Survey on Real Time moving object detection and back ground detection

for Intelligence Security System

2013

Survey on Real Time moving object detection and back ground detection for
Intelligence Security System
Prof. Praveen R. Barapatre

Mr. Rajkumar Bhimrao Pawar

Department Of Computer Engineering

SKN Sinhgad Institute of Technology, Kusgaon(Bk.)
A/p: Lonawala, Tal: Maval, Dist: Pune. 410401
MH, India.
E-mail: pravinbarapatre@hotmail.com

Department Of Computer Engineering

SKN Sinhgad Institute of Technology, Kusgaon(Bk.)
A/p: Lonawala, Tal: Maval, Dist: Pune. 410401
MH, India.
E-mail: rajpawar638@gmail.com

ABSTRACT

Moving object detection has been widely used in

surveillance system such as intelligent security
systems, airport security systems, video
monitoring systems, etc. Object tracking is one
of the most necessary steps for surveillance.
Moving object and background must be
separated as much as clear. Motion detection
typically refers to movement of objects as a
whole, e.g., movement of pedestrians or vehicles.
This pre processing can be done via Background
Subtraction and Temporal Differencing. A
popular object segmentation strategy is
background subtraction.
While traditional
methods of Segmentation using Thresholding,
Background subtraction and Background
estimation provide satisfactory results to detect
single objects, noise is produced in case of
multiple objects and in poor lighting conditions.
In this paper we will be focusing on important
features to build an intelligent surveillance
system and its architecture. A detailed analysis
of background subtraction methods and their
comparison is provided in this paper.
Keywords:
GMM,
Thresholding,
Segmentation, Loitering

I.

CCTV,

INTRODUCTION

In INDIA recently more focuses on public

security because of increasing crime day by day.
Like terrorists activity, rape, sexual harassment,

murder, theft etc. As we know, after the criminal

activity done, government deciding about to cover
more public area covering with a CCTV cameras
to improve the security. But in the traditional
systems [1] a network of passive cameras able of
only recording the moving images and converting
it to video files are deployed at the site of prime
importance. In case of accidents, terrorist attacks,
theft, malicious activities such video data comes
to be very important for analysis. But if traditional
system is used we require to watch the video data
again and again for long hours together because
collected data is very large and we dont know the
exact time when the malicious activity done. So
that it is much necessary to build an intelligent
system that would not only to provide a data as
per traditional system but also assist us to detect
malicious activities. It would be so easy if system
on its own recognizes some events as malicious
and extract only those frames which are providing
suspicious activity. The endless hours of watching
the video can be avoided [4].
II. BACKGROUND

REMOVAL

AND

OBJECT

TRACKING

Here we had shown how to isolate objects or

parts of objects from the rest of the image. In
video security, for example, the camera mostly
looks out on the same boring background, which
really isnt of interest. When something is left in
the scene that wasnt there before like a moving
object. We want to isolate those events and to be
able to ignore the endless hours when nothing is
changing (i.e background) .

ANNUAL STAFF CONVENTION - NATIONAL CONFERENCE 2013

Survey on Real Time moving object detection and back ground detection
for Intelligence Security System

2013

two-level. This means that each pixel is stored as a

single bit (0 or 1). The names black-and-white,
B&W, monochrome or monochromatic are often
used for this concept, but may also designate any
images that have only one sample per pixel, such
as grayscale images.
IV. PROPOSED SYSTEM
Fig.1: Background subtraction: original image (left) and
foreground image (right)

Moving object detection in real time is a

challenging task in visual surveillance systems. It
often acts as an initial step for further processing
such as classification of the detected moving
object. In order to perform more sophisticated
operations such as classification, we need to first
develop an efficient and accurate method for
detecting moving objects [6]. One of the simplest
and popular method for moving object detection is
a background subtraction method which often uses
background modeling, but it takes long time to
detect moving objects [6].

In the following architecture shown how camera

capturing frames for video and how our
intelligence system detects human normal
behaviors and malicious intents. If any suspicious
frame or activity done by human this system
traced it and start alarm to control crime. This
system is working with

III. Shadow removal

Performing the operation using a function on
each frame by 8*8 blocks wise and result is
compared with the variance threshold. If the result
is less than the variance threshold, it assumes as
shadow and it takes logic 0 otherwise it takes logic
1 [8]. Temporal difference method is very simple
and it can detect objects in real time, but it does
not provide robustness against illumination
change. The foreground extraction problem is
dealt with the change detection techniques, which
can be pixel based or region based. Simple
differencing is the most intuitive by arguing that a
change at a pixel location occurs when the
intensity difference of the corresponding pixels in
two images exceeds a certain threshold. A binary
image is a digital image that has only two possible
values for each pixel [6-7]. Typically the two
colors used for a binary image are black and white
though any two colors can be used. The color used
for the object(s) in the image is the foreground
color while the rest of the image is the background
color. Binary images are also called bi-level or

