Content uploaded by Ints Mednieks
Author content
All content in this area was uploaded by Ints Mednieks on Feb 28, 2023
Content may be subject to copyright.
ELEKTRONIKA IR ELEKTROTECHNIKA,ISSN 1392-1215,VOL.21,NO.5,2015
1Abstract—The paper is focused on the analysis of
classification possibilities of multisensor data with different
spatial resolutions using combined classifiers based on Bayes
approach with equal prior probabilities and on minimum of the
Mahalanobis distance. The task set up for the 2014 IEEE GRSS
Data Fusion Contest was chosen as an application example.
High resolution RGB image and lower resolution thermal
infrared image from the same urban area were processed to
perform classification of each higher resolution pixel.
Development of a fast and straightforward procedure was
targeted and combined classifiers are proposed for that,
exploiting spectral features from each data set separately. It is
shown that data fusion can be achieved using the proposed
classifiers and improvement of classification quality can be
obtained with respect to the cases where only one of the data
sets is used. The best classification results were obtained using
the combined Bayes- type classifier that provided overall
classification accuracy of about 95 % when the ground truth
pixels from the high resolution RGB image were used both for
design and testing.
Index Terms—Remote sensing, hyperspectral imaging,
image classification, data fusion.
I. INTRODUCTION
Remote sensing from airplanes and satellites has become a
widely used tool for solving tasks in management of natural
resources, urban planning, precision agriculture and other
areas. Different types of sensors are used for that, including
multispectral, hyperspectral, LiDAR, SAR, acquiring
different kinds of data. It is a common practice to employ
several sensors at once to obtain complementary information
from the same area. For example, LiDAR and multispectral
data from the same forest area are often collected to obtain
information for its inventory. LiDAR data can be processed
to obtain a height model of the stand, while spectral data can
be used to detect species or assess health of trees etc. Quite
often in this case, data from two different sources should be
used in a combined way to solve a specific task, i.e. data
fusion should be performed during processing. Data from
optical sensors are in general acquired in the form of three-
dimensional images where each pixel is related with its
spatial coordinates calculated from simultaneously collected
GPS information. Pixel size in this case depends on the
Manuscript received January 5, 2015; accepted June 26, 2015.
This research was performed within the project No.
2013/0031/2DP/2.1.1.1.0/13/APIA/VIAA/010 funded by the European
Regional Development Fund.
distance to the target, viewing angle and number of sensing
cells in the sensor. LiDAR and SAR data are usually pre-
processed to obtain images characterizing geographical areas
under study and are also registered to geographical
coordinates. Pixel size in this case can be chosen in the pre-
processing procedure but it is limited by the amount of
collected data within the spatial unit.
When data from sensors are obtained with different spatial
resolution, their fusion becomes a challenging task. It is also
crucial to perform precise registration of acquired images to
geographical coordinates. Otherwise pixels of these images
cannot be properly related with physical objects observed
and their combined use for analysis of these objects cannot
be performed correctly.
II.STATE OF THE ART
One of the major tasks in remote sensing is classification
of geographical areas or distinct objects represented by
acquired images. There are multiple classification
approaches developed for that [1]. If the data representing
each class can be interpreted as a sample realisation from the
multidimensional universe with Gaussian distribution, Bayes
classification approach is applicable and has shown good
results [2]. Therefore it is purposeful to consider Bayesian
approach to classification of multiresolution data obtained
from different sensors.
