18368aa9
David Mayerich
added a new set o...
|
1
2
3
4
5
6
7
8
9
|
# -*- coding: utf-8 -*-
"""
Created on Sun Jul 23 16:04:33 2017
@author: david
"""
import numpy
import colorsys
|
90c935e3
David Mayerich
updates
|
10
11
|
import sklearn
import sklearn.metrics
|
f275c01c
David Mayerich
finalized the pre...
|
12
13
14
|
import scipy
import scipy.misc
import envi
|
9b3cbdda
David Mayerich
changed the name ...
|
15
|
import hyperspectral
|
777fef6c
David Mayerich
added digital sta...
|
16
|
import random
|
59f31ea4
David Mayerich
fixed bugs in env...
|
17
18
|
import progressbar
import matplotlib.pyplot as plt
|
18368aa9
David Mayerich
added a new set o...
|
19
|
|
254e3bc8
David Mayerich
added a full clas...
|
20
21
22
23
24
25
26
27
28
29
30
31
32
33
|
#generate N qualitative colors and return the value for color c
def qualcolor(c, N):
dN = numpy.ceil(numpy.sqrt(N)).astype(numpy.int32)
h = c/N
sp = c/N * 2 * numpy.pi * dN + numpy.pi/2
s = numpy.sin(sp) * 0.25 + 0.75
vp = c/N * 2 * numpy.pi * dN
v = numpy.sin(vp) * 0.25 + 0.75
rgb = numpy.array(colorsys.hsv_to_rgb(h, s, v))
return rgb * 255
|
18368aa9
David Mayerich
added a new set o...
|
34
|
#generate a 2D color class map using a stack of binary class images
|
f275c01c
David Mayerich
finalized the pre...
|
35
36
|
#input: C is a C x Y x X binary image
#output: an RGB color image with a unique color for each class
|
71ee1c23
David Mayerich
added additional ...
|
37
|
def class2color(C):
|
18368aa9
David Mayerich
added a new set o...
|
38
39
|
#determine the number of classes
|
f275c01c
David Mayerich
finalized the pre...
|
40
41
42
43
|
nc = C.shape[0]
s = C.shape[1:]
s = numpy.append(s, 3)
|
18368aa9
David Mayerich
added a new set o...
|
44
45
|
#generate an RGB image
|
f275c01c
David Mayerich
finalized the pre...
|
46
|
RGB = numpy.zeros(s, dtype=numpy.ubyte)
|
18368aa9
David Mayerich
added a new set o...
|
47
48
49
|
#for each class
for c in range(0, nc):
|
254e3bc8
David Mayerich
added a full clas...
|
50
|
color = qualcolor(c, nc)
|
f275c01c
David Mayerich
finalized the pre...
|
51
|
RGB[C[c, ...], :] = color
|
18368aa9
David Mayerich
added a new set o...
|
52
|
|
538df2a2
David Mayerich
added function de...
|
53
54
55
56
|
return RGB
#create a function that loads a set of class images as a stack of binary masks
#input: list of class image names
|
f275c01c
David Mayerich
finalized the pre...
|
57
|
#output: C x Y x X binary image specifying class/pixel membership
|
538df2a2
David Mayerich
added function de...
|
58
|
#example: image2class(("class_coll.bmp", "class_epith.bmp"))
|
59f31ea4
David Mayerich
fixed bugs in env...
|
59
|
def filenames2class(masks):
|
6b2be991
sberisha
added utility fun...
|
60
61
62
63
64
65
66
67
|
#get num of mask file names
num_masks = len(masks)
if num_masks == 0:
print("ERROR: mask filenames not provided")
print("Usage example: image2class(('class_coll.bmp', 'class_epith.bmp'))")
return
|
f275c01c
David Mayerich
finalized the pre...
|
68
|
classimages = []
|
59f31ea4
David Mayerich
fixed bugs in env...
|
69
70
71
|
bar = progressbar.ProgressBar(max_value=num_masks)
for m in range(0, num_masks):
img = scipy.misc.imread(masks[m], flatten=True).astype(numpy.bool)
|
f275c01c
David Mayerich
finalized the pre...
