Chapter 2

Pixel-Based Change Detection Methods

Abstract In this chapter, we consider pixel-based change detection methods. First,

we provide well-known methods in the literature. Then, we propose two novel pixel-
based change detection methods.

Keywords Pixel-based change detection · Image differencing · Automated thresh-

olding · Percentile · Otsu’s method · Kapur’s algorithm · Image rationing · Image
regression · Least-squares · Change vector analysis (CVA) · Median filtering · Back-
ground formation · Fuzzy logic · Fuzzy xor

2.1 Image Differencing

In this technique, images of the same area, obtained from times t1 and t2 , are sub-
tracted pixelwise. Mathematically, the difference image is

Id (x, y) = I1 (x, y) − I2 (x, y), (2.1)

where I1 and I2 are the images obtained from t1 and t2 , (x, y) are the coordinates of
the pixels. The resulting image, Id , represents the intensity difference of I1 from I2 .
This technique works only if images are registered.
To interpret the difference image, we need to recall the quotation from Singh [1]:
“The basic premise in using remote sensing data for change detection is that changes
in land cover must result in changes in radiance values and changes in radiance due
to land cover change must be large with respect to radiance changes caused by other
factors.” Based on this principle, we can expect that intensity differences due to land
cover change resides at the tails of the difference distribution of the image. Assuming
that changes due to land cover are less than changes by other factors, we expect that
most of the difference is distributed around the mean. We can illustrate the difference
distribution as in Fig. 2.1.

M. İlsever and C. Ünsalan, Two-Dimensional Change Detection Methods, 7

SpringerBriefs in Computer Science, DOI: 10.1007/978-1-4471-4255-3_2,
© Cem Ünsalan 2012
8 2 Pixel-Based Change Detection Methods

Fig. 2.1 Distribution of a difference function. Significant changes are expected at the tails of the
distribution

For a zero mean difference distribution, we can normalize I2 as

σ1
I˜2 (x, y) = (I2 (x, y) − μ2 ) + μ1 , (2.2)
σ2

where I˜2 is the normalized form of I2 . μ1 , σ1 and μ2 , σ2 are the mean and the
standard deviation of I1 and I2 , respectively. After normalization, the mean and
standard deviation of the two images are equalized. Hence, the difference image will
have zero mean. Now, we can update Eqn. 2.1 as

Id (x, y) = |I1 (x, y) − I˜2 (x, y)|. (2.3)

To detect the change, we can apply simple thresholding to Id (x, y) as


1, Id (x, y) ≥ τ
T (x, y) = (2.4)
0, otherwise,

where the threshold τ is often determined empirically.

Since the threshold value in Eqn. 2.4 is important, various automated thresh-
old selection algorithms are proposed. Most of the time, the performance of these
algorithms is scene dependent due to the assumptions they are based on. Rosin and
Ioannidis [2] investigated the performance of several automated thresholding algo-
rithms using a large set of difference images calculated from an automatically created
ground truth database. They give results based on several measures for a complete
evaluation. In this study, we benefit from three different threshold selection methods.
These are percentile thresholding, Otsu’s method [3] and Kapur’s algorithm [4].
We will briefly explain these thresholding methods next. To this end, we make
some assumptions about Id as follows. Id is a grayscale image which is represented
by Ng gray levels, {1, 2, . . . , Ng }. The number of pixels at level i is denoted by n i
and the total number of pixels is N .
2.1 Image Differencing 9

The first thresholding method is based on the percentile [5]. It is a statistics of

ordinal scale data. Assume that A is a sorted array of pixel values of Id in ascending
order. Rank of the Pth percentile of Id is given by
 
P
R = ceil ×N (2.5)
100

where the ceil function rounds its argument to the nearest greater integer. Pth per-
centile is found by indexing A using that rank.
The second thresholding method is proposed by Otsu. It uses measures of class
separability in finding an optimal threshold value. Relative frequencies of pixel values
at level i are given by

