-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtest.py
144 lines (119 loc) · 4.8 KB
/
test.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
import tensorflow as tf
import os
import numpy as np
import cv2
import glob
#参数设置
learning_rate = 0.0001
train_size = 350
batch_size = 50
train_data_path = "train_data"
test_data_path = "test_data"
img_num = 2
n_classes = 2
batch_step = 0
test1 = "test_data\\test1.jpg"
#读取图片，利用canny算子进行边缘检测
img = cv2.imread('tupian/171921-005.bmp')
img_yuan = cv2.resize(img, (1024, 544))
img = cv2.cvtColor(img_yuan, cv2.COLOR_BGR2GRAY)
img = cv2.Canny(img, 50, 130)
#利用霍夫变换检测圆
circles = cv2.HoughCircles(img, cv2.HOUGH_GRADIENT, dp=1, minDist=40, param1=220, param2=30, minRadius=5, maxRadius=40)
#在保存切割的图像之前，清空文件夹内的内容
fileNames = glob.glob(test_data_path + r'\*')
for fileName in fileNames:
if fileName == test1:
pass
else:
os.remove(fileName)
#对图像进行标注，分割，保存
for circle in circles[0]:
x = int(circle[0])
y = int(circle[1])
r = int(circle[2])
img_rect = cv2.rectangle(img, (x-r, y-r), (x+r, y+r), (0, 0, 255))
img_rect_yuan = cv2.rectangle(img_yuan, (x-r, y-r), (x+r, y+r), (0, 0, 255))
img_save = img_rect[y-r:y+r, x-r:x+r]
img_save = cv2.resize(img_save, (64, 64))
cv2.imwrite("test_data/test" + str(img_num) + ".jpg", img_save)
cv2.imshow("ssi",img_rect_yuan)
cv2.imshow("sss",img_rect)
img_num += 1
print("请按任意键继续")
cv2.waitKey()
#提取测试图像
def get_test_images(test_data_path):
test_images = cv2.imread(test_data_path+"/test1.jpg")
test_images = test_images.reshape([1, 64*64*3])
for files in os.listdir(test_data_path):
files = cv2.imread(test_data_path+"/"+files)
files = files.reshape([1, 64*64*3])
test_images = np.vstack((test_images, files))
return test_images
#读取训练图像
def get_file(train_data_path,batch_step):
tubes = cv2.imread(train_data_path+"/1train76.jpg")
tubes = tubes.reshape([1, 64*64*3])
not_tubes = cv2.imread(train_data_path+"/0train209.jpg")
not_tubes = not_tubes.reshape([1, 64*64*3])
label_tubes = np.array([1])
label_not_tubes = np.array([0])
for file in os.listdir(train_data_path):
name = file.split("t")
if name[0] == "1":
file1 = cv2.imread(train_data_path+"/"+file)
file1 = file1.reshape([1, 64*64*3])
tubes = np.vstack((tubes, file1))
label_tubes = np.vstack((label_tubes, np.array([1])))
else:
file0 = cv2.imread(train_data_path+"/"+file)
file0 = file0.reshape([1, 64*64*3])
not_tubes = np.vstack((not_tubes, file0))
label_not_tubes = np.vstack((label_not_tubes, np.array([0])))
image_list = np.vstack((tubes, not_tubes))
label_list = np.vstack((label_tubes, label_not_tubes))
temp = np.hstack((image_list, label_list))
np.random.shuffle(temp)
images = temp[batch_step: batch_step + 50, 0:12288]
labels = temp[batch_step: batch_step + 50, 12288:]
labels = tf.one_hot(labels, 2, 1, axis=1)
labels = tf.reshape(labels, [-1, 2])
batch_step += 1
return images, labels
#设置占位符
xs = tf.placeholder(tf.float32, [None, 64*64*3])
ys = tf.placeholder(tf.float32, [None, 2])
image = tf.reshape(xs, [-1, 64, 64, 3])
##CNN，训练是否为假圆
conv1 = tf.layers.conv2d(inputs=image, filters=16, kernel_size=5, strides=1, padding="same", activation=tf.nn.relu) # ->64,64,16
pool1 = tf.layers.max_pooling2d(conv1, pool_size=2, strides=2) # ->32,32,16
conv2 = tf.layers.conv2d(pool1, 32, 5, 1, "same", activation=tf.nn.relu) # ->32,32,32
pool2 = tf.layers.max_pooling2d(conv2, 2, 2) # ->16,16,32
flat = tf.reshape(pool2, [-1, 16*16*32])
hidden1 = tf.layers.dense(flat, 64)
output = tf.layers.dense(hidden1, 2)
loss = tf.losses.softmax_cross_entropy(onehot_labels=ys, logits=output)
train_op = tf.train.AdamOptimizer(learning_rate).minimize(loss)
accuracy = tf.metrics.accuracy(labels=tf.argmax(ys, 1), predictions=tf.argmax(output, 1))[1]
sess = tf.Session()
init = tf.global_variables_initializer()
init_lo = tf.local_variables_initializer()
sess.run(init)
sess.run(init_lo)
#训练阶段
for step in range(train_size):
sum = -1
train_image, train_label = get_file(train_data_path,batch_step)
train_label = sess.run(train_label)
sess.run(train_op, {xs: train_image, ys: train_label})
loss_ = sess.run(loss, {xs: train_image, ys: train_label})
accuracy_ = sess.run(accuracy, {xs: train_image, ys: train_label})
print("step:%d" % step, "| loss:%.4f" % loss_, "| accuracy:%.4f" % accuracy_)
if step % 50 == 0:
test_image = get_test_images(test_data_path)
output_ = sess.run(output, {xs:test_image})
for i in range(len(output_[:, 1])):
if output_[i, 1] > output_[i, 0]:
sum += 1
print("试管数量为：%d" % sum)