-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathgradually_sequencelayer.py
346 lines (322 loc) · 10.7 KB
/
gradually_sequencelayer.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
import cv2
import numpy as np
import Mymodify
import os
inpath = "D:\\experiment\\pic\\batch\\"
outpath = "D:\\out\\try\\batch\\"
files = os.listdir(inpath)
# 按照从中心点右侧为0点逆时针一圈编号的相对位置进行标记8个矩阵
def localpixel(i1, j1, i, j, re, tag):
if i1 == i and j1 == j + 1:
re[i1, j1, 0] = tag
elif i1 == i - 1 and j1 == j + 1:
re[i1, j1, 1] = tag
elif i1 == i - 1 and j1 == j:
re[i1, j1, 2] = tag
elif i1 == i - 1 and j1 == j - 1:
re[i1, j1, 3] = tag
elif i1 == i and j1 == j - 1:
re[i1, j1, 4] = tag
elif i1 == i + 1 and j1 == j - 1:
re[i1, j1, 5] = tag
elif i1 == i + 1 and j1 == j:
re[i1, j1, 6] = tag
elif i1 == i + 1 and j1 == j + 1:
re[i1, j1, 7] = tag
# re[i, j, 8] 中心像素点 0是交叉 1是小 2是大 -1是两个区域的差值一样
# re[i, j, 9] 能分区 大区最小减去小区最大 不能分区 -1
# re[i, j, 0-7] 1是小 2是大 0是噪声没标记原始值 3是交叉
# 噪声点0，小边点1，大边点2，内部点3
for file in files:
if not file.endswith(".jpg"):
continue
file = file[:file.index(".")]
if len(file) == 0:
continue
src = file
raw = cv2.imread(inpath + src + ".jpg")
raw2 = cv2.cvtColor(raw, cv2.COLOR_BGR2GRAY)
# raw2 = cv2.bilateralFilter(raw2, 7, 50, 50)
np.savetxt(outpath + src + " gray" + ".csv", raw2, fmt="%d", delimiter=',')
cv2.imwrite(outpath + src + "gray" + ".jpg", raw2)
for th in range(2, 1, -2):
# 存储结果
re = np.zeros((raw2.shape[0], raw2.shape[1], 10))
# 存储噪声
noise_re = np.zeros((raw2.shape[0], raw2.shape[1]))
for i in range(1, raw2.shape[0] - 1):
for j in range(1, raw2.shape[1] - 1):
a, b, noise = Mymodify.Denoising(raw2, i, j)
num_a = []
num_b = []
num_noise = []
for k in a:
num_a.append(raw2[k[0]][k[1]])
for k in b:
num_b.append(raw2[k[0]][k[1]])
for k in noise:
num_noise.append(raw2[k[0]][k[1]])
num_a = sorted(num_a)
num_b = sorted(num_b)
num_noise = sorted(num_noise)
for k in noise:
noise_re[noise[0][0], noise[0][1]] += 1
if min(num_a) > max(num_b) and (min(num_a) - max(num_b)) > th:
# 可以分为两个部分，寻找中心像素点应该为哪个部分
mindiff1 = 255
mindiff2 = 255
for k in num_a:
if abs(int(raw2[i, j]) - k) < mindiff1:
mindiff1 = abs(int(raw2[i, j]) - k)
for k in num_b:
if abs(int(raw2[i, j]) - k) < mindiff2:
mindiff2 = abs(int(raw2[i, j]) - k)
if mindiff1 > mindiff2:
re[i, j, 8] = 1
elif mindiff1 < mindiff2:
re[i, j, 8] = 2
else:
re[i, j, 8] = -1
# 中心像素点区分度的大小
re[i, j, 9] = num_a[0] - num_b[-1]
# 分区之后的周围八个像素点分别相对位置八层的标记
for n in a:
tag = 2
i1 = n[0]
j1 = n[1]
localpixel(i1, j1, i, j, re, tag)
for n in b:
tag = 1
i1 = n[0]
j1 = n[1]
localpixel(i1, j1, i, j, re, tag)
elif min(num_b) > max(num_a) and (min(num_b) - max(num_a)) > th:
# 可以分为两个部分，寻找中心像素点应该为哪个部分
mindiff1 = 255
mindiff2 = 255
for k in num_a:
if abs(int(raw2[i, j]) - k) < mindiff1:
mindiff1 = abs(int(raw2[i, j]) - k)
for k in num_b:
if abs(int(raw2[i, j]) - k) < mindiff2:
mindiff2 = abs(int(raw2[i, j]) - k)
if mindiff1 > mindiff2:
re[i, j, 8] = 2
elif mindiff1 < mindiff2:
re[i, j, 8] = 1
else:
re[i, j, 8] = -1
