-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathpreprocess_square.py
181 lines (141 loc) · 6.35 KB
/
preprocess_square.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
import cv2
import numpy as np
import math
import os
show_Image = False
data_path = r"/home/pwrai/1124_train_photo"
target_path = "%s_preprocessed_square" % data_path
def panelAbstract(srcImage):
# read pic shape
imgHeight, imgWidth = srcImage.shape[:2]
imgHeight = int(imgHeight)
imgWidth = int(imgWidth)
# 二維轉一維
imgVec = np.float32(srcImage.reshape((-1, 3)))
criteria = (cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 10, 1.0)
flags = cv2.KMEANS_RANDOM_CENTERS
ret, label, clusCenter = cv2.kmeans(imgVec, 2, None, criteria, 10, flags)
clusCenter = np.uint8(clusCenter)
clusResult = clusCenter[label.flatten()]
imgres = clusResult.reshape(srcImage.shape)
# image轉成灰階
imgres = cv2.cvtColor(imgres, cv2.COLOR_BGR2GRAY)
cv2.imwrite("test.jpg", imgres)
# image轉成2維，並做Threshold
_, thresh = cv2.threshold(imgres, 127, 255, cv2.THRESH_BINARY_INV)
threshRotate = cv2.merge([thresh, thresh, thresh])
# 印出 threshold後的image
# if cv2.imwrite(r"./Photo/thresh.jpg", threshRotate):
# print("Write Images Successfully")
# 确定前景外接矩形
# find contours
_, contours, hierarchy = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
minvalx = np.max([imgHeight, imgWidth])
maxvalx = 0
minvaly = np.max([imgHeight, imgWidth])
maxvaly = 0
maxconArea = 0
maxAreaPos = -1
for i in range(len(contours)):
if maxconArea < cv2.contourArea(contours[i]):
maxconArea = cv2.contourArea(contours[i])
maxAreaPos = i
print("Contours:", len(contours))
if len(contours) > maxAreaPos:
objCont = contours[maxAreaPos]
else:
print("Error: abnormal contours")
return None # return error code
# cv2.minAreaRect生成最小外接矩形
rect = cv2.minAreaRect(objCont)
for j in range(len(objCont)):
minvaly = np.min([minvaly, objCont[j][0][0]])
maxvaly = np.max([maxvaly, objCont[j][0][0]])
minvalx = np.min([minvalx, objCont[j][0][1]])
maxvalx = np.max([maxvalx, objCont[j][0][1]])
if rect[2] <= -45:
rotAgl = 90 + rect[2]
else:
# 咖啡粉會執行else
rotAgl = rect[2]
if rotAgl == 0:
panelImg = srcImage[minvalx:maxvalx, minvaly:maxvaly, :]
else:
# 咖啡粉會執行else
_, dstRotBW = cv2.threshold(thresh, 127, 255, 0)
# 印出最小外接矩形
_, contours, hierarchy = cv2.findContours(dstRotBW, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
maxcntArea = 0
maxAreaPos = -1
for i in range(len(contours)):
if maxcntArea < cv2.contourArea(contours[i]):
maxcntArea = cv2.contourArea(contours[i])
maxAreaPos = i
x, y, w, h = cv2.boundingRect(contours[maxAreaPos])
# x，y是矩陣左上角的座標，w，h是矩陣的寬與高
# umsize代表 1pixel*umsize = 真實大小
umsize = 90000 / w
# print(umsize)
w = w / 8 # 寬度分為8等分
# 將沒有外圍輪廓的咖啡粉存入panelImg，固定照片大小，因此h以w代替
panelImg = srcImage[int(y + 2 * w):int(y + 6 * w), int(x +2 * w):int(x + 6 * w), :]
# 印出圖片真實大小
print("Image Size:", 4 * w * umsize, " um * ", 4 * w * umsize, " um")
return panelImg
def hist_equal_lab(img):
global show_Image
# Converting image. to LAB Color model
lab = cv2.cvtColor(img, cv2.COLOR_BGR2LAB)
# Splitting the LAB image to different channels
l, a, b = cv2.split(lab)
if show_Image:
cv2.namedWindow('l_channel', cv2.WINDOW_NORMAL)
cv2.imshow('l_channel', l)
cv2.namedWindow("a_channel", cv2.WINDOW_NORMAL)
cv2.imshow('a_channel', a)
cv2.namedWindow("b_channel", cv2.WINDOW_NORMAL)
cv2.imshow('b_channel', b)
# Applying CLAHE to L-channel
clahe = cv2.createCLAHE(clipLimit=3.0, tileGridSize=(8, 8))
cl = clahe.apply(l)
return cl
if __name__ == "__main__":
if not os.path.exists(target_path):
os.mkdir(target_path)
print("Target Path: ", target_path)
error_files = []
for root_Outer, dirs_Outer, files_Outer in os.walk(data_path, topdown=False):
for directory in dirs_Outer:
target_dir = os.path.join(target_path, directory)
if not os.path.exists(target_dir):
os.mkdir(target_dir)
print("Target Dir: ", target_dir)
for root, dirs, files in os.walk(os.path.join(root_Outer, directory), topdown=False):
for name in files:
img_path = os.path.join(root, name)
srcImage = cv2.imread(img_path)
(h, w) = srcImage.shape[:2]
center = (w/2, h/2)
for i in range(4):
rotation_matrix = cv2.getRotationMatrix2D(center, i*90, 1.0)
rstImage = cv2.warpAffine(srcImage, rotation_matrix, (w, h))
rstImage = panelAbstract(rstImage)
if rstImage is None:
error_files.append(img_path)
break
print(rstImage.shape)
rstImage = cv2.resize(rstImage, (1024, 1024), interpolation=cv2.INTER_LINEAR)
print(rstImage.shape)
rstImage = hist_equal_lab(rstImage)
# 印出結果
notcut_filename = '%s_result_%d' % (name.split('.')[0], i)
print('notcut_Filename: ' + notcut_filename)
print("Save in path: ", os.path.join(target_dir, notcut_filename))
for i in range(4):
for j in range(4):
cutImage = rstImage[int(i*256):int((i+1)*256), int(j*256):int((j+1)*256)]
filename = notcut_filename + '[%d][%d].%s' %(i,j, name.split('.')[-1])
print('new_Filename: ' + filename)
if cv2.imwrite(os.path.join(target_dir, filename), cutImage):
print("Write Images Successfully")
print("Error Files:", error_files)