Classification of multisensor data with different spatial
resolutions is usually performed by combining outputs of
separate classifiers each dealing with data from one sensor,
or designing a single classifier operating with a fused image
[3]. The first approach is simpler in general and there are
multiple studies following it [4]–[10]. Approaches to
combination of multiple individual classifiers including
Bayesian ones were considered by Xu et al. [4] with
application for handwriting recognition. Averaged Bayes
classifier was proposed to combine results of different
classifiers from the same data. Enhanced combination of
independent Bayesian classifiers was elaborated in [6],
based on original definition in [5]. As a result, an iterative
procedure was proposed using steps, similar to the
expectation-maximization algorithm. Other sophisticated
approaches include using of neural networks [7], Dempster-
Shafer theory [8] and fuzzy sets [10]. Authors of [3] and
[11] propose forming of the adequate mathematical models
for description of SAR images on the basis of probability
Classification of Multisensor Images with
Different Spatial Resolution
Aivars Lorencs1, Ints Mednieks1, Juris Sinica-Sinavskis1
1Institute of Electronics and Computer Science,
Dzerbenes St. 14, LV-1006 Riga, Latvia
mednieks@edi.lv
http://dx.doi.org/10.5755/ j01.eee.21.5.13333
81
ELEKTRONIKA IR ELEKTROTECHNIKA,ISSN 1392-1215,VOL.21,NO.5,2015
integral transforms of natural random values. However,
application of classification algorithms obtained in this way
has not provided excellent results with chosen training and
test sets.
a)
b)
Fig. 1. Illustration of processed multisensor data: (a) part of the RGB
image; (b) the same part of the band 10 image acquired by the thermal
infrared sensor, with ground truth for 7 classes.
Storvik et al. [12] proposed a Bayesian approach for the
case when pixels of lower resolution images precisely
overlap those from the higher resolution image and pixel
dimensions at lower resolutions are entire multiples of the
pixel dimensions at the higher (reference) resolution.
However, this situation can be achieved only in a single
multi-modal sensor.
III. TASK AND GOALS
A more common application case where the data were
acquired by two different sensors working in different
spectral ranges was considered within the 2014 IEEE GRSS
Data Fusion Contest (DFC) [13]. High resolution (~0.2 m ×
0.2 m pixel) RGB images and lower resolution (~1 m × 1 m
pixel) thermal infrared (TI) hyperspectral data in 84 bands
with wavelengths from 7.8 μm to 11.5 μm acquired from
urban area were presented for this contest. Classification of
higher resolution pixels into 7 land cover classes was
targeted and ground truth for all classes was presented to
facilitate that. In this paper we propose a classifier exploiting
DFC data from both sensors in its classification rule, present
and analyse the results obtained using different classifier
designs. The following goals were set up:
To design a classifier distinguishing all categories of
higher resolution pixels with high precision, i.e. low error
rate;
To provide a general classification approach not
exploiting specific features of the analysed scene i.e.
independent from image specifics;
To propose a solution requiring low computational
resources i.e. facilitate fast processing of large data sets.
Only “subset” images of the data set “grss_dfc_2014”
presented in the initial stage of the DFC were processed,
including all ground truth regions (see Fig. 1). Ground truth
regions were presented in a separate image with the same
spatial resolution as the RGB image and covered ~17 % of
the whole area of the “subset” image. To prepare ground
truth for the TI image, only pixels fully included in defined
ground truth regions were used, comprising ~12 % of the
whole area of the “subset” TI image.
IV. DESIGN OF SEPARATE CLASSIFIERS
To provide a solution of a defined task, a classifier
obtained using an appropriate data fusion method should be
designed. Our chosen approach presumes two stages: in the
first stage, two separate classifiers are designed, each using
one of the available images only; in the second stage, these
two classifiers are “mated” into a combined one, expecting
an increase of precision.
The vector of mean values
1 2 3
, , T
k k k k
μ
for
each class k,
1,7k
was calculated first from the ground
truth pixels of this class within the RGB image. After that,
covariance matrices for each class were calculated
1
1,
1
k
cT
k k k
k
c
Σ x μ x μ
(1)
where
k
c
is a number of pixels in the design set of the class
k;
x
is a column vector of the RGB intensity values of
pixel
.
Assuming that intensity distribution of pixels of the class k
in the RGB bands is represented by the random vector
1 2 3
, , T
k k k k
X X XX
, (1) actually gives us a point
estimate of the covariance matrix of this random vector.
To prepare a classifier for the TI image, 8 spectral bands
out of 84 were chosen, featuring lower image noise and
taken from different parts of spectral range of the TI sensor.
