-
Notifications
You must be signed in to change notification settings - Fork 12
/
match_features.py
142 lines (80 loc) · 3.97 KB
/
match_features.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
import numpy as np
import cv2
import os
class SiftKpDesc():
def __init__(self, kp, desc):
# List of keypoints in (x,y) crd -> N x 2
self.kp = kp
# List of Descriptors at keypoints : N x 128
self.desc = desc
class SiftMatching:
_BLUE = [255, 0, 0]
_GREEN = [0, 255, 0]
_RED = [0, 0, 255]
_CYAN = [255, 255, 0]
_line_thickness = 2
_radius = 5
_circ_thickness = 2
def __init__(self, img_1_path, img_2_path, results_fldr='', nfeatures=2000, gamma=0.8):
fname_1 = os.path.basename(img_1_path)
fname_2 = os.path.basename(img_2_path)
if not results_fldr:
results_fldr = os.path.split(img_1_path)[0]
self.result_fldr = os.path.join(results_fldr, 'results')
self.prefix = fname_1.split('.')[0] + '_' + fname_2.split('.')[0]
if not os.path.exists(self.result_fldr):
os.makedirs(self.result_fldr)
self.img_1_bgr = self.read_image(img_1_path)
self.img_2_bgr = self.read_image(img_2_path)
self.nfeatures = nfeatures
self.gamma = gamma
def read_image(self, img_path):
img_bgr = cv2.imread(img_path, cv2.IMREAD_COLOR)
return img_bgr
def get_sift_features(self, img_bgr, nfeatures=2000):
img_gray = cv2.cvtColor(img_bgr, cv2.COLOR_BGR2GRAY)
sift_obj = cv2.xfeatures2d.SIFT_create(nfeatures)
# kp_list_obj is a list of "KeyPoint" objects with location stored as tuple in "pt" attribute
kp_list_obj, desc = sift_obj.detectAndCompute(image=img_gray, mask=None)
kp = [x.pt for x in kp_list_obj]
return SiftKpDesc(kp, desc)
def match_features(self, sift_kp_desc_obj1, sift_kp_desc_obj2, gamma=0.8):
correspondence = [] # list of lists of [x1, y1, x2, y2]
for i in range(len(sift_kp_desc_obj1.kp)):
sc = np.linalg.norm(sift_kp_desc_obj1.desc[i] - sift_kp_desc_obj2.desc, axis=1)
idx = np.argsort(sc)
val = sc[idx[0]] / sc[idx[1]]
if val <= gamma:
correspondence.append([*sift_kp_desc_obj1.kp[i], *sift_kp_desc_obj2.kp[idx[0]]])
return correspondence
def draw_correspondence(self, correspondence, img_1, img_2):
if len(img_1.shape) == 2:
img_1 = np.repeat(img_1[:, :, np.newaxis], 3, axis=2)
if len(img_2.shape) == 2:
img_2 = np.repeat(img_2[:, :, np.newaxis], 3, axis=2)
h, w, _ = img_1.shape
img_stack = np.hstack((img_1, img_2))
for x1, y1, x2, y2 in correspondence:
x1_d = int(round(x1))
y1_d = int(round(y1))
x2_d = int(round(x2) + w)
y2_d = int(round(y2))
cv2.circle(img_stack, (x1_d, y1_d), radius=self._radius, color=self._BLUE,
thickness=self._circ_thickness, lineType=cv2.LINE_AA)
cv2.circle(img_stack, (x2_d, y2_d), radius=self._radius, color=self._BLUE,
thickness=self._circ_thickness, lineType=cv2.LINE_AA)
cv2.line(img_stack, (x1_d, y1_d), (x2_d, y2_d), color=self._CYAN,
thickness=self._line_thickness)
fname = os.path.join(self.result_fldr, self.prefix + '_sift_corr.jpg')
cv2.imwrite(fname, img_stack)
def run(self):
sift_kp_desc_obj1 = self.get_sift_features(self.img_1_bgr, nfeatures=self.nfeatures)
sift_kp_desc_obj2 = self.get_sift_features(self.img_2_bgr, nfeatures=self.nfeatures)
correspondence = self.match_features(sift_kp_desc_obj1, sift_kp_desc_obj2, gamma=self.gamma)
self.draw_correspondence(correspondence, self.img_1_bgr, self.img_2_bgr)
return correspondence
if __name__ == "__main__":
img_1_path = "/Users/aartighatkesar/Documents/Image-Mosaicing/input/p3/4.jpg"
img_2_path = "/Users/aartighatkesar/Documents/Image-Mosaicing/input/p3/5.jpg"
siftmatch_obj = SiftMatching(img_1_path, img_2_path, results_fldr='', nfeatures=2000, gamma=0.6)
siftmatch_obj.run()