-
Notifications
You must be signed in to change notification settings - Fork 1
/
scan.py
364 lines (263 loc) · 10.3 KB
/
scan.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
# -*- coding: utf-8 -*-
"""
OCR back-end server application.
Created on Mon Jul 10 11:00:00 2017
Author: Prasun Roy | CVPRU-ISICAL (http://www.isical.ac.in/~cvpr)
GitHub: https://github.com/prasunroy/ocr
"""
# imports
from __future__ import division
from __future__ import print_function
import cv2
import numpy
import os
from matplotlib import pyplot
# setup environment
# ---- image formats ----
extensions = ['.bmp', '.dib', '.jpeg', '.jpg', '.jpe', '.jp2', '.png', '.webp',
'.pbm', '.pgm', '.ppm', '.sr', '.ras', '.tiff', '.tif']
# ---- opencv version ----
opencv = int(cv2.__version__.split('.')[0])
################################################################################
# end-points of a 1D binary array
def endpoints(array):
end_0 = -1
end_1 = -1
if array[0] != 0:
end_0 = 0
if array[-1] != 0:
end_1 = len(array)
i = 0
j = len(array) - 1
while i <= j and (end_0 < 0 or end_1 < 0):
if array[i] == 0:
i += 1
elif end_0 < 0:
end_0 = i - 1
if array[j] == 0:
j -= 1
elif end_1 < 0:
end_1 = j + 1
if end_0 < 0:
end_0 = 0
if end_1 < 0:
end_1 = len(array)
return (end_0, end_1)
################################################################################
# read image
def imread(path, verbose=False):
image = None
if verbose: print('loading image.................. ', end = '')
if os.path.isfile(path):
extn = os.path.splitext(path)[-1]
if extn in extensions:
image = cv2.imread(path)
if verbose: print('done')
else:
if verbose: print('unsupported format')
else:
if verbose: print('not found')
return image
################################################################################
# preprocess image
def impreprocess(image, blur_kernel_size=(3, 3), thresh=100, verbose=False):
# convert image to grayscale
if verbose: print('converting colorspace.......... ', end = '')
image_gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
if verbose: print('done')
# apply guassian blurring to remove noise
if verbose: print('removing noise from image...... ', end = '')
image_blur = cv2.GaussianBlur(image_gray, blur_kernel_size, 0, 0)
if verbose: print('done')
# threshold image
if verbose: print('thresholding image............. ', end = '')
(_, image_th) = cv2.threshold(image_blur, thresh, 255,
cv2.THRESH_BINARY_INV)
if verbose: print('done')
return [image, image_gray, image_blur, image_th]
################################################################################
# scan image along rows
def imscan_rows(image, line_space_threshold=16, verbose=False):
# image height
im_rows = image.shape[0]
# initialize and populate accumulator
accu_rows = numpy.sum(image, axis=1, dtype='int') // 255
# find line segments along rows
zero_samp = 0
y_samples = []
for row in range(im_rows):
if accu_rows[row] == 0:
zero_samp += 1
elif zero_samp >= line_space_threshold:
y_samples.append(row - zero_samp // 2)
zero_samp = 0
elif len(y_samples) == 0:
y_samples.append(0)
zero_samp = 0
y_samples.append(im_rows - zero_samp // 2)
if verbose: print('found samples along y.......... {}'.format(y_samples))
return [accu_rows, y_samples]
################################################################################
# scan image along columns
def imscan_cols(image, y_samples=[], word_space_threshold=8, verbose=False):
# image width
im_cols = image.shape[1]
# number of detected lines
n_lines = len(y_samples) - 1
# initialize and populate accumulator
accu_cols_list = []
for line in range(n_lines):
accu_cols = numpy.sum(image[y_samples[line]:y_samples[line+1], :],
axis=0, dtype='int') // 255
accu_cols_list.append(accu_cols)
# find word segments along columns of each line segment
line_index = 0
x_samples_list = []
for accu_cols in accu_cols_list:
zero_samp = 0
x_samples = []
for col in range(im_cols):
if accu_cols[col] == 0:
zero_samp += 1
elif zero_samp >= word_space_threshold:
x_samples.append(col - zero_samp // 2)
zero_samp = 0
elif len(x_samples) == 0:
x_samples.append(0)
zero_samp = 0
x_samples.append(im_cols - zero_samp // 2)
x_samples_list.append(x_samples)
if verbose:
print('line {:2d} samples along x........ {}'
.format(line_index, x_samples))
line_index += 1