Fig.2 Generalized architecture of our system

a) Background Removal and segmentation
module

Step1: This routine will detect each frame at

real time from a video sequence and present its
status in form of matrix, its frame no, pixels, color
content and sample rate, etc. As we get features of
the each frame we can apply various processing
algorithms to these frames using image processing
algorithms. Mainly due to durability and
simplicity of its use we will be opting for Gaussian
models for background subtraction.
Step2: This will detect and remove background
features to detect the object in the scene.
Algorithms like GMM based will be a tool for

ANNUAL STAFF CONVENTION - NATIONAL CONFERENCE 2013

Survey on Real Time moving object detection and back ground detection
for Intelligence Security System
such removal. As shown in fig.3 we need to
remove unwanted background and separate
foreground data containing moving objects. Then
we apply segmentation algorithms to separate
moving objects from background.

Fig. 3: Architecture of proposed system.

Step 3: Non human objects are to be deleted
and human objects are to be maintained in frames.
Once we get a human object in frame we can track
the motion of such object in subsequent frames.
Loitering: Standing or roaming at a place more
than required and for more time than a preset
threshold.

2013

In the above figure shown a single person or no

interaction (see Fig. 2) consists of behaviors that
can be defined only by considering person(s),
which are not interacting with any other person or
vehicle. The person may be planning to do some
kind of malicious activity that will be much
harmful to the society or it may be possible he is
there for any positive activity. But precaution is
better than cure our system will detect the person
and take action like a criminal activity. To
maintain the peace in the society this kind of
surveillance system is much more useful. For
example, loitering, crowd counting, crowd flow
(behavior) analysis, and person talking on a cell
phone. Loitering is defined as the presence of an
individual in an area for a period of time longer
than a given time threshold. Methods for
automatically detecting loitering in real time
would enable deployed security to investigate suspicious individuals or to target loitering stations
for future investigation. Loitering is of special
interest to public transit systems since it is a
common practice of drug dealers, beggars, and
muggers, among others. In this survey, loitering
refers to a behavior that exclusively involves a
human [5].
V.

SEGMENTATION ALGORITHMS

In computer vision, segmentation refers to the

process of partitioning a digital image into
multiple segments (sets of pixels, also known as
super pixels). The goal of segmentation is to
simplify and/or change the representation of an
image into something that is more meaningful and
easier to analyze. Image segmentation is typically
used to locate objects and boundaries (lines,
curves, etc.) in images [3].

Fig.4 single-person or no interaction behavior.

Suspicious person (marked with an ellipse) loitering
for a long period of time without leaving in a bus.
ANNUAL STAFF CONVENTION - NATIONAL CONFERENCE 2013

Survey on Real Time moving object detection and back ground detection
for Intelligence Security System

Fig. 5: Segmentation: original image (left)

segmented image (right)

and

its

Thresholding is the simplest method of image

segmentation. From a grayscale image,
thresholding can be used to create binary images.
During the thresholding process, individual pixels
in an image are marked as "object" pixels if their
value is greater than some threshold value
(assuming an object to be brighter than the
background) and as "background" pixels
otherwise. This convention is known as threshold
above. Our aim is to choose best thresholding
algorithm for segmentation [1-3].
In practice, the histograms do not show two
clearly separated classes rather overlapping
classes. Many methods have been suggested in the
past for such kind of problem but still for
overlapping classes, it is hard to determine a
global threshold. Hence, attempts have been made
by proposing a new approach to determine the
global threshold for image segmentation. The
algorithm is found to produce satisfactory results
for images having histograms with bimodal feature
[9-10].

2013

intensity image; this has the effect of forming

valleys or basins (the low points) where there is
no texture and of forming mountains or ranges
(high ridges corresponding to edges) where there
are dominant lines in the image. It then
successively floods basins starting from user
specified (or algorithm-specified) points until
these regions meet. Regions that merge across the
marks so generated are segmented as belonging
together as the image fills up. In this way, the
basins connected to the marker point become
owned by that marker. We then segment the
image into the corresponding marked regions [9].
Meyer's watershed flooding algorithm
(1) A set of markers, pixels where the flooding
shall start, are chosen. Each is given a different
label.
(2) The neighboring pixels of each marked area
are inserted into a priority queue with a priority
level corresponding to the gray level of the
pixel.
(3) The pixel with the highest priority level is
extracted from the priority queue. If the
neighbors of the extracted pixel that have
already been labeled all have the same label,
then the pixel is labeled with their label. All
non-marked neighbors that are not yet in the
priority queue are put into the priority queue.
(4) Redo step 3 until the priority queue is empty.
The non-labeled pixels are the watershed lines
[13].
B. Advanced Background Subtraction

A. Advanced Segmentation Algorithms

a. Watershed algorithm:
This algorithm converts lines in an image into
mountains and uniform regions into valleys that
can be used to help segment objects. The
watershed algorithm first takes the gradient of the

Methods
a. Gaussian models

Running Gaussian average was first

proposed by Wren, Azarbayejani, Darell, pentland
in 1997. The main idea is to of actually fitting one
Gaussian distribution (, ) over the histogram.

