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
|
###############################################################################
# pytest_classify_s2.py: classify
# Author(s): Pieter.Kempeneers@ec.europa.eu
# Copyright (c) 2016-2019 European Union (Joint Research Centre)
# License EUPLv1.2
#
# This file is part of jiplib
###############################################################################
# History
# 2018/01/22 - Created by Pieter Kempeneers (pieter.kempeneers@ec.europa.eu)
# Change log
import argparse
import os
import math
import jiplib as jl
parser=argparse.ArgumentParser()
parser.add_argument("-input","--input",help="Path of the input raster dataset",dest="input",required=True,type=str)
parser.add_argument("-reference","--reference",help="Path of the reference raster dataset",dest="reference",required=True,type=str)
parser.add_argument("-model","--model",help="Path of the model output filename used for training",dest="model",required=False,type=str)
parser.add_argument("-training","--training",help="Path of the training vector dataset with raster information",dest="training",required=False,type=str)
parser.add_argument("-output","--output",help="Path of the classification output raster dataset",dest="output",required=False,type=str)
parser.add_argument("-sampleSize","--sampleSize",help="Sample size used for training svm or ann",dest="sampleSize",required=False,type=int,default=100)
parser.add_argument("-classifier","--classifier",help="classifier (sml, svm, ann)",dest="classifier",required=True,type=str,default="sml")
args = parser.parse_args()
# try:
if True:
jim=jl.createJim(args.input)
#preparation of reference
classDict={}
classDict['urban']=2
classDict['agriculture']=12
classDict['forest']=25
classDict['water']=41
classDict['rest']=50
classFrom=list(range(0,50))
classTo=[50]*50
for i in range(0,50):
if i>=1 and i<10:
classTo[i]=classDict['urban']
elif i>=11 and i<22:
classTo[i]=classDict['agriculture']
elif i>=23 and i<25:
classTo[i]=classDict['forest']
elif i>=40 and i<45:
classTo[i]=classDict['water']
else:
classTo[i]=classDict['rest']
print("jim.getProjection() {}".format(jim.getProjection()))
jim_ref=jl.createJim(filename=args.reference,dx=jim.getDeltaX(),dy=jim.getDeltaY(),ulx=jim.getUlx(),uly=jim.getUly(),lrx=jim.getLrx(),lry=jim.getLry(),t_srs=jim.getProjection())
jim_ref=jim_ref.reclass({'class':classFrom,'reclass':classTo})
if args.classifier == "sml":
print("SML classifier")
reflist=jl.JimList([jim_ref])
jim.d_band2plane();
# jim.train(reflist,{'method':'sml','model':args.model,'class':sorted(classDict.values()),'verbose':1})
# sml=jim.classify({'method':'sml','model':args.model,'verbose':1})
sml=jim.classifySML(reflist,{'class':sorted(classDict.values()),'verbose':1})
sml.write({'filename':'/tmp/sml_classes.tif'})
# #preparation of output
sml_class=sml.statProfile({'function':'maxindex'}).reclass({'class':list(range(0,sml.nrOfBand())),'reclass':sorted(classDict.values())})
if args.output:
sml_class.write({'filename':args.output})
sml_class.close()
else:
srcnodata=[0]
dict={'srcnodata':srcnodata}
if args.training:
trainingfn=args.training
dict.update({'oformat':'ESRI Shapefile'})
else:
dict.update({'oformat':'Memory'})
trainingfn='training.shp'
dict.update({'output':trainingfn})
dict.update({'class':sorted(classDict.values())})
sampleSize=-args.sampleSize #use negative values for absolute and positive values for percentage values
dict.update({'threshold':sampleSize})
dict.update({'bandname':['B02','B03','B04','B08']})
# dict.update({'band':[0,1,2,3]})
# dict.update({'verbose':2})
sample=jim.extractImg(jim_ref,dict)
if args.classifier == "svm":
#training
#explicitly define all band names to use for training
sample.train({'method':args.classifier,'label':'label','bandname':['B02','B03','B04','B08'],'model':args.model})
#define selection of band names to use for training
#sample.train({'method':'svm','label':'label','bandname':['B03','B08'],'model':args.model})
sample.close()
#classification
svm_class=jim.classify({'method':'svm','model':args.model})
#explicitly define all band indexes to use for classification
#svm_class=jim.classify({'method':'svm','model':args.model,'band':[0,1,2,3]})
#define selection of band indexes to use for classification
#svm_class=jim.classify({'method':'svm','model':args.model,'band':[1,3]})
if args.output:
svm_class.write({'filename':args.output})
svm_class.close()
elif args.classifier == "ann":
#training
sample.train({'method':args.classifier,'label':'label','model':args.model})
#explicitly define all band names to use for training
#sample.train({'method':'ann','label':'label','bandname':['B02','B03','B04','B08'],'model':args.model})
#define selection of band names to use for training
#sample.train({'method':'ann','label':'label','bandname':['B02','B04'],'model':args.model})
sample.close()
#classification
ann_class=jim.classify({'method':'ann','model':args.model})
#explicitly define all band indexes to use for classification
#svm_class=jim.classify({'method':'ann','model':args.model,'band':[0,1,2,3]})
#define selection of band indexes to use for classification
#ann_class=jim.classify({'method':'ann','model':args.model,'band':[0,2]})
if args.output:
ann_class.write({'filename':args.output})
ann_class.close()
else:
print("Error: classifier",args.classifier,"not implemented")
throw()
jim.close()
print("Success: classify")
try:
print("ok")
except:
print("Failed: classify")
jim.close()
|
