-
Notifications
You must be signed in to change notification settings - Fork 5
/
Copy pathresults_SIFT_SURF_RIFT_SCB_HOPC.py
127 lines (103 loc) · 5.25 KB
/
results_SIFT_SURF_RIFT_SCB_HOPC.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
"""
配准成功率计算：
输入实际的homo,gt homo,点；然后计算点的误差（我们平均选100个点）
"""
import numpy as np
from common.image_process import get_cores_point
from common.thermal_visible import ThermalVisble
from feature_based.SIFT_SURF_ORB_RIFT_SCB_HOPC import sift_registration,surf_registration,orb_registration,rift_registration,scb_registration,hardnet_registration,tfeat_registration
from template_based.CFOG import CFOG_registration
from TAMM_clean.TAMM import TAMM_registration
def vis_correspoint(homo,point_dict,threhold,imgpath):
#遍历每个点，计算其误差
CP_list=[]
num_correct=0
for point in point_dict:
gt_point=get_cores_point(point[0],point[1],homo)
cores_point=point_dict[point][0],point_dict[point][1]
error_point=np.linalg.norm(np.array(gt_point)-np.array(cores_point))
CP_list.append((point, gt_point, cores_point,error_point))
if error_point<threhold:
num_correct=num_correct+1
return CP_list,num_correct
#先都用黑白色的点，后面再调试
# if np.linalg.norm(np.array(gt_point)-np.array(cores_point))>threhold:
def point_generate(imgsize,num_point):
point_list=[]
step=int(imgsize/num_point)
start=int(step)
for index1 in range(start,imgsize,step):
for index2 in range(start,imgsize,step):
point=(index1,index2)
point_list.append(point)
return point_list
def one_img_process(thermal_tiff_path,visible_rgb_path,thermal_rgb_path,method,save_root,cfg=None):
# homo_save_path, error_save_path, point_save_path = save_root
#先缩放和裁剪，得到thermal和visible
thermal_visible_c = ThermalVisble(thermal_tiff_path=thermal_tiff_path,
thermal_rgb_path=thermal_rgb_path,
visible_rgb_path=visible_rgb_path,
scale={'thermal':2*1.15,'visible':1},
crop_size=1000,
attention_flag=False)
dict_method={'sift':sift_registration,
'surf':surf_registration,
'orb':orb_registration,
'scb':scb_registration,
'rift':rift_registration,
'hardnet':hardnet_registration,
'tfeat':tfeat_registration,
# 'hopc': hopc_registration,
'cfog':CFOG_registration,
'tamm':TAMM_registration
}
#在使用该方法得到homography和对应点
if method in ['sift','surf','orb','scb','rift','hopc']:
homo,correspoint=dict_method[method](thermal_visible_c.thermal_tiff,thermal_visible_c.visible_rgb_gray)
elif method in ['hardnet','tfeat']:
homo, correspoint = dict_method[method](thermal_visible_c.thermal_rgb, thermal_visible_c.visible_rgb)
elif method in ['cfog','tamm']:
homo, correspoint = dict_method[method](thermal_visible_c,cfg)
if homo is None:
homo=np.array([[1, 0, 0],
[0, 1, 0],
[0, 0, 1]], dtype=np.float64)
if method=='scb':
homo=np.array(homo).reshape(-1)
homo=[[homo[0],homo[1],homo[2]],[homo[3],homo[4],homo[5]]]
homo.append([0,0,1])
homo=np.array(homo)
print(thermal_tiff_path,method,homo)
thermal_visible_c.homo_warp(homo)
#保存，保存homo,保存对应点，保存可视化结果(缩放裁剪后的图像，配准后的图像)
save_dict={
'homo':'csv',
'visible_rgb':'png',
'thermal_rgb':'png',
'warp_thermal':'png',
'mosaic':'png'
}
if method in ['hardnet','tfeat','sift','surf','orb','rift','hopc']:
thermal_visible_c.draw_matchpoints(correspoint[0], correspoint[1], [], [])
save_dict['matchpoints_img']='jpg'
if method in ['cfog','tamm']:
thermal_visible_c.draw_matchpoints(correspoint[0], correspoint[1], correspoint[2], correspoint[3])
save_dict['matchpoints_img'] = 'jpg'
thermal_visible_c.result_save(save_root, save_dict)
def cfg_forcfog_tamm(method):
patch_size = 40 if method=='tamm' else 100
return {'bin_size':9,'patch_size':patch_size,'search_radius':50,'num_points':625,'thresh':2,'level_max':4,'fea':'CFOG'}
if __name__=='__main__':
#输入图像(原始图像)，输入方法，结果保存地址，得到结果(homo和可视化结果保存，缩放裁剪后的图像，配准后的图像)
method_list=['tamm']#['tfeat','hardnet','orb','sift','surf','rift','scb']
img_name_list = ['12-09-57-454']#'21-11-28-547'#'21-11-32-580',21-11-28-547,21-10-58-550,21-11-14-582,21-11-16-552,21-11-18-549,21-11-20-556,
for img_name in img_name_list:
thermal_path = f'test_img/{img_name}-radiometric.tiff'
thermal_rgb_path = f'test_img/{img_name}-radiometric.jpg'
visible_path = f'test_img/{img_name}-visible.jpg'
for method in method_list:
output_path = f'test_img/out/{method}/{img_name}'
one_img_process(thermal_path, visible_path,thermal_rgb_path,
method=method,
save_root = output_path,
cfg = cfg_forcfog_tamm(method))