International Journal of Innovative Technology and Exploring Engineering (IJITEE)
ISSN: 2278-3075, Volume-9 Issue-1, November 2019
887
Published By:
Blue Eyes Intelligence Engineering
& Sciences Publication
Retrieval Number: A4419119119/2019©BEIESP
DOI: 10.35940/ijitee.A4419.119119
Abstract: Object tracking is a troublesome undertaking and
significant extent in data processor perception and image
handling community. Some of the applications are protection
surveillance, traffic monitoring on roads, offense detection and
medical imaging. In this paper a recent technique for tracking of
moving object is intended. Optical flow information authorizes us
to know the displacement and speed of objects personate in a
scene. Apply optical flow to the image gives flow vectors of the
points to distinguishing the moving aspects. Optical flow is
accomplished by Lucas canade algorithm. This algorithm is
superior to other algorithms. The outcomes reveals that the intend
algorithm is efficient and accurate object tracking method. This
paper depicts a smoothing algorithm to track the moving object of
both single and multiple objects in real time. The main issue of
high computational time is greatly reduced in this proposed work.
Keywords: Computer Vision, Lucas Canade Algorithm, Object
Tracking Optical Flow Technique.
I. INTRODUCTION
Moving target tracking is to determine locations that the
same object from different frames in image sequences. When
moving object is correctly detected, the tracking is equivalent
to the characteristics matching problem. The characteristics
are created in consecutive image frames based on location,
velocity, shape, texture, color and so on. Moving object
tracking is a very considerable subject in the extent of data
processor vision. It is the premise of the object detection and
the cognizant behavior analysis. The robust object detection
is a necessary undertake for the precision of subsequent
object tracking and comportment analysis.
Many object detection algorithms have been intended, such
as temporal differencing, frame difference method
background subtraction method, optical flow method and so
on.
Temporal differencing [1] taken into account dissimilarity
in two succeeding frames with esteem to time. It is primarily
same as background subtraction technique to subtract the
image fetters target information through threshold value. It is
very quiet to implement but it is lack of significant pixels of
some types of moving object and cannot used in real time
implementations.
Revised Manuscript Received on November 06, 2019.
C.Karthika Pragadeeswari, Assistant Professor, Department of ECE,
Alagappa Chettiar Government college of Engineering and Technology,
Karaikudi 630003 India. Email: bk.karthika1969@gmail.com
G.Yamuna, Professor and Head, Department of ECE, Annamalai
University, Annamalainagar, 608002 India. Email: yamuna.sky@gmail.com
G.Yasmin Beham, PG Scholar, Department of ECE, Alagappa Chettiar
Government college of Engineering and Technology, Karaikudi 630003
India. Email: yasar14051996@gmail.com
Frame difference method [2] is a simple algorithm, in that
dispute between two successive frames obtained and the
object can be recognized by this method and is tracked by
template matching algorithm. This system is cost effective
and utilizes a supervision tool in distinct applications.
Background subtraction method [3] is used to subtract the
current frame from background taken as reference and
various background subtraction models are analyzed and
compared with their capability, reminiscence and fidelity. It
is the prime method for extended applications.
Optical flow technique characterized as the changes in the
movement of object with respect to brightness pattern in
image. Optical flow computes the motion between two
frames for every pixel in the frame which are taken at
different time intervals. Optical flow is extensively used
technique to track the moving objects efficiently. This
method is capable of providing complete movement
information and detects the moving object from the
background better than the other methods. However this
method is useful in object tracking. The main work of target
tracking is to select good target characteristics.
Determining an optical flow [4] is quiet important. A
technique for evaluating the optical flow instance is
demonstrated which appropriates that alters in location
analogous to brightness pattern. It readily varies at any place
in the image. A repeating performance is denoted which
satisfyingly projects the optical flow for several criterions in
image sequences. The algorithm is incredibly rapid and
additionally unfeeling towards quantization of contrast levels
and additive noise.
Prediction and Tracking of Moving Objects in Image
Sequences [5] that we utilize a forecast model for movement
of object velocity and area estimation got from Bayesian
hypothesis. Segmentation and optical flow tracking is used
for predicting future frames. The proposed algorithm
provides good moving object tracking capabilities. Such
capabilities are used for segmenting and estimating the
moving object velocity and segmentation in a future frame.
Optical Flow Based Moving Object Detection and Tracking
for Traffic Surveillance [6] that trusting on optical flow can
be through Horn Schunck algorithm. To evacuate clamors,
median filter is utilized and the undesirable objects are
excluded by applying Thresholding calculations in
morphological tasks. The outcomes prove that the framework
powerfully recognizes and tracks moving aspects in urban
video and evacuates the undesirable objects which are not
vehicles.
A Robust Algorithm for Real Time Tracking
With Optical Flow
C. Karthika Pragadeeswari, G.Yamuna, G. Yasmin Beham