Motion Object Detection Using BGS Technique

Integrated Intelligent Research (IIR) International Journal of Computing Algorithm
Volume: 05 Issue: 01 June 2016, Page No. 1- 4
ISSN: 2278-2397
1
Motion Object Detection Using BGS Technique
I.Jothipriya1
, K.Krishnaveni2
1
P.hd Scholar, Madurai Kammaraj University,Madurai.
2
Head, Department of Computer Science, Sri.SRNM College, Sattur, Virudhunagar Dist.
Email: dijpriya@gmail.com, kkveni_srnmc@yahoo.co.in
Abstract --- The detection of moving object is an important in
many applications such as a vehicle identification in a traffic
monitoring system,human detection in a crime branch.In this
paper we identify a vehicle in a video sequence.This paper
briefly explain the detection of moving vehicle in a video.We
introduce a new algorithm BGS for idntifying vehicle in a
video sequence. First, we differentiate the foreground from
background in frames by learning the background. Then, the
image is divided into many small nonoverlapped frames. The
candidates of the vehicle part can be found from the frames if
there is some change in gray level between the current image
and the background. The extracted background subtraction
method is used in subsequent analysis to detect a vehicle and
classify moving vehicle.
Keywords : Background Subtraction,Optical Flow,Temporal
Difference.
I. INTRODUCTION
Now a days the growth of vehicle in a road has necessary
thing.Identification of moving vehicle in a video plays an
important role in many applications such as traffic monitoring
system,accident detection,reduce congestion and finding
ambulance vehicle.The first step of finding the moving vehicle
is to take a video sequence.For that purpose we fix a camera on
a signal.It take a sequence of video.Videos are sequence of
images.Each image is called a frame.An image is divided into
two parts.First one is the background image and another is a
foreground.We planned a method BGS to finding the
foreground of an image. The paper is organized as
follows:Section II describes in brief motion detection
architecture and its approaches.Section III covers the object
detection methods.Section IV Describes comparative study of
object recognition methods. Section V introduces a BGS
Method.Section VI explain the proposed work.In Section VII
Decribes experimental results. Finally, we conclude with a
discussion in Section VIII.
II. OBJECT DETECTION
Object identification is performed to existence of items in video
and to accurately locate that object. Most visual surveillance
systems start with movement identification. Movement
recognition aims at segmenting regions corresponding to
moving objects from the rest of an image. Object detection is to
establish a relationship between objects or object parts in
successive frames and to extract sequential information about
objects such as path, posture, speed and direction. The detection
of moving object's area of adjust in the same image sequence
which captured at different intervals in a video[5]. In reality,
road traffic can be broadly classified into two categories,
identical and varied. Detected objects frame by frame in video
is a important and difficult task. It is a crucial part of
monitoring systems since without object track, the system could
not take out interrelated sequential information about objects
and higher level performance analysis steps would not be
possible. Moving object detection is the first step in video
analysis. It can be used in many regions such as traffic
monitoring and face tracking.
III. OBJECT DETECTION METHODS
Object Detection
 Temporal Differencing
 Frame Differencing
 Optical Flow
 Background Subtraction
Temporal Difference Method
Temporal Differencing is based on frame differentiation which
attempts to detect moving regions by making use of the
difference of continuous frames (two or three) in a video series.
This method is highly adaptive to fixed background. So, the
temporal difference is a simple method for detecting moving
objects in a fixed background. But if the environment is not
static, the temporal difference method will very sensitive to any
movement and is difficult to differentiate the accurate and false
movement. So the temporal difference method can only be used
to detect the possible object moving area which is for the
optical flow calculation to detect real object movement.[3] . It
has high flexibility with active changes. Temporal differencing
technique utilize the pixel-wise variation between two or three
successive frames in a video imagery to take out moving
regions from the background [2]. It has high flexibility with
active scene changes although it cannot mine all related pixels
of a foreground image mostly when the object moves slowly or
has consistent surface [3, 4].When a foreground object is not
moving, temporal differencing method cannot detect a change
between successive frames and it lose the detection of item.
Frame Differencing
Frame differencing is a pixel-wise differencing between two or
three consecutive frames in an image sequence to detect regions
corresponding to moving object such as human and vehicles.
The threshold function determines change and it depends on the
speed of object motion. If the speed of the object changes much,
then it’s difficult to maintain the quality of movementation. The
inter-frame differencing approach detect parts of moving
objects by compare two consecutive frames. But, it can identify
only differences in the background environment and, for that
reason,it detects only parts of a vehicle covering the background
in the previous frame. Despite some enhancing techniques this
approach cannot acceptably deal with reasonable traffic
conditions where vehicles might stay still for a long time.[1].
Optical Flow
In this method ,the pattern of clear motion of objects, surface
and boundaries in a video caused by the relative motion