Ng
ni 
pi = , pi ≥ 0, pi = 1. (2.6)
N
i=1

A threshold value at gray level k divides the histogram of Id into two classes.
Each class has its own probability of occurrence (total probability of its samples)
and own mean value. Evaluation function of the Otsu’s method is the between-class
variance given by
[μ Id ω(k) − μω ]2
σb2 = , (2.7)
ω(k)[1 − ω(k)]

where μ Id is the mean of Id ; ω(k) is the probability of the class which includes gray
levels up to k and μω is the mean of the class ω. The optimal threshold value k ∗
maximizes
σb2 (k ∗ ) = max σb2 (k). (2.8)
1≤k≤Ng

The last thresholding method we use is Kapur’s algorithm. Similar to Otsu’s

method, it divides the image histogram into two classes. It then utilizes the sum of
the entropy of these two classes as an evaluation function. The value which maximizes
this sum is taken as the optimal threshold value. For two classes A and B, Shannon
entropy of these classes are defined as


k
pi pi
H (A) = − ln , (2.9)
ω(k) ω(k)
i=1

Ng
 pi pi
H (B) = − ln , (2.10)
[1 − ω(k)] [1 − ω(k)]
i=k+1

where the histogram is divided at gray level k. The optimal threshold value k ∗
maximizes the sum φ(k) = H (A) + H (B) such that
10 2 Pixel-Based Change Detection Methods

Fig. 2.2 Images taken at two different times from a developing region of Adana

Fig. 2.3 Image differencing applied to the Adana image set. a The difference image b Thresholded
version

φ(k ∗ ) = max φ(k). (2.11)

1≤k≤Ng

To explain different change detection methods, we pick the Adana test image set
given in Fig. 2.2. The two images, taken in different times, in this set represent a
region with construction activity. These images are registered. Therefore, they can
be used for pixelwise change detection methods. The difference between these two
images is clearly seen. We will use this image set in the following sections also.
The difference image obtained from the Adana image set is as in Fig. 2.3a. This
image is color coded with the color scale given next to it. We also provide the
thresholding result in Fig. 2.3b. In thresholding, we benefit from Kapur’s method.
As can be seen, the thresholded image provides sufficient information about the
changed regions in the image.
Griffiths [6] used image differencing for detecting the change in urban areas. He
used Landsat TM data (with 30 m resolution), SPOT XS multispectral data (with
20 m resolution), and SPOT panchromatic data (with 10 m resolution) in his study.
He proposed using an urban mask to find changes in urban areas using image differ-
encing. Griffiths indicated that the mixture of buildings, streets, and small gardens in
urban areas produce a highly textured appearance compared with the much smoother
texture of arable fields. He used a standard deviation filter to quantify the texture. The
urban mask is multiplied by the difference image to eliminate non-urban areas. Fur-
thermore, he refined the results based on a previous study. In this technique, changes
that occur far from the urban areas are assumed to be non-urban change. This is
2.1 Image Differencing 11

because new urban development generally occurs at the periphery of existing urban
areas. Griffiths presented his results for each technique by visual interpretation.
Saksa et al. [7] used image differencing for detecting clear cut areas in boreal
forests. They tested three methods using Landsat satellite imagery and aerial
photographs as: pixel-by-pixel differencing and segmentation, pixel block-level dif-
ferencing and thresholding, pre-segmentation and unsupervised classification. In
the first method, they found the difference image. Then, they used a segmentation
algorithm to delineate the clear cut areas. In the second method, they included neigh-
boring pixels into the calculation of the difference image. Therefore, negative effects
of misregistration are reduced in the resulting image. In the third method, they first
segmented the images. Then, they obtained a segment-level image difference. They
labeled clear cut areas by using an unsupervised classification algorithm. Saksa et
al. concluded that, predelineated segments or pixel blocks should be used for image
differencing in order to decrease the amount of misinterpreted small areas.
Lu et al. [8] compared 10 binary change detection methods to detect land cover
change in Amazon tropical regions. They used Landsat TM (Thematic Mapper) data
in their study. In addition to band differencing, they tested a modified version of
image differencing where pixels are accepted as changed when majority of the bands
indicate change. For six-band Landsat TM data, if four of the bands indicate change
then the pixel value is labeled as changed. They reported that the difference of Landsat
TM band 5, modified image differencing, and principal component differencing
produced best results.

