-
Notifications
You must be signed in to change notification settings - Fork 0
/
template_dock.py
272 lines (223 loc) · 10.7 KB
/
template_dock.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
import sys
import os
import cv2
import numpy as np
from PyQt5.QtCore import *
from PyQt5.QtGui import *
from PyQt5.QtWidgets import *
from toggle_switch import Switch
class TemplatePanel(QWidget):
def __init__(self, template_filename:str=None, matching:bool=False, parent=None):
super().__init__(parent)
self.filename = template_filename
self.need_matching = matching
self.template_images = None
self.initAssets()
self.initUI()
def initAssets(self):
pass
def initUI(self):
pass
def matchTemplates(self, image:np.ndarray, templates:list):
base_h,base_w = image.shape[:2]
### Plain
hit_scores = np.zeros(len(templates), dtype=np.float32)
hit_locs = []
for i,template in enumerate(templates):
#print(template)
temp_h,temp_w = template.shape[:2]
if template is None or temp_h > base_h or temp_w > base_w:
hit_scores[i] = -1.0
hit_locs.append((-1,-1))
else:
background = image.astype(np.float32) - 128
object = template.astype(np.float32) - 128
score_image = cv2.matchTemplate(background, object, cv2.TM_CCORR_NORMED)
#score_image = cv2.matchTemplate(image, template, cv2.TM_CCORR_NORMED)
#score_image = cv2.matchTemplate(image, template, cv2.TM_CCOEFF_NORMED)
min_value,max_value,min_loc,max_loc = cv2.minMaxLoc(score_image)
hit_scores[i] = max_value
hit_locs.append(max_loc)
'''
### Multiprocessing
args_list = []
for template in templates:
args_list.append([image.copy(),template,cv2.TM_CCOEFF_NORMED])
score_images = None
score_locs = None
with multiprocessing.Pool(processes=4) as pool:
score_images = pool.starmap(cv2.matchTemplate, args_list) # for multiple arguments
score_locs = pool.map(cv2.minMaxLoc, score_images) # for single argument
# with multiprocessing.Pool(processes=2) as pool:
# score_images = pool.starmap(cv2.matchTemplate, args_list) # for multiple arguments
# with multiprocessing.Pool(processes=2) as pool:
# score_locs = pool.map(cv2.minMaxLoc, score_images) # for single argument
hit_scores = np.zeros(len(templates), dtype=np.float)
hit_locs = []
for i,[_,max_score,_,max_loc] in enumerate(score_locs):
hit_scores[i] = max_score
hit_locs.append(max_loc)
'''
###
i = np.argmax(hit_scores)
score = hit_scores[i]
loc = hit_locs[i]
return i,score,loc
def matchFeatures(self, image):
image_keypoints,image_descriptors = self.featureExtractor.detectAndCompute(image, None)
res = []
for template_dict in self.templateFeatures:
template = template_dict['image']
template_keypoints = template_dict['keypoints']
template_descriptors = template_dict['descriptors']
if template is not None and template_keypoints is not None and template_descriptors is not None:
matches = self.flannMatcher.knnMatch(template_descriptors, image_descriptors, k=2)
ratio_thresh = 0.5
good_matches = []
for m,n in matches:
if m.distance < ratio_thresh * n.distance:
good_matches.append(m)
template_points = []
image_points = []
for m in good_matches:
template_points.append(template_keypoints[m.queryIdx].pt)
image_points.append(image_keypoints[m.trainIdx].pt)
template_points,image_points = np.float32((template_points,image_points))
H,status = cv2.findHomography(template_points, image_points, cv2.RANSAC, 3.0)
h,w = template.shape[:2]
template_border = np.float32([[[0, 0], [0, h - 1], [w - 1, h - 1], [w - 1, 0]]])
image_border = cv2.perspectiveTransform(template_border, H)
#cv2.polylines(sceneImage, [np.int32(sceneBorder)], True, (255, 255, 0), 5)
res.append({'matches':len(good_matches), 'points':len(template_keypoints), 'border':image_border})
else:
res.append({'matches':0, 'points':len(template_keypoints), 'border':None})
return res
if __name__ == "__main__":
app = QApplication(sys.argv)
win = TemplatePanel()
win.show()
sys.exit(app.exec_())
copied = image.copy()
results = []
'''
##################################################################
for i,roi in enumerate(self.imageROIs):
y,yend,x,xend = [roi[1], roi[1]+roi[3], roi[0], roi[0]+roi[2]]
roi_image = image[y:yend,x:xend,:]
### Color Processing
foreground_mask,a_mask,b_mask,rect = self.processImage(roi_image, bin_range, a_range, b_range)
if foreground_mask is not None:
common_mask = None
if a_mask is not None and b_mask is not None:
common_mask = cv2.bitwise_and(a_mask, b_mask)
foregound_count = np.count_nonzero(foreground_mask == 255)
chroma_count = 0
if a_mask is not None:
chroma_count += np.count_nonzero(a_mask == 255)
