0.6 -
Recognition rate
Odo ©
ho 0"R 5 À
1 1 | : 1 |
o
N
T
e
2
a
e
Ll
A
1 1 i 2
5 10 15 20 25 30 35 40
Dimension
a
o
©
T
Recognition rate
o o o
o ^ oo
T I 1
o
u
1
e
>
Dimension
(c) 15 Training Samples
0.9}
0.8 -
0.7r
0.6r
Recognition rate
0.5r
04r
Dimension
(d) 20 Training Samples
Fig.1 The curves of recognition rate vs. Dimension
of different training samples of each approaches
Tab.1 The max recognition rates on Washington DC
Mall(%)
Training Size NPE RPC PCA SRDA
5x12 76.70 7911 7553 79.32
10x12 85.87 8340 8369 8721
15x12 890,57 3574 8552 8983
20x12 9067 3770 83762 915]
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B3, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
From Fig.1 and Tab.1, we can find that with increasing in
the number of the training samples, the recognition rate
increases. Also, we can notice that the SRDA shows the
best performance around all the methods, that is to say,
the SRDA, we proposed algorithm, always has the
highest resolution.
3.2 Experiments on AVIRIS Indian Pines
In this section, we conduct our experiments in following
steps:
1) There are 16 labels in this dataset. In order to conduct
the experiments reasonably, we reduce 6 labels whose
sample numbers are all less than 320. Thus we need
320*10 samples, and each sample contains 220 bands;
2) As described above, for the chosen 320 samples, 1(=5,
10, 15, 20) datum are randomly selected for training and
the rest 300 datum are used for testing;
3) The 5-nearest neighbor classifier is applied in PCA,
NPE and SRDA subspace. For PCA and SRDA, the
numbers of the subspace dimension are 40, while the
NPE's are 14, 40, 40 and 40. The curves of recognition
rate vs. dimension are shown in Fig.2. And the max
recognition rates of each method are also reported in
Tab.2.
0.55
0.45 -
2
S
= 04
S
t
2 0.35}
x whe NPE |
0.3} pos PCA |
SRDA
0.25}
0.2 LL L ends —
0 5 30 35 40
10. 15 20. | 25 Ls
(a) 5 TraimingiSamples
Recognition rate
0 5 10 15 20 25 30 85 40
(b) 10 TP#ieréirg Samples
Recognition rate
Tab.
From
SRDA
the PC
trainin
SRDA
sample
this is
And tk
$3 D
The e)
Indian
1) All