2.2 Image Rationing

Similar to image differencing, images are compared pixelwise in this method. There-
fore, images must be registered beforehand. The ratio image, used in this method, is
calculated by
I1 (x, y)
Ir (x, y) = . (2.12)
I˜2 (x, y)

In Eqn. 2.12, the Ir image takes values in the range [0, ∞). If the intensity values
are equal, it takes the value 1. To normalize the value of Ir , we can benefit from the
arctangent function as
 
I1 (x, y) π
Ir (x, y) = arctan − . (2.13)
I˜2 (x, y) 4

Now, ratio image takes values in the range [−π/4, π/4]. To threshold Ir , we
can benefit from the same methods as we did in the previous section. In Fig. 2.4a,
we provide the Ir image obtained from the Adana test image set. We provide the
thresholded version of this image in Fig. 2.4b. As in the previous section, we used
12 2 Pixel-Based Change Detection Methods

Fig. 2.4 Image ratio applied to the Adana image set. a The ratio image b Thresholded Version

Kapur’s method to obtain the optimal threshold value. As can be seen, the thresholded
image provides sufficient information on the changed regions in the image.

2.3 Image Regression

In image regression, the I2 image (obtained from t2 ) is assumed to be a linear function

of the I1 image (obtained from t1 ). Under this assumption, we can find an estimate
of I2 by using least-squares regression as

Iˆ2 (x, y) = a I1 (x, y) + b. (2.14)

To estimate the parameters a and b, we define the squared error between the measured
data and predicted data (for each pixel) as

e2 = (I2 (x, y) − Iˆ2 (x, y))2 = (I2 (x, y) − a I1 (x, y) − b)2 . (2.15)

The sum of the squared error becomes


N 
N
S= e2 = (I2 (xn , yn ) − a I1 (xn , yn ) − b)2 . (2.16)
n=1 n=1

Here, we assume that we have N observations. We want to find the parameters a

and b to minimize the sum of the squared error S. Therefore, we first calculate the
partial derivatives of S with respect to a and b as

∂S  N
= −2 (I2 (xn , yn ) − a I1 (xn , yn ) − b), (2.17)
∂b
n=1
2.3 Image Regression 13

Fig. 2.5 Image difference after regression applied to the Adana image set. a The difference image
b Threshold Version

∂S  N
= −2 [(I2 (xn , yn ) − a I1 (xn , yn ) − b)I1 (xn , yn )]. (2.18)
∂a
n=1

By equating Eqn. 2.17 and 2.18 to zero, we obtain two equations with two
unknowns as


N 
N 
N
0= I2 (xn , yn ) − a I1 (xn , yn ) − b, (2.19)
n=1 n=1 n=1

N 
N 
N
0= I2 (xn , yn )I1 (xn , yn ) − a I1 (xn , yn )2 − bI1 (xn , yn ). (2.20)
n=1 n=1 n=1

Solving these equations, we obtain

N N N
n n=1 I2 (x n , yn )I1 (x n , yn ) − n=1 I2 (x n , yn ) n=1 I1 (x n , yn )
a= N N , (2.21)
n n=1 I1 (xn , yn )2 − ( n=1 I1 (xn , yn ))2
N N
n=1 I2 (x n , yn ) − a n=1 I1 (x n , yn )
b= . (2.22)
N
We manually picked the observations (for n = 1, . . . , N ) from I1 and I2 (from the
unchanged areas). When we subtract I2 from Iˆ2 as Id (x, y) = I2 (x, y) − Iˆ2 (x, y),
we expect to find changes originating from land cover. When we apply this method to
the normalized I2 ( I˜2 ), we further eliminate the insignificant changes that still remain
after normalization. Consequently, this technique gives slightly better performance
compared to image differencing.
We provide the difference image obtained by image regression using our Adana
image test set in Fig. 2.5a. We provide the thresholded version of this image in
Fig. 2.5b. As in the previous sections, we benefit from Kapur’s method in threshold
selection. As can be seen, the change map obtained is similar to image differencing.
14 2 Pixel-Based Change Detection Methods