The following octet of TI bands was formed: (4, 14, 26, 36,
47, 57, 69, 78). After that, the same procedure as for the
RGB image was applied i.e. a vector of mean values
1 2 8
, ,..., T
k k k k
μ
for each class of pixels was
obtained and the covariance matrix calculated as follows
1
1,
1
k
cT
k k k
k
c
S y μ y μ
(2)
where
k
c
is the number of pixels in the design set of the
class k, formed for the TI image,
y
is the column vector of
the TI intensity values of pixel
taken from the chosen
bands.
With certain credibility we may assume that intensity
distribution of pixels of the class kin the RGB image is
represented by the random vector with Gaussian distribution
k
X
. The authors of [3] and [11] also consider this
82
ELEKTRONIKA IR ELEKTROTECHNIKA,ISSN 1392-1215,VOL.21,NO.5,2015
distribution as an adequate model for multidimensional
optical images. In this case, probability density function for
this vector can be approximately expressed by
1/2
3/2 1
2 exp ,
2
k k k
f M
xΣ x
(3)
where
k
Mx
is the square of the Mahalanobis distance
from vector
1 2 3
, , T
x x xx
to the vector of mean values
k
μ
.
Accepting hypothesis that intensity distribution of pixels
of the class kin the chosen 8 bands of the TI image is
represented by the random vector
k
Y
with Gaussian
distribution, probability density function for this vector can
be approximately expressed by
1/2
8/2 1
2 exp ,
2
k k k
g M
y S y
(4)
where
k
My
is the square of the Mahalanobis distance
from vector
1 2 8
, ,..., T
y y yy
to the vector of mean
values
k
μ
.
If the number of pixels that do not belong to any of the 7
classes is negligibly small, we can define two classifiers
which ignore these pixels.
We will denote by Wa classifier which classifies each
pixel in the RGB image as a pixel of class kif and only if the
intensity vector xmeets the following condition for each j
1.
k j
f f x x
(5)
In an analogous way we will define a classifier Vwhich
classifies each pixel in the TI image as a pixel of class kif
and only if its intensity vector ymeets the following
condition for each j
1.
k j
g g y y
(6)
Applying logarithmic operation to (5) and (6) we can
rewrite the classification rules in the following equivalent
form: classifier Wis a classifier which classifies a RGB
image pixel with intensity vector xas a pixel of class kif and
only if
ln ,
k
j k j
M M Σ
x x Σ
(7)
for each j. And, similarly, classifier Vis a classifier which
classifies a TI image pixel with intensity vector yas a pixel
of class kif and only if
ln ,
k
j k j
M M
S
y y S
(8)
for each j.
Apparently, classifiers Wand Vare qualified as Bayes
type classifiers. Our informative basis here allows to define
two other separate classifiers as well. Let us denote by W' a
classifier which classifies a RGB image pixel with intensity
vector xas a pixel of class kif and only if
,
k j
M Mx x
(9)
for each j. Similarly, classifier V' classifies a TI image pixel
with intensity vector yas a pixel of class kif and only if
,
k j
M M
y y
(10)
for each j.
V.DESIGN OF THE COMBINED CLASSIFIER
Combination of the separate classifiers is performed on
the basis of procedure relating each pixel from the RGB
image with a pixel from TI image so that the bigger pixel
from the TI image includes the larger part of the smaller
RGB image pixel. We will name such pixel from the TI
image the associated pixel. As the boundary pixels of the
ground truth polygons may have associated pixels only
partly corresponding to the defined ground truth areas, they
are eliminated from these polygons using the morphological
erosion and not used for design of the combined classifier.
Let us define the classification rule of the RGB image
pixels. We will denote by Ua classifier of RGB image
pixels which classifies a RGB image pixel
with intensity
vector xas a pixel of class kif and only if
ln ln ,
a a
j k j k
k k
j j
M M M M
x x y y
Σ S
Σ S
(11)
for each j, where
a
y
is the vector of the TI intensity values
of the pixel associated with pixel
. Classifier U' is
obtained by replacing the rule (11) with rule (12)
0,
a a
j k j k
M M M M
x x y y
(12)
for each j.