# 中心像素点区分度的大小
re[i, j, 9] = num_b[0] - num_a[-1]
# 分区之后的周围八个像素点分别相对位置八层的标记
for n in a:
tag = 1
i1 = n[0]
j1 = n[1]
localpixel(i1, j1, i, j, re, tag)
for n in b:
tag = 2
i1 = n[0]
j1 = n[1]
localpixel(i1, j1, i, j, re, tag)
else:
# 无法分为两个区域，有交集
# 中心像素点区分度的大小
re[i, j, 9] = -1
# 分区之后的周围八个像素点分别相对位置八层的标记
for n in a:
tag = 3
i1 = n[0]
j1 = n[1]
localpixel(i1, j1, i, j, re, tag)
for n in b:
tag = 3
i1 = n[0]
j1 = n[1]
localpixel(i1, j1, i, j, re, tag)
for n in noise:
tag = 3
i1 = n[0]
j1 = n[1]
localpixel(i1, j1, i, j, re, tag)
# 存储八层，对八邻域进行统计的结果
for i in range(8):
np.savetxt(outpath + src + " layer" + str(i) + ".csv", re[:, :, i], fmt="%d", delimiter=',')
# 对9层结果进行汇总标记
tag = np.zeros((raw2.shape[0], raw2.shape[1]))
for i in range(1, raw2.shape[0] - 1):
for j in range(1, raw2.shape[1] - 1):
for k in range(9):
if re[i, j][k] == 1:
tag[i, j] += 1
elif re[i, j][k] == 2:
tag[i, j] += 10
np.savetxt(outpath + src + " tag" + ".csv", tag, fmt="%d", delimiter=',')
# 对9层结果判断大小区域过渡区
tagre = np.zeros((raw2.shape[0], raw2.shape[1]))
for i in range(1, raw2.shape[0] - 1):
for j in range(1, raw2.shape[1] - 1):
if tag[i, j] % 10 != 0 and tag[i, j] // 10 % 10 == 0:
tagre[i, j] = 1 # 小边
elif tag[i, j] % 10 == 0 and tag[i, j] // 10 % 10 != 0:
tagre[i, j] = 2 # 大边
elif tag[i, j] % 10 == 0 and tag[i, j] // 10 % 10 == 0:
tagre[i, j] = 3 # 内部
elif tag[i, j] % 10 != 0 and tag[i, j] // 10 % 10 != 0:
tagre[i, j] = 0 # 矛盾过渡
np.savetxt(outpath + src + " tagre" + ".csv", tagre[:, :], fmt="%d", delimiter=',')
# 显示过渡点
guodu_show = np.zeros((tagre.shape[0], tagre.shape[1]))
for i in range(tagre.shape[0]):
for j in range(tagre.shape[1]):
if tagre[i, j] == 0:
guodu_show[i, j] = 255
cv2.imwrite(outpath + src + str(th) + "guodu" + ".jpg", guodu_show)
# 存储区分度
np.savetxt(outpath + src + " diff" + ".csv", re[:, :, 9], fmt="%d", delimiter=',')
# 存储中心像素点的分类
np.savetxt(outpath + src + " center" + ".csv", re[:, :, 8], fmt="%d", delimiter=',')
# 存储噪声点
np.savetxt(outpath + src + " noise" + ".csv", noise_re[:, :], fmt="%d", delimiter=',')
# 中值处理过渡点的函数
def smooth_mid_gray(tag, raw):
xxx = [-1, -1, -1, 0, +1, +1, +1, 0]
yyy = [+1, 0, -1, -1, -1, 0, +1, +1]
temp = np.zeros((raw2.shape[0], raw2.shape[1]))
for i in range(1, tag.shape[0] - 1):
for j in range(1, tag.shape[1] - 1):
if tag[i, j] > 0:
l = []
min_gray = 0
for k in range(8):
if tag[i + xxx[k], j + yyy[k]] == -1:
l.append(raw[i + xxx[k], j + yyy[k]])
if len(l) >= 3:
if len(l) % 2 == 0:
re1 = sorted(l)[int(len(l) / 2)]
re2 = sorted(l)[int((len(l) - 2) / 2)]
if abs(re1 - int(raw[i, j])) > abs(re2 - int(raw[i, j])):
raw[i, j] = re2
else:
raw[i, j] = re1
else:
raw[i, j] = sorted(l)[int(len(l) / 2)]
temp[i, j] = 1
for i in range(1, tag.shape[0] - 1):
for j in range(1, tag.shape[1] - 1):
if temp[i, j] == 1:
tag[i, j] = -1
return tag
# 均值处理内部点的函数
def smooth_gray(tag, raw):
xxx = [-1, -1, -1, 0, +1, +1, +1, 0]
yyy = [+1, 0, -1, -1, -1, 0, +1, +1]
flag = 0
for i in range(1, tag.shape[0] - 1):
for j in range(1, tag.shape[1] - 1):
sum = 0