Fig. 2.6 Unchanged and

changed pixel vectors in a 2-D
spectral space

2.4 Change Vector Analysis

Change Vector Analysis (CVA) is a technique where multiple image bands can be
analyzed simultaneously. As its name suggests, CVA does not only function as a
change detection method, but also helps analyzing and classifying the change. In
CVA, pixel values are vectors of spectral bands. Change vectors (CV) are calcu-
lated by subtracting vectors pixelwise as in image differencing. The magnitude and
direction of the change vectors are used for change analysis. In Fig. 2.6, a changed
pixel and an unchanged pixel are given in a two-band spectral space.
The change vector magnitude can indicate the degree of change. Thus, it can
be used for change and no-change classification. Under ideal conditions, such as
perfect image registration and normalization, unchanged pixel magnitudes must be
equal to zero. However, this is not the case in practice. Therefore, thresholding must
be applied to the change magnitude. While the change vector magnitude behaves
like a multi-band version of the image differencing, the change direction gives us
additional information about the change type. This is often more valuable than the
amount of change, since in most applications we are interested in a specific change
type.
In practice, the number of CV directions are uncountable. Therefore, it is nec-
essary to quantize the CV space and assign directions accordingly. A simple quan-
tization of CV directions can be achieved by dividing the space by its main axes.
In Fig. 2.7, a 2D CV space is quantized into four subsets (quadrants) by the axis
of band 1 and band 2. For three band images, subsets can be octants. CVs can be
assigned to subsets via signs of their components.
As mentioned earlier, CV directions can be used in classifying the change. By
using subsets, we can determine 2n classes of change for an n dimensional space. CVA
2.4 Change Vector Analysis 15

Fig. 2.7 Change vector space

is divided into four subsets

can also be applied to transformed data such as Kauth-Thomas Transformation (KTT)

(explained in Sect. 3.2) rather than to raw data. In the KTT space, a simultaneous
increase in the greenness feature and decrease in the brightness feature indicates gain
of vegetation. Therefore, in the change vector space of KTT bands, we can assign this
change class (change toward vegetation) to the subsets where greenness is positive
and brightness is negative.
CVA is used first by Malila [9] for change detection. He used the KTT with
CVA and reported results for change of forestation. He used change directions to
distinguish changes due to harvesting and regrowth. Johnson et al. [10] provided a
comprehensive investigation of CVA. They provided the details to the implemen-
tation of CVA after a functional description. They reported that, CVA can be used
in applications which require a full-dimensional data processing and analysis tech-
nique. They also found CVA to be useful for applications in which: the changes
of interest and their spectral manifestation are not well-known a priori; changes of
interest are known or thought to have high spectral variability; changes in both land
cover type and condition may be of interest.
In Fig. 2.8a, we provide the change vector magnitude image which can be used
for change and no-change classification. In Fig. 2.8b the change vector magnitude
image is thresholded by Kapur’s algorithm. This thresholded image also provides
sufficient information on changed regions.
16 2 Pixel-Based Change Detection Methods

0.8
(a) (b)
0.7
0.6
0.5
0.4
0.3
0.2
0.1

Fig. 2.8 CVA (in terms of magnitude value) applied to the Adana image set. a The magnitude
value b Threshold Version

2.5 Median Filtering-Based Background Formation

In this method, we detect changes by subtracting the multi-temporal images from

a reference image (background). This technique is widely used in video processing
for detecting and analyzing motion. Although well known in video signal processing
community, this is the first time median filtering is used for change detection in
satellite images.
For this method, several background formation techniques are proposed in the
literature. In this study, we use temporal median filtering for background formation.
It is defined as

Ibg (x, y) = Med(I1 (x, y), I2 (x, y), . . . , I N (x, y)), (2.23)

where Ibg is the background image and (I1 , I2 , . . . , I N ) are the images from times
(t1 , t2 , . . . , t N ).
Parameters of the median operator are pixel values from the same spatial location
(x, y). Median is calculated by choosing the middle element of the sorted array of
its parameters. This leads to the removal of outliers (impulsive or salt and pepper
noise) from the input pixel set. This characteristic of the median filter helps us to find
a pixel value for every spatial location which is equal or approximately equal to the
majority of the elements of the temporal pixel set. We expect to find the change by
subtracting each image from the background, thus enchanting deviations from the
median value.
We provide our multi-temporal images for background formation in Fig. 2.9. We
also provide the median filtering result in Fig. 2.10.
Using the extracted background image, we obtain four difference images as in
Fig. 2.11. We also provide the thresholded version of these images in Fig. 2.12. As
in the previous sections, we used Kapur’s algorithm in finding the threshold value.
As can be seen in Fig. 2.12, majority of the significant changes are gathered in the
first and last images in the series. These correspond to the two extremes of the time
interval.
2.6 Pixelwise Fuzzy XOR Operator 17

