-
Notifications
You must be signed in to change notification settings - Fork 6
/
BCD_test.py
181 lines (157 loc) · 8.06 KB
/
BCD_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
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
from model.BCDNET import BCDNET
from utils.EvaluationNew import Evaluation, Index
import torchvision.transforms as Transforms
import time
import glob
import cv2
import numpy as np
import torch
from model.BFE_DPN import BFExtractor
from tqdm import tqdm
import skimage.io as io
import matplotlib.pyplot as plt
if __name__ == "__main__":
print('Starting test...')
# Select device
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
# Define network
BFENet = BFExtractor(n_channels=3, n_classes=1)
BCDNet = BCDNET(n_channels=3, n_classes=1)
# copy model to device
BFENet.to(device=device)
BCDNet.to(device=device)
# Load model.pth
BFENet.load_state_dict(torch.load('BestmIoU_BFE_DPN_epoch91_mIoU_91.864527.pth', map_location=device))
BCDNet.load_state_dict(torch.load('BestmIoU_BCD_DPN_epoch91_mIoU_91.864527.pth', map_location=device))
# Test status
BFENet.eval()
BCDNet.eval()
trans = Transforms.Compose([Transforms.ToTensor()])
# Load dataset path
tests1_path = glob.glob('./samples/WHU/test/image1/*.tif')
tests2_path = glob.glob('./samples/WHU/test/image2/*.tif')
label_path = glob.glob('./samples/WHU/test/label/*.tif')
# Define evaluation index
IoU, c_IoU, uc_IoU, OA, Precision, Recall, F1 = 0, 0, 0, 0, 0, 0, 0
TPSum, TNSum, FPSum, FNSum, C_Sum_or, UC_Sum_or = 0, 0, 0, 0, 0, 0
num=0
f_acc = open('test_acc.txt', 'w')
f_time = open('test_time.txt', 'w')
with tqdm(total=len(label_path), desc='Test Epoch #{}'.format(num), ncols=160) as t:
for tests1_path, tests2_path, label_path in zip(tests1_path, tests2_path, label_path):
starttime = time.time()
# Save path
save_res_path = '.' + tests1_path.split('.')[1] + '_res.png'
save_res_path = save_res_path.replace('image1', 'results')
# Obtaining file name
# Accoring to your own directory to modify the position of the split character.
name = tests1_path.split('/')[5].split('.')[0]
# Read images
t1 = cv2.imread(tests1_path)
t2 = cv2.imread(tests2_path)
GT = cv2.imread(label_path)
label_img = cv2.cvtColor(GT, cv2.COLOR_BGR2GRAY)
test1_img = trans(t1)
test2_img = trans(t2)
test1_img = test1_img.unsqueeze(0)
test2_img = test2_img.unsqueeze(0)
# Copy tensor to device
test1_img = test1_img.to(device=device, dtype=torch.float32)
test2_img = test2_img.to(device=device, dtype=torch.float32)
# output prediction result
list = []
out1, feat1 = BFENet(test1_img)
out2, feat2 = BFENet(test2_img)
list.append(out1)
list.append(out2)
list.append(feat1)
list.append(feat2)
pred_Img = BCDNet(list)
# Get prediction image
pred = np.array(pred_Img.data.cpu()[0])[0]
pred[pred >= 0.5] = 255
pred[pred < 0.5] = 0
# print(num, tests1_path)
# Save result
cv2.imwrite(save_res_path, pred)
#################### Show result ############################
# io.imshow(pred)
"""
pred_ = pred.astype(np.uint8)
pred_ = np.expand_dims(pred_,2)
pred_ = np.repeat(pred_, 3, axis=2)
plt.figure(figsize=(10, 11))
plt.suptitle('Building Change Detection Model') # image
plt.subplot(2, 2, 1), plt.title('T1-time image')
plt.imshow(t1), plt.axis('off')
plt.subplot(2, 2, 2), plt.title('T2-time image')
plt.imshow(t2), plt.axis('off')
plt.subplot(2, 2, 3), plt.title('Ground truth image')
plt.imshow(GT), plt.axis('off')
plt.subplot(2, 2, 4), plt.title('Change detection map')