VI. RESULTS AND CONCLUSIONS
Testing of classifiers W,W',V,V',U, and U' on the basis
of the design sets provided the results presented in Table I.
To obtain comparable results, morphological erosion
introduced for design of the combined classifier was applied
for design of all classifiers. Kappa coefficient is based on
Cohen’s kappa measure and characterizes classifier's quality
in comparison with the ideal classifier. Classification results
of individual pixels obtained using the best classifier Uare
visualized in Fig. 2.
As it is seen from the table, Bayes type classifiers provide
better overall accuracy than their counterparts based on
Mahalanobis distance minimum principle in all cases. In
83
ELEKTRONIKA IR ELEKTROTECHNIKA,ISSN 1392-1215,VOL.21,NO.5,2015
addition, combined classifier provides better quality
measures than individual ones. Therefore we can conclude
that the proposed combination approach of separate
classifiers is fruitful.
In Fig. 2 one may notice that obtained pixel classification
errors are mainly related with intermixing of vegetation and
tree classes as well as erroneous classification of roads or
concrete roofs as grey roofs. Such errors were somewhat
expected due to spectral similarity of these classes and more
sophisticated classification approaches should be used to
distinguish between them.
TABLE I. CLASSIFICATION RESULTS.
W
W'
V
V'
U
U'
User's accuracy, %
road
96
53
92
91
96
87
trees
90
94
38
31
89
94
red roof
96
97
48
31
99
99
grey roof
83
94
57
77
91
96
concrete roof
97
96
19
59
97
96
vegetation
91
84
57
34
93
86
bare soil
94
90
54
50
97
97
Overall accuracy, %
93
83
55
54
95
91
Kappa
0.91
0.79
0.45
0.44
0.94
0.89
Fig. 2. Classification results of individual ground truth pixels obtained using the combined Bayes- type classifier U.
Analysis of the classification results lead to conclusion
that the assumption about the Gaussian distribution of the
intensity vectors of pixels within a land cover category is a
sufficiently adequate model of the real processed data.
Notwithstanding the already achieved, classification
accuracy can be probably improved by forming subclasses
within categories where the Bayes type classifier provides
relatively worse results. This could be planned as future
work. Another possible way to improve classification results
could be more thorough investigation of informativeness of
the spectral bands in TI image. However, in this work we
focused on development of the data fusion principle for
classification and mentioned improvement possibilities were
left out of its scope.
The authors of paper [12] have obtained good
classification results by designing classifiers based on more
sophisticated mathematical model, namely Markov random
field theory, describing distributions of the intensity vectors
of pixels. However, it cannot be substantiated that their
approach should be followed in particular application case.
Their assumption about the pixel sizes and locations in
images from different sensors is true in specific case which
cannot be related with our task. In addition, Cohen's kappa
measure achieved by these authors for real test data is not
higher than 0.9 and it is obtained for a different system of
pixel categories. The combined classifier proposed in this
paper seems much simpler and straightforward for
implementation.
ACKNOWLEDGMENT
The authors would like to thank Telops Inc. (Quebec,
Canada) for acquiring and providing the data used in this
study, the IEEE GRSS Image Analysis and Data Fusion
Technical Committee and Dr. Michal Shimoni (Signal and
Image Centre, Royal Military Academy, Belgium) for
organizing the 2014 Data Fusion Contest, the Centre de
Recherche Public Gabriel Lippmann (CRPGL, Luxembourg)
and Dr. Martin Schlerf (CRPGL) for their contribution of the
Hyper-Cam LWIR sensor, and dr. Michaela De Martino
(University of Genoa, Italy) for her contribution to data
preparation. The authors would like to thank reviewers for
their valuable suggestions as well as Madis Menke for his
assistance.
REFERENCES
[1] J. A. Richards, X. Jia, Remote Sensing Digital Image Analysis. An
Introduction. Springer, 439 p., 2006.