Fig. 2.9 Multi-temporal images used in background formation

Fig. 2.10 Background image

generated by median filtering

2.6 Pixelwise Fuzzy XOR Operator

The last method for pixelwise change detection is a novel contribution to the
community. In this method, the binary XOR operation is taken as a benchmark. Its
fuzzy version is used for change detection. Our rationale here is as follows. Assume
that we have two binary images (composed of only ones and zeros) and we want
to detect the changed pixels in these. Each pixel p(x, y) in a binary image B is
valued according to a characteristic function βB , which could also be called as the
“whiteness” function defined as
18 2 Pixel-Based Change Detection Methods

Fig. 2.11 Difference images for each sample generated by subtracting each sample from the back-
ground image

Fig. 2.12 Difference images are thresholded by Kapur’s algorithm


1, if B(x, y) is white
p(x, y) = βB (x, y) = (2.24)
0, otherwise.
2.6 Pixelwise Fuzzy XOR Operator 19

Between two pixels p1 and p2 , at the same (x, y) coordinates of the two binary
images B1 and B2 , the existence of a change can only mean that either “ p1 is white and
p2 is not” or “ p1 is not white and p2 is.” This directly implies the XOR operation in
binary logic. Hence the obvious solution to the change detection problem is XOR-ing
the two binary images pixelwise as

C(x, y) = B1 (x, y) ⊕ B2 (x, y). (2.25)

This operation gives ‘0’ for pixels having the same value in both images, and
gives ‘1’ for pixels having different values. Therefore, white pixels in the resulting
binary image C(x, y) represent the changed regions.
Unfortunately, this method cannot be applied to panchromatic or multispectral
satellite imagery (having pixel values in a certain range). In order to perform a
similar method on satellite imagery, we propose a fuzzy representation for these. We
also benefit from the combination of fuzzy and crisp (binary) operations.
Panchromatic images are composed of pixels with values p(x, y) in a certain
range. Normalizing these values and mapping them to the range [0, 1] effectively
translates the image into a fuzzy set, whose elements (pixels) have membership
grades in proportion to their “whiteness.” The membership grade g(x, y) of each
pixel p(x, y) in the grayscale image G is thus defined by the fuzzy membership
function μG as
⎧
⎪
⎪ 1.00, if G(x, y) is pure white
⎪
⎪
⎪
⎪
⎨... ...
g(x, y) = μG (x, y) = 0.50, if G(x, y) is gray (2.26)
⎪
⎪
⎪
⎪ ... ...
⎪
⎪
⎩0.00, if G(x, y) is pure black.

Comparison of two binary images involves the crisp question “Are these two pixels
different?.” Whereas a fuzzy comparison of two panchromatic images involves the
fuzzy question “How different are these two pixels?.” Also the question of “Above
what amount of difference shall the two pixels be labeled as changed?.” The amount
of difference between gray level values in the image domain directly corresponds
to the difference between the degrees of membership in the fuzzy domain. For this
particular application, the fuzzy complement (NOT) operation, defined as

ḡ(x, y) = μḠ (x, y) = 1 − g(x, y) (2.27)

and the algebraic representation of the fuzzy intersection (AND) operation, defined
as the multiplication of membership functions

μG 1 ∩G 2 = μG 1 (x, y)μG 2 (x, y) = g1 (x, y)g2 (x, y) (2.28)

were used to obtain a fuzzy difference metric [11].