|
72
|
classimages.append(img)
|
59f31ea4
David Mayerich
fixed bugs in env...
|
73
|
bar.update(m+1)
|
538df2a2
David Mayerich
added function de...
|
74
|
|
f275c01c
David Mayerich
finalized the pre...
|
75
76
|
result = numpy.stack(classimages)
sum_images = numpy.sum(result.astype(numpy.uint32), 0)
|
6b2be991
sberisha
added utility fun...
|
77
|
|
f275c01c
David Mayerich
finalized the pre...
|
78
79
80
|
#identify and remove redundant pixels
bad_idx = sum_images > 1
result[:, bad_idx] = 0
|
6b2be991
sberisha
added utility fun...
|
81
|
|
f275c01c
David Mayerich
finalized the pre...
|
82
|
return result
|
6b2be991
sberisha
added utility fun...
|
83
|
|
6b2be991
sberisha
added utility fun...
|
84
|
|
f275c01c
David Mayerich
finalized the pre...
|
85
86
87
|
#create a class mask stack from an C x Y x X probability image
#input: C x Y x X image giving the probability P(c |x,y)
#output: C x Y x X binary class image
|
6b2be991
sberisha
added utility fun...
|
88
|
def prob2class(prob_image):
|
f275c01c
David Mayerich
finalized the pre...
|
89
|
class_image = numpy.zeros(prob_image.shape, dtype=numpy.bool)
|
6b2be991
sberisha
added utility fun...
|
90
91
|
#get nonzero indices
nnz_idx = numpy.transpose(numpy.nonzero(numpy.sum(prob_image, axis=0)))
|
fad1a72c
David Mayerich
merged sebastian ...
|
92
|
|
6b2be991
sberisha
added utility fun...
|
93
94
95
96
97
98
|
#set pixel corresponding to max probability to 1
for idx in nnz_idx:
idx_max_prob = numpy.argmax(prob_image[:, idx[0], idx[1]])
class_image[idx_max_prob, idx[0], idx[1]] = 1
return class_image
|
f275c01c
David Mayerich
finalized the pre...
|
99
|
|
90c935e3
David Mayerich
updates
|
100
|
#calculate an ROC curve given a probability image and mask of "True" values
|
f275c01c
David Mayerich
finalized the pre...
|
101
102
103
104
105
106
107
108
109
110
111
112
|
#input:
# P is a Y x X probability image specifying P(c | x,y)
# t_vals is a Y x X binary image specifying points where x,y = c
# mask is a mask specifying all pixels to be considered (positives and negatives)
# use this mask to limit analysis to regions of the image that have been classified
#output: fpr, tpr, thresholds
# fpr is the false-positive rate (x-axis of an ROC curve)
# tpr is the true-positive rate (y-axis of an ROC curve)
# thresholds stores the threshold associated with each point on the ROC curve
#
#note: the AUC can be calculated as auc = sklearn.metrics.auc(fpr, tpr)
def prob2roc(P, t_vals, mask=[]):
|
90c935e3
David Mayerich
updates
|
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
|
if not P.shape == t_vals.shape:
print("ERROR: the probability and mask images must be the same shape")
return
#if a mask image isn't provided, create one for the entire image
if mask == []:
mask = numpy.ones(t_vals.shape, dtype=numpy.bool)
#create masks for the positive and negative probability scores
mask_p = t_vals
mask_n = mask - mask * t_vals
#calculate the indices for the positive and negative scores
idx_p = numpy.nonzero(mask_p)
idx_n = numpy.nonzero(mask_n)
Pp = P[idx_p]
Pn = P[idx_n]
|
6b2be991
sberisha
added utility fun...
|
132
|
|
90c935e3
David Mayerich
updates
|
133
134
135
136
137
138
|
Lp = numpy.ones((Pp.shape), dtype=numpy.bool)
Ln = numpy.zeros((Pn.shape), dtype=numpy.bool)
scores = numpy.concatenate((Pp, Pn))
labels = numpy.concatenate((Lp, Ln))
|
fad1a72c
David Mayerich
merged sebastian ...
