-
Notifications
You must be signed in to change notification settings - Fork 1
/
panorama2.py
181 lines (162 loc) · 6.22 KB
/
panorama2.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
import cv2
import numpy as np
import os
import math
def required_img(img):
gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
ret,thresh = cv2.threshold(gray,0,255,cv2.THRESH_BINARY)
image,contour,heic = cv2.findContours(thresh,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)
cnt = contour[0]
cnt = cnt.reshape(-1,2)
x = np.max(cnt[:,0],axis = 0)
y = np.max(cnt[:,1],axis = 0)
img2 = np.zeros((y,x,3),np.uint8)
img2 = img[0:y,0:x]
return img2
def black_remove(img):
gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
ret,thresh = cv2.threshold(gray,0,255,cv2.THRESH_BINARY)
image,contour,h = cv2.findContours(thresh,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)
cnt =contour[0]
epsilon = 0.1*cv2.arcLength(cnt,True)
approx = cv2.approxPolyDP(cnt,epsilon,True)
approx = arrange_points(approx)
x, y, w, h = cv2.boundingRect(cnt)
# img = img[y:y+h,x:x+p]
pts1 = np.float32(approx)
pts2 = np.float32([[x,y],[x,y+h],[x+w,y+h],[x+w,y]])
M = cv2.getPerspectiveTransform(pts1,pts2)
img =cv2.warpPerspective(img,M,(gray.shape[1],gray.shape[0]))
#a = cv2.rectangle(img,(x,y),(x+w,y+h),(0,0,255))
gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
ret,thresh = cv2.threshold(gray,0,255,cv2.THRESH_BINARY)
image,contour,h = cv2.findContours(thresh,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)
cnt =contour[0]
x,y,w,h = cv2.boundingRect(cnt)
img = img[y:y+h,x:x+w]
return img
def image_stitching(img1,img2,M1):
gray1 = cv2.cvtColor(img1,cv2.COLOR_BGR2GRAY)
gray2 = cv2.cvtColor(img2,cv2.COLOR_BGR2GRAY)
(h,w,c) = img1.shape
K = np.array([[807.3179989, 0, 307.83299049], [0, 804.22240703, 229.03518511], [0, 0, 1]])
img1 = cv2.copyMakeBorder(img1,50,50,300,300, cv2.BORDER_CONSTANT)
(h,w,c) = img2.shape
K = np.array([[807.3179989, 0, 307.83299049], [0, 804.22240703, 229.03518511], [0, 0, 1]])
img2 = cv2.copyMakeBorder(img2,50,50,300,300, cv2.BORDER_CONSTANT)
img3 = cv2.warpAffine(img2,M1,(gray2.shape[1]+gray1.shape[1],gray1.shape[0]))
temp = np.zeros_like(img3)
col = img3.shape[1]-img1.shape[1]
temp = cv2.resize(temp,(col,img3.shape[0]),interpolation = cv2.INTER_CUBIC)
print(np.shape(temp),np.shape(img1))
img1 = np.hstack((img1,temp))
gray3 = cv2.cvtColor(img3,cv2.COLOR_BGR2GRAY)
print('inside')
flag = np.array((gray1,gray3))
indx = np.argmax(flag,axis=0)
ind1 = np.where(indx ==0)
ind2 = np.where(indx==1)
img = np.zeros_like(img1)
img[ind1] = img1[ind1]
img[ind2] = img3[ind2]
print('middle')
return img
def cylindrical_warp(img,K):
foc_len = (K[0][0] +K[1][1])/2
cylinder = np.zeros_like(img)
temp = np.mgrid[0:img.shape[1],0:img.shape[0]]
x,y = temp[0],temp[1]
# print('img p=color',img[0,0])
theta= (x- K[0][2])/foc_len # angle theta
h = (y-K[1][2])/foc_len # height
p = np.array([np.sin(theta),h,np.cos(theta)])
p = p.T
p = p.reshape(-1,3)