ISSN: 2278-2397
2
between an viewer and the scence.To detect moving regions in
an image, optical flow [5] uses flow vectors of the moving
objects over time. It is used for motion-based segmentation and
tracking applications. It is a dense field of dislocation vectors
which defines the change of each pixel area. Optical flow is best
suited in occurrence of camera movement, but however most
flow calculation methods are computationally very complex and
responsive to noise.
IV. COMPARATIVE ANALYSIS OF OBJECT
DETECTION METHODS
In this section we analyze the previously discussed different
vehicle detection algorithms.Table I indicate the evaluation of
these techniques on the basis of different advantages and
disadvantages.By analyzing these algorithms we concluded that
background subtraction method(BGS) is far better than
others.BGS method is greater uses rate of detecting vehicles.
V. BGS METHOD
BGS Method
Background elemination is a technique for finding a foreground
object from its background. It also known as Foreground
Detection. Background elemination method describe the recent
image is subtract from a reference background image, which is
upgrad during a period of time. It works well only in static
cameras. The elemination leaves only non-static or new objects,
which include whole dark region of an object. This approach is
simple and computationally affordable for real-time systems,
but it is very sensitive to active vehicle changes from lightning
and irrelevant event etc. Therefore it is highly reliable for good
background maintenance model.[6] The problem with
background subtraction [8], [9] is to automatically update the
background from the input video frame and it should be able to
overcome the following problems:
 Movement in the background: Dynamic background
regions, such as lanes in a road,leaves and branches of
trees, or flowing water, should be recognized as part of the
environment.
 Illumination changes: The background model should be
capable of slow changes in light over a phase of time.
 Memory: The background module should not use much
resource, such as computing time and memory.
 Gloom:Gloom cast by moving object should be identified
as part of the background and not foreground.
 Mask: Moving object should be detected even if pixel
uniqueness are similar to those of the background.
 Bootstrapping: The background model must be maintained
even in the lack of instruction foreground object.
VI. PROPOSED WORK
BGS method is a technique for identifying vehicles on a road.In
this algorithm a pixel is marked as foreground
If [Cf(x,y) – Bf(x,y)] > T (1)
In this equation where Cf is the current frame,Bf(x,y) is the
background frame and T is the predefined threshold value.
After the background frame Bf(x,y) is obtained, subtract the
background frame Bf(x,y) from the current frame Cf(x,y). If the
pixel difference is greater than the set threshold value T, then
determines that the pixels occur in the moving vehicle,
otherwise, it is consider as the background pixels. The moving
vehicle can be detected after applying threshold operation . Its
expression is given below:
Df(x,y)=1 if difference is grater than T,Otherwise 0
Where Df(x,y) is the binary frame of different results,T is
dynamic which will be identified depends upon environment
changes.Therefore,we add a dynamic threshold ∆T to the
vehicle detection BGS method.It is expressed as
∆T = λ ∗
1
M ∗ N
∑ ∑ [C(i, j) − B(i, j)] (2)
M
j=1
N
i=1
In equation 2 ,where M x N is the size of each image, ΔT
reflects the overall changes in the environment.
Then,
Df(x,y)=1 if difference is grater than (T+ ∆T),Otherwise 0
It is effectively eliminate the impact of light changes.
BGS method follows four steps namely caputuring the input as
.avi file, perform preprocessing(morphological
operations),foreground identifying,vehicle detection. The
sematic flow diagram of the proposed method is shown in
figure I
Step 1:Caputuring the Input
Read the .avi file using VideorReader.The input video file
converted into images to detect vehicle.These images are
gathered into frames.Create the frame to each images.This
process will be continued for all the frames.
Step 2:Morphalogical Preprocessing
The input image undergoes a series of morphological
operations to detect the exact shape of the
object.Morphological image processing is a collection of non-
linear operations related to the shape or morphology of features
in an image. Morphological operations can also be applied to
greyscale images. It can be used in pre or post processing
(filtering, thinning, and pruning) for getting a representation or
description of the shape of objects/regions (boundaries,
skeletons convex hulls).
A)Opening
Opening consists of an erosion followed by a dilation and can
be used to eliminate all pixels in regions that are too small to
contain the structuring element. In other words, foreground
structures that are smaller than the structure element will
disappear. In figure 3 shows opening image.
B)Reconstruction
Morphological reconstruction to extract marked objects, find
bright regions surrounded by dark pixels, detect or remove
objects touching the image border, detect or fill in object holes,
filter out spurious high or low points, and perform many other
operations.In figure 3 shows recontrcted image.
Step 3:Background Elimination(Foreground Detection)
The background elimination method is used to finding of the
motion vehicle in the traffic system.The demonstration has set
up for the proposed system in the MATLAB software.Here
the current frame is initialized in the code and elimination is
done.After elimination the background image,current image