20 2 Pixel-Based Change Detection Methods

Fig. 2.13 Fuzzy XOR applied to the Adana image set. a Fuzzy AND (g1 (x, y)ḡ2 (x, y)) b Fuzzy
AND (ḡ1 (x, y)g2 (x, y)) c Thresholded version

In a manner similar to the binary case, the measure of change between two pixels
p1 and p2 is given by the degree of truth of the following statement: either “ p1
is lighter AND p2 is darker” OR “ p1 is darker AND p2 is lighter”; which can be
rephrased as, either “ p1 has a high membership grade AND p2 has a low membership
grade” OR “ p1 has a low membership grade AND p2 has a high membership grade.”
Considering that “having a low membership grade” is the opposite of “having a
high membership grade,” the former statement’s degree of truth is the complement of
the latter’s, and the degree of truth in “having a high membership grade” is equivalent
to the membership grade g(x, y) itself. Consequently, the above fuzzy rule can be
formulated as

C(x, y) = μ(G 1 ∩Ḡ 2 )∪(Ḡ 1 ∩G 2 ) (x, y) = (g1 (x, y)ḡ2 (x, y)) ∪ (ḡ1 (x, y)g2 (x, y)).
(2.29)
The fuzzy value C(x, y) represents the measure of change between two images
at the coordinate (x, y). The decision of a significant change can be made by means
of applying an appropriate threshold and converting C(x, y) to a crisp YES/NO
value. Experiments have shown that, the results from the two fuzzy AND operations
are distributed in a way that automatically indicates an appropriate threshold for
defuzzification. More explicitly, threshold values are obtained for both fuzzy AND
operations from τ = argmax(Ha ) + 2σa . Here, Ha is the histogram of the corre-
sponding fuzzy AND operation and σa is the standard deviation of the corresponding
fuzzy AND operation. In fact, applying this threshold and converting the fuzzy AND
results to a crisp binary value, and then combining them with the binary OR operator
yielded better results in detecting changed regions in satellite images. Therefore, the
2.6 Pixelwise Fuzzy XOR Operator 21

proposed method was eventually established as an ensemble of both fuzzy and binary
logic operations.
We provide the images obtained by fuzzy AND operations using our Adana image
set in Fig. 2.13a and Fig. 2.13b. We provide the thresholded version after finding the
C(x, y) in Fig. 2.13c. As can be seen, the resulting image shows the changed region
fairly well.

References

1. Singh, A.: Review article: Digital change detection techniques using remotely-sensed data. Int.
J. Remote Sens. 10(6), 989–1003 (1989)
2. Rosin, P.L., Ioannidis, E.: Evaluation of global image thresholding for change detection. Pattern
Recognition Lett. 24(14), 2345–2356 (2003)
3. Otsu, N.: A threshold selection method from gray-level histograms. IEEE Trans. Syst. Man
Cybern. 9(1), 62–66 (1979)
4. Kapur, J.N., Sahoo, P.K., Wong, A.K.C.: A new method for gray-level picture thresholding using
the entropy of the histogram. Comput. Vis. Graph. Image Process. 29(3), 273–285 (1985)
5. Devore, J.: Probability and Statistics for Engineering and Sciences. 6 edn. Thompson (2004)
6. Griffiths, G.H.: Monitoring urban change from Landsat TM and Spot satellite imagery by
image differencing. In: Proceedings of the 1988 International Geoscience and Remote Sensing
Symposium, vol. 1, (1988)
7. Saksa, T., Uuttera, J., Kolstrom, T., Lehikoinen, M., Pekkarinen, A., Sarvi, V.: Clear-cut detec-
tion in boreal forest aided by remote sensing. Scandinavian J. For. Res. 18(6), 537–546 (2003)
8. Lu, D., Mausel, P., Batistella, M., Moran, E.: Land-cover binary change detection methods
for use in the moist tropical region of the Amazon: A comparative study. Int. J. Remote Sens.
26(1), 101–114 (2005)
9. Malila, W.A.: Change vector analysis: An approach for detecting forest changes with Landsat.
In: LARS Symposia, p. 385 (1980)
10. Johnson, R.D., Kasischke, E.S.: Change vector analysis: A technique for the multispectral
monitoring of land cover and condition. Int. J. Remote Sens. 19(3), 411–426 (1998)
11. Klir, G.J., Yuan, B.: Fuzzy Sets and Fuzzy Logic Theory and Applications. Prentice Hall, New
York (1995)
http://www.springer.com/978-1-4471-4254-6