[2] R. Dinuls, G. Erins, A. Lorencs, I. Mednieks, J. Sinica-Sinavskis,
84
ELEKTRONIKA IR ELEKTROTECHNIKA,ISSN 1392-1215,VOL.21,NO.5,2015
“Tree species identification in mixed Baltic forest using LiDAR and
multispectral data”, IEEE Selected Topics Appl. Earth Observ.
Remote Sensing, vol. 5, no. 2, pp. 594–603, 2012. [Online].
Available: http://dx.doi.org/10.1109/JSTARS.2012.2196978
[3] A. Voisin, V. A. Krylov, G. Moser, S. B. Serpico, J. Zerubia,
“Supervised classification of multisensor and multiresolution remote
sensing images with a hierarchical copula-based approach”, IEEE
Trans. Geosci. Remote Sens., vol. 52, no. 6, pp. 3346–3358, 2014.
[Online]. Available: http://dx.doi.org/10.1109/TGRS.2013.2272581
[4] L. Xu, A. Krzyzak, C. Y. Suen, “Methods of combining multiple
classifiers and their applications to handwriting recognition”, IEEE
Trans. Syst., Man, Cybern., vol. 22, no. 3, pp. 418–435, 1992.
[Online]. Available: http://dx.doi.org/10.1109/21.155943
[5] Z. Ghahramani, H. C. Kim, “Bayesian classifier combination”,
Gatsby Computational Neuroscience Unit Technical Report No.
GCNU-T., London, UK, 9 p., 2003.
[6] E. Simpson, S. Roberts, I. Psorakis, A. Smith, “Dynamic Bayesian
combination of multiple imperfect classifiers”, Decision Making and
Imperfection, Berlin: Springer, pp. 1–38, 2013. [Online]. Available:
http://dx.doi.org/10.1007/978-3-642-36406-8_1
[7] J. A. Benediktsson, P. H. Swain, O. K. Ersoy, “Neural network
approaches versus statistical methods in classification of multisource
remote sensing data”, IEEE Trans. Geosci. Remote Sens., vol. 28,
no. 4, pp. 540–552, 1990. [Online]. Available: http://dx.doi.org/
10.1109/TGRS.1990.572944
[8] S. Foucher, M. Germain, J.-M. Boucher, G. B. Benie, “Multisource
classification using ICM and Dempster–Shafer theory”, IEEE Trans.
Instrum. Meas., vol. 51, no. 2, pp. 277–281, 2002 [Online].
Available: http://dx.doi.org/10.1109/19.997824.
[9] G. J. Briem, J. A. Benediktsson, J. R. Sveinsson, “Multiple classifiers
applied to multisource remote sensing data”, IEEE Trans. Geosci.
Remote Sens., vol. 40, no. 10, pp. 2291–2299, 2002. [Online].
Available: http://dx.doi.org/10.1109/TGRS.2002.802476
[10] M. Fauvel, J. Chanussot, J. A. Benediktsson, “Decision fusion for the
classification of urban remote sensing images”, IEEE Trans. Geosci.
Remote Sens., vol. 44, no. 10, pp. 2828–2838, 2006. [Online].
Available: http://dx.doi.org/10.1109/TGRS.2006.876708
[11] B. Storvik, G. Storvik, R. Fjortoft, “On the combination of
multisensor data using meta-Gaussian distributions”, IEEE Trans.
Geosci. Remote Sens., vol. 47, no. 7, pp. 2372–2379, 2009. [Online].
Available: http://dx.doi.org/10.1109/TGRS.2009.2012699
[12] G. Storvik, R. Fjortoft, A. H. S. Solberg, “A Bayesian approach to
classification of multiresolution remote sensing data”, IEEE Trans.
Geosci. Remote Sens, vol. 43, no. 3, pp.539–547, 2005. [Online].
Available: http://dx.doi.org/10.1109/TGRS.2004.841395
[13] 2014 IEEE GRSS Data Fusion Contest. [Online]: Available:
http://www.grss-ieee.org/community/technical-committees/data-
fusion/.
85