return [accu_cols_list, x_samples_list]
################################################################################
# draw boundaries on image
def imdraw_boundary(image, y_samples, x_samples_list,
color=(0, 255, 0), width=1):
# line boundaries
for y in y_samples:
cv2.line(image, (0, y), (image.shape[1]-1, y), color, width)
# word boundaries
i = 0
for x_samples in x_samples_list:
for x in x_samples:
cv2.line(image, (x, y_samples[i]), (x, y_samples[i+1]),
color, width)
i += 1
return image
################################################################################
# draw bounding boxes on image
def imdraw_bbox(image, image_th, y_samples, x_samples_list,
color=(0, 255, 0), width=1):
# number of detected lines
n_lines = len(y_samples) - 1
# process objects in image
image_rois = []
for line in range(n_lines):
x_samples = x_samples_list[line]
n_words = len(x_samples) - 1
for word in range(n_words):
row_0 = y_samples[line]
row_1 = y_samples[line+1]
col_0 = x_samples[word]
col_1 = x_samples[word+1]
subimage = image_th[row_0:row_1, col_0:col_1]
accu_rows = numpy.sum(subimage, axis=1, dtype='int')
accu_cols = numpy.sum(subimage, axis=0, dtype='int')
(bbox_row_0, bbox_row_1) = endpoints(accu_rows)
(bbox_col_0, bbox_col_1) = endpoints(accu_cols)
image_roi = subimage[bbox_row_0:bbox_row_1, bbox_col_0:bbox_col_1]
image_rois.append(image_roi)
offset_x = x_samples[word]
offset_y = y_samples[line]
cv2.rectangle(image, (bbox_col_0+offset_x, bbox_row_0+offset_y),
(bbox_col_1+offset_x, bbox_row_1+offset_y),
color, width)
return [image_rois, image]
################################################################################
# plot histogram
def plot_hist(accu_rows, accu_cols_list):
pyplot.figure()
pyplot.title('Horizontal Scan')
pyplot.xlabel('Frequency')
pyplot.ylabel('Row')
pyplot.gca().invert_yaxis()
pyplot.barh(range(len(accu_rows)), accu_rows)
line = 0
for accu_cols in accu_cols_list:
pyplot.figure()
pyplot.title('Vertical Scan of Line {}'.format(line))
pyplot.xlabel('Column')
pyplot.ylabel('Frequency')
pyplot.bar(range(len(accu_cols)), accu_cols)
pyplot.show()
return
################################################################################
# scan image by histogram
def imscanH(path, boundary_color=(0, 255, 0), boundary_width=1,
bbox_color=(255, 0, 0), bbox_width=1, plot=False, verbose=False):
# read image
image = imread(path, verbose=verbose)
# exit if read fails
if image is None:
return
# preprocess image
image_th = impreprocess(image, verbose=verbose)[-1]
# scan image along rows
[accu_rows, y_samples] = imscan_rows(image_th, verbose=verbose)
# scan image along columns
[accu_cols_list, x_samples_list] = imscan_cols(image_th, y_samples,
verbose=verbose)
# draw boundaries on image
image = imdraw_boundary(image, y_samples, x_samples_list,
boundary_color, boundary_width)
[image_rois, image] = imdraw_bbox(image, image_th,
y_samples, x_samples_list,
bbox_color, bbox_width)
# plot histogram
if plot:
plot_hist(accu_rows, accu_cols_list)
return [image, image_rois]
################################################################################
# scan image for contours
def imscanC(path, bbox_color=(0, 255, 0), bbox_width=1, verbose=False):
# read image
image = imread(path, verbose=verbose)
# exit if read fails
if image is None:
return
# preprocess image
image_th = impreprocess(image, verbose=verbose)[-1]
# find contours
if opencv == 3:
(_, contours, _) = cv2.findContours(image_th.copy(), cv2.RETR_EXTERNAL,
cv2.CHAIN_APPROX_NONE)
else:
(contours, _) = cv2.findContours(image_th.copy(), cv2.RETR_EXTERNAL,
cv2.CHAIN_APPROX_NONE)
# find bounding rectangle around each contour
bn_rects = []
for cntr in contours:
bn_rects.append(cv2.boundingRect(cntr))
# sort bounding rectangles from left to right
bn_rects.sort(key=lambda x: x[0])
# process each bounding rectangle
image_rois = []
for rect in bn_rects:
# attributes of bounding rectangle
x = rect[0]
y = rect[1]
w = rect[2]
h = rect[3]
# ignore tiny objects assuming them as noise
if h <= 8:
continue
# draw bounding rectangle on image
cv2.rectangle(image, (x, y), (x+w, y+h), bbox_color, bbox_width)
# extract region of interest from thresholded image using attributes of
# bounding rectangle
image_roi = image_th[y:y+h, x:x+w]
image_rois.append(image_roi)
return [image, image_rois]