plt.imshow(pred_), plt.axis('off')
plt.show()
time.sleep(0.5)
"""
##############################################################
endtime = time.time()
if num == 0:
f_time.write('each pair images time\n')
f_time.write(str(num) + ',' + str(starttime) + ',' + str(endtime) + ',' + str(
float('%2f' % (starttime - endtime))) + '\n')
# Accuracy Evaluation
monfusion_matrix = Evaluation(label=label_img, pred=pred)
TP, TN, FP, FN, c_num_or, uc_num_or = monfusion_matrix.ConfusionMatrix()
TPSum += TP
TNSum += TN
FPSum += FP
FNSum += FN
C_Sum_or += c_num_or
UC_Sum_or += uc_num_or
# Save val loss and accuracy
if num == 1:
f_acc.write('=================================================================================\n')
f_acc.write('|Note: (num, FileName, TP, TN, FP, FN)|\n')
f_acc.write('|Note: (ACC: FileName, OA, FA, MA, TE, mIoU, c_IoU, uc_IoU, Precision, Recall, F1)|\n')
f_acc.write('=================================================================================\n')
f_acc.write(str(num) + ',' + str(name) + '.tif' + ',' + str(TP) + ',' + str(TN) + ',' +
str(FP) + ',' + str(FN) + '\n')
num += 1
if num > 50:
Indicators = Index(TPSum, TNSum, FPSum, FNSum, C_Sum_or, UC_Sum_or)
IoU, c_IoU, uc_IoU = Indicators.IOU_indicator()
OA, Precision, Recall, F1 = Indicators.ObjectExtract_indicators()
FA, MA, TE = Indicators.CD_indicators()
# print("OA=", str(float('%4f' % OA)), "^^^^^", "mIoU=", str(float('%4f' % IoU)), "^^^^^", "c_mIoU=", str(float('%4f' % c_IoU)), "^^^^^", "uc_mIoU=", str(float('%4f' % uc_IoU)), "^^^^^", "Precision=",
# str(float('%4f' % Precision)), "^^^^^", "Recall=", str(float('%4f' % Recall)), "^^^^^", "mF1=", str(float('%4f' % F1)))
t.set_postfix({
'OA': OA,
'mIoU': '%.4f' % IoU,
'c_IoU': '%.4f' % c_IoU,
'uc_IoU': '%.4f' % uc_IoU,
'PRE': '%.4f' % Precision,
'REC': '%.4f' % Recall,
'F1': '%.4f' % F1})
t.update(1)
Indicators = Index(TPSum, TNSum, FPSum, FNSum, C_Sum_or, UC_Sum_or)
IoU, c_IoU, uc_IoU = Indicators.IOU_indicator()
OA, Precision, Recall, F1 = Indicators.ObjectExtract_indicators()
FA, MA, TE = Indicators.CD_indicators()
"""
print("OA=", str(float('%4f' % OA)), "^^^^^", "mIoU=", str(float('%4f' % IoU)), "^^^^^", "c_mIoU=",
str(float('%4f' % c_IoU)), "^^^^^", "uc_mIoU=", str(float('%4f' % uc_IoU)), "^^^^^", "Precision=",
str(float('%4f' % Precision)), "^^^^^", "Recall=", str(float('%4f' % Recall)), "^^^^^", "mF1=",
str(float('%4f' % F1)))
"""
f_acc.write('==========================================================================================================\n')
f_acc.write('|SumConfusionMatrix:| TP | TN | FP | FN |\n')
f_acc.write('|SumConfusionMatrix:|' + str(TPSum) + '|' + str(TNSum) + '|' + str(FPSum) + '|' + str(FNSum) + '|\n')
f_acc.write('==========================================================================================================\n')
f_acc.write('|TotalAcc:| OA | FA | MA | TE | mIoU | c_IoU | uc_IoU |Precision| Recall | F1 |\n')
f_acc.write('|TotalAcc:|' + str(float('%4f' % OA)) + '|' + str(float('%4f' % FA)) + '|' + str(float('%4f' % MA)) + '|' + str(float('%4f' % TE))
+ '|' + str(float('%4f' % IoU)) + '|' + str(float('%4f' % c_IoU)) + '|' + str(float('%4f' % uc_IoU)) + '|' +
str(float('%4f' % Precision)) + '|' + str(float('%4f' % Recall)) + '|' + str(float('%4f' % F1)) + '|\n')
f_acc.write(
'==========================================================================================================\n')
f_acc.close()
f_time.close()