-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathutils_focal_length.py
180 lines (153 loc) · 6.4 KB
/
utils_focal_length.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
# %% Definition of Classes and Functions
# value in mm is named without 'px'
# value in pixel is named with 'px'
import os
import pandas as pd
import numpy as np
def get_path():
basepath = os.path.abspath(__file__)
current_path = os.path.dirname(basepath)
return current_path
f = open(os.path.join(get_path(), 'cameraSensors.db'))
CAMSENSORDB = pd.read_csv(f,
names=['make', 'model', 'ccd_width', 'providers'],
header=None, sep=';')
class focal_length_helper:
'''A class for obtaining related informations about focal length
Attributes:
db_path: the path of the db file
img_path: the path of img
fl: focal length in mm
equivalent_fl: equvalent focal length in 35mm
img_width_px: image width in pixel
img_height_px: image height in pixel
'''
def __init__(self, img_path):
super().__init__()
self.img_path = img_path
self.fl, self.efl, self.img_width_px, self.img_height_px, self.make, self.model = self.obtain_exif_exif(
img_path)
def _ccd_width_db(self):
"""obtain ccd width based on make and model information in camera sensors database
Return:
ccd_width: ccd width in mm
"""
df = CAMSENSORDB
import re
self.make = str.split(self.make)[0]
# print(self.make, self.model)
df = df[df['make'].str.contains(self.make, flags=re.IGNORECASE)]
df = df[df['model'].str.contains(self.model, flags=re.IGNORECASE)]
# print(df)
self.ccd_width = df.iat[0, 2] # the third column
def _ccd_width_efl(self):
"""obtain ccd width based on equivalent focal length
1. EFL = FL * 43.27 / \sqrt(ccd width^2+ccd height^2)
2. ccd width/ccd height = image width/image height
Return:
ccd_width: ccd width in mm
"""
ccd_diagonal = 43.27 / (self.efl / self.fl)
self.ccd_width = np.sqrt(np.power(ccd_diagonal, 2) / (1 + np.power(self.img_height_px / self.img_width_px, 2)))
def obtain_focal_length_px(self, method='db'):
'''obtain focal length in pixel
Attributes:
image_path: the absolute path of the photo
method: the method to obtrain focal length in pixel, now support 'db'(default) or 'efl'
'''
if method == 'db':
self._ccd_width_db()
elif method == 'efl':
self._ccd_width_efl()
fl_px = self.img_width_px * self.fl / self.ccd_width
return fl_px
def obtain_focal_length_mm(self):
return self.fl
def obtain_intrinsic_matrix(self, method='db'):
'''Obtain the intrinsic matrix of a photo.
Attributes:
method: the method to obtain focal length in pixel, now support 'db'(default) or 'efl'
Returns:
k: an intrinsic matrix in numpy
'''
fl_px = self.obtain_focal_length_px(method)
k = np.array([[fl_px, 0, self.img_width_px / 2], [0, fl_px, self.img_height_px / 2], [0, 0, 1]])
return k
def obtain_intrinsics(self, method='db'):
'''Obtain the intrinsics of a photo.
Attributes:
method: the method to obtrain focal length in pixel, now support 'db'(default) or 'efl'
Returns:
fl_px: focal length in pixel
u_0:
v_0:
'''
fl_px = self.obtain_focal_length_px(method)
u_0 = self.img_width_px / 2
v_0 = self.img_height_px / 2
return fl_px, u_0, v_0
@staticmethod
def obtain_exif_exif(img_path):
'''Obtain exif using the pacakge exif. Install it using
```bash
pip install exif
```
Now is hided.
'''
from exif import Image
with open(img_path, 'rb') as image_file:
image = Image(image_file)
if not image.has_exif:
raise Exception('exif information is not exist!')
# for tag in dir(image):
# print(tag, '=', image.get(tag))
fl = image.focal_length
equivalent_fl = image.focal_length_in_35mm_film
img_width_px = image.pixel_x_dimension
img_height_px = image.pixel_y_dimension
make = image.make
model = image.model
return fl, equivalent_fl, img_width_px, img_height_px, make, model
@staticmethod
def __obtain_exif_exifread(img_path):
'''Obtain exif using the pacakge exifread. Install it using
```bash
pip install exifread
```
'''
import exifread
with open(img_path, 'rb') as image_file:
tags = exifread.process_file(image_file)
# for key in tags.keys():
# print(key, '=', tags[key].values)
fl = tags['EXIF FocalLength'].values[0].num / tags['EXIF FocalLength'].values[0].den
equivalent_fl = tags['EXIF FocalLengthIn35mmFilm'].values[0]
img_width_px = tags['EXIF ExifImageWidth'].values[0]
img_height_px = tags['EXIF ExifImageLength'].values[0]
make = tags['Image Make'].values
model = tags['Image Model'].values
return fl, equivalent_fl, img_width_px, img_height_px, make, model
@staticmethod
def extract_intrinsics_df(data_path, method='db'):
'''Extract the intrinsics of photos in the data_path into a dataframe for analysis.
Attributes:
data_path: the absolute paths of photos
method: the method to obtrain focal length in pixel, now support 'db'(default) or 'efl'
Returns:
A pands dataframe contains the pairs of filename and intrinsics
'''
df = pd.DataFrame(columns=['fl_px', 'u_0', 'v_0'], dtype=np.float16)
index = 0
for filename in os.listdir(data_path):
file_path = os.path.join(data_path, filename) # for every file
# print(file_path)
try:
intrinsics = focal_length_helper(file_path).obtain_intrinsics(method)
df.loc[index] = intrinsics # add to the dataframe
index += 1
except (AttributeError, KeyError, IndexError) as e:
# AttributeError, KeyError: when obtaining exif, some attributes are missing
# IndexError: when make and model is not contained in camera sensors db
# print(e)
continue
return df