-
Notifications
You must be signed in to change notification settings - Fork 0
/
pycis_lsd_wrapper.py
238 lines (219 loc) · 9.33 KB
/
pycis_lsd_wrapper.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
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
'''
PyCIS - Python Computational Inference from Structure
pylib/pycis_lsd_wrapper.py: Wraps the LSD functionality of pycis.c and lib/ files,
and provided switches for directing 2D/3D and centerline/edgeline calls.
Benjamin Feuge-Miller: [email protected]
The University of Texas at Austin,
Oden Institute Computational Astronautical Sciences and Technologies (CAST) group
*Date of Modification: December 27, 2021
#--------------------------------------------------------------------------------------
#PyCIS-LSD: An a-contrario detection sub-algorithm for extracting narrow lines within dense optical data cubes.
#Copyright (C) 2022, Benjamin G. Feuge-Miller, <[email protected]>
#
#PyCIS-LSD is free software: you can redistribute it and/or modify
#it under the terms of the GNU General Public License as published
#by the Free Software Foundation, either version 3 of the License,
#or (at your option) any later version.
#
#PyCIS-LSD is distributed in the hope that it will be useful,
#but WITHOUT ANY WARRANTY; without even the implied warranty of
#MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
#GNU Affero General Public License for more details.
#
#You should have received a copy of the GNU Affero General Public License
#along with this program. If not, see <http://www.gnu.org/licenses/>.
#
#**NOTICE**:
#PyCIS-LSD is modified from the source code of LSDSAR:
#"LSDSAR, a Markovian a contrario framework for line segment detection in SAR images"
#by Chenguang Liu, Rémy Abergel, Yann Gousseau and Florence Tupin.
#Pattern Recognition, 2019).
#https://doi.org/10.1016/j.patcog.2019.107034
#*Date of Modification: April 30, 2021*
#
#**NOTICE**:
#LSDSAR is modified from the source code of LSD:
#"LSD: a Line Segment Detector" by Rafael Grompone von Gioi,
#Jeremie Jakubowicz, Jean-Michel Morel, and Gregory Randall,
#Image Processing On Line, 2012. DOI:10.5201/ipol.2012.gjmr-lsd
#http://dx.doi.org/10.5201/ipol.2012.gjmr-lsd
#*Date of Modification: 27/06/2018*
#--------------------------------------------------------------------------------------
'''
## IMPORT NECESSARY LIBRARIES
import os
import numpy as np
import pylsd.pycis as pycis
import time
def get_size(I):
''' Get matrix size variables with necesary 0-setting '''
M = len(I)
if (I.ndim==1):
M=len(I)
N=0
O=0
elif (I.ndim==2) and (M>10):
M,N=np.shape(I)
O=0
elif (I.ndim==3) and (M>10):
M,N,O=np.shape(I)
else:
M,N,O=0,0,0
return M,N,O
def get_sizes(I,I0):
''' Get matrix size variables with necessary 0-settings '''
M,N,O = get_size(I)
#print('M %d, N %d, O %d'%(M,N,O))
M0,N0,O0 = get_size(I0)
#print('M0 %d, N0 %d, O0 %d'%(M0,N0,O0))
return M,N,O,M0,N0,O0
def my_unflatten(I,M,N,O):
''' Format image data for passing to pycis '''
if (O==0):
templist = np.zeros((M,N))
for i in range(M):
for j in range(N):
templist[i,j]=I[(i+M*j)]
if (O>0):
templist = np.zeros((M,N,O))
for i in range(M):
for j in range(N):
for k in range(O):
templist[i,j,k]=I[k+O*(i+M*j)]
I = templist
return I
def my_flatten(I,M,N,O):
''' Format image data for passing to pycis '''
if (O==0):
I = I.T.flatten().tolist()
if (O>0):
templist = np.zeros((M*N*O,))
for i in range(M):
for j in range(N):
for k in range(O):
templist[k+O*(i+M*j)] = I[i,j,k]
I = templist.tolist()
return I
def main(I,I0, folder='results',name='temp',
a=4.,d=.4,t=1.,scale=0.8,sigma=0.6,e=0,getp=1,p=[0,0,0,0,0,0,0,0,0,0,0,0],p2=[0,0,0,0,0,0,0,0,0,0,0,0],idxs=[-1,-1],shape=[4096,4096,25]):
'''
Run LSD pipeline for a test and conditioning image pair .
Reshape data and set input parameter vector.
Decide between output functions.
Input:
I : test image
I0: noise model (or prior edge lines)
folder: location for results
name: name for saving .png and .npy files
a: gradient-by-ratio parameter. Produces a (k*2+1)^d kernel for k=log(10)*a
eg a=1 produces a 7^d kernel, a=2 an 11^d kernel, a=3 a 15^d kernel, etc
d: density threshold for improving regions. Higher values enforce stronger linearity constraints
t: denominator factor for tolerance. Initial tolerance is 22.5deg, empirically optimal for 2D
p: markov kernel for estimating parallel alignments
(in edge case, build regions by parallel alignment and count alignments )
p2: markov kernel for estimating parallel alignments
(in edge case, build regions by parallel alignment and count alignments )
getp: flag for pipline - control output as markov kernel or lines. see pycis.c
scale: fraction of gaussian downsampled side lengths to input data (set to 1 for no downsampling)
sigma: std=sigma/scale for variance of gaussian downsampling via seperable 1D kernels
e: -log10(epsilon) NFA threshold. Choose 0 for epsilon=1 (default theory).