if b_mask is not None:
chroma_count += np.count_nonzero(b_mask == 255)
if common_mask is not None:
chroma_count -= np.count_nonzero(common_mask == 255)
chroma_ratio = int(chroma_count/foregound_count*100.0+0.5)
results.append('{0:03d}'.format(chroma_ratio))
# Eye candies
psuedo_image = np.zeros(roi_image.shape, dtype=np.uint8)
psuedo_image[foreground_mask == 255] = (255, 255, 255)
if a_mask is not None: psuedo_image[a_mask == 255] = (255, 0, 0)
if b_mask is not None: psuedo_image[b_mask == 255] = (0, 0, 255)
if common_mask is not None: psuedo_image[common_mask == 255] = (255, 0, 255)
roi_image = cv2.addWeighted(roi_image, 0.4, psuedo_image, 0.6, 0.0)
if rect is not None:
box = cv2.boxPoints(rect)
box = np.int0(box) # Points of 4 corners
cv2.drawContours(roi_image, [box], 0, (0,255,0), 1)
copied[y:yend,x:xend,:] = roi_image
else:
results.append('bad')
cv2.drawContours(copied, [np.array([[x,y],[xend-1,y],[xend-1,yend-1],[x,yend-1]])], 0, self.roiColors[i], 1)
'''
### Template matching
for i, roi in enumerate(self.imageROIs):
y, yend, x, xend = [roi[1], roi[1] + roi[3], roi[0], roi[0] + roi[2]]
roi_image = image[y:yend, x:xend, :]
yuv_image = cv2.cvtColor(roi_image, cv2.COLOR_RGB2YUV)
bin_image = cv2.inRange(yuv_image, bin_range[0], bin_range[1])
i, score, loc = self.matchTemplates(bin_image, [item['image'] for item in self.templateFeatures])
# Eye candies
template = self.templateFeatures[i]['image']
temp_h, temp_w = template.shape[:2]
template_mask = np.zeros(bin_image.shape, dtype=np.uint8)
template_mask[loc[1]:loc[1] + temp_h, loc[0]:loc[0] + temp_w] = template
psuedo_image = np.zeros(roi_image.shape, dtype=np.uint8)
psuedo_image[bin_image == 255] = (255, 255, 255)
psuedo_image[template_mask == 255] = (255, 0, 0)
roi_image = cv2.addWeighted(roi_image, 0.5, psuedo_image, 0.5, 0.0)
copied[y:yend, x:xend, :] = roi_image
if score <= 0.009:
results.append('bad')
else:
percent_score = int(score * 100.0 + 0.5)
results.append('{0:03d}'.format(percent_score))
# cv2.drawContours(copied, [np.array([[x,y], [xend-1,y], [xend-1,yend-1], [x,yend-1]])], 0, self.roiColors[i], 1)
###
'''
### Feature match
for i,roi in enumerate(self.imageROIs):
y,yend,x,xend = [roi[1], roi[1]+roi[3], roi[0], roi[0]+roi[2]]
roi_image = image[y:yend,x:xend,:]
yuv_image = cv2.cvtColor(roi_image, cv2.COLOR_RGB2YUV)
bin_image = cv2.inRange(yuv_image, bin_range[0], bin_range[1])
res = self.matchFeatures(bin_image)
print(res)
# Eye candies
template = self.templateFeatures[i]['image']
temp_h, temp_w = template.shape[:2]
template_mask = np.zeros(bin_image.shape, dtype=np.uint8)
template_mask[loc[1]:loc[1]+temp_h, loc[0]:loc[0]+temp_w] = template
psuedo_image = np.zeros(roi_image.shape, dtype=np.uint8)
psuedo_image[bin_image == 255] = (255, 255, 255)
psuedo_image[template_mask == 255] = (255, 0, 0)
roi_image = cv2.addWeighted(roi_image, 0.5, psuedo_image, 0.5, 0.0)
copied[y:yend, x:xend, :] = roi_image
if score <= 0.009:
results.append('bad')
else:
percent_score = int(score * 100.0 + 0.5)
results.append('{0:03d}'.format(percent_score))
cv2.drawContours(copied, [np.array([[x,y], [xend-1,y], [xend-1,yend-1], [x,yend-1]])], 0, self.roiColors[i], 1)
###
'''
'''
### OCR
for i,roi in enumerate(self.imageROIs):
y,yend,x,xend = [roi[1], roi[1]+roi[3], roi[0], roi[0]+roi[2]]
roi_image = image[y:yend,x:xend,:]
yuv_image = cv2.cvtColor(roi_image, cv2.COLOR_RGB2YUV)
bin_image = cv2.inRange(yuv_image, bin_range[0], bin_range[1])
with PyTessBaseAPI() as api:
pil_image = Image.fromarray(bin_image)
api.SetImage(pil_image)
boxes = api.GetComponentImages(RIL.TEXTLINE, True)
print('Found {} textline image components.'.format(len(boxes)))
for i, (im, box, _, _) in enumerate(boxes):
# im is a PIL image object
# box is a dict with x, y, w and h keys
api.SetRectangle(box['x'], box['y'], box['w'], box['h'])
ocrResult = api.GetUTF8Text()
conf = api.MeanTextConf()
print(u"Box[{0}]: x={x}, y={y}, w={w}, h={h}, "
"confidence: {1}, text: {2}".format(i, conf, ocrResult, **box))
# Eye candies
psuedo_image = np.zeros(roi_image.shape, dtype=np.uint8)
psuedo_image[bin_image == 255] = (255, 255, 255)
roi_image = cv2.addWeighted(roi_image, 0.4, psuedo_image, 0.6, 0.0)
copied[y:yend, x:xend, :] = roi_image
cv2.drawContours(copied, [np.array([[x,y], [xend-1,y], [xend-1,yend-1], [x,yend-1]])], 0, self.roiColors[i], 1)
'''
### Display results
if len(results) == 1:
self.resultLCD.display(results[0])
self.resultLCD.repaint()
else:
# TODO: append results to board, and display the decision.
pass