num = 0
if tag[i, j] > 0:
flag = 1
sum += raw[i, j]
num += 1
for k in range(8):
if tag[i + xxx[k], j + yyy[k]] == tag[i, j]:
sum += raw[i + xxx[k], j + yyy[k]]
num += 1
raw[i, j] = int(sum / num)
return flag
# ge = 1
# guodu = np.zeros((raw2.shape[0], raw2.shape[1]))
# for i in range(tagre.shape[0]):
# for j in range(tagre.shape[1]):
# if tagre[i, j] == 0:
# guodu[i, j] = 255
# for l in range(ge):
# temp = guodu.copy()
# for i in range(guodu.shape[0]):
# for j in range(guodu.shape[1]):
# if temp[i, j] == 0:
# temp[i, j] = -1
# for k in range(10):
# smooth_mid_gray(temp, raw2)
# for i in range(guodu.shape[0]):
# for j in range(guodu.shape[1]):
# if temp[i, j] == -1:
# temp[i, j] = 0
# cv2.imwrite(outpath + "temp" + str(th) + "____" + src + ".jpg", temp)
# cv2.imwrite(outpath + "4__raw" + str(th) + "____" + src + ".jpg", raw2)
# 平滑的过程 与 L0进行比较
# 迭代平滑
ge = 1
inner = np.zeros((raw2.shape[0], raw2.shape[1]))
for i in range(inner.shape[0]):
for j in range(inner.shape[1]):
if tagre[i, j] == 3:
inner[i, j] = 255
edge_big = np.zeros((raw2.shape[0], raw2.shape[1]))
for i in range(inner.shape[0]):
for j in range(inner.shape[1]):
if tagre[i, j] == 2:
edge_big[i, j] = 255
edge_small = np.zeros((raw2.shape[0], raw2.shape[1]))
for i in range(inner.shape[0]):
for j in range(inner.shape[1]):
if tagre[i, j] == 1:
edge_small[i, j] = 255
guodu = np.zeros((raw2.shape[0], raw2.shape[1]))
for i in range(inner.shape[0]):
for j in range(inner.shape[1]):
if tagre[i, j] == 0:
guodu[i, j] = 255
normal = np.zeros((raw2.shape[0], raw2.shape[1]))
for i in range(guodu.shape[0]):
for j in range(guodu.shape[1]):
if guodu[i, j] != 255:
normal[i, j] = raw2[i,j]
cv2.imwrite(outpath + "normal" + str(th) + "____" + src + ".jpg", normal)
# 平滑内部点
# for i in range(5):
# smooth_gray(inner, raw2)
# # cv2.imwrite(outpath + "1__raw" + str(th) + "____" + src + ".jpg", raw2)
# # 用平滑内部点处理大边
# for l in range(ge):
# temp = edge_big.copy()
# for i in range(guodu.shape[0]):
# for j in range(guodu.shape[1]):
# if inner[i, j] == 255:
# temp[i, j] = -1
# # 大边点为255>0
# # 内部点为 -1
# for k in range(5):
# smooth_mid_gray(temp, raw2)
# # cv2.imwrite(outpath + "2__raw" + str(th) + "____" + src + ".jpg", raw2)
# # 用平滑内部点处理小边
# for l in range(ge):
# temp = edge_small.copy()
# for i in range(guodu.shape[0]):
# for j in range(guodu.shape[1]):
# if inner[i, j] == 255:
# temp[i, j] = -1
# # 小边点为255>0
# # 内部点为 -1
# for k in range(5):
# temp = smooth_mid_gray(temp, raw2)
# cv2.imwrite(outpath + "3__raw" + str(th) + "____" + src + ".jpg", raw2)
for l in range(ge):
temp = guodu.copy()
for i in range(guodu.shape[0]):
for j in range(guodu.shape[1]):
if temp[i, j] == 0:
temp[i, j] = -1
# 过渡点为255>0
# 非过渡点为 -1
for k in range(15):
smooth_mid_gray(temp, raw2)
# 显示处理之后剩余过渡点
for i in range(guodu.shape[0]):
for j in range(guodu.shape[1]):
if temp[i, j] == -1:
temp[i, j] = 0
cv2.imwrite(outpath + "temp" + str(th) + "____" + src + ".jpg", temp)
cv2.imwrite(outpath + "4__raw" + str(th) + "____" + src + ".jpg", raw2)