image_points = K.dot(p.T).T
points = image_points[:,:-1]/image_points[:,[-1]]
points = points.reshape(img.shape[0],img.shape[1],-1)
cylinder = cv2.remap(img, (points[:, :, 0]).astype(np.float32), (points[:, :, 1]).astype(np.float32), cv2.INTER_LINEAR)
return cylinder
cap = cv2.VideoCapture('output.mp4')
img = []
K = np.array([[807.3179989, 0, 307.83299049], [0, 804.22240703, 229.03518511], [0, 0, 1]])
while(cap.isOpened()):
ret,frame = cap.read()
img = img + [frame]
cv2.imshow('frame',frame)
if cv2.waitKey(1) & 0xff == ord('q'):
break
cap.release()
cv2.destroyAllWindows()
print('Total Frames :',len(img))
orb = cv2.ORB_create(nfeatures=75)
i = 0
img1 = img[0]
img1 = cylindrical_warp(img1,K)
img1 = cv2.copyMakeBorder(img1,50,50,300,300, cv2.BORDER_CONSTANT)
canvas = np.zeros((img1.shape[0],img1.shape[1],img1.shape[2]),np.uint8)
canvas[0:img1.shape[0],0:img1.shape[1]] = img1
while i < len(img)-1:
img2 = img[i+1]
img2 = cylindrical_warp(img2,K)
img2 = cv2.copyMakeBorder(img2,50,50,300,300, cv2.BORDER_CONSTANT)
# Initiate SIFT detector
# find the keypoints and descriptors with SIFT
kp1, des1 = orb.detectAndCompute(img1,None)
kp2, des2 = orb.detectAndCompute(img2,None)
# BRUTE FORCE parameters
bf = cv2.BFMatcher(crossCheck=False)
matches = bf.knnMatch(des1,des2,k=2)
pt1 = []
pt2 = []
good = []
for m,n in matches:
if m.distance < 0.7*n.distance:
good.append(m)
#good = sorted(good,key = lambda x:x.distance)
#good = good[:15]
pt1 = np.float32([kp1[m.queryIdx].pt for m in good]).reshape(-1,1,2)
pt2 = np.float32([kp2[m.trainIdx].pt for m in good]).reshape(-1,1,2)
pt = (pt1 - pt2)
std = np.std(pt)
print(i,std)
if std > 50:
i = i + 1
continue
else:
#output = np.ones(img1.shape, dtype = 'uint8')
M, mask = cv2.estimateAffine2D(pt1, pt2, cv2.RANSAC, ransacReprojThreshold=0.4)
M1 = np.array([[1,0,M[0,2]],[0,1,M[1,2]]])
#print(M1)
stitch = image_stitching(canvas,img2,M1)
canvas = required_img(canvas)
#H,mask = cv2.findHomography(pt1,pt2,cv2.RANSAC,ransacReprojThreshold=4.0)
#out1 = cv2.warpAffine(img2, M1, (img1.shape[1]+img2.shape[1],img2.shape[0]))
#out1[0:img1.shape[0],0:img2.shape[1]] = img1
#out1 = cv2.warpPerspective(img2,H,(img1.shape[1]+ img2.shape[1],img1.shape[0]))
#case1 = np.logical_and(img1 == [0,0,0],out1 == [0,0,0])
#case2 = np.logical_and(img1 == [0,0,0],out1 != [0,0,0])
#case3 = np.logical_and(img1 != [0,0,0],out1 == [0,0,0])
#case4 = np.logical_and(img1 != [0,0,0],out1 != [0,0,0])
#case1index = np.where(case1)
#case2index = np.where(case2)
#case3index = np.where(case3)
#case4index = np.where(case4)
#output[case1index[0],case1index[1]] = [0,0,0]
#output[case2index] = out1[case2index]
#output[case3index] = img1[case3index]
#output[case4index] = img1[case4index]/2 + out1[case4index]/2
#cv2.imshow('image',img1)
#cv2.waitKey(0)
#cv2.imshow('image',img1)
#cv2.waitKey(0)
i = i + 1
continue
#except:
# i = i + 1
# continue
if 0xff == ord('q'):
break
cv2.imshow('output',stitch)
cv2.waitKey(0)
cv2.destroyAllWindows()