}, {}
for i in range(3):
angle_arc = (round(portion_angle*(2*i+1) + rotation_angle - 15 + 360) % 360) / 180 * math.pi;
left_corner['x'] = -min_ellipse[1][0] / 2.0 * math.sin(angle_arc) + min_ellipse_center[0]
left_corner['y'] = min_ellipse[1][0] / 2.0 * math.cos(angle_arc) + edge_circle.shape[0] - min_ellipse_center[1]
left_corner['y'] = edge_circle.shape[0] - left_corner['y']
angle_arc = round(portion_angle*(2*i+1) + rotation_angle) % 360;
if angle_arc <= 15 or angle_arc >= 345:
mid_point['x'] = round(min_ellipse_center[0])
mid_point['y'] = round(min_ellipse_center[1] - min_ellipse[1][0]/2.0 - tolerant_offset)
elif angle_arc >= 75 and angle_arc <= 105:
mid_point['x'] = round(min_ellipse_center[0] - min_ellipse[1][0]/2.0 - tolerant_offset)
mid_point['y'] = round(min_ellipse_center[1])
elif angle_arc >= 165 and angle_arc <= 195:
mid_point['x'] = round(min_ellipse_center[0])
mid_point['y'] = round(min_ellipse_center[1] + min_ellipse[1][0]/2.0 + tolerant_offset)
elif angle_arc >= 255 and angle_arc <= 285:
mid_point['x'] = round(min_ellipse_center[0] + min_ellipse[1][0]/2.0 + tolerant_offset)
mid_point['y'] = round(min_ellipse_center[1])
else:
mid_point['x'] = -1
mid_point['y'] = -1
angle_arc = (portion_angle*(2*i+1) + rotation_angle + 15) / 180.0 * math.pi
right_corner['x'] = -min_ellipse[1][0] / 2.0 * math.sin(angle_arc) + min_ellipse_center[0]
right_corner['y'] = min_ellipse[1][0] / 2.0 * math.cos(angle_arc) + edge_circle.shape[0] - min_ellipse_center[1]
right_corner['y'] = edge_circle.shape[0] - right_corner['y']
if mid_point['x'] != -1 :
if right_corner['x'] < left_corner['x'] and right_corner['x'] < mid_point['x']:
tmp_point['x'] = right_corner['x']
elif mid_point['x'] < left_corner['x'] and mid_point['x'] < right_corner['x']:
tmp_point['x'] = mid_point['x']
else:
tmp_point['x'] = left_corner['x']
if right_corner['y'] < left_corner['y'] and right_corner['y'] < mid_point['y']:
tmp_point['y'] = right_corner['y']
elif mid_point['y'] < left_corner['y'] and mid_point['y'] < right_corner['y']:
tmp_point['y'] = mid_point['y']
else:
tmp_point['y'] = left_corner['y']
if left_corner['x'] > right_corner['x'] and left_corner['x'] > mid_point['x']:
right_corner['x'] = left_corner['x']
elif mid_point['x'] > left_corner['x'] and mid_point['x'] > right_corner['x']:
right_corner['x'] = mid_point['x']
else:
pass
if left_corner['y'] > right_corner['y'] and left_corner['y'] > mid_point['y']:
right_corner['x'] = left_corner['x']
elif mid_point['y'] > left_corner['y'] and mid_point['y'] > right_corner['y']:
right_corner['y'] = mid_point['y']
else:
pass
else:
tmp_point = left_corner
w = abs(tmp_point['y'] - right_corner['y'])
h = abs(tmp_point['x'] - right_corner['x'])
measurement_rects.append({'x': round(min(tmp_point['x'], right_corner['x'])), 'y': round(min(tmp_point['y'], right_corner['y'])), 'w': round(w), 'h': round(h)})
cv2.circle(edge_circle, min_ellipse_center, 5, (255))
for rect in measurement_rects:
squared_distance = 0
for row in range(rect['y'], rect['y']+rect['h']):
for col in range(rect['x'], rect['x']+rect['w']):
if edge_circle[row,col] > 0:
squared_distance += (math.sqrt((col-min_ellipse_center[0])**2 + (row-min_ellipse_center[1])**2) - min_ellipse[1][0]/2)**2
print('squared distance {}'.format(squared_distance))
if squared_distance > 3000:
return True
return False
def check_gate(src_test_img, align_info, fitting_circle_radius=520, perpendicular_half_len=40, tangential_half_len=20):
test_img = copy.deepcopy(src_test_img)
circle_contour = metrology2D(test_img, align_info[0], align_info[1], fitting_circle_radius, perpendicular_half_len, tangential_half_len, align_info[2])
#circle_contour = metrology2D(test_img, align_info[0]+80, align_info[1]+20, fitting_circle_radius, perpendicular_half_len, tangential_half_len, align_info[2])
if circle_contour is None:
return True
min_ellipse = cv2.fitEllipse(np.array(circle_contour))
test_img = cv2.ellipse(test_img, min_ellipse, (255, 0, 0))
if DEBUG_MODE:
cv2.imwrite("test/circled.png", test_img)
if defect_gate(src_test_img, min_ellipse, fitting_circle_radius, (align_info[0]+80, align_info[1]+20), align_info[2]):
print('Slicing gate NG...')
return True
return False
class GateProcess(mp.Process):
def __init__(self, img_queue, out_queue):
super(GateProcess, self).__init__()
self.img_queue = img_queue
self.out_queue = out_queue
def run(self):
if self.img_queue is None or self.out_queue is None:
print('img_queue/out_queue is None.')
return
while True:
img = self.img_queue.get(block=True)
align_info = angle_test(img, 'painting', 90, True)
if align_info is not None:
ng = check_gate(img, align_info, 520, 40, 20)
print('check ng:{}'.format(ng))
self.out_queue.put(ng)
print('exit....')
def produce(q):
src_img = cv2.imread(prefix+"small.png")
while True:
q.put(src_img)
time.sleep(1)
if __name__ == "__main__":
if len(sys.argv) != 2:
print(f'usage: {sys.argv[0]} train|test|demo')
exit(0)
if sys.argv[1] == 'train':
angle_train(sys.argv[2], sys.argv[3], "circle_3ears_smooth", angle_range=[-60, 60], use_rot=True)
elif sys.argv[1] == 'test':
with open('templ_info/locations.json', 'r') as f:
infos = json.load(f)
print(infos)
info = infos['circle_3ears_smooth']
src_img = cv2.imread(prefix+"small.png")
align_info = angle_test(src_img, 'circle_3ears_smooth', 90, True)
src_img = cv2.imread(prefix+"small.png")
img = cv2.circle(src_img, (align_info[0], align_info[1]), 3, (0, 255, 0), -1)
align_info = (align_info[0]+info['off_x'], align_info[1]+info['off_y'], align_info[2])
print(align_info)
img = cv2.circle(img, (align_info[0], align_info[1]), 3, (0, 0, 255), -1)
cv2.imwrite('test/origin_point.png', img)
if align_info is not None:
check_gate(src_img, align_info, round(info['r']), 40, 20)
else:
img_queue = mp.Queue()
output_queue = mp.Queue()
producer = mp.Process(target=produce, args=(img_queue,))
producer.start()
#src_img = cv2.imread(prefix+"small.png")
#img_queue.put(src_img)
gate_process = GateProcess(img_queue, output_queue)
gate_process.start()
producer.join()
gate_process.join()
print(f'ng result: {output_queue.get()}')
#align_info = angle_test(src_img, 'painting', 90, True)
#if align_info is not None:
# check_gate(src_img, align_info, 520, 40, 20)