ANNUAL STAFF CONVENTION - NATIONAL CONFERENCE 2013

Survey on Real Time moving object detection and back ground detection
for Intelligence Security System
This was supposed to give the Background PDF.
This Background PDF was to be updated from
time to time for suppressing background changes.
This update was given by running average as
follows [3].
Generalized Gaussian Family model is most
representative background model. GGF model
was proposed by H. Kim, R. Sakamoto et all to
cope up with the background changes and
shadows using this model. Background was
modeled as pixel variation in a static scene over
time with GGF distribution defined [3].

2013

system people should be feel safe to survive in any

place any time.
Thus we have successfully defined our problem
statement to develop an Intelligent System for
Security and machine vision to detect abnormal
malicious intents. We have identified the
necessary steps and modules to achieve our
proposed system. In the light of the study we have
anticipated various algorithms and work done in
the fields of image processing and video
processing. We have identified important and core
technologies as Image segmentation, Background
removal, Object classification, motion tracking
and behavior analysis and have studied each
technology with its recent developments.

b. Bayesian Background Modeling

In probability theory and applications, Bayes'

theorem (alternatively Bayes' law or Bayes'
rule) relates the conditional probabilities P(A | B)
and P(B | A). It is commonly used in science and
engineering. The theorem is named for Thomas
Bayes.
Simple form:
For events A and B, provided that
P(B) 0

In a Bayesian inference step, the probability of

evidence B is constant for all models An. The
posterior may then be expressed as proportional to
the numerator:

REFERENCES
[1]

[2]

[3]

[4]

[5]

[6]

CONCLUSION
[7]

This paper and the study of related topic will

further benefit us in developing our system
successfully. Hopefully I expect this system will
minimize the crime rate in India. Using this

[8]

Niraj Gadhe, Dr: B.K . Lande, VJTI, IOSR Journal of

Engineering (IOSRJEN),Vol. 2 Issue 2, Feb.2012.
Integrating Color and Shape-Texture Features for
Adaptive Real-time Object Tracking by Junqiu
WANG and Yasushi YAGI.
Improved Adaptive Gaussian Mixture Model for
Background Subtraction by Zoran Zivkovic ICPR,
2004.
Intelligent System for detecting, Modeling,
Classification of human behavior using image
processing, machine vision and OpenCV Niraj .B.
Gadhe, Prof: Dr. B .K . Lande, Prof: Dr. B.B.Meshram,
ISSN: 2278 1323, Volume 1, Issue 4, June 2012.
Understanding Transit Scenes: A Survey on Human
Behavior-Recognition Algorithms Joshua Candamo,
Matthew Shreve, Member, IEEE, Dmitry B. Goldgof,
Fellow, IEEE, Deborah B. Sapper, and Rangachar
Kasturi, Fellow,IEEE Transactions On Intelligent
Transportation Systems, Vol. 11, No. 1, March 2010
An Efficient Real Time Moving Object Detection
Method for Video Surveillance SystemPranab Kumar
Dhar, Mohammad Ibrahim Khan, Image Processing
and Pattern Recognition Vol. 5, No. 3, September,
2012.
A. D. Sappa and F. Dornaika, An Edge-Based Approach to
Motion Detection, V. N. Alexandrov, G. D. van Albada, P.
M. A. Sloot and J. J. Dongarra, (eds.) LNCS, vol. 3991,
(2006), pp. 563570. Springer, Heidelberg.
Performance Comparison of Moving Object Detection
Techniques in Video Surveillance System Dr. Dheeraj

ANNUAL STAFF CONVENTION - NATIONAL CONFERENCE 2013

Survey on Real Time moving object detection and back ground detection
for Intelligence Security System

[9]

Agrawal, Nitin Meena, IJES , ISSN: 2319 1813

ISBN: 2319 1805, 20th January 2013.
Book on Image Segmentation and Compression Using
Hidden Markov Models by Jia Li, Robert M. Gray,
Springer, 01-Jan-2000.

[10]

Image Segmentation Using Thresholding and

Genetic Algorithm P. Kanungo, P. K. Nanda and U.
C. Samal .
[11] Dar-Shyang Lee, Member, IEEE Effective Gaussian
Mixture Learning for Video Background Subtraction
in Ieee Transactions On Pattern Analysis And
Machine Intelligence, Vol. 27, No. 5, May 2005.
[12] Gaussian Mixture Models, Yogesh Raja - Shaogang
Gong , Department Of Computer Science, Queen
Mary And Westfield College,England, { Jpmetal,Sgg}
@Dcs.Qmw.Ac.Uk.
[13] www.wikipedia.org

ANNUAL STAFF CONVENTION - NATIONAL CONFERENCE 2013

2013