ISSN: 2278-2397
3
and foreground image are show in figure 2.In that figure the
background is represented as black where as foreground
vehicle is white in color.
Step 4: Moving Vehicle Detected
Finally,moving vehicle is detected from background
environment.
VII. EXPERIMENTAL RESULT
Experimental results for moving vehicle detection using the
proposed method have been produced for many
frames.Here,we implement three different frames that represent
number of vehicles for traffic system.In figure 2 represented
as a background image for all frames,In figure 3 shows the
difference between frame 72 and its background elimination.
In figure 4 shows the difference between frame 35 and its
background elimination.
VIII. CONCLUSION
In this paper, a moving of vehicle is detected, based on BGS
method. We propose consistent BGS method which uses
thresholding to detect moving vehicle. This method is
successful in real time, and it works well for all frames from
our given .avi file.
References
[1] Rupali S.Rakibe, Bharati D.Patil, “Background
Subtraction Algorithm Based Human Motion
Detection”,International Journal of Scientific and
Research Publications, May 2013.
[2] Kinjal A Joshi, Darshak G. Thakore ―A Survey on
Moving Object Detection and Tracking in Video
Surveillance System International Journal of Soft
Computing and Engineering (IJSCE) ISSN: 2231-2307,
Volume-2, Issue-3, July 2012
[3] N. Paragios, and R. Deriche.. ―Geodesic active contours
and level sets for the detection and tracking of moving
objects.IEEE Trans. Pattern Analysis Machine
Intelligence. 22, 3, 266–280, 2000.
[4] S. Zhu and A. Yuille. ―Region competition: unifying
snakes, region growing, and bayes/mdl for multiband
image segmentation.‖ IEEE Trans. Pattern Analysis
Machine Intelligence 18, 9, 884–900, 1996.
[5] L. Vasu, “An effective step to real-time implementation of
accident detection system using image processing”, Master
of Science, Oklahoma State University, USA, (2010).
[6] Abhishek Kumar Chauhan, Deep Kumar,” Study of
Moving Object Detection and Tracking for Video
Surveillance”, IJARCSSE,2013
[7] M. Kalpana Chowdary , S. Suparshya Babu , S. Susrutha
Babu , Dr. Habibulla Khan‖,FPGA Implementation of
Moving Object Detection in Frames by Using Background
Subtraction Algorithm‖ International conference on
Comm.unication and Signal Processing, April 3-5, 2013.
[8] J. Xiao, H. Cheng, H. Sawhney, C. Rao, and M. Isnardi,
“Bilateral filtering- based optical flow estimation with
occlusion detection,” ECCV,2006, pp. 211–224
[9] Mr. Vishwadeep Uttamrao Landge, Mr. Rajan G.
Mevekari,Object Detectionand Object Classification using
Background Subtraction Method, International Journal of
Engineering Research & Technology (IJERT) ISSN:
2278-0181 Vol. 3 Issue 9, September- 2014.
Table I Comparative Analysis Of Vehicle Detection Methods
Methods Backgrou
nd
Subtractio
n
Tempora
l
Differen
ce
Frame
Differen
ce
Optical
Flow
Accuracy Fair Fair High Fair
Time Fair to
High
High Low to
Fair
High
Advantage It need
not much
of
memory
It
perform
for
dynamic
backdro
p
It
perform
for static
backdro
p
It
produce
motion
informa
tion
Disadvanta
ge
It does not
work
multi
environm
ent
It can
not
detect
changes
into
successi
ve
frames
It is
difficult
to
maintain
if the
speed of
object
changes
It
required
more
calculati
ons
Uses Rate 40% 10% 30% 20%
Figure 1 Sematic Flow Diagram of Vehicle Detection
Figure 2 Background Image

ISSN: 2278-2397
4
Figure 3:Difference between frame 72 and background
subtracted(foreground)
Figure 4:Difference between frame 35 and background
subtracted(foreground)
Figure 5 Morphalogical Preprocessing(Open,Reconstruct)

Motion Object Detection Using BGS Technique

Recommended

More Related Content

Similar to Motion Object Detection Using BGS Technique (20)

More from MangaiK4 (20)

Recently uploaded (20)

Motion Object Detection Using BGS Technique