|
139
|
return sklearn.metrics.roc_curve(labels, scores)
|
f275c01c
David Mayerich
finalized the pre...
|
140
|
|
71ee1c23
David Mayerich
added additional ...
|
141
|
#convert a label image to a C x Y x X class image
|
777fef6c
David Mayerich
added digital sta...
|
142
|
def label2class(L, background=[]):
|
71ee1c23
David Mayerich
added additional ...
|
143
|
unique = numpy.unique(L)
|
777fef6c
David Mayerich
added digital sta...
|
144
145
146
|
if not background == []: #if a background value is specified
unique = numpy.delete(unique, numpy.nonzero(unique == background)) #remove it from the label array
|
71ee1c23
David Mayerich
added additional ...
|
147
|
s = L.shape
|
777fef6c
David Mayerich
added digital sta...
|
148
|
s = numpy.append(numpy.array((len(unique))), s)
|
71ee1c23
David Mayerich
added additional ...
|
149
|
C = numpy.zeros(s, dtype=numpy.bool)
|
777fef6c
David Mayerich
added digital sta...
|
150
151
|
for i in range(0, len(unique)):
C[i, :, :] = L == unique[i]
|
71ee1c23
David Mayerich
added additional ...
|
152
153
|
return C
|
777fef6c
David Mayerich
added digital sta...
|
154
155
156
157
158
159
160
161
|
#randomizes a given mask to include a subset of n pixels in the original
def random_mask(M, n):
idx = numpy.flatnonzero(M)
new_idx = numpy.random.permutation(idx)
new_mask = numpy.zeros(M.shape, dtype=numpy.bool)
new_mask[numpy.unravel_index(new_idx[0:n], new_mask.shape)] = True
return new_mask
|
eccf10ff
David Mayerich
added comments fo...
|
162
163
164
165
|
#perform classification of an ENVI image using batch processing
# input: E is the ENVI object (file is assumed to be loaded)
# C is a classifier - anything in sklearn should work
# batch is the batch size
|
59f31ea4
David Mayerich
fixed bugs in env...
|
166
167
|
def envi_batch_predict(E, C, batch=10000):
|
254e3bc8
David Mayerich
added a full clas...
|
168
|
Fv = E.loadbatch(batch)
|
59f31ea4
David Mayerich
fixed bugs in env...
|
169
170
171
172
173
|
i = 0
Tv = []
plt.ion()
bar = progressbar.ProgressBar(max_value=numpy.count_nonzero(E.mask))
while not Fv == []:
|
ed3d7d30
David Mayerich
updated NWT file ...
|
174
|
Fv = numpy.nan_to_num(Fv) #remove infinite values
|
59f31ea4
David Mayerich
fixed bugs in env...
|
175
|
if i == 0:
|
254e3bc8
David Mayerich
added a full clas...
|
176
|
Tv = C.predict(Fv.transpose())
|
59f31ea4
David Mayerich
fixed bugs in env...
|
177
|
else:
|
254e3bc8
David Mayerich
added a full clas...
|
178
|
Tv = numpy.concatenate((Tv, C.predict(Fv.transpose()).transpose()), 0)
|
59f31ea4
David Mayerich
fixed bugs in env...
|
179
|
tempmask = E.batchmask()
|
9b3cbdda
David Mayerich
changed the name ...
|
180
|
Lv = hyperspectral.unsift2(Tv, tempmask)
|
59f31ea4
David Mayerich
fixed bugs in env...
|
181
182
183
184
|
Cv = label2class(Lv.squeeze(), background=0)
RGB = class2color(Cv)
plt.imshow(RGB)
plt.pause(0.05)
|
254e3bc8
David Mayerich
added a full clas...
|
185
|
Fv = E.loadbatch(batch)
|
59f31ea4
David Mayerich
fixed bugs in env...
|
186
|
i = i + 1
|
ed3d7d30
David Mayerich
updated NWT file ...
|
187
|
bar.update(len(Tv))
|