Robust to selection, but setting very large (e=6) can reduce spurious noise
and improve 2nd-order detection if there exists a statistically relevant number of detections
Output: one of the following:
inputv/inputvorth: updated settings vector with solved markov kernels
data1_name.npy: edge line detections
data2_name.npy: center line detections
Notes: I,I0 cannot be empty vectors. To set an 'empty image',
use a small matrix, e.g. 2x2 identity, and the function
will set the variables X,X0 appropriatly.
'''
if not os.path.exists(folder):
os.makedirs(folder)
'''
#global I3
#global I03
x1=I[0]; x2=I[1]; y1=I[2]; y2=I[3]
try:
I = I3[x1:x2, y1:y2, :]
except:
I = I3[x1:x2, :]
x1=I[0]; x2=I[1]; y1=I[2]; y2=I[3]
try:
I0 = I03[x1:x2, y1:y2, :]
except:
I0 = I03[x1:x2,:]
'''
## SET INPUT DIRECTIONS FOR VERSIONING - LSD CANNOT ACCEPT EMPTY IMAGES, BUT REQUIRES M/M0=0
M,N,O,M0,N0,O0 = get_sizes(I,I0)
## FLATTEN IMAGES FOR USE IN C, COPY FOR PLOTTING
I_full = np.copy(I)
I0_full= np.copy(I0)
I = my_flatten(I,M,N,O)
I0= my_flatten(I0,M0,N0,O0)
## SET INPUT PARAMETER VECTOR
p11= p[0];p01= p[1];p11_2= p[2];p01_2= p[3];p11_4= p[4];p01_4= p[5]
dp11=p2[0];dp01=p2[1];dp11_2=p2[2];dp01_2=p2[3];dp11_4=p2[4];dp01_4=p2[5]
pb11= p[6];pb01= p[7];pb11_2= p[8];pb01_2= p[9];pb11_4= p[10];pb01_4= p[11]
dpb11=p2[6];dpb01=p2[7];dpb11_2=p2[8];dpb01_2=p2[9];dpb11_4=p2[10];dpb01_4=p2[11]
alpha=a#4.
eps=10.**(-1.*e)#(1/1.)
density=d #0.4
angth=22.5/t
sizenum=np.sqrt(M**2.+N**2.)*5.
if sizenum>(10.**4):
sizenum=10.**4
if O>0:
sizenum=min(10.**6.,np.sqrt(M**2.+N**2.+O**2.)*5.)
inputv=[alpha,eps,density,sizenum,angth,
scale,sigma,shape[0],shape[1],shape[2],
p11, p01, p11_2, p01_2, p11_4, p01_4,
pb11, pb01, pb11_2, pb01_2, pb11_4, pb01_4]
inputvorth=[alpha,eps,density,sizenum,angth,
scale,sigma,shape[0],shape[1],shape[2],
dp11,dp01,dp11_2,dp01_2,dp11_4,dp01_4,
dpb11,dpb01,dpb11_2,dpb01_2,dpb11_4,dpb01_4]
## RUN LSD
#Get markov kernel if requested
markov = getp
if markov>1:
if markov==2:
pass
#print('\n--------------- MARKOV-PARALLEL: %s ---------------\n'%name,flush=True)
else:
pass
#print('\n--------------- MARKOV-ORTHOGONAL: %s ---------------\n'%name,flush=True)
time.sleep(1)
lines = pycis.pycis(I,M,N,O,I0,M0,N0,O0,inputv,inputvorth,markov)
del I, I0, alpha, density, angth, a,d,t,inputv,I_full,I0_full
#Return inputv
if (idxs[0]==-1) and (idxs[1]==-1):
return lines
else:
return (idxs,lines)
else:
#Find kernel and run estimation...
savename='%s/data1_%s'%(folder,name)
if markov==1: #markov = 1, run markov estimation plus edge detection
pass
#print('\n--------------- LSD + Markov: %s ---------------\n'%name,flush=True)
elif markov==0 and M0==0: #markov=0, edges with prior markov
pass
#print('\n--------------- LSD (Edge): %s ---------------\n'%name,flush=True)
elif markov==0 and M0>0: #markov=0, centerlines with prior markov
#print('\n--------------- LSD (Centers): %s ---------------\n'%name,flush=True)
savename = '%s/data2_%s'%(folder,name)
elif markov<0: #markov = 1, run markov estimation plus edge detection
pass
else:
print("ERROR: incompatible markov and X0 input .",flush=True)
quit()
time.sleep(1)
lines = pycis.pycis(I,M,N,O,I0,M0,N0,O0,inputv,inputvorth,markov)
if markov<0: #markov = 1, run markov estimation plus edge detection
return my_unflatten(np.asarray(lines),M,N,O)
#lines = [];
## SAVE LINE RESULTS
np.save(savename,lines)
del I, I0, alpha,density,angth, a,d,t,inputv,I_full,I0_full
#Return nothing - data is saved to file
return lines